Introduction
If one wishes to draw lessons on the effects of rapid extreme changes in the carbon cycle and global climate, the Palaeocene–Eocene thermal maximum (PETM) of 55 million years ago is a good place to start1, 2. But almost immediately we run into difficulties. We are not even sure what the root cause of this climatic turmoil was — a comet impact? Or did its origins lie closer to home? Two studies of magnetic particles in sediment cores from New Jersey published in Paleoceanography3, 4 address this debate. In doing so, they also expand the range of environmental consequences that stemmed from this extraordinary event.
Within a few thousand years of the start of the PETM, a huge injection of carbon dioxide depleted in 13C into the atmosphere and oceans caused the 13C/12C ratio in carbonate and organic compounds across the Earth to drop significantly and large amounts of calcium carbonate on the deep seafloor to dissolve5. At more or less the same time, temperatures spanning all latitudes, both at the Earth's surface and in the deep sea, rose by 5–8 °C. The hydrological cycle changed. On the land, mammals and plants undertook great migrations; in the sea, unusual plankton appeared in surface waters and much deep-sea fauna became extinct2. Over the following 80,000–200,000 years, the global carbon cycle and various Earth systems slowly recovered to something approaching their former state. But certain groups of organisms, such as the mammals and the foraminifera of the deep ocean, were affected forever2.
Where did all this excess carbon come from? One answer was suggested in 2003. Kent et al.6 discovered, preserved in three sediment drill cores from New Jersey, an exceptional abundance of single-domain magnetite (Fe3O4) grains, nominally between 30 and 100 nm in size, that were deposited on a continental shelf almost precisely during the PETM. Such magnetic nanoparticles are common in shallow marine sediment. Most derive from 'magnetotactic' bacteria, which precipitate uniform bullet-sized magnetite grains, typically in chains, within intracellular vesicles known as magnetosomes. These grains allow the bacteria to move along magnetic field lines7. But in their sediments, Kent et al. could find only a few lone nanoparticles of uniform shape. This observation led them to propose an origin other than bacteria, one that could also explain the huge carbon input at the beginning of the PETM — a cometary impact.
That conclusion was controversial, not least because it conflicted with other observations8, 9. Now, in the first of the new papers, Kopp et al.3 examine sediments spanning the PETM from one of the previously studied sites using several techniques, including transmission electron microscopy and ferromagnetic resonance spectroscopy. The latter technique has been applied only relatively recently to studies of ancient sediment, but can elucidate the shapes, dimensions and clustering of the small magnetite particles. The results indicate that sediment deposited before and after the PETM contains magnetite grains in a range of sizes, whereas the grains deposited during the event are of similar sizes, and some of them are in chains. This observation suggests a switch in the magnetite source: specifically, from an origin in continental rock to an origin in bacteria.
In the second paper, Lippert and Zachos4 conduct magnetic hysteresis and low-temperature demagnetization experiments on bulk sediments from a fourth New Jersey drill site. These authors document a high abundance of single-domain magnetite in sediment deposited during the PETM, just as Kent et al.6 did. But with extensive transmission electron microscopy analyses, they show that much of this magnetite comprises uniform bullet-shaped grains, 50–70 nanometres in length and often in chains, which is consistent with a bacterial source.
To date, the only 'evidence' for a comet impact at the start of the PETM has been the abnormally high amounts of single-domain magnetite in the New Jersey cores8, 9. But if, as the new results3, 4 suggest, there was no comet, we must seek a terrestrial source for the abnormal carbon input at that time. Of the possibilities, the amount of material involved — between 2,000 and 5,000 gigatonnes of carbon substantially depleted in 13C (ref. 5) — strongly implicates the release and oxidation of seafloor methane5, 10 or terrestrial organic compounds11.
