Black carbon in Paleocene–Eocene boundary sediments: A test of biomass combustion as the PETM trigger

https://doi.org/10.1016/j.palaeo.2008.06.010Get rights and content

Abstract

Several explanations have been proposed to account for the dramatic perturbation to Earth's carbon cycle that led to extreme warming, ocean acidification, mass extinction, and a major negative carbon isotope excursion (CIE) at the Paleocene–Eocene (P/E) boundary. One reason that the ultimate cause for this event remains ambiguous is the difficulty in directly testing these multiple hypotheses. We conducted a test of one proposed mechanism, the “wildfire hypothesis,” that suggests that combustion of a significant fraction of abundant Paleocene peat deposits within several thousand years could release enough CO2 to Earth's atmosphere to account for all of these observations. Soot and graphitic black carbon (hereafter collectively referred to as GBC), both combustion byproducts, are widely distributed (through smoke) by subaerial combustion, and are remarkably resistant to diagenetic degradation. Using samples provided by the Ocean Drilling Program, we measured GBC concentrations across the P/E boundary in two important Paleocene–Eocene Thermal Maximum (PETM) sections to test a prediction of the wildfire hypothesis, namely that the combustion of Paleocene organic matter at the onset of the PETM would produce a peak in sedimentary GBC. In the Bass River section (New Jersey margin), GBC concentrations are uniformly high (121–358 ppm), consistent with the core's continental margin setting, but there is no evidence of increased GBC flux at the PETM. Furthermore, carbon isotopic ratios of GBC display the negative CIE seen in PETM organic matter, indicating that this material was sourced from contemporaneous (P/E boundary) biomass, not geologically old carbon. At site 1210 (Shatsky Rise), GBC concentrations were uniformly below our method's detection limit (0.5 ppm), whereas the wildfire hypothesis would predict high GBC concentrations (∼ 3000) ppm accompanying the event. Together, these data appear to refute but cannot conclusively rule out biomass burning as the cause for the PETM event.

Introduction

The Paleocene–Eocene Thermal Maximum (PETM) is the most prominent abrupt climatic event in the Cenozoic Era (Zachos et al., 2001). During the PETM, high latitude sea surface temperature (SST) increased by up to 10 °C, while tropical SST and deep ocean temperatures rose by 4° to 5 °C (Zachos et al., 2003). Accompanying this warming were a mass extinction of benthic foraminifera (Kennett and Stott, 1991), rapid shifts in the ranges of terrestrial flora (Wing et al., 2005) and fauna (Bowen et al., 2002), changes in the global hydrological cycle (Bowen et al., 2004), shifts in ocean circulation patterns (Nunes and Norris, 2006), rapid shoaling of the calcite compensation depth (Zachos et al., 2005), and a prominent negative carbon isotope excursion (CIE) that affected all parts of the global exogenic carbon cycle. Recovery from this event was gradual, taking ∼ 170 kyr (Rohl et al., 2007). The CIE, which developed in only a few kyr (Farley and Eltgroth, 2003), had a magnitude of at least − 2.5‰ (Zachos et al., 2003), indicating that the PETM was caused by the rapid injection of a large amount of isotopically light carbon into the ocean/atmosphere system (Dickens et al., 1995). Possible sources of isotopically light carbon include catastrophic release of methane hydrates (CH4–H; δ13C =  60‰) from the continental slope seafloor (Dickens et al., 1995), the impact of a carbonaceous bolide (δ13C = −45‰; Kent et al., 2003), and the release of organic carbon (Corg; δ13C = −25‰) either by combustion of Paleocene peatlands (Kurtz et al., 2003) or by bacterially mediated oxidation from a desiccated epicontinental seaway (Higgins and Schrag, 2006). In this paper we present a test of the Kurtz et al. (2003) “wildfire hypothesis” by measuring black carbon concentrations and isotope ratios in Ocean Drilling Program sediment cores that span the Paleocene–Eocene boundary.

