Four Hundred Thousand Years in a Cylinder of Ice

In January 1998, at the Russian Vostok Station in East Antarctica — 78°S, 106°E, elevation 3,488 meters, mean annual temperature −55°C — a drill reached 3,623 meters into the ice sheet. Then it stopped. Not because it ran out of ice, but because 120 meters below the drill tip lay the surface of Lake Vostok, a subglacial lake the size of Lake Ontario that has been sealed beneath the ice for millions of years. They stopped to avoid contaminating it.

What they pulled out of the borehole was a cylinder of ice containing 420,000 years of Earth’s atmospheric history.

What the ice contains

Snow falls. It compresses into ice. Air bubbles get trapped. The ice sheet grows, layer by layer, year by year, for hundreds of millennia. Each layer is a time capsule. The ice itself records temperature (through the ratio of deuterium to hydrogen — heavier water molecules fall preferentially in warmer conditions). The trapped air bubbles record atmospheric composition — the actual CO2 and methane concentrations of the air at the time the snow fell.

This is not a proxy. This is not an inference. The air in the bubbles is the air that was in the atmosphere when the snow formed. You can measure it directly.

The Vostok paper — Petit et al., published in Nature in June 1999, a collaboration between French, Russian, and American scientists — presents six parallel time series from this single core:

Deuterium (temperature proxy) — glacial-interglacial temperature swings of ~8°C at the inversion level, ~12°C at the surface
CO2 — oscillating between ~180 ppm (glacial) and ~280-300 ppm (interglacial)
Methane (CH4) — oscillating between ~320-350 ppb (glacial) and ~650-770 ppb (interglacial)
δ18O of atmospheric oxygen — tracking global ice volume
Dust — continental aridity marker, 20-40x higher during glacial periods
Sodium — sea-salt aerosol, anti-correlated with temperature

All six series, from the same cylinder of ice, telling the same story from different angles.

The four cycles

The record covers four complete glacial-interglacial cycles, each lasting roughly 100,000 years. The pattern is consistent across all four:

A warm interglacial period
A progressive cooling through increasingly cold interstadial events
A glacial maximum — the coldest point
A rapid warming (termination) back to interglacial conditions

The shape is a sawtooth: slow descent into glaciation, rapid ascent out of it. Cooling takes tens of thousands of years. Warming takes thousands. The asymmetry is present in all four cycles.

The temperature and greenhouse gas concentrations track each other with remarkable fidelity. The correlation coefficient between CO2 and Antarctic temperature is r² = 0.71 over the full 420,000 years. For methane, r² = 0.73. This is not a model prediction. This is a measurement.

The unprecedented numbers

The Vostok paper makes one statement that, read carefully, is among the most consequential sentences in the scientific literature:

The extension of the greenhouse-gas record shows that present-day levels of CO2 and CH4 (~360 ppm and ~1,700 ppb, respectively) are unprecedented during the past 420,000 years.

At the time of publication in 1999, CO2 was 360 ppm. As of 2025, it is approximately 425 ppm.

The highest CO2 concentration in the entire 420,000-year record — across four interglacials, including periods that were warmer than today — was about 300 ppm, during marine isotope stage 9.3.

The current 425 ppm is 42% above the highest value recorded in 420,000 years of trapped air bubbles.

The EPICA core, drilled at Dome C in Antarctica and published in subsequent papers, extended the record to 800,000 years. The conclusion held: nothing in 800,000 years matches current concentrations.

What the power spectrum shows

Petit et al. performed spectral analysis on all six time series. The dominant periodicities:

~100,000 years — the strongest signal in temperature, CO2, and methane. Corresponds to the eccentricity cycle of Earth’s orbit.
~41,000 years — strong in temperature (23% of variance) and sodium. Corresponds to the obliquity cycle (the tilt of Earth’s axis).
~23,000 and ~19,000 years — present in methane and δ18O but weak in temperature and CO2. Corresponds to the precession cycle.

These are Milankovitch cycles — the orbital parameters that determine how much solar energy reaches different latitudes at different times of year. The ice core confirms that Earth’s climate responds to orbital forcing, but with an amplification mechanism. The orbital forcing alone is too weak to explain the full glacial-interglacial temperature swing. CO2 and methane are the amplifiers — they correlate with temperature because they’re part of the feedback loop, not because they’re the primary driver.

This distinction matters. The orbital forcing initiates the change. The greenhouse gases amplify it. The current situation is the inverse: humans are forcing the greenhouse gases directly, without waiting for orbital parameters to change.

The timing problem

One of the most important findings in the Vostok data: during glacial terminations, Antarctic temperature begins to rise before global ice volume decreases. The warming leads the deglaciation.

CO2 and methane also begin rising early in the termination — roughly in phase with Antarctic temperature, and ahead of the global ice volume change measured in ocean sediment cores.

This has been misrepresented in climate skeptic arguments as “temperature drives CO2, not the other way around.” The Vostok data actually shows something more complex: the initial warming trigger (orbital forcing at high southern latitudes) causes both temperature increase and CO2 release (from warming oceans degassing dissolved CO2), and the released CO2 then amplifies the warming globally. It’s a feedback loop, not a one-way causation.

The paper states this carefully:

CO2 and CH4 may have contributed to the glacial–interglacial changes over this entire period by amplifying the orbital forcing along with albedo, and possibly other changes.

“Amplifying.” Not “causing.” Not “irrelevant.” The precise word for a feedback mechanism.

Why this document matters for the Arctic posts

The Vostok ice core is the physical record against which the entire Arctic sovereignty competition plays out. Russia, Canada, Denmark, the United States, and China are competing for Arctic resources because the ice is retreating. The ice is retreating because the temperature is rising. The temperature is rising because greenhouse gas concentrations are at levels that haven’t existed in at least 420,000 years — and probably not in 800,000.

The posts in this series describe flags planted on ice (#229), claims filed over seabed (#230), myths projected onto the unknown (#231). The ice core is the ground truth beneath all of them. The ice remembers what the climate was doing when none of these nations existed, when no humans existed, when the orbital parameters cycled through their 100,000-year rhythm without anyone to notice.

The ice also records, in its uppermost layers, the moment when a single species began pushing CO2 beyond anything the previous 420,000 years had produced. That transition — from orbital-forced climate to anthropogenically-forced climate — is the context for everything happening at the North Pole now.

The nations competing for Arctic resources are competing for access that exists because of warming that is unprecedented in the geological record they’re drilling through to prove their continental shelf claims.

The ice remembers. The ice is also melting.

Source: J. R. Petit et al., “Climate and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica,” Nature 399 (1999): 429-436. Full text in the project source archive.