Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 277–289

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology

journal homepage: www.elsevier.com/locate/palaeo

Chemostratigraphic age model for the Tornillo Group: A possible link

between fluvial stratigraphy and climate
Clément P. Bataille a,b,⁎, Dylana Watford b, Stephen Ruegg b, Alexander Lowe b, Gabriel J. Bowen b
a
University of North Carolina, Geological Sciences, 104 South Road, Mitchell Hall, NC 27599, USA
b
University of Utah, Department of Geology and Geophysics, Salt Lake City, UT 84112, USA

a r t i c l e i n f o a b s t r a c t

Article history: The expression of global climate trends (106 to 107 years) and events (105 years) in terrestrial sedimentary sec-
Received 7 January 2016 tions can only be assessed from long, continuous continental records. Such records are rare due to the paucity of
Received in revised form 9 June 2016 well-dated terrestrial deposits covering millions of years. This study uses isotope chemostratigraphy to develop
Accepted 16 June 2016
an age model for fluvial sedimentary rocks of the Tornillo Group (Tornillo Basin, TX) and to decipher the potential
Available online 18 June 2016
expression of global climate variations in this sub-tropical paleo-fluvial system. The carbon isotope ratio (δ13C) of
Keywords:
Tornillo Group pedogenic carbonate correlates well with δ13C variations from benthic foraminifera and suggests
Cretaceous deposition between ~69 Ma and ~52 Ma for the studied stratigraphic interval. Higher sediment accumulation
Paleogene rates occurred in the greenhouse period of the early Eocene than during the cooler middle Paleocene. Carbon iso-
K/Pg boundary tope excursions associated with Eocene hyperthermals were not identified, but their predicted stratigraphic po-
Hyperthermals sitions coincide with thick sand bodies. We interpret the long-term sediment accumulation rates and rapid shifts
Paleosols in facies distribution as reflecting change in sediment supply in the basin, possibly driven by fluctuations in pre-
Carbon isotope ratio cipitation seasonality and intensity driven associated with global temperature changes. Carbon and oxygen iso-
Pedogenic carbonate
tope ratios from pedogenic carbonates display a strong correlation throughout the section, interpreted as an
expression of coupling between the hydrological cycle and soil processes. Those observations suggest that global
climate variations, along with tectonic and eustatic sea-level changes, may be an important control of strati-
graphic variations of the Tornillo Group at multiple timescales.
© 2016 Published by Elsevier B.V.

1. Introduction timescales of 103 to 105 years (Blum and Törnqvist, 2000). At larger
timescale (N105 years), fluvial system dynamic and sedimentation are
Recognition and interpretation of fluvial responses to climate in the thought to be primarily controlled by tectonic processes (Blum and
geological record remains challenging because of the overprinting of au- Törnqvist, 2000). Identifying a potential influence of global climate for
togenic and allogenic controls on stratigraphic architecture at different these longer timescales requires continuous and well-dated terrestrial
timescales (Abels et al., 2013; Allen et al., 2014; Atchley et al., 2004; fluvial sequences.
Blum and Törnqvist, 2000; Hajek et al., 2012; Kraus, 2002; Olsen et al., In this work, we aim to compare well-recognized global temperature
1995; Paul Wright and Marriott, 1993; Shanley and McCabe, 1994). Nu- trends of the Cretaceous and Paleogene with the stratigraphic variations
merous investigators have attempted to correlate observed changes in observed within the fluvial sediments of the Tornillo Group in Big Bend
fluvial systems with known or inferred changes in climate (see review National Park (TX) to identify potential correlation between global cli-
in Blum and Törnqvist, 2000). For instance, rainfall amount and season- mate change and sedimentation (Fig. 1). A prerequisite for this work
ality is a critical control of sediment supply in sub-tropical basins (Cecil is to improve the chronostratigraphic framework of the Tornillo Group
et al., 2003), but unambiguous examples of this relationship within the (Rapp et al., 1983; Schiebout et al., 1987) to provide the temporal reso-
sedimentary record are limited. Our understanding of the interactions lution and accuracy necessary for investigating this correlation. We use
of climate and fluvial stratigraphy comes primarily from modern fluvial carbon isotope chemostratigraphy in combination with sedimentologic,
system (Cecil et al., 2003) or stratigraphic sequences spanning biostratigraphic, and magnetostratigraphic data to refine the age model
of the Tornillo Group and to assess the temporal continuity of sedimen-
tation. We then compare age constrained stratigraphic observations
with existing global temperature records of the early Paleogene to ex-
⁎ Corresponding author at: University of North Carolina, Geological Sciences, 104 South
Road, Mitchell Hall, NC 27599, USA. plore associations between climate and sedimentation in the Tornillo
E-mail address: [email protected] (C.P. Bataille). Basin.

http://dx.doi.org/10.1016/j.palaeo.2016.06.023
0031-0182/© 2016 Published by Elsevier B.V.
278 C.P. Bataille et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 277–289

Fig. 1. Location of field area, modified from Lehman (1986) A. Map showing the geographic location of the Tornillo Basin and associated structural features, the geographic extent of
Tornillo Group exposures and the location of the Tornillo Flat locality; B. Cross-section of Tornillo Group strata along the ʻA-A’ line in Fig. 1A and location of the Tornillo Flat locality.

