Greenhouse Effect & Earth’s Energy Balance · Guide · glidepathscience.org
GuideGreenhouse Effect & Earth’s Energy Balance
Contents
This Guide explains the ideas and equations behind every Walkthrough step, Concept Challenge question and Predict round. Part 1 prepares the Intro track and Part 2 the Advanced track; each takes about ten minutes. The worked examples use Mars and other cases that are not the sim’s own problems.
Part 1 · The essentials
Math you’ll need: algebra, powers and fourth roots.
A planet settles where energy in equals energy out
A planet gains energy one way, by absorbing sunlight, and loses it one way, by radiating infrared to space. Nothing else crosses the vacuum around it. If it absorbs more than it radiates, it warms; a warmer planet radiates more; and it settles at the temperature where the two are equal. Every temperature in this sim, with or without an atmosphere, is found by writing down that balance and solving it.
The sim draws one square meter of the planet’s surface, with a column of atmosphere above it, averaged over the whole globe and a whole day. Each arrow carries energy, and is as wide as the watts per square meter it carries.
Sunlight in: S, the factor of four and the albedo
The Sun radiates a power $L$ = 3.828 × 10²⁶ W in all directions. At a distance $d$ it is spread over a sphere of area $4πd^2$, so each square meter facing the Sun receives
$S = L/(4πd^2)$
At Earth this solar constant is about 1361 W/m² (measured by satellites; Kopp and Lean, 2011). Farther out it falls as $1/d^2$.
A planet catches sunlight on a disk, the shadow it casts, of area $πR^2$. But it spins, and the energy is shared by its whole surface, $4πR^2$. Averaged over day and night, each square meter receives $S/4$, about 340 W/m² at Earth. Half of the four is night; the other half is that most of the day side meets sunlight at a slant.
Some of that is reflected straight back by clouds, ice, snow and bright ground. The fraction reflected is the albedo $α$, about 0.30 for Earth. Reflected light never heats anything, so the planet absorbs
$F_{in} = (1 − α)S/4$
Infrared out, and the bare-rock temperature
Any surface glows. A blackbody at temperature $T$ (in kelvins) radiates $σT^4$ from each square meter, with $σ$ = 5.67 × 10⁻⁸ W/(m²·K⁴), the Stefan–Boltzmann law from Blackbody Radiation. Because of the fourth power, doubling the power radiated raises the temperature only by $2^{1/4}$ = 1.19.
With no atmosphere, the surface radiates straight to space, and the balance is
$σT_e^4 = F_{in}$, so $T_e = (F_{in}/σ)^{1/4}$
For Earth, $F_{in}$ = 238 W/m² gives $T_e$ ≈ 255 K, about −18 °C. Earth’s mean surface temperature is about 288 K (15 °C): some 33 K warmer than a bare rock would be. That difference is the greenhouse effect. Seen from space, though, Earth still sends out about 239 W/m², the glow of a 255 K body, so $T_e$ is called the effective emission temperature.
Two wavelengths: visible in, infrared out
Wien’s law puts a blackbody’s peak at $λ_{max} = b/T$, with $b$ = 2.898 × 10⁻³ m·K. The Sun’s surface, at 5772 K, peaks near 0.50 μm, in visible light; Earth, at about 255 K, near 11 μm, in the thermal infrared. The two spectra barely overlap. A gas can therefore be nearly clear to sunlight and still absorb the planet’s own glow, and water vapor and carbon dioxide do just that.
One layer that absorbs infrared
Add one layer of atmosphere, clear to sunlight but opaque to infrared. It absorbs all of the surface’s infrared, warms, and in balance radiates at its own temperature, $σT_1^4$, from both faces: up to space and down to the surface. The downward part is the back radiation. The layer does not trap energy for good; it passes it on, half each way.
Write the balances. At the top, the planet must still send $F_{in}$ to space, and with an opaque layer all of it comes from the layer: $σT_1^4 = F_{in}$, so the layer sits at $T_e$. The surface absorbs sunlight plus back radiation, $F_{in} + σT_1^4 = 2F_{in}$, and must radiate that much:
$σT_s^4 = 2F_{in}$, so $T_s = 2^{1/4}T_e$
For Earth that is 303 K, now too warm. Stacking $N$ opaque layers repeats the argument: each adds one more $F_{in}$, and $T_s = (N + 1)^{1/4}T_e$.
