Chance played a role in determining whether Earth stayed habitable

The study addresses why Earth remained continuously habitable for billions of years despite factors that should have driven it outside habitable bounds, notably the 30% increase in solar luminosity (the Faint Young Sun problem) and the short residence time of carbon in the surficial system, which implies potential for rapid climate failure. Comparative planetology (Venus too hot, Mars too cold and possibly formerly habitable) underscores that habitability can be lost. Two explanatory classes are reviewed: (1) stabilising mechanisms (e.g., biotic regulation via Gaia and geochemical negative feedbacks such as silicate weathering on continents, islands, or seafloor) and (2) observer selection (weak anthropic principle), wherein rare long-term habitability arises by chance among many planets. The prevailing view emphasizes stabilising feedbacks, but the potential role of chance is often omitted. This paper develops a general, minimal climate-habitability model to test whether long-term habitability outcomes reflect mechanism, chance, or both.

The paper synthesizes prior work on the Faint Young Sun paradox and planetary climate stability. It reviews: (a) stellar evolution and increasing solar luminosity; (b) Venus and Mars as examples of lost or absent present-day habitability; (c) the carbon cycle's potential for rapid disequilibrium leading to snowball or runaway greenhouse states; (d) stabilising hypotheses including Gaia (life-regulated climate), continental/island/seafloor silicate weathering as negative feedbacks, and their debated efficacy based on proxy records (e.g., Cenozoic cooling, PETM recovery attributed to organic carbon burial); and (e) the weak anthropic principle proposing observer selection among many planets. It also notes exoplanet diversity and abundance, implying a large statistical ensemble in which rare outcomes can occur.

A minimalist, general planetary climate-habitability model was built to evaluate the persistence of thermal habitability over 3 billion years (By). The model has a single state variable, surface temperature T, governed by dT/dt = f(T) + φ, where f(T) is the net temperature-dependent feedback (warming/cooling rate) and φ is a long-term climate forcing. Habitability is defined solely by temperature staying within a habitable range. Key elements: (1) Habitable temperature range: Tmin = −10 °C, Tmax = +60 °C (aimed at allowing liquid water and enzyme stability for complex life). (2) Randomly configured feedbacks: For each planet, the number of nodes N is drawn from a uniform integer distribution U_i(2,20). Node temperatures are evenly spaced between Tmin and Tmax. Each node’s feedback strength f_i is drawn from N(0,100) °C ky−1; f(T) between nodes is linearly interpolated. This yields random mixtures of stabilising (negative slope) and destabilising (positive slope) regions, possible attractors, and potential runaway zones. (3) Long-term forcing: Each planet has a φ drawn from N(0,50) (°C ky−1) By−1, representing net secular drivers (e.g., stellar luminosity changes, geophysical evolution). (4) Perturbations: Each run experiences instantaneous temperature perturbations representing events such as large eruptions, impacts, orbital changes, stellar flares, etc. Perturbation magnitudes: small N(2,1) °C, medium N(8,4) °C, large N(32,16) °C, with random sign. Expected numbers per 3 By are planet-specific and drawn as λ_S ~ U_i(4000,20000), λ_M ~ U_i(40,400), λ_L ~ U_i(0,5), with actual counts from Poisson distributions P(λ). Perturbation times are random. (5) Initial conditions and success criterion: Each run starts at T0 ~ U(Tmin, Tmax). A run is a success if T remains within [Tmin, Tmax] for 3 By; otherwise it fails upon first exit. (6) Integration: Matlab ode23s with adaptive timesteps; integration restarts at each perturbation. (7) Experimental designs: (a) Hypothesis tests contrasting chance-alone (H1) vs mechanism-alone (H2) using planets run 1000 times without long-term forcing and with modest perturbations, including specifically unfavorable, neutral (no feedbacks), and optimal stabilising configurations. (b) Large ensemble: 100,000 distinct planets each run 100 times with independent perturbations and initial temperatures to map success-rate distributions. (c) Rerun analysis: 1,000 planets each run twice with different initial conditions and perturbations to quantify repeatability. (8) Sensitivity analyses: 37 variants spanning feedback distributions, node counts, interpolation methods, time-varying feedbacks, alternative habitable ranges and their evolution, perturbation frequencies/magnitudes, forcing distributions and trends, and required habitability duration; the principal qualitative result was robust across all variants.

