Climate physics · Settled physics
Light in, heat out, and gases that hold some back.
The Earth is warm because of a neat trick played by the air. Sunlight comes down easily and heats the ground. The ground then tries to send that heat back out to space, but it leaves as a kind of invisible light called infrared, and some gases in the air catch it.
Those gases, mostly carbon dioxide and water vapour, soak up the infrared and glow it back out in all directions, including back down. So some of the heat that was leaving gets turned around and sent to the surface again. That returned warmth is the greenhouse effect, and without it the Earth would be a frozen ball, roughly thirty degrees colder.
Here is the part that matters. Add more of those gases and you catch more of the outgoing heat, so the planet settles at a warmer temperature. That is what burning coal, oil and gas does: it puts extra carbon dioxide into the air, which holds back a little more heat every year.
The name is a bit of a fib, by the way. A real glass greenhouse mostly works by stopping warm air from blowing away. The gases in the sky work by catching infrared. Same result, warmth kept in, but a different mechanism.
Start with a simple rule: in the long run, the energy Earth soaks up from the Sun has to equal the energy it radiates back to space. If more comes in than goes out, the planet warms until the two match again. That balance point is what sets the temperature.
Sunlight arrives mostly as visible light, and the atmosphere is nearly transparent to it, so it reaches the ground and warms it. Everything warm radiates, and at Earth's temperature that radiation comes out as infrared, a longer wavelength than visible light. The catch is that the atmosphere is not transparent to infrared. Certain molecules absorb it strongly.
Which molecules, and why? A gas absorbs infrared only if its molecule can flex or bend in a way that shifts its electrical charge. Carbon dioxide, water vapour, methane and nitrous oxide all can, so they are greenhouse gases. Nitrogen and oxygen, which make up most of the air, are simple two-atom molecules that cannot, so they let infrared straight through. This is why a gas that is a tiny fraction of the atmosphere can still matter enormously.
A greenhouse gas absorbs an outgoing infrared photon and later re-emits one in a random direction. About half heads back down. The surface therefore receives energy from the Sun and from the warm atmosphere above it, and settles hotter than sunlight alone would leave it. This natural effect adds roughly \(33^{\circ}\text{C}\), which is the difference between a habitable planet and a frozen one.
Adding carbon dioxide strengthens this. More gas means the atmosphere holds back a larger share of the outgoing infrared, the balance tips, and the surface warms until enough infrared escapes again to restore it. The extra trapping from a given change in concentration is called the radiative forcing.
Two temperatures, not one. Balance the absorbed sunlight against outgoing radiation for a bare Earth and you get an effective radiating temperature of about \(255\,\text{K}\), roughly \(-18^{\circ}\text{C}\). The actual surface sits near \(288\,\text{K}\), about \(+15^{\circ}\text{C}\). The gap is the greenhouse effect. The key is that the \(255\,\text{K}\) is the temperature of the layer that actually radiates to space, which is high up and cold, not the ground.
Raising the emission level. The clean way to think about added carbon dioxide is this. Space only sees infrared emitted from the altitude where the atmosphere finally becomes thin enough to let it out. Add more gas and that effective emission level moves higher, where the air is colder. A colder layer radiates less, so for a moment less energy leaves than arrives, and the whole column warms until the emission level is back at the temperature needed to balance the books. The surface, tied to that column, ends up hotter.
Why the forcing is logarithmic. The strongest absorption bands of carbon dioxide are already nearly saturated, so each extra molecule adds less than the last. The result is that radiative forcing grows roughly with the logarithm of concentration: each doubling adds about the same amount, close to \(3.7\,\text{W}\,\text{m}^{-2}\), rather than each added part per million adding a fixed slug of heat.
Feedbacks do most of the work. Carbon dioxide alone is a modest lever. What amplifies it is feedback. A warmer atmosphere holds more water vapour, itself a strong greenhouse gas, roughly doubling the direct effect. Melting ice and snow darken the surface so it absorbs more sunlight. Clouds push both ways and remain the biggest uncertainty. The underlying radiative physics is settled and testable; the size of the feedbacks is where the harder science and the projections live.
Related: Entropy & the Second Law · next: Photosynthesis · or go back to all topics.