2. Global climate variations during the early Paleogene greenhouse period of the Early Eocene Climatic Optimum (Zachos
et al., 2001). Numerous short-term transient climate events also oc-
Numerous proxy studies indicate that the early Paleogene was a curred during this interval of time and are linked to major perturbations
sustained interval of warm climate, with temperatures warmer than of the carbon cycle (Zachos et al., 2001). The Paleocene Eocene Thermal
preindustrial conditions in both marine and terrestrial settings (Huber Maximum (PETM) is the best-studied of these events (see review by
and Caballero, 2011). Hot global temperature drove an intensified hy- McInerney and Wing, 2011), but other events of smaller magnitude,
drological cycle in the early Paleogene due to the enhanced transport e.g., the Eocene Thermal Maximum 2 and 3 (ETM 2 and ETM 3), have
of water vapor from moisture divergence zones into convergence also been recognized globally (Abels et al., 2012; Abels et al., 2016;
zones (Huber and Caballero, 2011). Superimposed on this sustained Lourens et al., 2005). These large global temperature fluctuations of
greenhouse condition were global warming and cooling trends (106 to the early Paleogene, occurring over multiple timescales, likely contrib-
107 years) and events (104 to 105 years) that are well documented in uted to significant variations in rainfall amount and intensity because
stable isotope records from marine carbonates (Littler et al., 2014; of the dependence of the hydrological cycle on global temperature
Westerhold et al., 2011). The middle to late Maastrichtian (70 Ma– (Held and Soden, 2006).
66 Ma) constituted the termination of a long-term cooling trend
preceding the Cretaceous/Paleocene boundary (K/Pg) (Cramer et al., 3. Geological setting of the Tornillo Basin and stratigraphy of the
2009; Friedrich et al., 2012). The early Paleocene was marked by a Tornillo Group
rapid warming, reaching a maximum around 62 Ma, followed by a
long-term global cooling culminating between 59 Ma and 57 Ma 3.1. Geological setting of the Tornillo Basin
during the Paleocene Carbon Isotope Maximum (Littler et al., 2014;
Westerhold et al., 2011; Zachos et al., 2001). After 57 Ma, a gradual The geological names and definitions of the sedimentary units stud-
warming trend coincided with a decrease in δ13C values of marine car- ied in this paper are still debated (Lehman, 1988; Schiebout et al., 1987;
bonate and culminated between 54 Ma and 52 Ma with the super- Schiebout et al., 1988). Here we use the original classification defined by
 C.P. Bataille et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 277–289 279

