Transient ocean warming and shifts in carbon reservoirs during the early Danian

Abstract

A long-standing question in Paleogene climate concerns the frequency and mechanism of transient greenhouse gas-driven climate shifts (hyperthermals). The discovery of the greenhouse gas-driven Paleocene–Eocene Thermal Maximum (PETM; ∼ 55 Ma) has spawned a search for analogous events in other parts of the Paleogene record. On the basis of high-resolution bulk sediment and foraminiferal stable isotope analyses performed on three lower Danian sections of the Atlantic Ocean, we report the discovery of a possible greenhouse gas-driven climatic event in the earliest Paleogene. This event – that we term the Dan-C2 event – is characterized by a conspicuous double negative excursion in δ¹³C and δ¹⁸O, associated with a double spike in increased clay content and decreased carbonate content. This suggests a double period of transient greenhouse gas-driven warming and dissolution of carbonates on the seafloor analogous to the PETM in the early Paleocene at ∼ 65.2 Ma. However, the shape of the two negative carbon isotope excursions that make up the Dan-C2 event is different from the PETM carbon isotope profile. In the Dan-C2 event, these excursions are fairly symmetrical and each persisted for about ∼ 40 ky and are separated by a short plateau that brings the combined duration to ∼ 100 ky, suggesting a possible orbital control on the event. Because of the absence of a long recovery phase, we interpret the Dan-C2 event to have been associated with a redistribution of carbon that was already in the biosphere. The Dan-C2 event and other early Paleogene hyperthermals such as the short-lived early Eocene ELMO event may reflect amplification of a regular cycle in the size and productivity of the marine biosphere and the balance between burial of organic and carbonate carbon.

Introduction

Recent findings from deep-sea sections at Central Pacific (Shatsky Rise), equatorial Atlantic (Demerara Rise) and Southeast Atlantic (Walvis Ridge) oceans have revealed the existence of several prominent carbonate dissolution levels interspersed in lower Paleogene carbonate sediments (Bralower et al., 2002, Edgar et al., 2007, Hancock and Dickens, 2005, Kroon and Zachos, 2007, Lourens et al., 2005, Petrizzo, 2005). The carbonate dissolution levels are interpreted to have been caused by abrupt climate changes associated with increased greenhouse gas levels (Bralower et al., 2002, Lourens et al., 2005, Zachos et al., 2005). These findings suggest that the Paleocene–Eocene Thermal Maximum (PETM; Kennett and Stott, 1991, Zachos et al., 1993) was not a unique event, but was the most severe of a series of early Paleogene hyperthermals (Kroon and Zachos, 2007, Nicolo et al., 2007), amongst them the ELMO event in the early Eocene (Lourens et al., 2005). This important discovery needs further confirmation by multiple records from different oceanic settings to gain a global picture and explore the driving mechanisms of hyperthermals. Here, we report evidence for an early Danian (early Paleocene) transient warming event that bears a number of the hallmarks of a greenhouse gas-driven climate change.

The early Danian has been well documented as a time of high turnover rate in pelagic ecosystems following the Cretaceous/Paleogene (K/Pg) mass extinction event (Coxall et al., 2005, D'Hondt, 2005, D'Hondt et al., 1996a, Gerstel et al., 1987, Olsson et al., 1999, Smit, 1982). At the time of the mass extinction, a wide range of geochemical evidence indicates that the flux of organic matter to the deep ocean collapsed (Hsü et al., 1982, Hsü and McKenzie, 1985, Stott and Kennett, 1989, Zachos et al., 1989). The range of evidence includes a global reduction in surface-to-deep carbon isotopic (δ¹³C) gradient and inter-basin differences in δ¹³C (Arthur et al., 1987, D'Hondt et al., 1998, Stott and Kennett, 1989, Zachos and Arthur, 1986, Zachos et al., 1989), both of which apparently persisting for several million years after the K/Pg boundary interval (Adams et al., 2004, D'Hondt et al., 1998). Previous stable isotope studies also indicate a general decrease in global ocean δ¹³C during the early Danian that may reflect carbon sequestration in the deep ocean (Shackleton and Hall, 1984, Stott and Kennett, 1990, Zachos et al., 1989). It is still debated whether these geochemical characteristics represent the response to a long-term, multimillion-year, collapse of oceanic productivity (Hsü and McKenzie, 1985, Zachos et al., 1989) or whether oceanic productivity recovered quickly, within years of the K/Pg mass extinction (D'Hondt et al., 1998, Coxall et al., 2005). In this latter hypothesis, δ¹³C gradients remained low for million of years not because of a lack of productivity (Broecker and Peng, 1982), but because a reduced fraction of total productivity was exported to the deep sea (Adams et al., 2004, D'Hondt et al., 1998). The reduction in carbon export is possibly due to extinction of fecal pellet producers or a shift to smaller-celled primary producers (D'Hondt et al., 1998, D'Hondt, 2005, Thomas, 2007).

Most previous studies on the early Danian have relied on widely spaced samples to estimate mean climate states. Because of the long-standing controversies over the Chicxulub extraterrestrial impact-theory (Alvarez et al., 1980) for the K/Pg boundary (see for example Arenillas et al., 2006, Keller et al., 2003, MacLeod et al., 2006, Morgan et al., 2006, Smit, 1999), much work has focused either on the late Maastrichtian or on the K/Pg event itself and comparatively few attempts have been made to assess the recovery from the mass extinction. Furthermore, because of the low sedimentation rates in the known lower Danian deep-sea sequences, there is a lack of high resolution, open-ocean isotope records documenting the short-term climate trends within the long recovery period. As a result, little is known about the state of the recovering carbon cycle and it is unclear whether transient climate events occurred during the early Danian. Here, our primary objective is to investigate the large-scale changes in climates and associated carbon cycle dynamics during the early Danian. It requires that we step beyond discussion of mean states and closely examine short-term stable oxygen and carbon isotopes variations in open-ocean records. The framework we present here is based on highly time-resolved whole sediment bulk and foraminiferal shell based δ¹⁸O and δ¹³C changes from the North Atlantic Ocean, over the ∼ 400 ky that followed the K/Pg boundary mass extinction. In addition, a comparison is made with bulk sediment δ¹³C results obtained from the South Atlantic Ocean.