🔬 physics

Coupled atmospHere Interior modeL Intercomparison (CHILI). I. Evolutionary Modelling -- Primordial Magma Oceans of Earth and Venus

The CHILI project presents the first intercomparison of coupled interior-atmosphere evolution codes for Earth and Venus, revealing that while nominal models align with empirical constraints, significant divergences in predicted magma ocean lifetimes and atmospheric compositions arise from sensitivities in volatile partitioning, energy transport, and mantle geodynamics.

Original authors: Harrison Nicholls, Joshua Krissansen-Totton, Tim Lichtenberg, Laura Schaefer, Keiko Hamano, Maxime Maurice, Henri Samuel, Alexandra Papesh, Carlos Ortiz-Quintana, Junellie Perez, Yamila Miguel, Denis

Published 2026-06-24

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Harrison Nicholls, Joshua Krissansen-Totton, Tim Lichtenberg, Laura Schaefer, Keiko Hamano, Maxime Maurice, Henri Samuel, Alexandra Papesh, Carlos Ortiz-Quintana, Junellie Perez, Yamila Miguel, Denis Sergeev, Philipp Baumeister, Spanan Dash, Leoni Janssen, Jonathan Keathley, Alexandre de Larminat, Emmanuel Marcq, Lena Noack, Hugo Pelissard, Bo Peng, Emma Postolec, Ramses Ramirez, Mariana Sastre, Andrea Zorzi

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the early Solar System as a giant, chaotic kitchen where rocky planets like Earth and Venus were just born. They weren't the cool, solid rocks we see today; they were glowing, bubbling balls of molten rock, essentially global "lava oceans."

This paper, titled CHILI (Coupled atmospHere Interior modeL Intercomparison), is like a massive taste-test competition between eight different chefs (computer models). Each chef is trying to simulate how these lava oceans cooled down to become the solid planets we know. The goal? To figure out why Earth became a habitable paradise while Venus turned into a scorching hellscape, even though they started out as twins.

Here is the breakdown of what the paper found, using simple analogies:

1. The Great Simulation Contest

The researchers didn't just use one computer program. They gathered eight different "recipes" (models) from scientists around the world. Each recipe had slightly different ingredients and cooking instructions:

Some chefs focused heavily on how the atmosphere traps heat (the "blanket").
Others focused on how the molten rock moves and swirls inside the planet (the "convection").
Some assumed the planet lost gas to space quickly, while others thought it held onto its gases tightly.

They ran these recipes for both Earth and Venus to see how long the "lava oceans" would last before turning into solid crust.

2. The Results: A Tale of Two Planets

For Earth:
Most of the chefs agreed that Earth's lava ocean cooled down relatively quickly. They predicted it solidified within 4 million years. This fits well with what we know from ancient rocks and minerals found on Earth today. It's like a pot of soup that cools down fast enough to form a skin on top, allowing water to eventually condense and life to begin.

For Venus:
The results for Venus were much more chaotic. While Earth's chefs mostly agreed, the Venus chefs argued wildly.

Some models said Venus cooled quickly, similar to Earth.
Others said Venus got stuck in a "stuck-on-low" mode. Because Venus is closer to the Sun, the heat from the star kept the lava ocean boiling for a very long time—up to 50 million years in some simulations.
Think of it like a car engine: Earth's engine cooled down and stopped running. Venus's engine kept revving because the heat from the Sun (the "traffic") kept it from cooling off, trapping the planet in a molten state for much longer.

3. Why Did the Chefs Disagree?

The paper found that the differences in the results didn't come from the amount of "ingredients" (like hydrogen or carbon) the planets started with. Instead, the disagreements came from how the chefs calculated the physics.