Maxwell et al. (1967), with the Javelina Formation (Fm), the Black Peaks mudstones at the top (Atchley et al., 2004). Several factors had been
Fm and the Hannold Hill Fm constituting the Tornillo Group (Fig. 1). The previously invoked to explain these cyclical variations in stratigraphy
Tornillo Group was deposited in the Tornillo Basin, an asymmetric fore- of the Tornillo Group: tectonically driven changes in the rate of subsi-
land basin formed during the Laramide orogeny and bounded by the dence of the basin, or in the rate of uplift of the parent rocks, associated
Chihuahua Tectonic Belt and the Marathon Uplift. Sediments accumu- with Laramide orogeny uplift phases (Lehman, 1991; Schiebout et al.,
lated in the tectonically active Tornillo Basin during Cretaceous and 1988); climatically driven changes in the rate of sediment supply
Paleogene time. The basin is surrounded on several sides by folds and (Lehman, 1991; Schiebout et al., 1988); and eustatic sea level changes
thrust faults that were active during sediment accumulation, and controlling base level (Atchley et al., 2004; Lehman, 1991).
the Tornillo Group internally shows evidence of intra- and inter-
formational deformation that indicate active tectonism occurred within
3.3. Existing chronostratigraphy for the Tornillo Group
the basin itself (Lehman and Busbey, 2007). The Tornillo Basin fill is
composed of marine, paralic and fluvial sediments recording the gradu-
Evidence from radiometric dating, magnetostratigraphy, biostratig-
al (106 to 107 years) withdrawal of the Cretaceous Interior Seaway
raphy and chemostratigraphy suggests a Lancian age for the Javelina
(Lehman, 1991). The Tornillo Group tops this regressive sedimentary
Fm (Busbey et al., 1989; Lehman and Coulson, 2002; Lehman et al.,
sequence and was deposited at increasing distance from the shoreline
2006; Sankey, 2001; Sankey, 2005), late Puercan to Torrejonian ages
by a meandering fluvial system flowing southeast into the Laramide
for the lower Black Peaks Fm (Lawson, 1976; Standhardt, 1995),
Foreland Sea (Lehman, 1991; Maxwell et al., 1967; Schiebout et al.,
Tiffanian to potentially Wasatchian ages for the upper Black Peaks Fm
1987). The Tornillo Basin was disrupted by magmatic activity in the
(Rapp et al., 1983; Schiebout, 1974; Schiebout et al., 1987) and
Eocene (Chisos and Rosillos Mountains) and segmented into fault-
Wasatchian age for the Hannold Hill Fm (Rapp et al., 1983; Schiebout,
bounded exposures as result of Neogene Basin-and-Range extensional
1974; Wilson, 1967). The Tornillo Group spans a maximum age range
faulting (Lehman, 1991; Lehman and Busbey, 2007; Maxwell et al.,
of 72 Ma to 50.3 Ma, but the coarse resolution of this age model gives
1967).
rise to several key uncertainties. No definitive evidence of the K/Pg
boundary has been recovered in the Tornillo Group (e.g. iridium layer,
3.2. Stratigraphy of the Tornillo Group
spherules), making the presence and exact stratigraphic position of
this event ambiguous (Lehman, 1990). No early Puercan fauna have
The Tornillo Group is exposed in several places in Big Bend National
been recovered from the Black Peaks Fm, bringing into question the
Park and is overlain non-conformably by the Eocene Canoe Fm (Lehman
presence of early Paleocene strata (Lehman and Busbey, 2007;
and Busbey, 2007). Exposures of the Tornillo Group have been loosely
Standhardt, 1995). Several sandstones of the Tornillo Group (e.g. Exhibit
correlated using marker beds and biostratigraphy (Lehman and
Sandstone; log-jam) display clear evidence for erosional downcutting,
Busbey, 2007). The Javelina Fm and Black Peaks Fm display vertically
indicating the possibility of large unconformities throughout the section
aggrading floodplain mudstones, often developed into colorful
(Lehman, 1988; Schiebout et al., 1987; Schiebout et al., 1988).
paleosols, alternating with laterally aggrading channel sandstones
(Lehman, 1990). The Javelina Fm displays olive, purple, and light gray
mudstones and has a high proportion of thick tabular sandstones 4. Material and methods
which form characteristic ridges. The Black Peaks Fm is characterized
by a high proportion of recessively weathering mudstones with dark 4.1. Field work
gray, black, and red colors (Lehman, 1990). The contact between the Ja-
velina Fm and Black Peaks Fm is conformable and placed at the top of This work focuses on the west Tornillo Flat exposures (Fig. 1) be-
the highest ridge-forming sandstone (Lehman, 1990). The lower and cause of the extensive outcrops and large amount of stratigraphic,
upper sections of the Black Peaks Fm are separated by a distinctive mul- biostratigraphic, magnetostratigraphic and chemostratigraphic data
tistory sandstone complex yielding silicified fossil stumps commonly available in this area (Rapp et al., 1983; Schiebout, 1974; Schiebout
designated as the “log jam” sandstone (Lehman and Busbey, 2007; et al., 1987; White and Schiebout, 2004; White and Schiebout, 2008).
Wheeler, 1991). The Hannold Hill Fm exposures are limited to the To bracket the stratigraphic section, we identified the contacts between
Tornillo Flat area. The Hannold Hill Formation may have originally the Javelina Fm and Black Peaks Fm, Black Peaks Fm and Hannold Hill
been extensive over a broader area, but was removed by erosion prior Fm and Hannold Hill Fm and Canoe Fm using the Big Bend field guide
to deposition of the Canoe Formation in the rest of the park (Lehman (Lehman and Busbey, 2007) and maps and field notes from Thomas
and Busbey, 2007). The Hannold Hill Fm is dominated by ridge- Lehman (Turner et al., 2011). A series of stratigraphically continuous
forming conglomeratic sandstone beds, with distinctive extra-basinal sub-sections were measured along the Tornillo Creek bed and the Ex-
gravel inter-bedded with brown, pink and reddish mudstone intervals hibit Ridge using a 1.5 m Jacob's staff. In most cases, stratigraphic corre-
(Runkel, 1988). The contact between the Black Peaks Fm and Hannold lation between sub-sections was easily accomplished by tracing
Hill Fm is non-conformable but the amount of erosion associated distinctive marker beds (Fig. 2 and Appendix A).
with the basal Exhibit sandstone remains debated (Lehman, 1988; During field work, the section lines described in Rapp et al. (1983)
Schiebout et al., 1987; Schiebout et al., 1988). and Schiebout (1974) were followed to anchor the biostratigraphic
The deposits of the Javelina Fm and Black Peaks Fm are typical of a and magnetostratigraphic results of these authors within the strati-
meandering fluvial depositional system with laterally aggrading chan- graphic sections measured in this work. Data collected by these authors
nel sandstones alternating with stacked sequence of meter-scale, were tied to the stratigraphic column using a series of laterally persis-
upward-fining overbank deposits (Atchley et al., 2004; Lehman, 1988; tent marker beds (e.g. black paleosols and major sandstone units). Sev-
Lehman, 1990; Lehman, 1991). Meter-scale depositional cycles occur- eral dinosaur fossil sites producing Lancian assemblages were also
ring within decameter-scale depositional cycles are interpreted as included in the composite column. These sites were discovered within
representing successive episodes of avulsion of a channel migrating the upper Javelina Fm at “Sauropod Hills” south of the Canoe Syncline,
within the floodplain (Kraus, 1999). The Javelina Fm and Black Peaks ~ 500 m from the base of the composite section and within the same
Fm also contain stacked hectometer aggradational cycles characterized structural block (Hunt and Lehman, 2008; Lehman and Busbey, 2007;
by an upward increase in paleosol maturity, associated with changes Lehman and Coulson, 2002). The stratigraphic column detailed in
in fluvial deposition regime from braided (or meandering), multi- Lehman and Busbey (2007) was tied to our composite section using
story sandstones with limited overbank mudstones at the base, stratigraphic unit thickness and type to establish approximate correla-
to meandering, single-story sandstones with abundant overbank tions (Appendix A).
280 C.P. Bataille et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 277–289

