Exoplanet demographics: occurrence rates, habitability zone, atmospheric characterization

Intuition Beginner

Since the first exoplanet around a Sun-like star was found in 1995, astronomers have confirmed more than 5,000 worlds orbiting other stars. They come in types absent from our solar system: hot Jupiters (gas giants scorched by their stars on days-long orbits), super-Earths (rocky planets two to ten times Earth's mass), and mini-Neptunes (small gas planets with no local analog). The census grows every month, drawn from transit and radial-velocity surveys that watch hundreds of thousands of stars.

NASA's Kepler telescope stared at one patch of sky for four years and revealed how common small planets are. Perhaps twenty to fifty percent of Sun-like stars host a roughly Earth-sized world in the habitable zone — the orbital band where a planet could keep liquid water on its surface. Around the smaller, cooler M dwarfs, the habitable zone sits closer in, making these dim red stars prime hunting grounds for temperate worlds.

The James Webb Space Telescope now peers through the thin ring of atmosphere backlit during a planet's transit. Its spectrographs split the starlight into colours, recording fingerprints of water vapour, carbon dioxide, methane, and other gases. Some of these are candidate biosignatures — gases that, on Earth at least, are produced by life. The search has only just begun, but the first atmospheres of small, temperate worlds are now within reach.

Visual Beginner

Planet type	Radius vs Earth	Description	Example
Hot Jupiter	~10x	Gas giant on a days-long orbit	51 Peg b
Super-Earth	1.25-2x	Rocky, heavier than Earth	Kepler-442b
Mini-Neptune	2-4x	Small gas envelope, no analog	Kepler-11f
Gas giant	>6x	Jupiter analog on a wide orbit	HR 8799 b

Star type	Habitable zone	Detection ease	Flare risk
M dwarf	Close in (weeks-months)	Easiest	High
K dwarf	Moderate (months)	Moderate	Low
G (Sun-like)	~1 AU (year-long)	Harder	Low

Worked example Beginner

Example 1: Counting Earth analogues

Kepler suggests that roughly thirty percent of Sun-like stars host an Earth-sized planet in the habitable zone. Within 50 light-years of the Sun there are about 1,000 Sun-like stars. Multiplying, $0.30\times 1000=300$, so several hundred such worlds should sit in our cosmic neighbourhood, even before counting the far more numerous M-dwarf systems.

The takeaway: Earth-sized, potentially temperate planets are not rare. The challenge is reading their atmospheres, not finding the planets themselves.

Example 2: Where is the habitable zone?

A planet stays warm enough for liquid water when it receives about the same stellar flux Earth gets. Flux falls off with the square of distance, so the habitable-zone distance scales with the square root of the star's luminosity. For a dim M dwarf with $0.01$ times the Sun's luminosity, $\sqrt{0.01}=0.1$, placing the habitable zone near $0.1$ AU and giving a year of only a few weeks.

The takeaway: cooler stars pack their habitable zones in tight, which is why transiting temperate planets turn up most often around M dwarfs.

Check your understanding Beginner

Formal definition Intermediate+

Occurrence rates

The occurrence rate $\eta$ is the average number of planets per star in a specified bin of period $P$ and planet radius $R_p$. The headline numbers are ${\eta }_{\oplus }$, the fraction of Sun-like stars hosting a roughly Earth-sized planet in the habitable zone, and ${\eta }_{\mathrm{Jup}}$, the corresponding fraction hosting a Jupiter analog on a wide orbit. Kepler yields ${\eta }_{\oplus }$ in the range 10-50 percent depending on the precise bin boundaries and detection thresholds, while ${\eta }_{\mathrm{Jup}}$ is of order 1 percent. Occurrence is a strong function of both stellar type and planet parameters: M dwarfs host more close-in small planets than Sun-like stars, and the radius distribution is bimodal ^{[source pending]}.

The radius valley

Fulton et al. (2017) showed that the planet radius distribution is bimodal, with a deficit — the radius valley — between roughly 1.5 and 2.0 $R_{\oplus }$. The two peaks correspond to super-Earths (rocky, with thin or no envelopes) and mini-Neptunes (with thick hydrogen-helium envelopes). The valley's location is consistent with photoevaporation: stellar high-energy flux strips the envelopes of the smallest gaseous planets, turning mini-Neptunes into bare super-Earths. Core-powered mass loss, driven by the planet's own cooling luminosity, produces a similar imprint, and distinguishing the two mechanisms is an active research problem ^{[source pending]}.

