A consistent chronostratigraphic framework is required to understand the
effect of major paleoclimate perturbations on both marine and terrestrial
ecosystems. Transient global warming events in the early Eocene, at 56–54 Ma, show the impact of large-scale carbon input into the ocean–atmosphere
system. Here we provide the first timescale synchronization of continental
and marine deposits spanning the Paleocene–Eocene Thermal Maximum (PETM) and
the interval just prior to the Eocene Thermal Maximum 2 (ETM-2). Cyclic
variations in geochemical data come from continental drill cores of the
Bighorn Basin Coring Project (BBCP, Wyoming, USA) and from marine deep-sea
drilling deposits retrieved by the Ocean Drilling Program (ODP). Both are
dominated by eccentricity-modulated precession cycles used to construct a
common cyclostratigraphic framework. Integration of age models results in a
revised astrochronology for the PETM in deep-sea records that is now
generally consistent with independent
Early Eocene greenhouse climate on Earth was punctuated by transient global
warming events (Cramer et al., 2003; Kirtland Turner et al., 2014). The
Paleocene–Eocene Thermal Maximum (PETM or Eocene Thermal Maximum 1, ETM-1) at
55.93 Ma (Kennett and Stott, 1991; Koch et al., 1992; Bowen et al., 2001;
Zachos et al., 2005; Bowen et al., 2015) is the most pronounced hyperthermal
event, with a
The hyperthermal events in outcrops and ocean drill cores can be identified
by the characteristic negative carbon isotope excursions (CIEs), although
these differ in magnitude (McInerney and Wing, 2011; Sluijs and Dickens,
2012; Bowen, 2013). The CIEs are interpreted to be due to massive inputs of
The hyperthermals provide important evidence for understanding the dramatic
and long-lasting consequences of a rapid massive input of CO
Key records for studying the early Eocene climate and hyperthermals come from
carbonate-rich deep-sea drill cores from Walvis Ridge in the South Atlantic
(Zachos et al., 2005) and terrestrial fluvial deposits with paleosols in the
Bighorn Basin in Wyoming, USA (Koch et al., 1992; Bowen et al., 2001, 2015;
Abels et al., 2012, 2016). Deep-sea records have much lower sedimentation
rates, on the order of cm kyr
Because records from both realms show precession-dominated cyclicity, it should be possible to correlate and synchronize them. In summer 2011, the Bighorn Basin Coring Project (BBCP) drilled 900 m of overlapping cores from three sites covering the interval of the PETM and Elmo events (Clyde et al., 2013). The BBCP retrieved continuous unweathered material for common multiproxy studies. The main purpose of this report is to establish high-resolution age models for the BBCP drill cores based on cyclostratigraphy and integrate existing age models from outcrops. Second, these new BBCP age models will be combined with deep-sea records to synchronize and improve the available astronomical age model for the PETM and Elmo interval. The new age models will allow other studies to compare multiple proxy records from both realms at unprecedented temporal accuracy. Our study also allows us to consider whether the mammalian turnover called Biohorizon B in the Bighorn Basin fossil faunas (Schankler, 1980; Chew, 2009, 2015; Chew and Oheim, 2013) is synchronous with changes in deep-sea calcareous nannofossil assemblages (Agnini et al., 2007) significantly prior to the Elmo event.
The BBCP drilled late Paleocene to early Eocene fluvial deposits including the PETM at PCB and Basin Substation (BSN), and the Elmo interval at Gilmore Hill (GMH) (Fig. 1; Clyde et al., 2013). Two overlapping holes were drilled at BSN down to 138.4 mbs (BSN11-1A) and 138.6 mbs (BSN11-1B). At PCB two overlapping holes were drilled to 130 mbs (PCB11-2A) and 245.1 mbs (PCB11-2B). At GMH one hole was drilled down to 202.4 mbs (GMH11-3A) and a second down to 66.7 mbs (GMH11-3B). All cores were split and processed according to IODP standards that included visual core description, color and line scanning, sampling for post-party investigations at home laboratories, and archiving at the Bremen Core Repository during a BBCP Science Party at MARUM, University Bremen, Germany, in January 2012 (Clyde et al., 2013).