Dissociation of methane gas hydrate is a compelling explanation for the PETM because its unusual isotopic composition gives it the leverage to cause the CIE by addition of as little as 1100 Gt of carbon to the exogenic carbon cycle (Dickens et al., 1995). However, recent PETM simulations indicate that at least 4500 Gt of carbon are required to account for observed warming and shoaling of the CCD (Zachos et al., 2003, Zachos et al., 2005). Because Corg is less isotopically negative than CH4, much more carbon would be required to produce the observed CIE (Dickens, 2001), making oxidation of isotopically normal Corg more consistent with these features of the PETM. Although there was likely a greater mass of carbon stored as living terrestrial biomass during the Paleocene than there is today (Beerling, 2000), this reservoir alone could not supply enough carbon to have caused the PETM. However, the Paleocene is also recognized as a time of significant peat accumulation worldwide. Kurtz et al. (2003) used sedimentary carbon and sulfur isotope records to argue that ∼ 15,000 Gt of Corg accumulated as peat during this time. If wildfires in one or more areas of Paleocene peat accumulation rapidly oxidized a large portion of this carbon and returned it to the exogenic carbon cycle, the release of massive amounts of Corg may have been sufficient to have caused the PETM. Initial support for this mechanism comes from models of peat combustion which show that sustained burning is capable of causing negative carbon isotope excursions throughout the exogenic carbon cycle (Finkelstein et al., 2006), and the discovery of abundant charcoal within the PETM interval of the Cobham Lignite Bed in England (Collinson et al., 2003).

Peatland wildfires in Indonesia during the El Niño-induced drought of 1997 released 0.81–2.57 Gt of carbon to the atmosphere (Page et al., 2002), graphically demonstrating the capacity of peat combustion to transfer huge amounts of Corg back into the exogenic carbon cycle. Although human activities played a part in triggering the Indonesian wildfires of 1997 (Page et al., 2002), natural processes can convert peatlands from carbon sinks to carbon sources. When peatlands dessicate, they are easily subject to ignition by lightning strike, volcanic activity or even self-ignition during degradation (Scott, 2000, Svensen et al., 2003). For this reason, peat fires are common during dry years, and once fires have begun, the underlying peat can burn even when water content is 200% to 400% of organic matter content (Kajii et al., 2002). The wildfire hypothesis calls upon some combination of warmer temperatures, decreased precipitation, regional uplift and a lowering of sea level, allowing the desiccation and subsequent ignition of Paleocene peatlands.

In addition to CO2, biomass burning produces a host of other byproducts, including gases, aerosols and coarse particulates. Black carbon (BC) is one such byproduct formed during combustion (Kuhlbusch and Crutzen, 1995). BC consists of a continuum of materials ranging from large fragments of slightly charred biomass to microscopic soot and graphitic black carbon (hereafter collectively referred to as GBC), which form by the condensation of gases produced during combustion (Masiello, 2004). If the P/E boundary was indeed a time of extreme wildfire activity, such fires would have produced enormous amounts of BC from across this continuum (Collinson et al., 2003). Of particular interest to our investigation are GBC particles, which can act as a long-range tracer of paleowildfire activity.

GBC is characterized by micron- to submicron-sized particles which may be transported thousands of km once they become airborne (Masiello, 2004). Large fires produce smoke plumes which can circumnavigate the globe (Duncan et al., 2003, Damoah et al., 2004), and smoke particles in remote locations often originate in distant regions. For example, GBC in arctic haze is dominated by particles from southeast Asia (Koch and Hansen, 2005), while in the remote equatorial Pacific GBC originates from terrestrial biomass burning in the tropics (Andreae et al., 1984). Therefore, GBC particles from an event such as the wildfires proposed for the P/E boundary would be distributed, and eventually deposited in marine sediments, across the globe.

Two additional characteristics of GBC make it useful as a tracer of paleowildfire activity. First, GBC is highly refractory. Long-term records of GBC accumulation have been extended into the Middle Cretaceous for the North Atlantic Ocean (Summerhayes, 1981) and into the earliest Cenozoic for the Pacific Ocean (Herring, 1985), while others have presented GBC from the Cretaceous–Tertiary boundary as evidence of fires ignited by the K/T boundary impact (Wolbach et al., 1988, Scott et al., 2000). Second, carbon isotopic fractionation during burning is minor, typically < 2‰ (Bird and Grocke, 1997, Turney et al., 2006). This has allowed δ13C of GBC to be used to indicate the biomass source of GBC (Bird and Grocke, 1997).

As a test of the wildfire hypothesis, we present the first characterization of GBC from P/E boundary sediments, from two ODP sites representing distinct sedimentary environments. One prediction of the “wildfire hypothesis” is that evidence of increased production of GBC from P/E boundary wildfires should be preserved, perhaps globally, as elevated GBC levels in the sedimentary record of the P/E boundary. A second prediction is that this GBC should have carbon isotope ratios consistent with pre-P/E boundary peat rather than contemporaneous biomass.

