A topic in the Open Knowledge Graph — a free, open map of 15,290 topics and the order to learn them in.
Ice cores preserve continuous records of snow accumulation, temperature, and atmospheric composition (trapped air) dating back 800,000+ years. δ¹⁸O and δD in ice reflect past temperature via fractionation during precipitation; trapped air bubbles contain CO₂ and CH₄ at levels of past atmospheres. Dust and cosmogenic isotope ratios (e.g., Be-10) provide information about atmospheric circulation and solar activity. Ice cores from Greenland and Antarctica span multiple glacial-interglacial cycles and reveal abrupt climate changes (Dansgaard-Oeschger events, Heinrich events).
Examine core data from Greenland and Antarctica side-by-side; note asynchrony in temperature shifts (e.g., Younger Dryas warming in Greenland, continued cooling in Antarctica) and interpret in terms of ocean circulation changes.
Ice cores are not infinitely precise; dating uncertainty and layer-counting ambiguity increase with depth. Also, δ¹⁸O is affected by both temperature and precipitation patterns (moisture source, distillation), complicating interpretation.
From your study of paleoclimate proxies, you know that scientists reconstruct past climates using indirect indicators preserved in natural archives. Ice cores are among the most powerful of these archives because they preserve *two independent records simultaneously*: the ice itself records temperature and precipitation, while tiny air bubbles trapped between snowflakes as they compressed into ice preserve actual samples of the ancient atmosphere. No other proxy provides direct measurements of past atmospheric composition.
The temperature record relies on isotopic fractionation. Water molecules containing the heavier oxygen isotope ¹⁸O (or deuterium, ²H) evaporate less readily and condense more readily than those with the lighter ¹⁶O (or ¹H). As moisture travels from warm ocean sources toward the poles, it progressively loses heavy isotopes through precipitation along the way — a process called Rayleigh distillation. The colder the climate, the more depleted the remaining vapor (and the resulting polar snow) becomes in heavy isotopes. By measuring the ratio δ¹⁸O or δD in each layer of an ice core, scientists can estimate the temperature at the time that snow fell. More negative values indicate colder conditions; less negative values indicate warmer periods.
The trapped air bubbles tell a complementary story. As snow accumulates and compresses into firn and then solid ice, air pockets are sealed off from the atmosphere. These bubbles preserve the actual concentrations of CO₂, CH₄, and other greenhouse gases at the time of trapping. The EPICA Dome C core from Antarctica extends this record back over 800,000 years, revealing a striking pattern: CO₂ and temperature rise and fall together through glacial-interglacial cycles, with CO₂ ranging between about 180 ppm (glacial) and 280 ppm (interglacial). Additional information comes from dust layers (indicating dry, windy conditions and the extent of continental ice sheets), volcanic ash and sulfate layers (marking eruptions that can be cross-dated), and cosmogenic isotopes like ¹⁰Be (reflecting solar activity and cosmic ray flux).
One of the most dramatic discoveries from ice cores is the existence of abrupt climate changes. Greenland cores reveal Dansgaard-Oeschger events — rapid warmings of 8–15°C occurring within decades, followed by gradual cooling over centuries. Heinrich events, identified by layers of ice-rafted debris in North Atlantic sediments and correlated with cold phases in Greenland cores, indicate massive iceberg discharges from the Laurentide Ice Sheet. Comparing Greenland and Antarctic cores reveals a "bipolar seesaw": when Greenland warms abruptly, Antarctica cools, and vice versa — a pattern explained by reorganizations of the Atlantic meridional overturning circulation. These discoveries transformed our understanding of climate, showing that the climate system is capable of rapid, nonlinear shifts, not just the slow orbital pacing predicted by Milankovitch theory alone.