Location map for BBCP (Bighorn Basin Coring Project, Wyoming, USA),
ODP Leg 208 (Sites 1260 and 1263), and ODP Leg 113 (Site 690) on a 56 Ma
paleogeographic reconstruction in Mollweide projection (from
Here we present the results of processing and interpreting line scan images
and color reflectance data for the BSN, GMH, and PCB sites, with the records
from PCB being presented initially in Bowen et al. (2015). All cores from
PCB, BSN, and GMH were X-ray fluorescence (XRF) scanned over the course of 2012 at MARUM,
University of Bremen, and we use the iron (Fe) intensity data here as well.
XRF data were collected every 2 cm down-core using XRF core scanner 3
(Avaatech serial no. 12) at MARUM, University of Bremen, over a 1.2 cm
Pedogenic carbonate nodules for isotope analysis were identified as discrete,
small (
Polecat Bench PCB-A
All data and tables from this study are available through open access in the PANGAEA
database (
Weathering resulted in brighter, yellowish colors within the upper
Core images, color reflectance, and Fe data have been utilized to correlate
between parallel holes for PCB, BSN, and GMH drill cores. As normal shipboard
routine for multiple holes drilled by IODP, cores of the BBCP were offset
from the original drilling depth (mbs) and subsequently combined. The
correlation of the two PCB records and the resulting composite can be found
in Bowen et al. (2015, their Fig. 1). Lithological logs, core images, color
reflectance CIE
Evolutionary Power Spectral Density values
were calculated for the XRF core scanning Fe
intensities and color reflectance
Basin Substation BSN-A
At PCB, the dominant cycles in Fe intensity data are 8.2, 3.45, 1.2, 1.02, and
0.58 m long (Fig. 2), and in the
At BSN, a range of dominant cycles in Fe intensity are apparent (6.8, 3.3, 2,
1.7, 1.25, 1.05, and 0.58 m; Fig. 3). In the
At GMH the dominant cycles in Fe intensity and
PETM age model for Polecat Bench (PCB) and ODP Sites 690, 1262, 1263, 1265, 1266, and 1267. From left to right: precession cycle number of the extracted cycles as in Röhl et al. (2007) and this study, the equivalent depth in the Polecat Bench outcrop and the BBCP drill core, the respective depth of the assigned precession cycles at ODP sites, the relative age to the onset of the PETM assuming 21 kyr duration for each precession cycle counted, and finally the absolute age using 55.930 Ma as the age of the onset of the PETM as in Westerhold et al. (2007). Bold fond marks the onset of the PETM as identified in each drill core.
Gilmore Hill GMH-A
The obtained data allow the construction of a cyclostratigraphy for drill cores from GMH and PCB. We correlated GMH and PCB records with outcrop successions to allow full integration of Bighorn Basin surface-derived data with the drill cores.
For the PCB drill cores, a correlation to a composite outcrop (Abdul Aziz et al., 2008) is already available (Bowen et al., 2015), identifying well-known local marker beds consisting of red to purplish paleosols. The rhythmic stacking pattern of the paleosols is driven by orbital precession cycles and was successfully used to establish a cyclostratigraphy for the PETM at PCB (Abdul Aziz et al., 2008) roughly consistent with independent age models from deep-sea cores (Farley and Eltgroth, 2003; Röhl et al., 2007). For the PCB cores a set of pedogenic age models was developed, calculating time based on thickness, sediment type, and paleosol maturity of individual stratigraphic units (Bowen et al., 2015), but no cyclostratigraphy similar to the outcrop has been developed so far.