Here are the main sources of confusion:

The Atmosphere Blanket: Some models thought the atmosphere was a thick, heavy wool blanket that trapped heat perfectly, keeping the lava hot for eons. Others thought it was a thin sheet that let heat escape easily.
The Gas Mix: Did the lava release mostly water vapor, or mostly carbon dioxide? Some models said the lava released a mix that created a super-greenhouse effect, while others said it released gases that cooled the planet faster.
The "Mushy" Phase: When rock is cooling, it doesn't go from liquid to solid instantly. It goes through a "mushy" stage (like chocolate that is half-melted). Different models handled this mushy stage differently, which changed how fast the planet could lose its heat.
The Escape Route: Some models assumed the planet was losing its atmosphere to space (like a balloon slowly deflating), while others assumed it kept everything. This changed how much heat was trapped inside.

4. The Big Takeaway

The paper concludes that we don't have a single, perfect recipe yet.

While all the models agree that Earth and Venus started as molten balls and eventually cooled, the speed and the exact path they took depend heavily on which computer code you use. The differences in the code's math are currently more important than the differences in the planets' starting ingredients.

Why does this matter?
If we want to understand other planets orbiting distant stars (exoplanets), we need to know which "chef" is telling the truth. If we use the wrong recipe, we might think a distant planet is a habitable ocean world when it's actually a molten lava ball, or vice versa.

The authors say that to solve this, we need to test these computer models against real-world data (like ancient rocks from Earth and future missions to Venus) to figure out which physics rules are the most accurate. Until then, our understanding of how rocky planets evolve is a bit like trying to bake a cake without knowing exactly how hot the oven really is.

Technical Summary: Coupled atmospHere Interior modeL Intercomparison (CHILI). I. Evolutionary Modelling – Primordial Magma Oceans of Earth and Venus

Problem Statement
Earth and Venus represent divergent evolutionary outcomes from initially molten "magma ocean" states, yet their early thermal and geophysical histories remain poorly constrained. While numerical simulations are the primary tool for reconstructing these histories and interpreting exoplanet observations, existing coupled interior-atmosphere codes employ heterogeneous approximations and simplifications. These discrepancies introduce significant uncertainties in predicting planetary cooling timescales, atmospheric compositions, and volatile budgets. The CHILI project aims to benchmark these models against the "ground truth" of the Solar System's terrestrial planets to identify critical model sensitivities and limitations.

Methodology
This study (Part I of CHILI) presents a controlled intercomparison of eight distinct evolutionary codes: GOOEY, NEONGOOEY, PROTEUS, PACMAN, LINCS, MOAI, PlanAtMO, and an early/late variant of MOAI. The models simulate the coupled evolution of the interior and atmosphere for "nominal" Earth and Venus scenarios.

Initial Conditions: All simulations initialize at a stellar age of 50 Myr with a fully molten state, a fixed core radius (55% of planetary radius), and identical initial volatile inventories: $4.7 \times 10^{20}$ kg of hydrogen (equivalent to three Earth oceans) and $2.73 \times 10^{20}$ kg of carbon.
Parameter Grid: Beyond the nominal cases, the study explores a grid of nine scenarios varying hydrogen (1, 5, 10 oceans) and carbon inventories to assess sensitivity to volatile budgets.
Model Constraints: Simulations neglect clouds and aerosols, enforce a constant 10% Bond albedo, and disable Rayleigh scattering (except in MOAI). Termination criteria vary by model but generally occur when the mantle melt fraction drops below a specific threshold (ranging from 1% to 5%).
Physics Variations: The codes differ in their treatment of radiative transfer (line-by-line vs. correlated-k vs. gray), atmospheric escape (fractionating vs. non-fractionating, energy-limited vs. diffusion-limited), mantle convection (boundary layer theory vs. mixing-length), rheology, and volatile partitioning (equilibrium vs. dynamic melt trapping).

Key Contributions and Results
The intercomparison reveals that while all models qualitatively agree on the rapid solidification of Earth's magma ocean (within $\sim$ 4 Myr for nominal cases), they exhibit substantial quantitative divergence in cooling timescales and atmospheric evolution.

