An Ancient Carbon Mystery

See allHide authors and affiliations

Science  08 Dec 2006:
Vol. 314, Issue 5805, pp. 1556-1557
DOI: 10.1126/science.1136110

About 55 million years ago, Earth experienced a period of global warming that lasted ∼170,000 years (1). This climate event—the Paleocene-Eocene Thermal Maximum (PETM)—may be the best ancient analog for future increases in atmospheric CO2. But how well do we understand this event?

Temperature records from the tropics to the poles indicate that at the start of the PETM, global temperatures increased by at least 5°C in less than 10,000 years (2). The rise in surface temperature was associated with changes in the global hydrological cycle (3) and a large decrease in the 13C/12C ratio of marine (4) and terrestrial carbonates (5) and of organic carbon (3). This carbon isotopic excursion indicates that changes in the global carbon cycle were linked to global warming.

Furthermore, the ocean's carbonate compensation depth—the depth above which carbonate accumulates on the sea floor—rose substantially at the start of the carbon isotope excursion (5). This change is consistent with ocean acidification associated with a rapid influx of CO2. Although the change in ocean chemistry was not uniform throughout the ocean (6, 7), the confluence of isotopic and sedimentological data supports the conclusion that atmospheric CO2 was the primary greenhouse gas driving the PETM. Yet, the source of the CO2 remains a mystery.

Biological responses to global warming during the PETM include changes in the ecology of marine organisms, a mass extinction of benthic foraminifera (4, 8), and a global expansion of subtropical dinoflagellates at the earliest onset of the event (9). Global warming also coincides with the appearance of modern orders of mammals (including primates), a transient dwarfing of mammalian species, and a migration of large mammals from Asia to North America (8).

According to one hypothesis, the PETM was caused by the release of ∼2000 PgC from the destabilization of methane hydrates (which would subsequently oxidize to form CO2) (10). However, it is unlikely that methane was the sole source of warming. For example, the size of the methane hydrate reservoir at the end of the Paleocene was probably much smaller than it is today (11), and the magnitude of the sustained warming and the change in the carbonate compensation depth are compatible with a much greater mass of carbon than originally estimated (6). To account for larger carbon inputs, other sources have been invoked, including the oxidation of terrestrial (12) and marine (13) organic carbon and/or volcanic outgassing and thermal decomposition of organic matter (14). There is no single satisfactory explanation.

But whatever the source, the carbon input responsible for the PETM must have been massive. Given a global temperature sensitivity range of 1.5 to 4.5°C per doubling of the atmospheric CO2 concentration and global mean annual temperatures perhaps 5°C warmer than during recent pre-industrial times, estimates for pre-PETM atmospheric CO2 concentrations range from 600 to 2800 parts per million (ppm), broadly consistent with estimates from proxy data (15). Starting from these conditions, an increase of 750 to 26,000 ppm of atmospheric CO2 would be required to account for an additional 5°C rise in global temperature, which implies an addition of 1500 to 55,000 PgC to the atmosphere alone (see the first figure).

CO2 input during the PETM.

The amount of additional atmospheric CO2 responsible for the PETM warming depends on the pre-PETM atmospheric CO2 concentration and the climate sensitivity to CO2 doubling. To determine pre-PETM atmospheric CO2 concentrations (blue line), we assumed pre-PETM global mean annual temperature 5°C warmer than during recent pre-industrial times, when atmospheric CO2 concentrations were 280 ppm. To determine PETM atmospheric CO2 concentrations (orange line), we assumed a 5°C warming during the PETM and a surface ocean 5× saturated with respect to calcite.

Sustaining this concentration for tens of thousands of years implies partial equilibration with the carbonate system in the ocean, indicating a total release of 5400 to 112,000 PgC (see the second figure), with 3900 to 57,000 PgC of released carbon residing in the ocean (and with additional carbon supplied by the dissolution of carbonates). The extraordinary magnitude of these estimates is evident when compared against the 5000 PgC estimated for conventional fossil fuel resources available today.

Carbon release during the PETM.

The amount of carbon needed to explain a 5°C change in global mean temperature depends on pre-PETM CO2 conditions (see the first figure) and the climate sensitivity to CO2 doubling (including associated system feedbacks). The source of carbon released (and climate sensitivity) can be estimated from the carbon isotopic composition of the released carbon and the δ13C excursion it produced. For example, assuming a carbon isotope excursion of −3 to −5‰, carbon from methane (with an average δ13C value of −60‰, green bar) would imply a carbon input of 1800 to 3500 PgC and a climate sensitivity of 6.8 to 7.8°C per CO2 doubling. Terrestrial/marine organic carbon refers to organic carbon derived from the primary production of terrestrial and/or marine plants.

The input of carbon responsible for the PETM altered the stable carbon isotopic composition of the Eocene oceans and atmosphere. Marine carbonate records indicate a carbon isotope excursion between −2.5 and −3 per mil (‰), but records from ancient soil carbonates and plant organic matter reveal a much larger change of over −5‰ (3, 5). Explanations have been presented to account for these isotopic differences (5), but this evidence can also suggest that the global carbon isotope excursion was larger than determined from marine carbonates (3).

These details may appear esoteric, but to determine the mass and source of carbon responsible for the >5°C warming during the PETM, we must match the magnitude of the carbon isotope excursion with the mean global temperature sensitivity to CO2 and associated climate feedbacks (see the figures). One conclusion from this approach is that CO2 derived from methane hydrates could only have caused the PETM if the climate sensitivity to CO2 was much higher than currently assumed. Yet carbon sources other than methane, such as the oxidation of primary terrestrial and/or marine organic carbon, together with commonly accepted estimates of climate sensitivity, would require extremely large carbon inputs to explain the warming. Thus, the PETM either resulted from an enormous input of CO2 that currently defies a mechanistic explanation, or climate sensitivity to CO2 was extremely high.

The next challenges are to constrain the magnitude and rate of carbon input (and that of other greenhouse gases) and to develop realistic models for the cause of this anomalous, but clearly CO2-induced global warming event. Solving this mystery will allow us to determine whether the PETM is a true analog for future climate change.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
View Abstract

Stay Connected to Science

Navigate This Article