Cyclostratigraphy for Polecat Bench. From left to right: PCB-A (red)
and PCB-B (blue) soil nodule carbon isotope data (Bowen et al., 2015),
The cyclostratigraphy for PCB cores presented here is based on cycle counting
of both the
For each of the precession cycles we assume a constant duration of 21 kyr as
done in previous deep-sea cyclostratigraphic models (Röhl et al., 2007;
Westerhold et al., 2007) setting the onset of the PETM as 0. Relative ages
with respect to the onset of the PETM are given in Table 1. For absolute ages
we use the age for onset of the PETM as 55.930 Ma (Westerhold et al., 2007, 2015)
and add or subtract the relative age. According to
the filter of the
Cyclostratigraphy for Gilmore Hill. Left to right: GMH-A (red) soil
nodule carbon isotope data (Table S9) for stratigraphic reference,
Age model for Gilmore Hill (GMH) drill core and regional outcrops. From left to right: labeling ID for identified precession cycles (see text for details), the respective depth of the assigned precession cycles in the Upper Deer Creek, the Deer Creek, and the Gilmore Hill sections as well as the Gilmore Hill BBCP drill core. Precession cycle number as defined at ODP sites counting precession cycles (Westerhold et al., 2007) followed by the relative and absolute age to the onset of the PETM assuming 21 kyr duration for each precession cycle counted, using 55.930 Ma for the onset of the PETM (Westerhold et al., 2007) and using 54.050 Ma for the absolute age of ETM-2 (Westerhold et al., 2017a).
Extensive outcrop work and cyclostratigraphic interpretations are also
available for terrestrial sediments across the ETM-2 (Elmo, Lourens et al.,
2005) in the Bighorn Basin (Clyde et al., 1994, 2007; Abels et al., 2012,
2013, 2016; D'Ambrosia et al., 2017). Alluvial sedimentary
cycles before and after the Elmo are shown to be precession forced (Abels et
al., 2012). The GMH cores cover an interval prior to the ETM-2, which is cut
out by a sandstone channel complex (Clyde et al., 2013). To establish a
cyclostratigraphy for the GMH drill cores, precession and half-precession
cycles were extracted from the
This study not only provides a high-resolution cyclostratigraphic age model for the BBCP drill cores but also improves existing age models for the interval spanning the PETM and prior to ETM-2. BBCP drill core and Bighorn Basin outcrops are an ideal basis for revision of the PETM age model that is mainly derived from deep-sea sediments with low sedimentation rates. Dissolution of carbonate, particularly at the onset of the PETM, hampers the establishment of a complete chronology for the PETM from marine cores (Kelly et al., 2005, 2010). Uncertainty in correlation between terrestrial successions from the Bighorn Basin and marine records also makes it difficult to evaluate if the terrestrial Biohorizon B is synchronous with major changes in marine calcareous nannofossil assemblages prior to ETM-2 (Agnini et al., 2007). Bighorn Basin deposits of the Fort Union and Willwood formations with their extremely high sedimentation rates allow very detailed insight into changes in biota on land. Using our new astrochronology we can test if these biotic turnovers were coeval in marine and terrestrial ecosystems.
Overview for the Paleocene–Eocene Thermal Maximum (PETM) data from deep-sea records and the terrestrial Polecat Bench (PCB) drill core against age. Core images and lithology log (Gingerich, 2006) for PCB, core images of ODP Sites 1262, 1267, 1266, 1265, 1263, and 690 (aligned from left to right according to the water depth from deep to shallow), defined phases of events in the PETM (Röhl et al., 2007) on the new age model, extracted precession cycles using a Gaussian filter of the PCB XRF Fe intensity data, stable carbon isotope data from PCB soil nodules (Bowen et al., 2015), the deep-sea benthic foraminifera and bulk sediment (690 – Bains et al., 1999; Leg 208 – Zachos et al., 2005), and carbonate content (690 – Farley and Eltgroth, 2003; Leg 208 – Zachos et al., 2005). Letters indicate horizons as identified by Zachos et al. (2005) adjusted to the new age model for the deep-sea sites.
Precession-related variations in XRF core scanning data (Röhl et al.,
2000, 2007) and the concentration of extraterrestrial
At the Walvis Ridge ODP sites, the top of the clay layer coincides with the
top of the initial rapid recovery of the CIE (Recovery phase I in Murphy et
al., 2010). To correlate deep-sea and terrestrial records, the onset and the
top of the initial rapid recovery of the CIE are commonly used (McInerney and
Wing, 2011). The PCB cyclostratigraphy indicates that the duration of this
interval mentioned above covers six precession cycles or
Overview of data for the interval prior to the Eocene Thermal Maximum 2 (ETM-2) from deep-sea records and the terrestrial Gilmore Hill (GMH) drill core against age. Core images for GMH-A and B, core images of ODP Sites 1262, 1263, and 690 (aligned from left to right according to the water depth from deep to shallow), XRF Fe core scanning data from 1262 (red), 1263 (black), 690 (grey) (Westerhold et al., 2007), and GMH, extracted Gaussian filter of the GMH XRF Fe intensity data, stable carbon isotope data of soil nodules from the Gilmore Hill area (black – Gilmore Hill section; Abels et al., 2012, and D'Ambrosia et al., 2017; blue – GMH drill core), and the deep-sea benthic foraminifera (1262 – Littler et al., 2014) and bulk sediment (690 – Cramer et al., 2003; 1262 – Zachos et al., 2010). Position of Biohorizon B is after Abels et al., 2012, and D'Ambrosia et al., 2017 (black bar represents best estimate; gray bars represent conservative estimate – see text for discussion); the change in calcareous nannofossils (gray bar and text box) at ODP Site 1262 from Agnini et al. (2007).
