What We Publish

An Ancient Carbon Mystery

Mark Pagani, Ken Caldeira, David Archer, and James C. Zachos · Science 314, 1556-1557 · 2006

Key finding. Producing more than 5 degrees Celsius of Paleocene-Eocene warming requires a total carbon release of roughly 5,400 to 112,000 petagrams — against about 5,000 petagrams in conventional fossil fuel resources today — unless climate sensitivity was far higher than the usual 1.5 to 4.5 degrees Celsius per doubling, and no known reservoir can supply carbon on that scale.

A contour plot with the carbon isotopic composition of the released carbon, from 0 to minus 70 per mil, on the vertical axis and the amount of carbon released, from 1 to over 100 thousand petagrams, on a logarithmic horizontal axis. A second horizontal axis across the top marks climate sensitivity to carbon dioxide doubling from 1.5 to 9.5 degrees Celsius, and a second vertical axis on the right marks the carbon isotope excursion from minus 1 to minus 7 per mil. Two green horizontal bands mark the isotopic signature of methane near minus 60 per mil and of terrestrial and marine organic carbon near minus 25 per mil. Curved bands for isotope excursions of minus 3 and minus 5 per mil sweep across the plot. Where the methane band meets those curves, the implied carbon release is small but the implied climate sensitivity lies near the far left of the top axis.
The isotopic composition of the released carbon and the size of the carbon isotope excursion together fix both how much carbon was released and what climate sensitivity is required. Methane, with its very light signature, implies a small release but a sensitivity of 6.8 to 7.8 degrees Celsius per doubling; isotopically heavier sources are compatible with conventional sensitivity only at releases no known reservoir can supply. Second figure, "Carbon release during the PETM" from Pagani, Caldeira, Archer, and Zachos (2006), Science 314, 1556-1557. Reproduced under author reuse rights. Extracted from the published PDF and resized for web display.

What question did this research address?

About 55 million years ago the Earth warmed by at least 5 degrees Celsius in less than 10,000 years and stayed warm for roughly 170,000 years. This event, the Paleocene-Eocene Thermal Maximum, is often described as the best ancient analogue for a future rise in atmospheric carbon dioxide.

A large drop in the carbon-13 to carbon-12 ratio of marine and terrestrial carbonates, and a rise in the ocean's carbonate compensation depth, both point to a rapid influx of carbon dioxide as the cause. The source of that carbon has never been identified. This Perspective asked what constraint the size of the warming, combined with the size of the carbon isotope excursion, places on the answer.

What did we find?

Assuming pre-event global mean temperature 5 degrees Celsius above pre-industrial and a climate sensitivity in the standard 1.5 to 4.5 degrees Celsius per doubling range, pre-event atmospheric carbon dioxide falls between 600 and 2,800 parts per million, which is broadly consistent with proxy estimates.

Producing a further 5 degrees Celsius of warming from that starting point requires adding 750 to 26,000 parts per million of carbon dioxide, or 1,500 to 55,000 petagrams of carbon to the atmosphere alone.

Sustaining that concentration for tens of thousands of years implies partial equilibration with the ocean carbonate system, so the total release must have been 5,400 to 112,000 petagrams of carbon, with 3,900 to 57,000 petagrams ending up in the ocean.

The leading hypothesis — release of about 2,000 petagrams from destabilized methane hydrates — cannot easily work. The late Paleocene hydrate reservoir was probably much smaller than today's, and matching a carbon isotope excursion of 3 to 5 per mil with methane's very light isotopic signature implies an input of only 1,800 to 3,500 petagrams, which in turn demands a climate sensitivity of 6.8 to 7.8 degrees Celsius per doubling.

Isotopically heavier sources such as oxidized terrestrial or marine organic carbon are compatible with conventional climate sensitivity, but only at carbon inputs large enough that no proposed reservoir accounts for them.

The two constraints cannot both be relaxed. Either the Paleocene-Eocene Thermal Maximum followed an enormous carbon input with no mechanistic explanation, or the climate sensitivity to carbon dioxide was extremely high.

Why does it matter?

Pairing the carbon isotope excursion with the temperature change turns a descriptive record into a quantitative test — it forces any proposed source of the carbon to be consistent with both its isotopic composition and its climatic effect, which rules out explanations that satisfy one and ignore the other.

Whether the Paleocene-Eocene event is a true analogue for the present depends on the answer. If the warming required a carbon input an order of magnitude larger than all conventional fossil fuel resources, the event bounds nothing about the coming century; if instead it reflects a very high climate sensitivity, it is directly relevant and the implication is unwelcome.

Citation

Mark Pagani, Ken Caldeira, David Archer, and James C. Zachos (2006). An Ancient Carbon Mystery. Science 314, 1556-1557.

Related