The Sun heats the surface. Greenhouse gases slow the rate at which that heat escapes to space, so the surface settles warmer. The equation used to deny this is the one that proves it.
What σ is. It sets how much power a perfect emitter radiates per square metre at a given temperature. It is not the Boltzmann constant k = 1.38×10⁻²³ J/K; rather σ is built from k with Planck’s constant h and the speed of light c: σ = 2π⁵k⁴ / (15h³c²). Stefan measured the law in 1879; Boltzmann derived it in 1884.
Why the fourth power. A warming body radiates harder for two reasons at once: it emits more photons per second (their number rises roughly as T³), and each photon carries more energy as the spectrum shifts to shorter wavelengths (Wien’s law), so the average photon energy rises roughly as T. Multiply them: T³ × T = T⁴. Formally it is what you get by integrating Planck’s blackbody curve over all wavelengths. The steepness matters: a small change in temperature shifts the radiated power a great deal.
Reading the stages. Reflection of incoming sunlight appears only at stage 3, and clouds do it, not CO₂. Greenhouse gases let sunlight through and act on the outgoing infrared alone. Values are W/m²; stage 3 figures are approximate global averages.
The reason CO₂ can warm without blocking the Sun is spectral. The two bodies radiate in almost separate parts of the spectrum, and CO₂’s absorption bands fall under one of them, not the other.
The spectral mismatch is the engine. The Sun radiates where CO₂ is transparent; Earth radiates where CO₂ absorbs. CO₂’s dominant band is at 15 µm, with a strong band at 4.3 µm and weak near-infrared bands near 2.0 and 2.7 µm. Curves are normalised to equal height so both are visible.
| Absorbed sunlight | 240 W/m² |
| Surface emission at 288 K, σT24 | 390 W/m² ↑ |
| Measured back-radiation, σT14 | ≈ 333 W/m² ↓ |
| Net surface radiative loss | ≈ 57 W/m² |
| Effective temperature without GHGs | −18°C (255 K) |
| Surface temperature with GHGs | +15°C (288 K) |
| Greenhouse effect | 33°C |
Throughout, T2 > T1, so Q stays positive: net heat always flows surface to atmosphere to space. The second law is never broken. CO₂ being an emitter is not the opposite of insulation; the σT14 term is the atmosphere emitting, and that is what slows the loss.
Comparison is Earth with versus without greenhouse gases, same albedo and rotation, not the Moon, whose extreme cold owes more to having no atmosphere at all and a fortnight-long night. The two-body form treats the atmosphere as a single shell at T1; the real atmosphere radiates from many layers and a flat cartoon does not close the budget exactly. None of that changes the result.