Section snippets

Study sites

We selected cores from two ODP sites for our study, both of which have served as important sites for previous PETM-related research. Due to the widespread distribution of Paleocene peatlands in the tropics and the middle and high latitudes of both hemispheres (Scotese, 2001) we did not attempt to locate our sites relative to a single potential paleowildfire location, but instead chose sites based on disparate sedimentary settings. A continental margin site with rapid accumulation of

Methods

We adopted the method of Gelinas et al. (2001) for isolation and quantification of GBC. Briefly, this method uses HCl and HF to demineralize samples, followed by several treatments with increasingly concentrated trifluoroacetic acid to hydrolyze the majority of Corg in the sample. The remaining material is then heated to 375 °C for 24 h in a well ventilated muffle furnace, thermally oxidizing any remaining non-BC Corg, while BC loss is ≤ 5% (Gustafsson et al., 1997). At the end of this

Results

We measured GBC in 21 samples from the Bass River core covering a ∼ 14 m core interval (364.8–351.0 m) roughly centered around the onset of the CIE at 357.0 m. Sampling interval was ∼ 10 cm from 356.4–357.7 m, increasing to as much 2 m with increasing distance from the CIE. Samples weighed ∼ 2 g prior to extraction. GBC was detected in all samples, with concentrations ranging from 121–358 ppm, varying significantly between adjacent sample depths (Fig. 1). Although the highest concentration of GBC

Discussion

If combustion of Corg caused the carbon isotope excursion and associated global warming of the PETM, the mass of material burned must have been enormous (Table 2). We use a simple carbon cycle box model to determine the mass of Corg combustion needed to induce the observed CIE at the P/E boundary. The model consists of a single box representing the exogenic carbon cycle with a steady state mass of 40,000 Gt carbon with δ13C = 0.5‰. The input carbon flux is constant at 0.6 Gt/yr, such that the

Conclusions

Chemothermal oxidation provides a reasonably robust, reproducible method of searching for evidence of combustion byproducts in Paleocene–Eocene boundary sediments, thus providing a test of the wildfire hypothesis as a mechanism for the PETM. Our results, GBC concentrations and carbon isotope ratios from two sites (Bass River and 1210B) differ sharply from the predictions of the hypothesis and thus appear to refute, if not rule out, wildfires as a cause of the PETM. GBC is completely

Acknowledgments

We thank Bob Michener and Mark Altabet for their assistance with elemental analysis and mass spectrometry of our GBC extracts, Yves Gelinas for assistance with our extraction procedure and Carrie Masiello for sharing her expertise on the subject of black carbon. This research used samples and data provided by the Ocean Drilling Program (ODP). ODP is sponsored by the U.S. National Science Foundation (NSF) and participating countries under the management of Joint Oceanographic Institutions (JOI),

References (51)

  • MasielloC.A.

    New directions in black carbon organic geochemistry

    Marine Chemistry

    (2004)
  • MiddelburgJ.J. et al.

    Black carbon in marine sediments

    Marine Chemistry

    (1999)
  • ScottA.C.

    The Pre-Quaternary history of fire

    Palaeogeography Palaeoclimatology Palaeoecology

    (2000)
  • ScottA.C. et al.

    Fire across the K–T boundary: initial results from the Sugarite Coal, New Mexico, USA

    Palaeogeography Palaeoclimatology Palaeoecology

    (2000)
  • TurneyC.S.M. et al.

    Carbon isotope fractionation in wood during carbonization

    Geochimica Et Cosmochimica Acta

    (2006)
  • BowenG.J. et al.

    Mammalian dispersal at the Paleocene/Eocene boundary

    Science

    (2002)
  • BowenG.J. et al.

    A humid climate state during the Palaeocene/Eocene thermal maximum

    Nature

    (2004)
  • CollinsonM.E. et al.

    Cobham Lignite Bed and Penecontemporaneous Macrofloras of Southern England: A Record of Vegetation and Fire Across the Paleocene–Eocene Thermal Maximum

    (2003)
  • ColosimoA.B. et al.

    Evidence for lysocline shoaling at the Paleocene/Eocene Thermal Maximum on Shatsky Rise, Northwest Pacific

  • Cramer, B.S., 1998. The Late Paleocene Thermal Maximum, Bass River, NJ: Neritic response to a runaway greenhouse event....
  • CramerB.S. et al.

    An exceptional chronologic, isotopic, and clay mineralogic record of the latest Paleocene thermal maximum. Bass River, NJ, ODP 174AX

    Bulletin De La Societe Geologique De France

    (1999)
  • DamoahR. et al.

    Around the world in 17 days—hemispheric-scale transport of forest fire smoke from Russia in May 2003

    Atmospheric Chemistry and Physics

    (2004)
  • DickensG.R.

    Modeling the global carbon cycle with a gas hydrate capacitor: significance for the Late Paleocene Thermal Maximum

  • DickensG.R. et al.