Neither of the two latest studies fully addresses this issue. But they raise a fascinating question: why do abundant remains of magnetotactic bacteria precisely mark the PETM in the sediment sequences of the ancient New Jersey shelf? Magnetite-precipitating bacteria generally occupy a specific environmental niche in which surrounding water has little dissolved oxygen, a high supply of iron (Fe2+), and little or no hydrogen sulphide, which causes magnetite to dissolve12. Such conditions sometimes occur in open bodies of water, notably in stratified estuaries and lakes whose lower reaches lack oxygen. But a far more common setting is the 'suboxic' horizon of marine sediment sequences that receive large amounts of organic carbon and iron-bearing minerals.
Here, the bacteria occupy an intermediate depth interval with ideal water chemistry: the series of microbially mediated reactions that occurs as organic sediments are buried has already stripped most dissolved oxygen and nitrate and consumed manganese and iron oxides and hydroxides, releasing dissolved Fe2+; and the reduction of sulphate has not yet become dominant, so hydrogen sulphide levels are low. Such bacterial magnetite is rarely preserved in the geological record7: as the sediment is continuously buried deeper, the small particles dissolve in deeper pore waters containing hydrogen sulphide12.
The New Jersey sediments are mostly clay and quartz derived from the adjacent land mass. The 5–10-m-thick, particularly clay-rich interval spanning the PETM accumulated at much faster rates (more than 10 cm per millennium) than sediment deposited before or after the event (less than 2 cm per millennium)6, 13. The simplest explanation for the magneto-fossil record involves this pulse of terrestrial sediment3, 4, 8 (Fig. 1). Before and after the PETM, sediment accumulated slowly, and with a limited supply of solid organic carbon and iron. The resulting deep penetration of dissolved oxygen and low generation of dissolved Fe2+ would have meant minimal bacterial magnetite production. During the PETM, however, the supply of organic carbon and iron increased significantly with the increased levels of sediment deposition. This deposition occurred on top of pore waters that still contained dissolved oxygen, so the sediments did not later come into contact with anoxic pore waters rich in hydrogen sulphide, as is the case for normal steady-state conditions. The result was an expanded suboxic zone in which magnetite was not dissolved. Lower levels of dissolved oxygen in the water column because of warming or freshening of surface waters, or both13, might have further enhanced the formation of this suboxic zone3, 4.
Figure 1: Laying down the sediments of New Jersey.
Kopp et al.3 and Lippert et al.4 both furnish evidence that the large quantities of single-domain magnetite (Fe3O4) found in sediment cores in New Jersey from the Palaeocene–Eocene thermal maximum (PETM) 55 million years ago are of bacterial origin. Under standard, steady-state conditions, bacterially produced magnetite forms in pore waters with low dissolved oxygen and a high supply of dissolved Fe2+, but then dissolves as it is gradually buried deeper in pore waters rich in hydrogen sulphide. The result is just a thin layer of magneto-fossils in young sediment7. During the PETM, great quantities of terrigenous sediment, including organic carbon, iron oxides and iron hydroxides, might have been delivered to the continental margin, perhaps because of enhanced wet and dry seasons on the adjacent land. These rapidly accumulating sediments landed on a previously deposited, oxygen-rich interval of sediment, allowing a thick suboxic zone to develop without underlying hydrogen sulphide. This suboxic zone may have been enhanced by lower dissolved oxygen in the water column, caused by warmer conditions and brackish surface waters. The net consequence was a rapidly deposited sediment section with the abundant magnetofossils observed in the two papers3, 4.
Full size image (38 KB)As both papers mention, a third study of the magnetic properties of sediment spanning the PETM from three sections deposited on the shelf and upper slope of New Zealand is in the pipeline14. This work lacks detailed transmission electron microscopy and ferromagnetic resonance analyses, but shows, on the basis of magnetic-hysteresis measurements, an anomalous abundance of magnetite precisely across the PETM. In light of other recent literature13, 15, the explanation might be that the massive carbon input and extreme global warming of the PETM induced fundamental changes in the precipitation and the delivery of river-borne sediment to continental margins1. Meanwhile, the search for a fully satisfactory origin of that great carbon input goes on.
