The basic mechanism
The Earth receives energy from the sun as visible light (plus a bit of ultraviolet and infrared). The surface absorbs this energy. Hot surfaces emit infrared radiation — invisible heat radiation. Some of that infrared escapes back to space, cooling the planet. Some is absorbed by gases in the atmosphere and re-emitted, including downward, warming the surface.
That re-absorption and re-emission of infrared is the greenhouse effect. Without it, Earth would average about −18°C. With it, Earth averages about +15°C. The difference — 33°C — is the whole reason the planet is hospitable to liquid water and the life that depends on it.
Why specific gases
Different gases interact with different wavelengths of light. The way to tell what each gas does is to measure its absorption spectrum — which wavelengths it absorbs and which it transmits.
The atmosphere is mostly nitrogen (78%) and oxygen (21%). These are nearly transparent to both visible light and infrared. They don't contribute to the greenhouse effect at all.
The greenhouse gases — those that absorb infrared — are minor by mass:
- Water vapour (variable, typically 1-4% of air by volume).
- Carbon dioxide (~420 ppm today, was ~280 before industrialization).
- Methane (~1.9 ppm, but ~80× more potent than CO₂ per molecule over 20 years).
- Nitrous oxide (~0.3 ppm).
- Various other trace gases (chlorofluorocarbons, ozone).
These molecules have shapes and bond arrangements that let them vibrate at frequencies in the infrared range. When an infrared photon's frequency matches one of these vibration modes, the molecule absorbs it. A short time later, the molecule re-emits a photon — in a random direction. About half goes upward (toward space), half downward (toward surface).
The result: outgoing infrared is partly recycled. The surface temperature has to rise to push out as much energy as comes in. That higher equilibrium temperature is what we live in.
The actual mechanism
It's worth getting this right. A common simplification: "CO₂ acts like a blanket trapping heat." Useful, but not quite accurate.
A better picture: outgoing infrared from Earth's surface can take two paths:
Direct escape. If the wavelength is in a "window" where no atmospheric gases absorb it, the photon zips through to space and is lost. The atmosphere is transparent in some bands.
Absorbed and re-emitted. In the bands where greenhouse gases absorb, the photon is absorbed within a kilometer or so. It's later re-emitted — but from a higher, colder atmospheric layer.
The total infrared emission to space is set by the temperature of the layer that's actually emitting. Adding CO₂ raises the effective emission altitude — the photons end up being released from a slightly higher, slightly colder layer. Colder layers emit less. To maintain the energy balance (incoming = outgoing), the surface and lower atmosphere have to warm up.
This is the technically correct picture. The naive "blanket" analogy gets the right answer but not the right mechanism.
How we know it's real
The greenhouse effect has been demonstrated in many ways:
Lab experiments (Tyndall, 1859). John Tyndall measured the infrared absorption of various gases in glass tubes. CO₂, water vapour, methane all absorbed strongly; nitrogen and oxygen barely did. The exact physics behind today's climate science was published 165 years ago.
Spectroscopy. Satellites looking down at Earth see the infrared spectrum emerging from the atmosphere. Specific wavelength bands are demonstrably dimmer (less energy escaping) precisely at the absorption bands of CO₂, water vapour, methane, etc. This has been measured since the 1970s.
Surface-based measurements. Detectors at the surface measure the downward infrared radiation. These show clear signatures of CO₂ and water vapour re-emission, increasing measurably over decades as those gases have increased.
Planetary comparisons. Mars has very little CO₂ and is brutally cold despite getting only 40% less sunlight than Earth (would be ~−5°C with no atmosphere, is actually ~−60°C). Venus has 96.5% CO₂ in a much thicker atmosphere and is over 460°C at the surface despite a comparable distance from the sun. The greenhouse effect explains both extremes.
Geological records. Ancient climates correlate with reconstructed atmospheric CO₂. When CO₂ was higher (millions of years ago, during certain warm periods), temperatures were higher. When it was lower (during ice ages), temperatures were lower. The relationship has been consistent through the Phanerozoic eon.
Why is CO₂ singled out for climate change?
If water vapour contributes more to the greenhouse effect (about 50%) than CO₂ (about 20%), why do we focus on CO₂?
The answer is persistence and forcing vs feedback:
- CO₂ stays in the atmosphere for centuries to millennia. It takes a long time to be removed by natural processes. Adding more CO₂ has a long-lasting effect.
- Water vapour reaches equilibrium with temperature in days. Warmer atmosphere holds more water vapour — but only because the temperature got warmer first. Water vapour amplifies any warming, but doesn't cause warming on its own.
So CO₂ is the forcing: a direct cause of warming. Water vapour is the feedback: amplification on top of whatever forcing is happening. Both are real, but they play different roles, and CO₂ is what humans are changing directly.
This is also why methane gets attention despite being even more potent: humans are releasing it in large amounts, and it persists for a decade or two — long enough to drive warming, even though shorter-lived than CO₂.
What "doubling" means
Climate sensitivity is usually defined as: how much does the equilibrium temperature rise when CO₂ is doubled?
Pre-industrial CO₂ was about 280 ppm. Doubling to 560 ppm is somewhere we're on track to reach this century if emissions continue. The best estimates of climate sensitivity put the resulting warming at 2.5 to 4°C — most likely around 3°C.
This is the result of: direct CO₂ forcing (about 1°C worth on its own), plus all the feedbacks (water vapour, ice-albedo, lapse rate, clouds). The feedbacks roughly triple the direct effect.
The uncertainty isn't in the direct CO₂ physics — that's been well-pinned-down for over a century. It's in the feedbacks, particularly the cloud feedback. Models broadly agree on the direction (positive total feedback, amplifying warming) but disagree on the magnitude.
What it doesn't mean
Two common misconceptions to clear up:
It doesn't mean Earth would be "uninhabitable" without intervention. Earth has been warmer than today in the geological past (the Eocene, for instance, was several degrees warmer with no polar ice). The concern is the rate of change — humans are pushing the system faster than ecosystems and species can adapt.
It doesn't mean a warmer planet is uniformly warmer everywhere. Polar regions warm faster than the tropics. Some areas get wetter, some drier. Sea level rises. Weather extremes (heat waves, droughts, intense storms) increase. The averages move; the variations move more.
Want to internalize climate physics further? NerdSip can generate a 5-minute personalized course on greenhouse gases and climate sensitivity.
The takeaway
The greenhouse effect is the partial recycling of outgoing infrared radiation by gases that absorb at infrared wavelengths. It's why Earth averages +15°C instead of −18°C. It works through wavelength-specific absorption and re-emission by molecules like CO₂, water vapour, and methane. We've measured it in the lab, in the field, from satellites, and across planets. Increasing atmospheric CO₂ shifts the energy balance, requiring the surface to warm to maintain it. The basic physics has been understood since the 1850s; the modern climate-change concern is a quantitative question about how much warming, how fast, and what that does to ecosystems, agriculture, and infrastructure.