Who moved first — temperature or CO₂? Ice cores and paleoclimate

In Antarctic ice, local warming often appears before carbon dioxide rises. That observation is real. Treating it as proof that CO₂ cannot drive global temperature is not. The records describe a coupled system: triggers, carbon feedbacks, and greenhouse amplification — not a one-line slogan.

Ice cores · Temperature · CO₂

1. The question, without the slogan

Long climate swings before the industrial era line up closely with atmospheric CO₂. The serious question is order and mechanism:

Did temperature and CO₂ always move together?
When they did not, which led — and in which place on Earth?
Does a lag in one record mean CO₂ is only a passive follower of oceans and life?
What does modern climatology actually treat as established?

Correlation is strong. Causation runs both ways. Lead and lag depend on which temperature series you use and which interval you examine. That is what the data support — not a campaign line from either side of the public argument.

2. What Antarctic ice cores show

Antarctic cores (Vostok, EPICA Dome C and related records) remain the best continuous joint archive of local temperature proxies and trapped-air CO₂ across the last roughly 800,000 years.

Glacial CO₂ sat near ~180 ppm; interglacial peaks near ~280 ppm. Temperature proxies and CO₂ co-vary strongly across that window. At many glacial terminations, Antarctic temperature rises before CO₂ — often by centuries, sometimes approaching on the order of a thousand years, with estimates that depend on the termination and on how the ice-age–gas-age difference is handled. Later high-resolution work finds smaller lags or near-synchrony at some change points and clearer lags at others.

A local Antarctic lead does not mean CO₂ cannot warm the climate. It means the first nudge out of a glacial in the far south need not be carbon dioxide. Orbital forcing, ice-sheet dynamics, and Southern Ocean circulation are the usual candidates for that initial move.

Figure 1. Schematic of the frequently cited Antarctic pattern: local temperature proxy rises; CO₂ follows. This is a diagram of the idea, not a plot of one named core.

3. Global temperature is not Antarctic temperature

This is the pivot of the research literature — and the point public debate most often skips.

Shakun et al. (2012, Nature) assembled a global surface-temperature stack from 80 proxy records for the last deglaciation (~20–10 ka). Their results matter because they change the thermometer:

Global mean temperature correlates with CO₂ and, over much of the deglaciation, lags CO₂ — carbon rises first in the global comparison.
Antarctic warmth can still lead, consistent with hemispheric heat redistribution when Atlantic overturning weakens: earlier southern warmth, delayed northern warmth.
CO₂ supplies a global, in-phase greenhouse forcing. Model experiments in that study support carbon dioxide as a principal amplifier of global deglacial warming once carbon leaves the ocean–climate system.

So the slogan “temperature always leads CO₂” is usually an Antarctic claim dressed up as a planetary one.

Figure 2. Same deglacial window, different series: Antarctic temperature can lead; global mean temperature tends to track or lag CO₂ (after the logic of Shakun et al. 2012). Schematic.

4. The physical loop

A standard process sequence for a glacial termination runs roughly as follows:

Orbital change alters seasonal insolation, especially Northern Hemisphere summers relevant to ice sheets.
Ice and ocean circulation respond; the Southern Ocean and Antarctic region can warm early.
Ocean carbon is released — solubility, stratification, sea ice, and biology remain active research topics. Atmospheric CO₂ rises by tens of ppm.
Greenhouse forcing from that CO₂ adds global warming, including the large northern ice-sheet retreat that defines the end of the ice age.
Further feedbacks — albedo, additional ocean and land carbon — reinforce the shift.

In that loop, temperature can lead CO₂ regionally at the start, while CO₂ can lead or pace global mean temperature as the main greenhouse amplifier. Marine and terrestrial biology participate in the carbon budget; the dominant glacial–interglacial CO₂ swing is still attributed primarily to ocean carbon reorganization, not a simple story in which flora and fauna alone “raise” the interglacial.

Figure 3. Coupled sequence: trigger → carbon release → greenhouse amplification → feedbacks. Both “temperature led” and “CO₂ drove global warming” can be true at different steps.

5. Deeper time is a different laboratory

Beyond the ice-core window, CO₂ is reconstructed from proxies — boron isotopes, paleosols, stomata, alkenones and related methods — and uncertainty bands widen.

The Paleocene–Eocene Thermal Maximum (~56 Ma) is a geologically rapid carbon release with global warming of order ~5–6°C in recent estimates. The event is treated as greenhouse-driven. Carbon sources (volcanic and/or sedimentary) are debated; the direction of the argument is “carbon injection, climate response,” on a timescale still longer than the modern industrial rise.