Habitable-zone definitions

Kasting, Whitmire, and Reynolds (1993) defined the habitable zone by the limits on surface liquid water for an Earth-like atmosphere. The inner edge is set by the moist greenhouse, where water vapour dominates the upper atmosphere and is photolysed and lost to space, and by the runaway greenhouse, where oceans evaporate entirely. The outer edge is set by the maximum greenhouse limit, beyond which even a thick carbon-dioxide atmosphere cannot keep the surface above freezing because of CO2 condensation and Rayleigh scattering. Conservative and optimistic boundaries bracket this range; recent-Venus and early-Mars limits widen the optimistic zone ^{[source pending]}.

Atmospheric characterization methods

Three observational modes probe exoplanet atmospheres. Transmission spectroscopy divides the transit depth across wavelengths, reading the additional absorption imprinted by the atmosphere at the terminator. Secondary-eclipse spectroscopy measures the dayside thermal emission (or reflected light) that vanishes when the planet passes behind the star. Phase-curve monitoring tracks the brightness change through the full orbit, mapping the longitude-temperature structure and heat redistribution. Molecules recovered include H2O, CO2, CO, CH4, Na, K, and, in photochemical atmospheres, SO2; biosignature candidates are O2, O3, and the CH4/CO2 disequilibrium pair ^{[source pending]}.

Key result: habitable-zone boundaries and occurrence-rate estimation Intermediate+

The demographic and habitability questions reduce to two analytic cores: where the habitable zone sits as a function of stellar luminosity, and how many planets inhabit it once the survey's missed detections are corrected for. Deriving both fixes the framework that the master tier refines statistically.

Habitable-zone scaling. A planet in radiative equilibrium absorbs stellar flux on its cross-section and re-radiates over its full surface. Equating the two gives the equilibrium temperature

$$T_{\mathrm{eq}}={\left[\frac{(1-A)L_{\star }}{16\pi \sigma d^2}\right]}^{1/4},$$

where $A$ is the Bond albedo, $L_{\star }$ the stellar luminosity, and $d$ the orbital distance. Holding $T_{\mathrm{eq}}$ fixed at a habitable-zone edge immediately gives the scaling $d\propto \sqrt{L_{\star }}$. Writing the flux at the edge as $S_{\mathrm{eff}}$ times the solar constant, the habitable-zone distance in astronomical units is

$$d_{\mathrm{HZ}}[\mathrm{AU}]=\sqrt{\frac{L_{\star }/L_{\odot }}{S_{\mathrm{eff}}}}.$$

For the conservative Kopparapu et al. (2013) limits around the Sun, $S_{\mathrm{eff},\mathrm{inner}}\approx 1.07$ (moist greenhouse) and $S_{\mathrm{eff},\mathrm{outer}}\approx 0.36$ (maximum greenhouse), giving $d_{\mathrm{HZ}}\in [0.97,1.70]$ AU. For an M dwarf with $L_{\star }/L_{\odot }=0.01$ the whole zone contracts to roughly $0.10$-$0.17$ AU, with periods of a few to a few tens of days ^{[source pending]}. This single scaling explains why temperate transiting planets cluster around M dwarfs: the habitable zone sits close in, raising both the transit probability $R_{\star }/a$ and the transit repeat rate.

Occurrence-rate estimation. A survey detects only a censored subset of the true planet population. The inverse detection efficiency method (IDEM) weights each detected planet by the inverse of the probability that the survey would have found it. Grouping planets into a bin of period and radius, the occurrence-rate estimator is

$$\hat{\Gamma }=\frac{1}{N_{\star }}\sum _{j\in \mathrm{bin}}\frac{1}{{\eta }_{\det }(P_j,R_j;{\text{star}}_j)},$$

where $N_{\star }$ is the number of surveyed stars, the index $j$ runs over the detected planets in the bin, and ${\eta }_{\det }$ is the per-star detection completeness evaluated at that planet's parameters. The completeness ${\eta }_{\det }$ folds together the geometric transit probability, the instrumental signal-to-noise threshold, the survey window function, and the vetting efficiency; it is measured empirically by injecting synthetic transits into the light curves and recovering them. Summing the inverse-completeness weights and normalising by $N_{\star }$ corrects for the planets that were present but missed, yielding the average number of planets per star in the bin ^{[source pending]}.

