This simulation reconstructs roughly 800,000 years of climate history as recorded in deep polar ice cores such as Vostok and EPICA Dome C. A synthetic signal is built from three Milankovitch orbital components and converted to a δ¹⁸O temperature proxy using the relationship Δδ¹⁸O ≈ 0.7·ΔT, with a paired atmospheric CO₂ curve. The result shows the characteristic sawtooth of slow glaciations and rapid terminations.
The controls let you adjust the ice accumulation rate, slide the time window start (0–700 ka) and span (50–800 ka), and toggle overlays for eccentricity (100 ka), obliquity (41 ka), precession (23 ka), the CO₂ curve and the δ¹⁸O temperature trace. Reading these archives is how palaeoclimatologists quantify natural glacial cycles and benchmark today's CO₂ rise against the deep past.
What does this ice core simulation actually show?
It plots a reconstructed climate record spanning about 800,000 years before present. The blue line is a δ¹⁸O temperature proxy, the dashed red line is atmospheric CO₂, and optional thin lines show the three Milankovitch orbital cycles that pace the ice ages. Shaded bands mark colder glacial intervals.
What is δ¹⁸O and why is it used as a thermometer?
δ¹⁸O is the ratio of heavy oxygen-18 to oxygen-16 in the ice, measured in parts per thousand (per mil). During cold periods the heavier isotope is preferentially retained, so δ¹⁸O tracks temperature. This simulation uses the approximation Δδ¹⁸O ≈ 0.7·ΔT to convert the proxy into degrees.
What are Milankovitch cycles?
They are slow, predictable changes in Earth's orbit and tilt that alter how sunlight is distributed across the planet. The three components here are eccentricity (≈100,000-year orbital shape), obliquity (≈41,000-year axial tilt) and precession (≈23,000-year wobble). Their combined forcing paces the rhythm of glacial and interglacial periods.
The accumulation slider scales the ice build-up rate from 0.5x to 2.0x. The window-start slider moves the left edge of the view from 0 to 700 ka, and the span slider sets how many thousand years are displayed (50–800 ka). Together they let you zoom into a single termination or view the whole 800 ka record.
In real ice cores, atmospheric CO₂ changes follow temperature changes by several hundred years during deglaciations, because warming oceans and shifting circulation release stored carbon. The panel notes a representative lag of roughly 800 years; the simulation reproduces this by tying CO₂ to the temperature proxy with a small offset.
Glacial cycles are asymmetric: ice sheets grow slowly over tens of thousands of years but collapse rapidly during a termination, giving a sawtooth shape. The model captures this non-linear ice-volume response so that glaciations build gradually while interglacial transitions appear abrupt.
It is a synthetic, physically motivated reconstruction, not the raw Vostok or EPICA measurements. The orbital periods, the δ¹⁸O–temperature scaling, the 180–280 ppm CO₂ range and the glacial sawtooth all reflect published ice-core science, but the exact point values are generated for illustration.
Across the past 800,000 years, ice cores show atmospheric CO₂ swinging between roughly 180 ppm during the coldest glacial maxima and about 280–300 ppm during warm interglacials. The simulation keeps CO₂ within this band, which highlights how far modern levels now exceed any natural value in this record.
It reports which orbital cycle most strongly governs the chosen window. For the last roughly 800,000 years the ~100,000-year eccentricity rhythm dominates, so wide spans report ~100 ka, while narrower views emphasise the faster obliquity (41 ka) and precession (23 ka) beats.
They provide a direct, year-by-year archive of past temperature and greenhouse-gas concentrations, letting scientists measure natural variability and the tight coupling between CO₂ and temperature. That long baseline is essential context for judging how unusual and rapid present-day, human-driven warming really is.