Broader Cenozoic reconstructions associate warmer deep-time climates with higher CO₂ within proxy error. That is correlation plus radiative physics — not a single ice-core lag slogan extended backwards without limit.

6. What is established — and what is not

Strongly supported. CO₂ is a greenhouse gas; raising it increases radiative forcing. Pleistocene CO₂ and climate co-vary; CO₂ is a major feedback and amplifier of orbital-scale change. For the last deglaciation, global temperature stacks show CO₂ rising with or ahead of global mean warming, while Antarctica can warm earlier for ocean-circulation reasons.

Still refined. Exact centuries of lead and lag at each termination; the quantitative split among Southern Ocean mechanisms; the precise carbon inventory of events such as the PETM; how sensitivity scales between glacial, interglacial, and high-CO₂ greenhouse states.

Not established. That flora and fauna alone paved the way for CO₂ as the primary glacial–interglacial control. That a past regional lag cancels the radiative effect of the modern, human-driven CO₂ rise. The modern rise is larger in magnitude than the classic ice-age swing and far faster, and it draws on fossil carbon, not only ocean outgassing.

Figure 4. Scale and rate: glacial ~180 ppm, interglacial ~280 ppm, modern ~420+ ppm — with the industrial rise compressed into decades, not millennia.

7. Public debate: what the lag slogan leaves out

In public argument, the ice-core lag is often used as a single stroke: temperature rose first, therefore CO₂ cannot be a cause, therefore concern about modern CO₂ is misplaced. The chain fails at each step.

It swaps a regional series for a global claim. Antarctic ice is a superb archive of Antarctic conditions. It is not a substitute for a global temperature stack. When the thermometer is widened, the lead/lag story changes (Shakun et al. 2012 and related work).

It treats a feedback as a nullity. In a coupled system, a variable can lag at the start and still amplify most of the warming. That is ordinary systems behaviour, not a trick of language. Paleoclimate research has long described CO₂ as an amplifier of orbital-scale change; the lag data constrain timing, they do not delete the greenhouse effect.

It ignores mechanism and rate. Deglacial CO₂ rose mainly as the ocean–climate system reorganised over centuries to millennia. The modern rise is dominated by fossil carbon on a timescale of decades. Pointing to a glacial lag does not reassign the source or the speed of today’s increase.

It invites a mirror-image error. The opposite slogan — that CO₂ always moved first and biology is irrelevant — is also too neat. Orbital forcing, ice, and ocean circulation matter. Honest reading keeps the triggers and the amplifier in the same frame.

Research literature treats the lag as a timing constraint on mechanisms. Public rhetoric often treats it as a verdict on greenhouse physics. Those are different claims. Only the first is what the ice actually warrants.

8. Direct answers

Question	Evidence-based answer
Did long pre-industrial and prehistoric climate changes coincide with CO₂ changes?	Often yes in the ice-core era; generally yes in deeper time within proxy uncertainty.
Were temperature changes always caused by CO₂ rise?	No. Triggers can be orbital, ice, and ocean. CO₂ is a major amplifier — and for global mean in the last deglaciation, a leading greenhouse driver once released.
Did temperature always rise first so that life could raise CO₂?	Antarctic temperature often moves early; global mean need not. Biology participates; ocean processes dominate the classic ~100 ppm glacial–interglacial CO₂ swing.
Does modern climatology have established findings?	Yes on greenhouse physics, co-variation, and the amplifier role. Refined, not slogan-simple, on exact leads and lags by region and interval.

Sources (minimum set)

Shakun, J.D. et al. (2012). Global warming preceded by increasing carbon dioxide concentrations during the last deglaciation. Nature 484, 49–54.
Ice-core CO₂ and temperature phasing: Monnin et al.; Parrenin et al.; later EPICA / Dome C phasing studies in Climate of the Past and related journals.
Process context: IPCC paleoclimate chapters; ocean-carbon and Southern Ocean mechanism literature on glacial–interglacial CO₂.
PETM and deep-time CO₂: Tierney et al. and related proxy–model studies on warming magnitude and carbon release.

Figures 1–3 are explanatory schematics for orientation, not reproductions of published data plots. Figure 4 uses rounded textbook levels for glacial, interglacial, and modern CO₂.

Who Moved First — Temperature or CO₂?