Fig. 2. Composite stratigraphic column for the Tornillo Group at west Tornillo Flat compiled from a series of measured sub-sections (Fig. 1). Arrows indicates uncertainty in the correlation
between sub-sections. The location of each sub-section is indicated on Fig. A1. Additional details about the correlation uncertainty are given in Appendix A.

4.2. Pedogenic carbonate nodules origins. The depth of nodule formation could not be easily estimated
due to the difficulty of identifying paleosol tops in these aggradational
Pedogenic carbonates from paleosols have been widely used as a floodplain deposits, where pedogenesis occurred nearly continuously.
substrate for δ13C chemostratigraphy in terrestrial environments
(Cerling, 1984; Cerling and Quade, 1989; Cerling et al., 1992; Sheldon 4.3. Carbon isotope chemostratigraphy
and Tabor, 2009). Carbon in pedogenic carbonate nodules is derived
from soil CO2, which is a mixture of atmospheric CO2 and organic- A fundamental assumption underlying the use of pedogenic carbon-
respired CO2. While many factors can influence the δ13C values of soil ate nodules as a substrate for carbon isotope chemostratigraphy is that
CO2 (e.g., soil drainage, soil respiration rates, organic matter δ13C their δ13C variation is primarily controlled by variations in global atmo-
value), data from modern and ancient soils show that δ13C values of sphere/ocean δ13C values, and can thus be pattern-matched with δ13C
soil carbonates faithfully record secular and transient δ13C changes in values from marine carbonates (Cerling, 1984). To apply carbon isotope
the exogenic carbon cycle (Cerling, 1984; Sheldon and Tabor, 2009). chemostratigraphy, the calcite material analyzed needs to be precipitat-
Similarly, δ18O values of pedogenic carbonates provide insights ed at the time of soil formation and not during recrystallization events
into soil water δ18O variations and/or soil temperature fluctuations or from recycled material. Change in soil texture and drainage, as well
(Cerling, 1984; Sheldon and Tabor, 2009). Most of the mudstones as conditions and depth of nodule formation or biologically-driven
from the Javelina Fm, Black Peaks Fm and Hannold Hill Fm are δ13C variations in soil organic matter, can also impact the δ13C values
pedogenically modified, as indicated by lack of primary depositional of pedogenic carbonates (Breecker et al., 2009; Cerling and Quade,
textures, strong color variegation, presence of bioturbation features, 1989). The δ13C values from Tornillo Group pedogenic carbonates
and abundant cm-scale pedogenic carbonate nodules. Pedogenic car- were screened to identify samples with isotope values that were likely
bonate nodules were collected in paleosols throughout the section and influenced significantly by factors other than global exogenic atmo-
used as substrate for δ13C and δ18O analysis. At least two nodules were sphere/ocean δ13C variations. Tests were developed to assess the poten-
recovered at each sampling site. GPS coordinate, stratigraphic position, tial influence of: 1) soil type and soil drainage on δ13C values by
color and soil texture of each sampled paleosol were recorded (Fig. 1, searching for correlation between the isotopic composition of pedogen-
Table S1). Nodules were collected in situ from freshly exposed matrix ic carbonates and paleosol color and grain size, 2) condition of pedogen-
by digging ~ 50 cm deep trenches. Nodules present in sandstones or ic carbonate nodule formation and diagenetic history by categorizing
channel lags were not collected due to their potential exogenous nodules as a function of their size, shape, morphology and mineralogy
 C.P. Bataille et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 277–289 281