A leaky layer, and a change of albedo
A real atmosphere lets some infrared through. A layer that absorbs a fraction $ε$ of the infrared reaching it also emits only $εσT^4$ from each face (a good absorber is an equally good emitter). Solving its balances gives
$T_s = T_e[2/(2 − ε)]^{1/4}$
which runs from $T_e$ at $ε$ = 0 (bare rock) to $2^{1/4}T_e$ at $ε$ = 1. With $ε$ ≈ 0.78, Earth’s surface comes out at 288 K: one leaky layer is a fair cartoon of the real infrared blanket.
Three things set the temperature a planet settles at: the sunlight reaching it ($S$), the fraction it reflects ($α$), and how strongly its atmosphere absorbs and re-radiates infrared ($ε$, $N$). Raising the albedo lowers $F_{in}$ and with it every temperature in the column.
Part 2 · Going further
Math you’ll need: Part 1, a small linear system, a derivative and a first-order differential equation.
The emission temperature and the emission height
In balance the outgoing longwave radiation (OLR) equals $F_{in}$, so $T_e = (\mathrm{OLR}/σ)^{1/4}$ is fixed by the absorbed sunlight alone, whatever the atmosphere does inside. IPCC AR6’s global energy budget (Chapter 7, Fig. 7.2) has about 340 W/m² arriving, 100 W/m² reflected and 239 W/m² leaving as infrared, with the surface emitting about 398 W/m² and receiving about 342 W/m² of back radiation.
In the real troposphere temperature falls with height at the lapse rate $Γ$ ≈ 6.5 K/km. The air that space “sees” is the air at $T_e$, at a height $z_e = (T_s − T_e)/Γ$, about 5 km. This gives a second picture of the greenhouse effect: more absorber raises the emission level into colder air, the planet radiates less, and the column warms until it radiates enough again.
For one gray layer the absorptivity that gives a surface temperature $T_s$ is $ε = 2[1 − (T_e/T_s)^4]$; the layer itself then sits at $2^{−1/4}T_s$.
N gray layers, Kirchhoff’s law and a window
Kirchhoff’s law: at each wavelength a body in equilibrium emits exactly as well as it absorbs. A gray layer of absorptivity $ε$ therefore emits $εσT_i^4$ from each face. With $N$ such layers, a fraction $w$ of the surface’s infrared passing every layer untouched (the window, near 8–12 μm), and fluxes measured as $x_i = σT_i^4$, every balance is linear in the $x_i$: each layer absorbs $ε$ of the band flux reaching it from below and from above and emits $ε(1 − w)x_i$ each way. The engine solves these $N + 1$ equations directly and checks that energy closes at every level.
Three results are worth knowing. Opaque layers give $T_s = (N + 1)^{1/4}T_e$, which climbs fast and overestimates any real atmosphere. Leaky layers add slowly, because transmissions multiply: two layers of $ε$ = 0.5 pass $(1 − 0.5)^2$ = 25 % of the surface’s infrared, so they warm the surface less than one opaque layer. A window lets the surface cool directly to space; with one layer opaque outside it,
$T_s = T_e[2/(1 + w)]^{1/4}$ (window)
Radiative forcing
Add absorber and, before anything warms, the outgoing infrared falls: colder air higher up now stands in for more of the surface’s glow. That drop in net flux at the top, with every temperature held, is the radiative forcing $ΔF$. For one gray layer, OLR $= (1 − ε)σT_s^4 + εσT_1^4$, so raising $ε$ by $Δε$ gives $ΔF = Δε\,σ(T_s^4 − T_1^4)$, about 3.9 W/m² for $ε$ from 0.78 to 0.80.
CO₂’s strongest band is already opaque at its center, so more CO₂ mainly widens it, and the forcing grows as a logarithm. The simplified expression of Myhre et al. (1998), used in the IPCC’s third and fifth assessments, is
$ΔF = 5.35\ln(C/C_0)$
giving 3.71 W/m² for doubled CO₂. AR6 assesses the effective radiative forcing of doubled CO₂ at 3.93 ± 0.47 W/m², which includes the atmosphere’s fast adjustments.
The response: the Planck feedback, other feedbacks and the time to settle
A forcing leaves the planet taking in more than it sends out. As it warms by $ΔT$ its outgoing flux rises by about $λΔT$, until the books balance again:
$ΔT = ΔF/λ$
If only the temperature changed, $λ$ would be the Planck response, the derivative of $σT^4$ at the emission temperature: $λ_0 = 4σT_e^3$, about 3.7 W/(m²·K) in this model (AR6 assesses 3.22 W/(m²·K) for the real atmosphere’s vertical structure). Warming also changes the atmosphere: more water vapor absorbs more infrared, less ice reflects less sunlight, and clouds shift. AR6 assesses these at about +1.30 (water vapor and lapse rate together), +0.35 (surface albedo) and +0.42 (clouds) W/(m²·K), which offset part of the Planck response and leave a net $λ$ ≈ 1.16 W/(m²·K) (very likely 0.51 to 1.81). A smaller $λ$ means a larger response; AR6’s best estimate of the equilibrium warming for doubled CO₂ is 3 K, very likely 2 to 5 K.
