Earth's Energy Balance: The Sunlight In, Heat Out Ledger

Dividing the Sun by four

Start at the top of the atmosphere with the solar constant: the flux of sunlight crossing a surface held face-on to the Sun at Earth's average distance of one astronomical unit. Satellites like SORCE and TSIS-1 pin it at about 1361 W/m², drifting by only ~1 W/m² over the 11-year sunspot cycle. That is the flux on the sunlit disk. But Earth is not a flat disk — it is a rotating sphere, and it catches sunlight on its cross-sectional area (πR⊕²) while radiating heat from its entire surface (4πR⊕²).

The ratio of those two areas is exactly 4. So to get the sunlight averaged over the whole planet, day and night, you divide the solar constant by four: 1361 ÷ 4 ≈ 340 W/m². This single geometric fact — dividing the Sun by four — is why the tropics broil and the poles freeze, and why the global-average number is so much gentler than the searing 1361 W/m² a spacecraft feels in direct sunlight.

Of that 340 W/m² arriving, not all sticks. About 30% is reflected straight back to space — that fraction is Earth's albedo. Bright clouds and the clear atmosphere bounce back roughly 77 W/m²; the surface (ice, deserts, ocean glint, snow) reflects another ~23 W/m². The remaining ≈240 W/m² is absorbed — some in the atmosphere, most at the surface. That 240 W/m² is the entire budget the planet has to work with, and everything the climate does is an accounting of where it goes.

How the planet pays it back: the ledger of outgoing heat

Energy in must equal energy out, or the temperature runs away. Earth cannot reflect its way to balance — reflection only handles the sunlight it never absorbed. The absorbed 240 W/m² has to be shed as thermal infrared radiation, invisible longwave photons pouring off into the cold of space. When incoming absorbed sunlight equals outgoing longwave radiation (OLR), the books close.

Here is the subtlety that trips everyone up. A 240 W/m² blackbody, by the Stefan–Boltzmann law, radiates at a temperature of about 255 K, or −18 °C — a frozen wasteland. Yet Earth's surface averages a comfortable 288 K (15 °C). The 255 K figure is real, but it is the temperature of the layer that actually radiates to space, which sits several kilometers up in the atmosphere. The surface below is warmer because of what happens in between.

Follow the energy at the surface and you find a much busier ledger than "sunlight in, heat out." The ground absorbs ~163 W/m² of sunlight but radiates a whopping ≈398 W/m² of infrared upward (the observed global-mean surface emission; a uniform 288 K surface alone would give ~390 W/m² by Stefan–Boltzmann, and the real, uneven surface radiates a touch more). It stays warm because the atmosphere rains most of that heat right back down as back radiation, ≈340 W/m². The surface also loses energy without radiation at all:

• Latent heat — evaporating water carries away ≈86 W/m², released again high up when clouds form.
• Sensible heat — warm air and thermals convect away ≈20 W/m².

These flows are large and largely cancel; the tiny leftover that leaks to space is what keeps the whole system in balance.

The greenhouse effect: a blanket, not a mirror

The gap between the 255 K a bare rock would show and the 288 K we enjoy — about 33 K of warmth — is the greenhouse effect, and it is the single reason Earth has liquid oceans instead of a global glacier. The mechanism is often mangled, so here is the honest version.

The atmosphere is nearly transparent to incoming visible sunlight (peak wavelength ~0.5 µm) but opaque to outgoing infrared (peak ~10 µm). Greenhouse gases — water vapor first, then carbon dioxide, methane, nitrous oxide, ozone — absorb the surface's upwelling IR and re-emit it in all directions, including back down. The planet must still radiate 240 W/m² to space to balance the Sun, but because the opaque gases force that final emission to happen from a cold, high altitude, the surface underneath has to be warmer to push enough radiation up through the blanket. Add more greenhouse gas and you raise the effective emission altitude into even colder air; the surface warms further to compensate.

Two persistent misconceptions worth killing:

• It is not a mirror. The greenhouse effect does not reflect sunlight — albedo does that. Greenhouse gases act on the outgoing infrared, slowing heat's escape.
• It is not a real greenhouse. A garden greenhouse warms mostly by stopping convection (trapping warm air behind glass). The atmospheric effect is genuinely radiative — an unfortunately durable misnomer coined in the 19th century.

Without it, Earth's average temperature would be that lethal −18 °C. Venus, with a runaway CO₂ atmosphere, shows the other extreme: a surface near 460 °C, hotter than Mercury's, despite being farther from the Sun.

Reading the books from orbit: the imbalance

For thousands of years the ledger balanced to within a hair — incoming and outgoing matched, and the climate held steady. We now know it no longer does, and we can measure the shortfall directly from space. NASA's CERES instruments (Clouds and the Earth's Radiant Energy System), flying since 1997 on Terra, Aqua, and later satellites, watch reflected sunlight and emitted infrared across the whole globe. They reveal that Earth currently absorbs slightly more than it radiates: a positive energy imbalance of roughly 0.9–1.0 W/m², with the 2005–2019 mean estimated at 0.90 ± 0.15 W/m² and recent years running higher still.