From rates to yields. Multiplying $\hat{\Gamma }$ by a target-list size forecasts how many habitable-zone terrestrial planets a future mission will recover, and the same scaling sets the observing-time budget for atmospheric follow-up. The estimator's precision is limited by small-number statistics in the rarest bins (giant planets, long periods) and by systematic errors in the completeness model, which the master tier addresses with hierarchical Bayesian methods.

Exercises Intermediate+

Advanced results Master

Bayesian occurrence-rate estimation

The IDEM estimator of the intermediate tier is unbiased in the limit of large samples but becomes biased and brittle in sparsely populated bins, which is exactly where the rarest and most sought-after planets (Earth analogues, giant planets on wide orbits) reside. The modern treatment is hierarchical Bayesian inference: a parametric occurrence-rate model $\Gamma (P,R_p\mid \theta )$ is placed over the period-radius plane, the survey's detection completeness and vetting efficiency enter as a forward model, and the posterior $p(\theta \mid \{\text{detections}\})$ is sampled with Markov-chain Monte Carlo. Approximate Bayesian computation extends this to cases where the likelihood is intractable, simulating synthetic catalogues through the full detection pipeline and accepting parameter draws whose synthetic yield matches the observed count. Detection completeness itself is measured by injection/recovery: tens of thousands of synthetic transits are planted in the real light curves, passed through the vetting pipeline, and the fraction recovered as a function of period, radius, and stellar magnitude yields the empirical completeness surface ${\eta }_{\det }(P,R_p;m)$ ^{[source pending]}.

Radius-valley formation mechanisms

Two leading mechanisms compete to explain the bimodal radius distribution. In the photoevaporation model, the star's high-energy (X-ray and extreme-ultraviolet) flux heats the upper atmosphere of a gaseous mini-Neptune; the envelope expands and hydrodynamic escape strips it over hundreds of megayears, leaving a bare rocky core if the planet's escape velocity is low enough. In the core-powered mass-loss model, the planet's own cooling luminosity, augmented by the stellar bolometric flux, drives boil-off of the primordial envelope. The two mechanisms predict subtly different valley locations as functions of stellar mass and system age, and current data — from Kepler, K2, and the California-Kepler Survey — are consistent with both contributing, with photoevaporation dominating at early times around active stars and core-powered loss operating over longer timescales ^{[source pending]}.

Atmospheric retrieval and its degeneracies

Atmospheric retrieval inverts an observed spectrum against a forward radiative-transfer model to infer the temperature profile, chemical abundances, and cloud properties. Forward models integrate line-by-line opacities (H2O, CO2, CO, CH4, Na, K, and others) over a pressure-temperature profile, with cloud and haze layers adding grey or wavelength-dependent opacity. Retrieval codes — PLATON, ATMO, NEMESIS, petitRADTRANS — explore the parameter space with nested sampling (MultiNest, PyMultiNest) or MCMC, returning posterior distributions for each quantity. Transmission spectra suffer two chronic degeneracies: cloud and haze opacity flattens spectral features, suppressing the molecular fingerprints and capping the information content; and the reference-radius degeneracy entangles the assumed planetary radius at a reference pressure with the inferred mean molecular weight and scale height, so that a cloud deck at depth can masquerade as a high mean-molecular-weight atmosphere. Breaking these degeneracies requires combining transmission with emission or phase-curve data and priors on the bulk density from mass and radius measurements ^{[source pending]}.

JWST early results and emission spectroscopy

JWST has transformed atmospheric characterisation. The detection of carbon dioxide in the hot Saturn WASP-39 b by the Early Release Science programme was the first unambiguous CO2 feature in an exoplanet spectrum, and the simultaneous sulphur dioxide detection signalled photochemistry driven by stellar ultraviolet. The sub-Neptune K2-18 b provoked a contested claim of dimethyl sulphide (DMS), a potential biological tracer, in its hydrogen-rich atmosphere; the debate turned on the statistical significance of weak features at near-infrared wavelengths and on the assumed atmospheric chemistry, illustrating how retrieval degeneracies and model priors can dominate weak-signal claims. Thermal emission and phase-curve measurements with JWST's MIRI map the dayside-nightside temperature contrast and the hotspot offset, quantifying the efficiency of heat redistribution: a strongly irradiated hot Jupiter with efficient advection shows a small contrast and a large eastward offset, while a bare rock radiating re-radiatively shows a stark day-night dichotomy. Direct high-resolution spectroscopy with ground-based Extremely Large Telescopes (ELT/METIS) will resolve individual Doppler-shifted molecular lines for temperate planets around the nearest M dwarfs, a complementary route to the space-based low-resolution transit spectra ^{[source pending]}.

