Paleoclimate: In 1989, a Soviet drill team at Vostok Station in

In 1989, a Soviet drill team at Vostok Station in East Antarctica pulled ice from a depth of 2,083 meters that had fallen as snow roughly 160,000 years ago. Trapped inside were air bubbles — actual samples of the Pleistocene atmosphere, preserved at -55°C for longer than Homo sapiens has existed as a species. When researchers cracked those bubbles and measured the CO₂ inside, they found concentrations of about 190 parts per million. The preindustrial baseline was 280. Today's reading is above 420. That single data point, recovered from a hole in the ice, compressed the entire arc of glacial-interglacial climate change into a number you can hold in your head.

NASA

Nothing major moved on the paleoclimate research front this week, but the underlying science is worth a careful look on its own terms.

What a Proxy Actually Is

NOAA NESDIS / STAR

NASA SDO

The instrumental record — thermometers, rain gauges, weather stations — extends back roughly 150 years in any meaningful global coverage. Before that, reconstructing climate means reading materials that were shaped by climate without being designed to record it. These are proxies.

Ice cores are the most information-dense proxies available. The ratio of oxygen isotopes (¹⁶O to ¹⁸O) in each annual layer shifts predictably with temperature: colder air carries proportionally less of the heavier isotope, so a layer laid down during a cold year has a different isotopic signature than one from a warm year. Stack enough layers, and you have a continuous temperature record going back 800,000 years at Dome C in Antarctica — eight glacial cycles, each one a controlled experiment in how Earth's climate system responds to orbital forcing.

Tree rings — the field is dendrochronology — work on a shorter timescale but with annual precision. A wide ring means a good growing season; a narrow one means stress, usually from drought or cold. By cross-dating overlapping ring sequences from living trees, dead wood, and structural timbers in old buildings, researchers have assembled continuous chronologies extending back more than 10,000 years in some regions. The bristlecone pines of the White Mountains in California contribute some of the oldest living material to these records, individual trees exceeding 4,000 years in age.

Speleothems — cave formations like stalagmites — grow in annual layers too, and their isotopic composition reflects the temperature and moisture of the rainfall that percolated through the rock above them. Because caves are geographically fixed and often preserve material spanning hundreds of thousands of years, speleothems are particularly valuable for pinning down the timing of abrupt climate transitions with uranium-thorium dating accurate to within decades across deep time.

The Younger Dryas Problem

The most instructive stress test for all of this proxy evidence is a climate event called the Younger Dryas, which ran from approximately 12,900 to 11,700 years before present. At the end of the last ice age, as the planet was warming, temperatures in Greenland plunged roughly 10°C in what the ice core record suggests was a matter of decades — possibly less. Then, about 1,200 years later, temperatures rebounded by a similar magnitude, with some ice core evidence suggesting the transition happened within a human lifetime.

The mechanism most widely accepted involves the collapse of freshwater discharge into the North Atlantic from melting ice sheets, which disrupted the thermohaline circulation that moves heat poleward. What makes the Younger Dryas scientifically productive is that it appears in multiple independent proxy systems simultaneously: Greenland ice cores, European pollen records, cave formations in Ireland and China, and marine sediment cores from the North Atlantic seafloor all register the same anomaly at the same time. That convergence is what separates a genuine climate signal from noise in a single archive.

The Younger Dryas also illustrates why paleoclimate matters for contemporary climate science. If a model cannot reproduce a 10°C regional shift that we know happened, that model's projections for future abrupt change carry less weight. Proxy records are the only empirical data set large enough and old enough to validate model behavior across a full range of climate states — not just the narrow corridor of the last century and a half.

The Holocene Baseline

The Holocene — the interglacial period that began after the Younger Dryas, roughly 11,700 years ago — is the climate state in which human civilization developed. Agriculture, cities, written language: all of it emerged during a period of relative thermal stability that the proxy record shows is unusual by Pleistocene standards.

Paleoclimate reconstructions of the Holocene, including the widely cited work behind the "hockey stick" temperature graph, draw on exactly this toolkit: tree rings for the past few millennia, ice cores and sediment cores to extend the baseline further back. The shape of that record — a long, gradual cooling from a Holocene thermal maximum around 6,000 years ago, then a sharp upturn in the industrial era — is not a product of any single proxy. It is the convergence of dozens of independent archives pointing the same direction.

On the Radar

• If you want a concrete entry point into this field, the NOAA Paleoclimatology database (ncei.noaa.gov/products/paleoclimatology) hosts raw proxy data from ice cores, tree rings, and sediment records — actual downloadable datasets, not summaries.
• Watch for publications from the PAGES 2k Consortium, which coordinates multi-proxy reconstructions of the past 2,000 years and updates its synthesis as new data comes in.
• The next time you see a climate projection extending to 2100 or beyond, check whether the underlying model has been validated against a known past climate event — the Last Glacial Maximum (about 21,000 years ago) and the mid-Holocene are the two standard benchmarks.