CO₂ makes up approximately 0.042% of the atmosphere. Less than half of half of half of nothing. And it turns out that’s enough to rewrite the planet’s climate history. If that doesn’t strike you as strange, you haven’t thought about it properly.
Before getting into the physics, there’s a current piece of data that deserves a moment’s attention: in May 2026, the concentration of CO₂ in the atmosphere exceeded 432 parts per million (ppm) for the first time at its seasonal peak, reaching an average of 432.00 ppm according to Scripps and 432.3 ppm according to NOAA, in measurements from the Mauna Loa Observatory [1]. Ralph Keeling, director of the Scripps CO₂ programme, lamented the lack of progress upon facing yet another historic high: “Atmospheric CO₂ has continued its relentless rise over the past year, reaching another all-time record and pushing us further into a world of high CO₂ concentrations. I wish I had better news” [1]. Before industrialisation, the level was approximately 280 ppm. We have added more than 152 ppm in less than two centuries.
Now comes the physics.
The Earth emits light you can’t see
The Sun sends energy in the form of short-wave radiation: visible light, some ultraviolet. The Earth absorbs it, heats up, and in turn emits energy. But because it’s much cooler than the Sun, it emits that energy as infrared radiation — long-wave, invisible to the human eye, what we colloquially call “heat.” So far, nothing remarkable. The planet receives energy, emits energy, and in equilibrium both flows cancel each other out.
The problem is that some gases in the atmosphere are transparent to incoming solar radiation but opaque to outgoing infrared. They act as a one-way valve: they let incoming energy through but obstruct outgoing energy. That is the greenhouse effect. It’s not a theory or a story. It’s basic physics, measurable in a laboratory since the nineteenth century.
Radiative forcing is the quantitative way of describing this imbalance: how many additional watts per square metre (W/m²) the atmosphere retains relative to pre-industrial equilibrium. According to the IPCC AR6, the total accumulated radiative forcing from 1750 to 2019 is approximately 2.7 W/m² [2]. That might sound like little, but we’re talking about every square metre of the entire planet — at a planetary scale, it’s enormous: equivalent to leaving a small space heater running on every square metre of the Earth’s surface, non-stop, for decades.
The industrial era is generally considered to have begun around 1760 in England, which is why most pollution control measures attempt to use that date as their baseline.
Why CO₂ and not nitrogen
Here comes the part people usually skip — and which changes everything. Nitrogen and oxygen, which together make up 99% of the air, do not absorb infrared radiation. CO₂, methane (CH₄), water vapour, and nitrous oxide (N₂O) do. Why?
It all comes down to how molecules vibrate. To trap heat (infrared radiation), a molecule needs its electric charges to shift position as it moves.
The most abundant gases in the air, such as nitrogen (N₂) or oxygen (O₂), are symmetrical: they vibrate in a balanced way, so heat passes straight through them. CO₂, by contrast, is an elongated molecule that bends and vibrates asymmetrically. In doing so, it creates an electrical imbalance that traps the Earth’s heat at precisely the frequency at which our planet emits it most strongly.
Arrhenius calculated it by hand in 1896
Here’s the fact that always astonishes me. In 1896, more than 120 years ago, Swedish physicist Svante Arrhenius published in the Philosophical Magazine the first quantitative model of the greenhouse effect [6]. No computers. No instruments for directly measuring atmospheric CO₂. Working with infrared emission data from the Moon — provided by astronomer Samuel Langley as a proxy for estimating gas absorption coefficients. He performed between 10,000 and 100,000 calculations by hand [7], computing temperature variations across 10-degree latitude bands, across four seasons, for five different CO₂ levels. He concluded that doubling atmospheric CO₂ would raise global temperature by 5 to 6°C [7].
Current climate models estimate that sensitivity at between 2.5 and 4°C [3]. Arrhenius overshot — he ignored clouds, ocean circulation, and several feedback mechanisms — but NASA acknowledges that his calculations were “surprisingly accurate” given the state of knowledge at the time [8]. One man, pencil in hand, in the nineteenth century, without fully knowing what he was calculating, arrived at the correct order of magnitude for one of the most complex physical phenomena on the planet. It’s an impressive fact, and a slightly unsettling one.
Where Arrhenius did get it wrong was the timescale: he naively assumed it would take millennia to double CO₂. He vastly underestimated the speed at which humans would burn through natural resources [8].
The first person to connect CO₂ with climate was a woman who wasn’t allowed to present her own work
John Tyndall is widely credited as the father of climate science for his 1859 experiments on infrared absorption. Deservedly so, in part. But in 1856 — three years earlier — American scientist Eunice Newton Foote had published an experiment in which she filled glass cylinders with different gases, exposed them to sunlight, and measured their temperature. CO₂ heated up the most. Her conclusion: “An atmosphere of that gas would give to our Earth a high temperature.” [9] Foote published this in the American Journal of Science and Arts. However, it was not she who presented her findings at that year’s annual AAAS conference: physicist Joseph Henry presented them on her behalf, because women were not permitted to present their own work at such forums [9]. Tyndall never mentions her. Whether he knew of her work, he never acknowledged it [10]. History once again demonstrating the barriers women faced in having their work recognised, a product of the prevailing sexism of the era.