Biosignature frameworks, false positives, and technosignatures

A biosignature is an observation — a gas, a surface reflectivity feature, or a disequilibrium combination — whose presence is most plausibly explained by life. Current frameworks, distilled in the "Earth Through Time" concept, ask what a remote observer would detect for Earth at each geologic eon: an anoxic Archean atmosphere would show methane and carbon dioxide without oxygen, a Proterozoic atmosphere would show weak oxygen and ozone alongside methane, and only the modern atmosphere shows the strong O2-O3/CH4 disequilibrium. The central difficulty is the false positive. Around M dwarfs, stellar ultraviolet photolyses CO2 and H2O to produce abiotic O2 and O2 columns large enough to mimic a photosynthetic signal; the detection of CO as a by-product, or the absence of reducing gases, serves as a discriminant. Surface reflectivity features such as the vegetation red edge are additional but fragile discriminants. Technosignatures — radio emission (SETI), waste-heat infrared excess from Dyson spheres or megastructures, or atmospheric industrial pollutants such as CFCs — extend the search to technological life, with the advantage that false positives are rarer but the signals far more speculative ^{[source pending]}.

Drake equation, M-dwarf flares, and climate dynamics

The Drake equation $N=R_{\star }{f}_p{n}_e{f}_l{f}_i{f}_{c}L$ decomposes the number of detectable civilisations into a product of astrophysical and biological factors. Exoplanet demographics now supply empirical values for the first three: the star-formation rate $R_{\star }$, the planet-habitable-zone fraction $f_p{n}_e$ (effectively ${\eta }_{\oplus }$ for Sun-like stars and its M-dwarf analogue), and increasingly the fraction of habitable-zone planets that retain atmospheres. The remaining factors — the fractions leading to life, intelligence, and communicative technology, and the longevity $L$ — remain unconstrained by astronomy, and the distinction between habitable (the right physical conditions) and inhabited (life actually present) is the unbridgeable gap the equation frames. For M-dwarf planets the habitability question is sharpened by two complications. First, intense stellar flares and coronal mass ejections during the prolonged pre-main-sequence phase can erode atmospheres through hydrodynamic escape and sputtering, so the relevant quantity is not whether a planet sits in the habitable zone but whether any atmosphere survives the host's active youth. Second, habitable-zone M-dwarf planets are likely tidally locked, presenting a permanent dayside and nightside; climate modelling suggests that a thick enough atmosphere or ocean can redistribute heat and avoid atmospheric collapse on the nightside, but thin atmospheres freeze out, making atmospheric retention and heat-transport efficiency the twin gatekeepers of M-dwarf habitability ^{[source pending]}.

Connections Master

This unit is the demographic and characterisation capstone of the exoplanet chapter. It presupposes the detection machinery of 28.05.02 pending — the transit depths, radial-velocity semi-amplitudes, and direct-imaging contrasts that produce the raw catalogues — and the qualitative habitability framing of the overview unit 28.05.01, replacing each loose statement there with a quantified occurrence rate or habitable-zone boundary. The intermediate-tier derivation of $d_{\mathrm{HZ}}\propto \sqrt{L_{\star }}$ is the practical echo of the radiative-equilibrium arguments developed in the stellar-structure unit 28.02.02 pending, while the equilibrium-temperature expression inherits the Stefan-Boltzmann flux balance that underpins the stellar-evolution and stellar-endpoints units 28.02.03 pending, 28.02.04 pending.

The inferential machinery extends well beyond astronomy. Occurrence-rate estimation from a censored, detection-efficiency-limited sample is a problem in survival analysis and hierarchical Bayesian inference, sharing its mathematical core with medical reliability and ecological capture-recapture statistics; atmospheric retrieval is an application of inverse problem theory and Bayesian nested sampling; and the Drake equation is a dimensional decomposition of an uncertain product that recurs in risk analysis and Fermi estimation. The unit is therefore a consumer of statistical inference, radiative transfer (line-by-line opacities inherited from Earth-atmosphere and planetary-atmosphere work), and comparative planetology, the last anchored by the solar-system ground truth of 28.01.01 and 28.01.02 pending.

