← Gregory MundayA small planet, a bigger questionTerrain lab ↗

Reflect.

Turn a planet. Change its light. Follow its climate.

One orbit,
thousands of small worlds.

BUILDING YOUR PLANET
Earth coastlines · Jin ocean / C45 land
Preparing spherical cells…Year 1 · January
0.00 model yearsDaily physics · one turn / model year

The planet’s response

Warming vs reference+0.00 °CSensitivity scenarios: 0.00 to 0.00°C
Imbalance change vs reference+0.00 W/m²CO₂ forcing: +0.00 W/m²
Global mean temperature— °CSensitivity scenarios: 14.00 to 14.00°CHeat change: 0.00 ZJ
ECS 2.5–4°C / doubling · central 3°C

Preparing the reference orbit. All calculations stay in your browser.

A small piece of the planet

Choose Inspect and click the globe, or enter coordinates below.

Blank albedo uses the supplied C45 land / Jin ocean scheme. The multiplier scales the global scattering dial; it is not a cloud fraction. Edits retain the experiment’s temperature and history.

A small model with explicit assumptions

Sunlight on a sphere

The reduced latitude–longitude grid has about 100 km spacing, with fewer cells towards each pole. Every mean and power balance uses each cell’s spherical area. Coastlines use the public-domain Natural Earth land mask, sampled at half-degree resolution. Vegetation, snow and sea ice are illustrative prescribed fields; small islands and inland waters are simplified.

Kepler’s equation determines the orbit, its changing speed, and solar distance. Incoming radiation varies as 1361/r² W/m²; axial tilt sets solar declination. Daily mean sunlight includes polar day and night. Perihelion is near 3 January; its solar longitude is held at 282.94°. The calendar starts in today’s season. The globe makes one decorative turn per model year (10 seconds at default speed); physical insolation still averages a real 24-hour day. Display lighting is illustrative.

Learned land retains the supplied C45 coefficients, normalisation and 16-point Gaussian quadrature. Eight-point daylight quadrature averages its black-sky kernels; white-sky kernels determine diffuse reflection. Snow, ice and physical C45 inputs are prescribed, with no automatic climate feedback.

Open ocean uses Jin et al. (2011), ported from SpeedyWeather’s gm/albedo branch: Fresnel reflection, the roughness correction, clear-sky diffuse reflection, volume scattering and wind-driven foam. Cox–Munk roughness is diagnosed from the global 10 m wind dial. Daylight quadrature averages direct ocean albedo; sea ice mixes with albedo 0.6. The final mixture is bounded to 0–1.

Jin et al. (2011) ↗
Pinned SpeedyWeather implementation ↗
Natural Earth data ↗
Seasonal insolation reference ↗

Scattering, without clouds

The global dial scales a smooth, heterogeneous per-tile scattering field. Each cell divides sunlight into a direct beam and a scattered contribution. The scattered field spreads with a spherical heat kernel, a Gaussian approximation with the selected width.

Pairwise equal and opposite power transfers conserve total incoming radiation, wrap around longitude, and cross polar rows. Positive stable substeps prevent negative light. Scattering changes local direct/diffuse albedo and where the light arrives. It does not add heat diffusion, clouds, atmospheric absorption or backscatter to space.

Brushes apply a circular Gaussian falloff. Albedo edits prescribe reflectivity rather than changing heat capacity or vegetation. Inspect any tile to edit its albedo and scattering multiplier precisely.

FaIR’s three-layer temperature response

The global temperature response uses the deterministic three-layer thermal equations from FaIR. Surface, upper-ocean and deep-ocean temperatures exchange heat. The forcing is 5.35 ln(C/280) plus the area-weighted change in absorbed sunlight from your surface, scattering and orbital edits. Daily matrix-exponential updates integrate these equations without importing Python or the full FaIR package.

ECS scenarios are 2.5, 3 and 4°C per doubling, using the IPCC AR6 assessed likely range and best estimate. Heat capacities and ocean exchanges are illustrative FaIR example parameters with conservative exchange; this is not a calibrated FaIR posterior ensemble. The shaded range spans sensitivity scenarios, not a probabilistic 17th–83rd percentile transient forecast.

Playback starts automatically at the dated NOAA CO₂ estimate shown beside the dial. All three thermal layers begin in equilibrium with that concentration; the reference seasonal cycle is initialised periodically. This is an idealised present-CO₂ equilibrium, not a historical spin-up or an observed present-day temperature. The absolute baseline is prescribed.

The unchanged 280 ppm reference retains the original Earth-like orbit, surfaces, wind and scattering. Its prescribed baseline averages 14°C. The main imbalance measures the experiment’s change relative to that reference. Total imbalance and integrated heat also include seasonal storage. Regional temperature patterns use a separate approximate local response adjusted to match the FaIR global mean.

Surface fields stay prescribed: no interactive clouds, circulation, ice feedback or carbon cycle. Edits preserve the ongoing temperatures and history. This lightweight model is intended for exploration and future temperature-goal gameplay.

FaIR energy balance model ↗
IPCC AR6 sensitivity assessment ↗
Logarithmic CO₂ forcing ↗