    Dissociation of oceanic methane hydrate as a cause of the carbon-isotope excursion at the end of the Paleocene

    Paleoceanography

    (1995)
  • DuncanB.N. et al.

    Indonesian wildfires of 1997: impact on tropospheric chemistry

    Journal of Geophysical Research-Atmospheres

    (2003)
  • Cited by (33)

    • Decreased soil carbon in a warming world: Degraded pyrogenic carbon during the Paleocene-Eocene Thermal Maximum, Bighorn Basin, Wyoming

      2021, Earth and Planetary Science Letters
      Citation Excerpt :

      For example, in the Arctic, and even under wetter PETM conditions (Pagani et al., 2006), Denis et al. (2017) inferred from biomarker evidence that increased fire occurrence was linked to a climate-driven ecological shift to angiosperm-dominated vegetation. In addition, at different sites, evidence suggested that fire occurrence increased (Collinson et al., 2009), did not appear to change (Moore and Kurtz, 2008), or decreased (Boucsein and Stein, 2009) during the PETM relative to the Paleocene. A fifth study examined PAHs in a PETM section in the Piceance Basin in Colorado, USA, but similar to what we observed in the Bighorn Basin, organic matter degradation made it difficult to discern how wildfire occurrence changed during the PETM (Denis et al., 2021).

    • Enhanced magnetization of the Marlboro Clay as a product of soil pyrogenesis at the Paleocene–Eocene boundary?

      2017, Earth and Planetary Science Letters
      Citation Excerpt :

      The few specific charcoal results available for the CIE are even more ambiguous. High but variable concentrations of graphitic black carbon were found across the CIE interval at the Bass River site (Fig. 1) but there was no distinctive high concentration of charcoal at the CIE onset level (Moore and Kurtz, 2008). But in England there is a high abundance of charcoal in the CIE in the Cobham Lignite Bed described as unparalleled elsewhere in the Paleogene of southern England (Collinson et al., 2003).

    • Fire and ecosystem change in the Arctic across the Paleocene–Eocene Thermal Maximum

      2017, Earth and Planetary Science Letters
      Citation Excerpt :

      Based on our study and the few other studies of fire occurrence during the PETM, changes in fire occurrence varied by location. At IODP site 1210 (Shatsky Rise) in the west-central Pacific, there was no evidence of fire occurrence; at the New Jersey margin (Bass River section) there was evidence of fire, but no clear change in occurrence during the PETM (Moore and Kurtz, 2008). In England, fire occurrence decreased and was associated with a change in vegetation; in this case, a shift from an herbaceous fern and woody angiosperm fire-prone system to less ferns and woody plants, and increased wetland plants (Collinson et al., 2009).

    • Paleosol-based paleoclimate reconstruction of the Paleocene-Eocene Thermal Maximum, northern Argentina

      2017, Palaeogeography, Palaeoclimatology, Palaeoecology
      Citation Excerpt :

      The negative CIE is attributed to a large release of isotopically depleted carbon into the ocean-atmosphere system. Many mechanisms have been proposed for this release including the dissociation of methane hydrates from the seafloor (e.g. Dickens et al., 1995), explosive volcanism (e.g. Bralower et al., 1997), comet impact (e.g. Kent et al., 2003), and worldwide wildfires (e.g. Moore and Kurtz, 2008). Considering these hypothesized causes, modelers have estimated the amount and timing of carbon released to the atmosphere.

    • Long-term eustatic cyclicity in the Paleogene: a critical assessment

      2016, Proceedings of the Geologists' Association
      Citation Excerpt :

      The full strength of the Cenozoic glaciation was reached in the last few millions of years of this era (Ehlers and Gibbard, 2007), but some glaciation persisted through the Paleogene Period in the Southern Hemisphere, where relatively short-termed glacial advances occurred (Zachos et al., 2001; Miller et al., 2005a; Ruban et al., 2012), and the first ice sheet appeared in the Northern Hemisphere in the middle Eocene (Polyak et al., 2010). Two remarkable events were the Paleocene–Eocene Thermal Maximum and the early Eocene climatic optimum (Zachos et al., 2001; Kent et al., 2003; Cohen et al., 2007; Pearson et al., 2007; Weijers et al., 2007; Moore and Kurtz, 2008; Retallack, 2008; Gornitz, 2009; Roberts et al., 2009; Westerhold et al., 2009; Eberle et al., 2010; Hodgson et al., 2011; Bowen, 2013; Chew and Oheim, 2013; Hyland and Sheldon, 2013; Pujalte et al., 2014). There were some other potentially significant events such as the early Paleocene hyperthermal (Ali, 2009; Bornemann et al., 2009).

    View all citing articles on Scopus
    View full text