• Role of chance and mechanism: The distribution of outcomes differs markedly from both pure chance (H1) and pure mechanism (H2); both planetary properties and chance events together determine long-term habitability. A runs test on ordered planet results rejects random ordering (p < 0.001), ruling out H1. - Hypothesis checks with specific planets: An unfavorable-feedback planet failed in all 1000 runs (consistent with H2). A no-feedback planet (random walk) had 0/1000 successes, supporting that chance alone is insufficient over 3 By. An optimally stabilising planet succeeded in all 1000 runs, consistent with H2. - Large ensemble (100,000 planets x 100 runs = 10 million runs): About 8,710 planets (~9%) succeeded at least once; only 1 planet succeeded in all 100 repeats. Success rates spanned the entire spectrum (from 1% to higher values), not just 0% or 100%, indicating partial but not absolute control by intrinsic feedbacks. The overall fraction of successful runs was 0.0145 (1.45%). The distribution of times to climate failure was distinct from expectations under H1 or H2. - Rerun repeatability (“rerunning the tape”): In a sample of 1,000 planets, 15 succeeded on the first run and 10 on the second, with 6 overlapping; extended analysis indicated an average 39% overlap—if a planet succeeded once, there was a 39% chance it would succeed again on a repeat run and a 61% chance it would not. - Example trajectories: Planets exhibited diverse feedback landscapes (single/multiple basins, runaway zones) and responses to perturbations/forcings; some remained habitable for the full 3 By despite large perturbations, others failed quickly due to initial conditions or perturbation-induced basin jumps. - Robustness: Across 37 sensitivity analyses covering a wide set of biological, geophysical, geochemical, and astronomical uncertainties, the central result—outcomes intermediate between H1 and H2—remained unchanged, though absolute success rates varied.

Findings imply that Earth’s unbroken multi-billion-year habitability was likely contingent rather than inevitable. Small differences in perturbation histories (e.g., timing and magnitude of super-eruptions, asteroid impacts) or long-term forcings could have led to climate exit from habitable bounds. Geological evidence of near-catastrophic episodes (e.g., Snowball Earth events) aligns with a system close to boundaries at times. The ensemble perspective suggests that many Earth-like exoplanets, even with similar initial endowments, may fail to remain habitable due to chance sequences of perturbations, explaining why long-term habitability might be rare among numerous rocky worlds. While the model is intentionally simplified, extensive sensitivity tests support the inference that both stabilising mechanisms and stochastic events shape habitability trajectories. Independent models with different structures should be applied to further test these conclusions.

The study introduces a general, parsimonious climate-habitability model and large-ensemble simulation framework showing that Earth’s long-term habitability most plausibly reflects a combination of intrinsic climate-stabilising mechanisms and chance. Results are inconsistent with explanations based solely on either random chance or deterministic stabilisation. This reframes the habitability problem: even planets with some stabilising feedbacks may fail under adverse perturbation sequences, and some planets may succeed partly due to good fortune. The conclusions are robust across extensive sensitivity tests. Future research should develop and compare independent modelling approaches with richer process detail, integrate expanding exoplanet observations, refine constraints on feedback strengths and secular forcings, and better quantify distributions of perturbation histories and required durations for the evolution of intelligence.

The model is highly idealised: a single state variable (surface temperature) represents climate; habitability is thermal only and ignores other environmental constraints (e.g., atmospheric composition, ocean chemistry, radiation). Feedbacks are represented statistically rather than mechanistically; f(T) is linearly interpolated between randomly assigned nodes. The habitable temperature bounds (−10 to +60 °C) and required duration (3 By) are assumed and may not be universal. Long-term forcings are simplified and drawn from a symmetric distribution; actual stars and planets may have biases. Perturbation statistics and magnitudes are approximations. Runs consider the same astronomical location per planet across repeats. These simplifications reduce realism and may omit important dynamics, though sensitivity analyses suggest the main qualitative conclusions are robust.