A single watt per square meter over the whole 5.1 × 10¹⁴ m² of Earth's surface adds up to about 5 × 10¹⁴ watts of continuous heating — on the order of 500 terawatts, or roughly 25 times humanity's entire primary power consumption. Where does it go? The bookkeeping is remarkably lopsided:

• ≈89% into the oceans, which have enormous heat capacity and store the surplus mostly in the upper few hundred meters.
• ≈5% warming the land, ~1–2% the atmosphere.
• The remainder melting ice — Arctic sea ice, Greenland, Antarctica, mountain glaciers.

Crucially, the space-based radiation measurements (CERES) and the entirely independent measurements of where the heat went — Argo profiling floats logging ocean warming, satellite altimetry watching the sea level rise from thermal expansion, and gravity missions like GRACE and GRACE-FO weighing melting ice — agree. Two separate ways of reading the ledger converge on the same imbalance. That agreement is one of the sturdiest results in climate science.

Feedbacks: what turns the dial

The imbalance is not a fixed number; it responds to the climate it is warming, through feedbacks that can amplify or damp the change. These are the levers on both sides of the ledger.

On the reflection side (albedo): As the planet warms, bright reflective ice and snow melt, exposing dark ocean and tundra that absorb far more sunlight. This ice-albedo feedback is strongly amplifying and is a big reason the Arctic is warming several times faster than the global average. Clouds cut both ways — low bright clouds cool by reflecting sunlight, high thin cirrus warm by trapping infrared — and the net cloud feedback is the single largest uncertainty in the whole budget.

On the heat-trapping side (greenhouse): The most powerful feedback of all involves water vapor. Warmer air holds more of it (roughly 7% more per degree Celsius, by the Clausius–Clapeyron relation), and water vapor is itself a greenhouse gas — so warming begets more warming. Against these amplifiers stands the great stabilizer: the Planck feedback. Because radiated power grows as the fourth power of temperature (Stefan–Boltzmann), a hotter Earth radiates disproportionately harder, which is what ultimately caps the warming and lets a new balance be reached.

The bottom line, honestly stated: the amplifying feedbacks (water vapor, ice-albedo) outweigh the damping ones, so a given nudge — like doubling CO₂, which directly forces about 4 W/m² — produces more warming than the raw physics alone would give. Exactly how much more (the "climate sensitivity") remains the most-debated number in the field, with a likely range of about 2.5–4 °C per CO₂ doubling.

Two centuries of learning to weigh the planet

The idea that a planet's warmth is a balance sheet is old and hard-won. In the 1820s Joseph Fourier reasoned that Earth should be far colder than it is if it merely balanced sunlight, and speculated the atmosphere was trapping heat — the first inkling of a greenhouse effect. In the 1850s Eunice Newton Foote and, independently, John Tyndall showed in the lab that carbon dioxide and water vapor absorb infrared radiation, giving the effect a physical mechanism. In 1896 Svante Arrhenius did the first quantitative calculation of how much surface temperature would change if atmospheric CO₂ doubled — landing on a value (about 5–6 °C) in the right ballpark of the modern range, over a century ago and by hand.

Measuring the budget from above had to wait for the space age. The first satellite radiation-budget experiments flew on Explorer 7 in 1959; the Earth Radiation Budget Experiment (ERBE) launched in the 1980s; and the modern CERES era began with the Tropical Rainfall Measuring Mission in 1997 and Terra in 1999. In parallel, the Argo float array — reaching its full ~3,000-float deployment by 2007 — finally let us weigh the ocean's heat directly, and GRACE (2002–2017) and GRACE-FO (2018–) let us weigh vanishing ice from orbit by its gravity.

The synthesis diagrams that summarize all of this — the famous Kiehl–Trenberth energy budget (1997, updated by Trenberth, Fasullo, and Stephens through the 2010s) — are now textbook fixtures: a single flowchart of arrows carrying 340 in, 100 reflected, 240 out, and every latent and radiative flux in between. What began as Fourier's puzzle over a too-warm planet is now a ledger read continuously from space, accurate to a fraction of a watt — precise enough to catch a planet quietly running a fever.

The two temperatures of Earth: the bare-rock number physics predicts versus the number the greenhouse effect delivers.

Quantity	Bare planet (no greenhouse gases)	Real Earth
Absorbed sunlight	≈240 W/m²	≈240 W/m²
Radiating temperature to space	255 K (−18 °C)	255 K (−18 °C), but from high in the atmosphere
Surface temperature	255 K (−18 °C)	288 K (15 °C)
Surface upward IR	≈240 W/m²	≈398 W/m²
Liquid oceans?	Mostly frozen	Yes
Difference	—	+33 K of warmth from the atmosphere