and, 3) diagenetic or recycling history by comparing the isotopic com- 440 m and covers the upper Javelina Fm and the entire Black Peaks
positions of different mineral phases from the pedogenic carbonates. Fm and Hannold Hill Fm. The bottom 95 m of the composite stratigraph-
The results of these tests are reported below (detailed results in Appen- ic column is assigned to the Javelina Fm. The contact between the Jave-
dix B) and were used to eliminate data from specific samples that are lina Fm and the underlying Aguja Fm was not identified because at west
likely affected by these local processes. The analysis provides strong Tornillo Flat the lower Javelina Fm is covered by Quaternary deposits,
support that the remaining data reflect a global signal of δ13C change thus the composite section starts part way up the Javelina Fm. The con-
suitable for chemostratigraphic correlation. tact between the Javelina Fm and Black Peaks Fm is conformable and
corresponds to the top of the last ridge-forming sandstone at 95 m in
4.4. Nodule characterization and sampling the composite section. The contact is easily identified as it is character-
ized by a clear stratigraphic transition between ridge-forming sand-
Nodules were sliced, polished using a lapidary wheel and 600-grit stones of the Javelina Fm and colorful, recessively weathering
silica carbide powder, washed and dried. The color, mineral composi- mudstones of the Black Peaks Fm. The Black Peaks Fm corresponds to
tion, and fabric of each nodule were recorded under a binocular micro- levels 95 m to 377 m in the composite stratigraphic column. The Black
scope (Appendix B). For each nodule, a mounted dental drill was used to Peaks Fm was separated into two lithostratigraphic sub-sections: the
collect 150 μg (±10%) of primary micritic carbonate for δ13C and δ18O lower Black Peaks Fm between 95 m and 185 m and the upper Black
analysis. Some nodules displayed visually distinct micrite phases Peaks Fm between 185 m and 377 m. The separation between these
(Appendix B). For 66 individual nodules, these phases were analyzed sub-sections corresponds to the “log jam” sandstone, which occurs be-
to assess internal variability. Most of the nodules contained both prima- tween 185 m and 200 m (Fig. 3C). The contact between the Black
ry micrite and diagenetic spars. Incorporation of spar in the micrite sam- Peaks Fm and Hannold Hill Fm is unconformable and corresponds to
ple can strongly affect the isotope results because of the distinct isotopic the erosive base of the Exhibit Ridge sandstone at 377 m (Fig. 3E). The
composition of this material. To avoid contamination, nodules were stratigraphic interval between 377 m and 440 m covers the entire
drilled in areas distant from any visible spar or recycled material and Hannold Hill Fm and is unconformably overlain by the Big Yellow Sand-
only drilled superficially to avoid potential inadvertent contamination stone Member of the Canoe Fm (Fig. 3F).
when drilling below the slabbed surface. Where possible, diagenetic The Tornillo Group displays some clear stratigraphic patterning with
spar was also drilled and analyzed for δ13C and δ18O values to character- some intervals dominated by fine-grained overbank deposits (0 m–
ize this secondary phase. 30 m, 95 m–155 m, 215 m–265 m, 280 m–322 m and 326 m to
377 m) and others dominated by coarse-grained channel deposits
4.5. Isotope analysis (30 m–95 m, 155 m–215 m, 265 m–280 m and 377 m–440 m)
(Fig. 3A). Intervals dominated by overbank deposits are characteristic
The collected carbonate powder was placed into an Exitainer vial of a meandering fluvial system, with vertically aggrading mudstone
and reacted with 100% phosphoric acid at 75 °C for at least 1 h. The and rare fluvial channels (Fig. 3D). The mudstones from these intervals
CO2 generated from this reaction was chromatographically purified are often pedogenically modified (Fig. 3B). They display purple, red or
using a GasBench II device and analyzed on a Delta-Plus gas-source purple-red B-horizons, evidence of bioturbation, mottling and centime-
mass spectrometer. Data from each analysis batch were corrected ter sized slickenslides. Most yield millimeter to decimeter-sized pedo-
for peak-area effects and calibrated to international reference scales genic carbonate nodules. Such paleosols are equivalent to modern
using a two-end member linear calibration equation based on mea- vertisols or alfisols (Lehman, 1990). Other paleosols within these inter-
sured values for co-analyzed reference carbonates (Carrera Marble, vals display a uniform black color with meter scale slickenslides, manga-
δ13 C = 2.09‰, δ 18 O = − 1.85‰, and LSVEC, δ 13C = − 46.6‰, nese nodules and abundant gypsum sheets. While gypsum is a product
δ18 O = − 26.7‰). Analytical precision, determined as the standard of pyrite oxidation and not an original pedogenic feature, its presence
deviation of multiple measurements of a secondary reference calcite does suggest organic-rich, anoxic conditions occurred during pedogen-
within each analysis batch, was generally b0.1‰ for both carbon and esis. These paleosol proprieties resemble modern histosols (Lehman,
oxygen isotope ratios. 1990). Above 328 m, the upper Black Peaks Fm displays a clear change
in stratigraphy. Black layers are no longer present and the stratigraphy
4.6. Chemostratigraphic correlation, age model and deposition rates becomes dominated by thick (5–10 m), bright red paleosols with no dis-
tinct horizonation. In this interval pedogenic carbonate nodules are rare
Available biostratigraphic and magnetostratigraphic data were used and millimeter-sized. In the Hannold Hill Fm, sediments are coarser and
in conjunction with isotope chemostratigraphy from carbonate nodules paleosols are rare and weakly developed. Some paleosols of the
to refine the age model for the Tornillo Group. Tie-points were identi- Hannold Hill Fm yield centimeter-sized pedogenic carbonate nodules.
fied between the Tornillo Group composite section and the stacked Pa- Intervals dominated by coarse-grained channel deposits are more
leogene marine records from ODP sites 1209 (Westerhold et al., 2011) variable stratigraphically, featuring channel sandstones interbedded
and 1262 (Littler et al., 2014) (Fig. 3C and D). Those tie-points represent with yellow, tan and brown siltstones (Fig. 3A). Siltstones and mud-
either 1) the beginning or the end of a magnetochron or 2) a recogniz- stones are usually weakly pedogenically modified and do not yield ped-
able chemostratigraphic marker constrained by magneto- and biostrat- ogenic carbonate nodules. Some of the yellow mudstones present in
igraphic data. The chronology from the marine stacks, based on orbital these interval yield gypsum sheets and/or radial fibrous nodules. The
tuning of 405 kyr eccentricity cycles expressed in the δ13C record, was stratigraphy of these sequences varies laterally due to the incision and
used to establish an absolute chronology for the Big Bend section. lateral migration of channels. The Hannold Hill Fm is characterized by
Using this information allowed us to calculate non-decompacted an increased proportion of conglomeratic sandstones. The Big Yellow
mean sediment accumulation rates for each interval of the Big Bend sec- Sandstone Member has a sheet-like geometry with common channel
tion by linear interpolation between tie-points. lags, and marks a transition from the meandering stream deposits char-
acteristic of the Tornillo Group to a braided stream deposit characteristic
5. Results of the Canoe Fm (Lehman, 1991) (Fig. 3F).

