What do Antarctic ice cores (EPICA, Vostok) show for atmospheric CO2 during the Last Glacial Maximum (~21,000 years ago)?

1. What the Vostok and EPICA ice cores actually measure

Vostok and EPICA obtain ancient air trapped in bubbles within Antarctic ice and measure CO2 concentration directly from those bubbles, producing long, continuous records of atmospheric CO2 over hundreds of thousands of years. The EPICA Dome C and Vostok cores together underpin high‑resolution composite CO2 records for 650–800 kyr and show repeated glacial–interglacial swings in CO2 ^{[5] [4]}.

2. Numbers for the Last Glacial Maximum

The cores indicate low glacial CO2 values during the LGM — commonly cited near ~180 ppm — and agreement across multiple publications and summaries that LGM CO2 was about 180 ppm compared with ~280 ppm in preindustrial times, i.e., a ~100 ppm difference to preindustrial or ~75–90 ppm change across the deglaciation depending on the precise interval used ^{[1] [2] [3] [6]}.

3. How much did CO2 rise during deglaciation?

Dome Concordia (EPICA) gas records show an increase of about 76 ppmv in CO2 over the transition from the LGM to the Holocene, occurring over several millennia in discrete steps ^[3]. Broader summaries and models describe glacial–interglacial swings on the order of ~75–90 ppm, with natural ocean‑atmosphere processes invoked to explain the change ^{[2] [7]}.

4. Timing and the “CO2 lag” debate

Ice cores show close covariation of temperature proxies and CO2, but dating differences between ice and trapped gas introduce uncertainty in lead/lag analysis. Some reconstructions and recent reanalyses place CO2 increases lagging Antarctic temperature by up to several centuries or more — earlier estimates cited lags of a few hundred years, and some analyses reported possible lags of up to ~1,400 years in particular segments — but newer work reduces that uncertainty and finds much smaller lags in several intervals ^{[8] [9]}. Reporting emphasizes that dating complexities (gas age vs ice age differences) are the reason for variable lag estimates ^{[8] [9]}.

5. Why CO2 was lower during the LGM — competing explanations

Current literature and modeling point to multiple, interacting causes: colder ocean temperatures increasing oceanic CO2 storage, changes in ocean circulation and biological productivity (including iron‑dust fertilization of the Southern Ocean), and altered ocean chemistry and alkalinity. Studies and model syntheses estimate cold oceans account for roughly half the CO2 reduction, with dust/biological effects and other processes explaining much of the remainder ^{[2] [7] [10]}.

6. Limits and uncertainties to keep in mind

Ice‑bubble CO2 is direct but averaged over the gas trapping period and may smooth short‑term variability; deeper, low‑accumulation sites like Vostok have larger uncertainty in gas/ice age differences and potential temporal offsets of up to centuries to >1,000 years in some intervals ^[11]. EPICA Dome C provides higher temporal reach and has been the basis for many recent analyses, yet chronology uncertainty grows with depth and requires modeling and independent markers ^{[4] [5]}.

7. Broader context and modern comparison

Across the last 800,000 years, CO2 never exceeded ~300 ppm until the industrial era; modern CO2 is far higher (~420+ ppm in recent years), and the post‑industrial rise is much faster than natural changes at the LGM–Holocene transition ^{[4] [12]}. Ice‑core data thus show that the LGM was characterized by unusually low CO2 (~180 ppm) compared with interglacials, and natural deglacial mechanisms increased CO2 by roughly three‑quarters to a hundred ppm over millennia ^{[1] [6] [3]}.

If you want, I can extract the specific LGM CO2 time slice from the EPICA and Vostok published tables and produce a short chart of the commonly reported values and ranges cited above (data source citations included).