Historical and philosophical context Master

The habitable-zone concept long predates the exoplanet era. Huang Su-Shu in 1959 and, more influentially, Kasting, Whitmire, and Reynolds in 1993 framed the zone by the moist and runaway greenhouse inner limits and the maximum-greenhouse outer limit, anchored in one-dimensional radiative-convective climate models rather than in any observed planet. The Kasting et al. (1993) boundaries, refined by Kopparapu et al. (2013) with updated absorption coefficients, remain the standard reference even as three-dimensional general-circulation models now test the atmospheric-collapse and cloud-feedback assumptions that the one-dimensional calculation could not.

The demographic revolution came with Kepler. Petigura, Howard, and Marcy (2013) applied a forward-modelling completeness correction to the Kepler catalogue and reported an eta-Earth of roughly 20 percent for Earth-sized planets in the habitable zones of Sun-like stars, with the rate climbing toward M dwarfs. Fulton et al. (2017) then exposed the bimodal radius distribution and the radius valley, re-framing super-Earths and mini-Neptunes as the products of envelope loss rather than two distinct formation channels. Together these results converted the question "are Earth-sized planets common?" from speculation into a measured quantity, with error bars.

Atmospheric characterisation has moved in discrete leaps. Charbonneau et al. (2002) detected sodium in the atmosphere of HD 209458 b with HST, the first exoplanet atmospheric measurement; water vapour followed in several hot Jupiters over the next decade. JWST's detection of carbon dioxide in WASP-39 b in 2022 opened the era of unambiguous molecular spectroscopy for transiting planets, and the contested K2-18 b DMS claim of 2023 foreshadowed the interpretive battles that will accompany every future temperate-world spectrum. The philosophical stakes — the distinction between habitable and inhabited, and between a biosignature and a false positive — are now concrete observational questions rather than abstract speculation, and the Drake equation, once a vessel for ignorance, has had its first three factors filled in by measurement.

Bibliography Master

1. Kasting, J. F., Whitmire, D. P., and Reynolds, R. T. (1993). "Habitable zones around main sequence stars." Icarus, 101, 108-128.

2. Kopparapu, R. K., Ramirez, R., Kasting, J. F., et al. (2013). "Habitable zones around main-sequence stars: new estimates." Astrophysical Journal, 765, 131.

3. Petigura, E. A., Howard, A. W., and Marcy, G. W. (2013). "Prevalence of Earth-size planets orbiting Sun-like stars." Proceedings of the National Academy of Sciences, 110, 19273-19278.

4. Fulton, B. J., Petigura, E. A., Howard, A. W., et al. (2017). "The California-Kepler Survey. III. A gap in the radius distribution of small planets." Astronomical Journal, 154, 109.

5. Owen, J. E. and Wu, Y. (2017). "The evaporation valley in the Kepler planets." Astrophysical Journal, 847, 29.

6. Ginzburg, S., Schlichting, H. E., and Sari, R. (2018). "Core-powered mass loss and the evaporation valley." Monthly Notices of the Royal Astronomical Society, 476, 759-765.

7. Charbonneau, D., Brown, T. M., Noyes, R. W., and Gilliland, R. L. (2002). "Detection of an extrasolar planet atmosphere." Astrophysical Journal, 568, 377-384.

8. Seager, S. (2010). Exoplanet Atmospheres: Physical Processes. Princeton University Press.

9. Madhusudhan, N. (2019). "Exoplanetary atmospheric retrievals." Exoplanet Atmospheres: Proceedings of the International Astronomical Union, 14, 2-10.

10. The JWST Transiting Exoplanet Community Early Release Science Team (2023). "Identification of carbon dioxide in an exoplanet atmosphere." Nature, 614, 649-652.

11. Madhusudhan, N., Sarkar, S., Constantinou, S., et al. (2023). "Carbon-bearing molecules in a possible Hycean atmosphere." Astrophysical Journal, 956, L13.

12. Perryman, M. (2018). The Exoplanet Handbook (2nd ed.). Cambridge University Press.

13. Carroll, B. W. and Ostlie, D. A. (2017). An Introduction to Modern Astrophysics (2nd ed.). Pearson, Chapter 23.

14. Winn, J. N. and Fabrycky, D. C. (2015). "The occurrence and architecture of exoplanetary systems." Annual Review of Astronomy and Astrophysics, 53, 409-447.

15. Des Marais, D. J., Harwit, M. O., Jucks, K. W., et al. (2002). "Remote sensing of planetary properties and biosignatures on extrasolar terrestrial planets." Astrobiology, 2, 153-181.