5.1. Stratigraphy 5.2. Pedogenic carbonate nodules

A composite stratigraphic column was compiled from the different Eight hundred and eleven pedogenic carbonate nodules were col-
sub-sections measured at west Tornillo Flat (Fig. 2). The section spans lected throughout the Tornillo Flat section, representing a minimum of
282 C.P. Bataille et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 277–289

Fig. 3. Photographs of selected sub-sections at Tornillo Flat (Fig. 1). A. Exposure of upper Javelina strata at TF4-2012 sub-section with mudstone-dominated intervals displaying purple
paleosol horizons and a sand-dominated interval with amalgamated channel sandstones and downcutting, B. Paleosol profile in the Lower Black Peaks strata at TF5-2012 sub-section,
with horizons characterized by slickensides (Bss), carbonate nodule accumulation (Bk), and coarser-grained weakly pedogenically modified sediment (C), C. Upper “log jam”
sandstone and transition between lower and upper Black Peaks Formation at TF3-2012 sub-section, D. Exposure of upper Black Peaks strata with vertically aggrading black and purple
paleosols and laterally aggrading channel facies at TF2-2012 sub-section, E. Uppermost Black Peaks Formation, with thick red paleosol sequence, and lower Hannold Hill Formation,
separated by the Exhibit sandstone at TF2-2010 sub-section, F. Upper Hannold Hill Formation and Big Yellow Sandstone of the Canoe Formation at TF3-2014.

