A full planet climate model from Blue Marble Space and the University of Colorado Boulder finds that an old assumption about how fast rock weathering removes carbon dioxide was too pessimistic, extending the habitability window later in the Sun's life.
Earth's habitability deadline just moved, not because the physics changed, but because a better computer model says the old assumption about how fast rock weathering pulls carbon dioxide out of the air was too pessimistic.
A new study by Jacob Haqq-Misra of the Blue Marble Space Science Institute and Eric Wolf of the University of Colorado Boulder runs a full three-dimensional Earth-system simulation across two scenarios: one that pins global temperature to today's climate, and one with an active carbon cycle. The result, as Ars Technica reports, pushes the window in which Earth stays friendly to complex life roughly a billion years later than older one-dimensional models predicted.
The endpoint is unchanged. The Sun is still on track to swell into a red giant and engulf the inner solar system in about five billion years. What shifted is the schedule for the long, slow warming that precedes that finale.
For decades, the standard habitability calculation has been a one-dimensional column model: a single averaged stack of ocean and atmosphere that tracks how the Sun's rising brightness heats the planet over time. Useful, but coarse. The new model simulates the full planet in three dimensions, which lets it capture how heat, rain, and rock weathering vary by latitude and season rather than averaging them into a single number.
That distinction matters because the slow thermostat on Earth is not the Sun. It is the carbonate-silicate cycle. Rain erodes silicate rock, pulling CO2 out of the air and locking it into carbonate that rides ocean crust down to the mantle. Volcanoes eventually return that carbon to the atmosphere. On long timescales, the strength of that feedback, how much faster rocks weather when temperatures rise, sets how much CO2 the planet can hold before everything overheats.
Older 1D models assumed a tight coupling between temperature and weathering. Haqq-Misra and Wolf tested a weaker coupling, in line with recent geochemical evidence, and let the 3D climate physics decide what follows. With that assumption, the weathering feedback does less to stabilize the climate, but the climate itself stays cooler than the 1D model predicted at any given solar brightness. The net effect is more time before oceans boil and the surface sterilizes.
The authors are careful to flag what this does and does not show. The new timeline is a model output, not a measurement. It rests on a single physical assumption, the temperature sensitivity of silicate weathering, that the literature is still debating. Run the model with a stronger coupling and the schedule snaps back closer to the older estimates. The "extra time" headline is a statement about a specific pair of scenarios, not a revised fact about the planet.
It is also worth noting what the 1D models got right. They were the best available simplification for an era without the computing power to run a full Earth-system model to deep time. The new paper does not dismiss that lineage; it builds on it. Where it differs is in the resolution, and in the willingness to let a single, well-motivated assumption, weaker weathering feedback, propagate through a 3D system to its conclusion.
The practical upshot is less about Earth's future, which remains terminal, and more about how habitability forecasts are made at all. The lesson is that long-range predictions are revisable when the underlying assumptions shift, and that the assumptions are where the science is most active. For anyone who studies exoplanet climates, that is the headline: a billion-year habitability window is not a fixed number, it is a function of what you plug in.