Cooling Timescales and Solidification:
- Earth: Nominal Earth models predict solidification within 4 Myr, consistent with empirical constraints from Hadean zircons. However, cooling timescales are more sensitive to the specific model physics (e.g., mantle viscosity, melting curves) than to the initial volatile inventory.
- Venus: Venus scenarios show greater variance. Several models (LINCS, PROTEUS, PlanAtMO) predict "Type II" behavior, where the planet enters a state of global radiative equilibrium (runaway greenhouse) that sustains a magma ocean for up to 49 Myr until atmospheric escape strips the greenhouse gases. Other models (MOAI, GOOEY) predict direct solidification ("Type I").
Atmospheric Composition and Volatile Partitioning:
- Carbon Speciation: Early atmospheres are predicted to be CO/CO $_2$ dominated. The ratio depends heavily on the mantle's redox state ( $f_{O2}$ ) and surface temperature. Models like PROTEUS (fixed $f_{O2}$ ) predict CO $_2$ -rich atmospheres, while MOAI (dynamic redox) predicts CO-dominated atmospheres.
- Hydrogen Storage: A critical source of inter-model disagreement is the treatment of hydrogen storage in the solid mantle. PACMAN and MOAI incorporate "dynamic melt trapping," sequestering significant hydrogen in the solidified mantle, leading to thinner, CO $_2$ -dominated atmospheres. PROTEUS, which neglects solid-phase storage, predicts massive H $_2$ O-dominated atmospheres (>450 bar).
- Escape: Atmospheric escape is a major point of disagreement. LINCS predicts significant hydrogen loss (Type II behavior), while GOOEY predicts negligible loss. The study finds that while escape drives mantle oxidation in some models, its role in regulating thermal evolution timescales is secondary to internal geodynamics and radiative transfer.
Radiative Transfer and Geodynamics:
- OLR Variance: Outgoing Longwave Radiation (OLR) varies by orders of magnitude between models due to differences in radiative transfer schemes (e.g., inclusion of collision-induced absorption, stratospheric temperature profiles). These differences in OLR-T $_{surf}$ relationships drive the divergent cooling rates more significantly than variations in initial carbon budgets.
- Mantle Rheology: Differences in viscosity parametrization and melting curves (solidus/liquidus) significantly impact the rate of heat transport. PROTEUS, for instance, predicts slower cooling due to its treatment of convective suppression by settling cumulates and specific viscosity laws.

Significance and Claims
The paper asserts that the "zoo" of magma ocean models currently yields divergent evolutionary pathways not solely due to unknown initial conditions, but primarily due to differing physical parametrizations.

Model Sensitivities: The study identifies volatile partitioning (specifically deep hydrogen storage), atmospheric escape efficiency, mantle viscosity, and melting curves as the most critical sensitivities driving model divergence.
Implications for Exoplanets: Because Earth and Venus serve as the only "ground truth" for benchmarking, these discrepancies limit the ability to interpret exoplanet observations (e.g., from JWST). The authors argue that validating these specific physical treatments is a prerequisite for meaningful interpretation of exoplanet data.
Future Directions: The paper concludes that future model improvements must focus on accurate radiative transfer (including continuum absorption), robust parametrization of melt trapping and rheology, and the inclusion of unmodelled physics such as tidal heating and cloud formation. The authors note that these findings set the stage for Part II (TRAPPIST-1 exoplanets) and Part III (static atmosphere intercomparison) of the CHILI project.

The authors maintain a modest stance, acknowledging that no single model is "correct" and that the current intercomparison highlights epistemological uncertainties in the underlying physics rather than providing a definitive solution. The primary value lies in quantifying these uncertainties to guide future observational and theoretical efforts.

1. The Great Simulation Contest

2. The Results: A Tale of Two Planets

3. Why Did the Chefs Disagree?

4. The Big Takeaway

More like this