190 paleosols (Table S1). The density of carbonate nodule sampling was particularly within the upper Javelina Fm. These nodules have spherical
strongly influenced by stratigraphy. Carbonate nodules were rare or ab- shapes, vary in size from 7 mm to 5.6 cm, and display radial fibrous in-
sent in black and yellow paleosols and in the thick red paleosols of the ternal texture (Appendix B).
upper Black Peaks Fm, which limits the resolution of the chemo-
stratigraphic record in those intervals.
Two types of pedogenic carbonate nodules were distinguished based 5.3. Isotope measurements
on their morphology and stratigraphic associations. The first type is
found in colorful, pedogenically-modified overbank deposits. These Carbon and oxygen isotope measurements from pedogenic carbon-
nodules have irregular, spherical, ellipsoidal or tubular shapes and ates of the Tornillo Group are listed in Table S1. The average standard
vary in size from 2 mm to 10.5 cm. They are dominated by micritic cal- deviations of multiple δ13C and δ18O measurements of individual
cite with varying concentrations of occluded clay and silt grains from nodules are 0.20‰ and 0.28‰, respectively (n = 134). The average
the mudstone matrix, giving a homogeneous to mottled appearance. standard deviations of δ13C and δ18O measurements from multiple
Some of these nodules also contain veins of macrocrystalline calcite or nodules collected at the same sampling locality are 0.04‰ and 0.27‰
sparite, which can represent up to ~30% of the total calcite. The second (n = 8). These values are small relative to the range of isotopic varia-
type of carbonate nodule is found in yellow and gray mudstones, tions observed among sampling sites.
 C.P. Bataille et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 277–289 283