As a result of the updated orbital chronology, the duration of the clay layer
is now estimated to be
The updated duration of the CIE in the deep sea using cyclostratigraphy is now in
agreement with the
Direct comparison of the
Eccentricity-modulated precession cycles dominate geochemical records in the interval prior to ETM-2 and have been used to establish cyclostratigraphic age models (Westerhold et al., 2007; Zachos et al., 2010; Littler et al., 2014). To synchronize marine and terrestrial records (GMH), we simply adopted the cyclostratigraphic ages from the deep-sea sections (Fig. 8, Table 2). Remarkably, the same eccentricity-related amplitude modulation of XRF Fe data can be observed for both records. Low-amplitude precession cycles are present from 54.65 to 54.48 Ma followed by high-amplitude cyclicity around 54.44 Ma, implying a climate system feedback to modulations in precession affecting both realms (Fig. 6). Variations in the XRF Fe data of the deep sea are most likely driven by changes in carbonate deposition. In the terrestrial sediments these variations have been attributed to large-scale reorganization of the fluvial system driven by astronomically forced changes in the hydrological cycle (Abels et al., 2012). Comparing the stable carbon isotope curves from deep-sea benthic foraminifera, bulk carbonate sediment, and soil nodules (Table S9) shows similar congruent variations, even outside the extraordinary hyperthermal events (e.g., ETM-2, Abels et al., 2016). Consistent small-scale variability is clearly linked to changes in the global carbon cycle.
Knowing that the age model for the deep-sea and the Bighorn Basin records are
synchronous, we can test for a temporal relation between Biohorizon B in the
Bighorn Basin (Clyde et al., 2007) and biotic changes in calcareous
nannofossils in deep-sea records. The best constraint on the stratigraphic
position of Biohorizon B comes from the Gilmore Hill section where it falls
between locality MP167 (LAD of
In sediments from ODP Site 1262 (Walvis Ridge), a series of biotic events,
representing rapid evolutionary change, are documented in the calcareous
nannofossil assemblage by the radiation of a “second generation” of apical
spine-bearing sphenolith species (e.g.,
Sedimentary records of Fe intensities, core images, and color reflectance
data were used to build composite records for the Bighorn Basin Coring
Project drill cores from Gilmore Hill, Basin Substation, and Polecat Bench.
Eccentricity-modulated precession-scale cyclicity observed in the
high-resolution data allowed the construction of cyclostratigraphic age
models for GMH and PCB spanning a 500 kyr interval prior to the ETM-2 and a
500 kyr interval across the PETM. The established orbital chronology of the
drill core data is not only consistent with previous age models from
outcrops but also helps to improve the cyclostratigraphic age model for the
PETM in deep-sea records. Synchronization and integration of all records
define a duration of
The data reported in this paper are tabulated in the
Supplement and archived in PANGAEA under
The supplement related to this article is available online at:
The authors declare that they have no conflict of interest.
We thank Editor Yves Godderis and the two anonymous reviewers for their effort and critical comments improving the paper. This research used samples and data provided by the International Ocean Discovery Program (IODP). IODP is sponsored by the US National Science Foundation (NSF) and participating countries. Financial support for this research was provided by the Deutsche Forschungsgemeinschaft (DFG) and the National Science Foundation (EAR0958821, EAR0958583, EAR1261312). The article processing charges for this open-access publication were covered by the University of Bremen. Edited by: Yves Godderis Reviewed by: two anonymous referees