The warming is not instant. With an ocean mixed layer of depth $h$ and heat capacity $C = ρc_ph$ per square meter, $C\,dT/dt = ΔF − λΔT$, so $ΔT$ approaches $ΔF/λ$ as $1 − e^{−t/τ}$ with
$τ = C/λ$
What the layer model leaves out
Radiation alone would make the lower atmosphere’s temperature fall with height faster than air can stably sustain. Instead warm, moist air rises: convection and the latent heat of evaporation carry about a quarter of the surface’s loss upward and hold the profile near the lapse rate. Radiative–convective models add exactly that to the layers. Also left out: sunlight absorbed in the atmosphere, clouds’ own physics, the spectrum beyond one window, the deep ocean, which slows the approach to balance over centuries, and geography, seasons, day and night.
Reference
Common mistakes
Part 1
- Using $S$ for the average sunlight: a sphere gets $S/4$.
- Taking the reflected part, $αS/4$, as the absorbed part.
- Working in degrees Celsius in $σT^4$: use kelvins.
- Doubling the temperature when the power doubles: $T$ goes as the fourth root.
- Saying the layer “traps” heat for good: it radiates all it absorbs, half up, half down.
- Expecting the outgoing infrared to stay low after a new balance: it returns to $F_{in}$.
Part 2
- Adding absorptivities of layers: transmissions multiply.
- Using log base 10 in $5.35\ln(C/C_0)$.
- Thinking a deeper ocean changes the final warming: it changes only $τ$.
The equation sheet
| Part 1 | |
|---|---|
| solar constant | $S = L/(4πd^2)$ |
| absorbed | $F_{in} = (1 − α)S/4$ |
| bare rock | $σT_e^4 = F_{in}$ |
| Wien’s law | $λ_{max} = b/T$ |
| one opaque layer | $T_s = 2^{1/4}T_e$ |
| one gray layer | $T_s = T_e[2/(2 − ε)]^{1/4}$ |
| N opaque layers | $T_s = (N + 1)^{1/4}T_e$ |
| Part 2 | |
|---|---|
| fit one layer | $ε = 2[1 − (T_e/T_s)^4]$ |
| window | $T_s = T_e[2/(1 + w)]^{1/4}$ |
| CO₂ forcing | $ΔF = 5.35\ln(C/C_0)$ |
| Planck response | $λ_0 = 4σT_e^3$ |
| response | $ΔT = ΔF/λ$ |
| time to settle | $τ = C/λ$, $C = ρc_ph$ |
$S$ solar constant (W/m²), $L$ the Sun’s luminosity, $d$ distance, $α$ albedo, $F_{in}$ absorbed sunlight per m² of surface, $σ$ = 5.670 × 10⁻⁸ W/(m²·K⁴), $T_e$ emission temperature, $T_s$ surface temperature, $b$ = 2.898 × 10⁻³ m·K, $ε$ a layer’s absorptivity (and emissivity), $N$ number of layers, $w$ window fraction, $C/C_0$ CO₂ concentration ratio, $λ$ feedback parameter (W/(m²·K)), $C$ heat capacity per m², $ρ$ and $c_p$ seawater’s density and specific heat, $h$ mixed-layer depth, $τ$ time constant.
Further reading
- IPCC, 2021: Climate Change 2021: The Physical Science Basis, Working Group I, Chapter 7, “The Earth’s Energy Budget, Climate Feedbacks and Climate Sensitivity” (Forster et al.): the energy budget (Fig. 7.2), forcing, feedbacks (Table 7.10) and climate sensitivity.
- G. Myhre, E. J. Highwood, K. P. Shine and F. Stordal, “New estimates of radiative forcing due to well mixed greenhouse gases,” Geophysical Research Letters 25, 2715–2718 (1998).
- G. Kopp and J. L. Lean, “A new, lower value of total solar irradiance: Evidence and climate significance,” Geophysical Research Letters 38, L01706 (2011).
- This sim’s Field Manual (H): the model and what it leaves out.
Greenhouse Effect & Earth’s Energy Balance v1.1 · © 2026 Dr. David Meier · CC BY-NC-SA 4.0 · glidepathscience.org