Seasonal insolation · CO₂ forcing · Albedo feedback · Ice extent trends
This simulator models Arctic sea-ice dynamics on a simplified polar grid. Each cell can be open ocean or ice, and its state depends on the local energy balance: incoming solar radiation (which varies with season and latitude), reflected radiation (governed by albedo — ice reflects ~60 % of sunlight while open water absorbs ~94 %), and outgoing longwave emission modified by greenhouse forcing. The critical insight is the ice–albedo positive feedback: as ice melts, darker ocean absorbs more heat, accelerating further melt — potentially triggering an abrupt tipping point.
Arctic sea-ice extent has declined by roughly 13 % per decade (September minimum) since satellite observations began in 1979. The ice–albedo feedback is one of the strongest amplifying mechanisms in the climate system. During the Eocene thermal maximum (~50 Ma), CO₂ exceeded 1 000 ppm and the Arctic was ice-free, with crocodiles living above the Arctic Circle. Some models project an ice-free Arctic summer before 2050 under high-emission scenarios.
This simulation models a 40×60 grid of Arctic cells, each holding an ice fraction and a temperature anomaly that evolve month by month under a simplified local energy balance. Absorbed solar energy depends on seasonal insolation and each cell's albedo (ice ≈0.62, open ocean ≈0.06), while outgoing longwave radiation is damped by a logarithmic CO₂-forcing term, ln(CO₂/280)/ln(2). The resulting temperature tendency drives freezing below −2 °C and melting above 0 °C, with simple diffusion smoothing each cell toward its neighbours. Because melting exposes darker, more heat-absorbing ocean, the model reproduces the ice–albedo feedback loop that can tip Arctic ice into runaway retreat.
A polar grid where each of the 2,400 cells tracks its own ice fraction (0–1) and temperature. Ice-covered cells appear white-blue, open ocean is shown by temperature-tinted dark blue. An optional line chart plots total ice extent (fraction of the grid still frozen) across simulated decades, making feedback tipping points visible.
Drag the CO₂ Level slider (280–1200 ppm) to change greenhouse forcing, and Solar Multiplier (80–120%) to mimic Milankovitch-style insolation changes. Sim Speed (1×–10×) sets how many simulated years pass per second. Toggle Grid overlay to see individual cells, and Extent chart to show/hide the ice-extent history plot. Pause/Reset restart the run from a fresh randomised ice distribution.
Real satellite records since 1979 show September Arctic sea-ice minimum extent declining by roughly 13% per decade — a trend this simulation's feedback loop qualitatively reproduces. During the Eocene thermal maximum, atmospheric CO₂ exceeded 1,000 ppm and the Arctic was ice-free enough for crocodiles to live near the pole.
Ice reflects most incoming sunlight (high albedo, around 0.6), while open ocean absorbs most of it (low albedo, around 0.06). When ice melts, the darker ocean surface absorbs more solar energy, warming the water further and melting more ice. This self-reinforcing loop is one of the strongest amplifiers in the Arctic climate system, and it is exactly what the simulation's per-cell albedo calculation reproduces.
The CO₂ slider feeds into a greenhouse forcing factor computed as 1 + 0.7·log₂(CO₂/280). Higher CO₂ increases this factor, which reduces the modelled outgoing longwave radiation from each cell. Since each cell's temperature tendency depends on absorbed solar energy minus outgoing radiation, raising CO₂ tips more cells toward net warming, favouring melt over freeze.
Each row represents a latitude band, with row 0 modelling the pole and the last row modelling the edge of the domain near the equator. Initial ice probability scales with latitude raised to a power (lat^1.2), so cells nearer the pole start with a higher chance of being ice. Seasonal insolation is also weighted by latitude, so polar cells receive less average sunlight and tend to stay colder.
Each cell has its own temperature anomaly, updated from the local energy balance every simulated month. If a cell's temperature drops below −2 °C, its ice fraction increases; if it rises above 0 °C, ice fraction decreases. Between those thresholds the cell's ice cover is roughly stable, which is why a small band of temperatures corresponds to the marginal ice zone where extent is most sensitive to forcing.
It is a simplified conceptual model, not a general circulation model. It captures the core mechanisms of seasonal insolation, albedo feedback and greenhouse forcing on a coarse 40×60 grid with basic neighbour diffusion, which is enough to reproduce qualitative behaviours like feedback amplification and tipping points. It omits ocean heat transport, cloud feedbacks, ice thickness and dynamics, so absolute numbers should be read as illustrative rather than predictive.