1. MS-23, Woods Hole Oceanographic Institution, Woods Hole, Massachusetts 02540-1541, USA
2. Fachbereich Geowissenschaften, Universität Bremen, Postfach 330 440, 28334 Bremen, Germany

Correspondence to: Richard D Norris¹ Correspondence and requests for materials should be addressed to R.D.N. (e-mail: Email: rnorris@whoi.edu).

Abstract

Current models of the global carbon cycle lack natural mechanisms to explain known large, transient shifts in past records of the stable carbon-isotope ratio (¹³C) of carbon reservoirs^{1, 2}. The injection into the atmosphere of 1,200–2,000 gigatons of carbon, as methane from the decomposition of sedimentary methane hydrates, has been proposed to explain a ¹³C anomaly^{3, 4} associated with high-latitude warming¹ and changes in marine^{5, 6, 7} and terrestrial⁸ biota near the Palaeocene–Eocene boundary, about 55 million years ago. These events may thus be considered as a natural 'experiment' on the effects of transient greenhouse warming. Here we use physical, chemical and spectral analyses of a sediment core from the western North Atlantic Ocean to show that two-thirds of the carbon-isotope anomaly occurred within no more than a few thousand years, indicating that carbon was catastrophically released into the ocean and atmosphere. Both the ¹³C anomaly and biotic changes began between 54.93 and 54.98 million years ago, and are synchronous in oceans and on land. The longevity of the ¹³C anomaly suggests that the residence time of carbon in the Palaeocene global carbon cycle was 120 thousand years, which is similar to the modelled response after a massive input of methane^{3, 4}. Our results suggest that large natural perturbations to the global carbon cycle have occurred in the past—probably by abrupt failure of sedimentary carbon reservoirs—at rates that are similar to those induced today by human activity.

1. MS-23, Woods Hole Oceanographic Institution, Woods Hole, Massachusetts 02540-1541, USA
2. Fachbereich Geowissenschaften, Universität Bremen, Postfach 330 440, 28334 Bremen, Germany

Correspondence to: Richard D Norris¹ Correspondence and requests for materials should be addressed to R.D.N. (e-mail: Email: rnorris@whoi.edu).