Could warming become non-linear at critical tipping points in the climate system?
This is conjecture, but grounded in the following: the radiative forcing of CO₂ is logarithmic — which technically means each additional ppm matters slightly less than the previous one, giving us a degree of breathing room.
However, the climate system is not just CO₂: there are feedbacks — water vapour, ice albedo, permafrost methane — that can amplify warming in a non-linear way once certain thresholds are crossed. For example, as permafrost thaws, it releases enormous quantities of methane, estimated to be 28 times more damaging than CO₂.
Another problem is the planet’s ice surface. Ice reflects 80% of solar heat, but as the poles melt, the reflective surface shrinks — and is replaced by ocean water, which absorbs 90% of solar heat. Precisely the opposite.
Then there is the destruction of ancient forests and wildfires. The world’s great forests — the Amazon, the Siberian taiga — are being damaged to the point of no return. These forests are carbon sinks, absorbing enormous quantities of CO₂, far more than any young forest can. Their destruction also affects the climate directly, triggering droughts and heatwaves that leave forests more vulnerable to fires and disease.
When a forest burns or dies from drought, it stops absorbing CO₂ and simultaneously releases all the CO₂ stored over centuries in its wood in a single event.
The IPCC AR6 estimates an equilibrium climate sensitivity of 2.5 to 4°C for a doubling of CO₂ [3], but the upper end of that range assumes feedbacks that are not yet well quantified. The relevant question is not whether CO₂ causes warming — that has been settled since 1896 — but whether the complete climate system can shift to a different regime in a relatively abrupt way. That’s where the physics of the greenhouse effect reaches its limits, and the non-linear dynamics of complex systems begin.
Conclusion
The physics of the greenhouse effect is not complicated. It’s uncomfortable. A gas that makes up less than 0.05% of the atmosphere modifies the planet’s energy balance because its molecular structure, by a quantum coincidence, absorbs precisely the wavelengths the Earth emits. Arrhenius calculated it by hand 130 years ago. A woman intuited it three years before Tyndall and was never allowed to speak in public about it. And now we’re at 430 ppm and rising at a rate without precedent in the last 800,000 years of ice core records [42].
You don’t need to understand quantum physics or be particularly sharp to find this disturbing.
References (assessed with our reliability rating)
[1] Scripps Institution of Oceanography / NOAA Global Monitoring Laboratory, “Annual Carbon Dioxide Peak Passes Another Milestone”, June 2025. Scripps UCSD & NOAA GML. Reliable
[2] IPCC Sixth Assessment Report (AR6), Working Group I, Technical Summary, 2021. Chapter 7: Earth’s Energy Budget, Climate Feedbacks and Climate Sensitivity. Reliable
[3] IPCC AR6, WGI, Chapter 7: equilibrium radiative forcing for a doubling of CO₂ estimated at 3.71 W/m²; equilibrium climate sensitivity (ECS) assessed at between 2.5 and 4°C. Reliable
[4] Jeevanjee, N. et al., “A Simple Spectroscopic Model for CO₂ Radiative Forcing”, Journal of Climate, 2021; Romps, D. et al., 2022. Discussed in: “Physicists Pinpoint the Quantum Origin of the Greenhouse Effect”, Quanta Magazine, August 2024. With reservations
[5] Wordsworth, R. et al., “Fermi Resonance and the Quantum Mechanical Basis of Global Warming”, The Planetary Science Journal, published March 2024. arXiv:2401.15177. With reservations
[6] Arrhenius, S., “On the Influence of Carbonic Acid in the Air upon the Temperature of the Ground”, Philosophical Magazine and Journal of Science, Series 5, Vol. 41, April 1896, pp. 237–276. Reliable
[7] Crawford, E., “Arrhenius’ 1896 Model of the Greenhouse Effect in Context”, Ambio, 1997. Also: The Climate Historian, Substack, January 2024. With reservations
[8] NASA Earth Observatory, “An Introduction to Climate Modeling”. nasa.gov/features/ModelingIntro. With reservations
[9] Foote, E., “Circumstances affecting the heat of the Sun’s rays”, The American Journal of Science and Arts, 2nd Series, Vol. XXII, November 1856, pp. 382–383. Reviewed in NOAA Climate.gov, “Happy 200th birthday to Eunice Foote”, 2019. Reliable
[10] Jackson, R., “Eunice Foote, John Tyndall and a question of priority”, Notes and Records: The Royal Society Journal of the History of Science, 2019. doi:10.1098/rsnr.2018.0066. Reliable
[4] and [5] are marked with reservations not due to methodological concerns but because they are recent results not yet extensively replicated, though published in solid journals. [7] carries reservations due to the inclusion of a Substack source without peer review.