The δ13C values from pedogenic carbonates (nodule averages) range water saturated conditions relative to other paleosols of the Black
from −16.4‰ to −3.3‰, and δ18O values range between −13.2‰ and Peaks Fm (White and Schiebout, 2008). We interpret the lower δ13C
3.3‰ (Fig. 4). δ13C and δ18O values of microcrystalline calcite are limited and δ18O values as representing a stronger influence of soil-respired
to only part of this range, clustering between −13.2‰ and −8.4‰ and CO2 and smaller influence of soil-water evaporation in the poorly
−6.4‰ and − 2.5‰, respectively. A strong correlation exists between drained soils in which the nodules formed. Although the average offsets
δ13C and δ18O values of micrite samples (R2 = 0.52). are small, our anlaysis suggests that the isotopic values of pedogenic
The average δ18O value of spar samples is much lower than that of carbonate nodules from black paleosols are not directly comparable to
micrite samples (−8.6‰ and −4.3‰, respectively; t-test, p = 1E-14), those from well-drained paleosols, and may not be robustly correlated
likely reflecting the higher temperature of formation of the diagenetic with δ13C values from marine carbonates. Data from these nodules
spars. Isotopic values for spar samples are widely dispersed, which were also excluded from the chemostratigraphic record (Appendix B).
could be a reflection of either multiple episodes of recrystallization dur- The C and O isotopic composition of pedogenic carbonates from the
ing burial or contamination of sparry calcite by micrite calcite during Tornillo Flats section, following screening for outliers and compilation of
microdrilling. Although the majority of micrite samples have isotopic a composite section (see Discussion), show strong and coherent strati-
compositions that cluster as distinct from the spars, several micrite sam- graphic variation (Fig. 5). Patterns of δ13C and δ18O are strongly correlat-
ples displayed δ18O values which fall out of the micrite isotopic cluster ed throughout the section (Fig. 4). Values drop from a maximum near
(Fig. 4). Those outliers are possibly associated with contamination by the base of the section to a distinct minimum within the lowermost
spar material during microdrilling. Given this possibility, a numerical part of the Black Peaks Fm, with a substantial sampling gap cutting
screening approach similar to that described by Bowen et al. (2015) out the record from much of this lowest ~ 100 m of the composite
was applied to remove such contaminated outliers from the dataset section. Values gradually increase through the Black Peaks Fm, with a
that may not primarily record the isotopic composition of ancient soil broad maximum occurring across the upper ~1/3 of the Formation. Sev-
CO2. Data from individual nodules were excluded from the composite eral coherent, short-term fluctuations in δ13C values are superimposed
record if their δ13C and δ18O values fell outside of the range defined by on this broader trend. Within the uppermost ~ 40 m of the Black
the mean ± four standard deviations of the surrounding six samples Peaks, δ13C values in particular, begin to decline again toward a second-
(three stratigraphically above and three below) (Fig. 4) (Appendix B). ary minimum within the Hannold Hill Fm.
The average δ13C and δ18O values of radial fibrous nodules (−7.1‰
and − 9.2‰, respectively) differ from those of pedogenic carbonates 6. Discussion
(− 10.4‰ and − 4.3‰, respectively; t-test, p = 3E-05 and 0.0001).
The origin of the morphologically distinct radial nodules is unclear, 6.1. Age model
but they are often associated with stratigraphic intervals dominated
by channel facies. The distinct isotopic values suggest that data from fi- 6.1.1. Javelina Formation
brous nodules are not directly comparable to those derived from other
pedogenic carbonate nodules, and thus may not be robustly correlated 6.1.1.1. Existing age model. Dinosaur biostratigraphy, magneto-
with δ13C values from marine carbonates. These nodules were thus ex- stratigraphy and U-Pb dating have given a Lancian age (69 Ma to
cluded data from the chemostratigraphic record (Appendix B). 66 Ma) for the Javelina Fm (Hunt and Lehman, 2008; Lehman and
Nodule micrite values did not vary as a function of paleosol color Busbey, 2007; Lehman et al., 2006). The base of the composite section
with the exception of those from black paleosols, which had slightly from this work is roughly equivalent to the TMM 41361 site, which
lower mean values (−11.0‰ and −5.1‰ for δ13C and δ18O respective- yielded a Torosaurus utahensis and is thought to be ~69 Ma. Sites TMM
ly) than those of nodules from non-black paleosols (− 10.3‰ and 41047, TMM 41398 and TMM41436 correlate approximately to the
−4.3‰; t-test, p = 0.02 and 0.01). Soil drainage can have a significant ~70 m level of the composite section (Fig. 5). Those sites yielded Lancian
impact on the isotopic composition of pedogenic carbonates by modify- assemblages including Quetzacoaltus, Almausaurus sanjuanensis and
ing the proportion of isotopically distinct soil respired CO2 and atmo- Tyrannosaurus. The last white sandstone at the top of the Javelina Fm
spheric CO2 within the soil (Mintz et al., 2011). The black paleosols are is potentially late Cretaceous in age because it has yielded Lancian
characterized by high content of smectite, gypsum sheets and/or man- assemblages in other exposures of the Tornillo Group (Pterodactyl
ganese nodules, and are interpreted as indicating more reduced and Ridge, Dogie mountains and Dawson Creek sections) (Lehman and
Busbey, 2007).

6.1.1.2. Refined age model. The average δ13C values from the Javelina Fm
are higher than the average δ13C values from the lower Black Peaks Fm.
Those higher values can be correlated with the high δ13C values of ben-
thic foraminifera for the late Maastrichtian, from 69 Ma to 66 Ma
(Cramer et al., 2009). At higher resolution, the first 30 m of the compos-
ite section display δ13C variations which correlate well the Mid-
Maastrichtian Event (MME), the coolest period of the late Cretaceous,
occurring between 69 Ma and 68 Ma (Cramer et al., 2009). In the ben-
thic foraminifera the high δ13C values of the MME are separated by a
brief but pronounced double negative inflection in the δ13C record
(Cramer et al., 2009; Voigt et al., 2012). We observe similar negative
inflections between 20 m and 30 m, where δ13C values reach − 11‰.
We thus interpret the first 30 m of the composite section to have
been deposited between ~ 69 Ma (tie point 1 at 0 m) and ~ 67.5 Ma
(tie point 2 at 30 m), giving a similar age to that determined by
chemostratigraphy for the upper Javelina Fm in the Dawson Creek
area (Nordt et al., 2003). The MME occurs within the C31n
Fig. 4. Bivariate plot of carbon and oxygen isotope values for carbonate material collected
at west Tornillo Flat. The dashed line corresponds to the linear regression model between
magnetochron, and collecting magnetostratigraphic data within the
the carbon and oxygen isotope values of micrite samples that passed the data screening Javelina Fm could further resolve the age model of this interval at
procedures (green triangles). West Tornillo Flat. The absence of pedogenic carbonate nodules within
284 C.P. Bataille et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 277–289

Fig. 5. Stratigraphy and age model at west Tornillo Flat, correlated with global chronostratigraphic data. A. Composite stratigraphic section at west Tornillo Flat, with magnetostratigraphic
and biostratigraphic data anchored to the composite section (Rapp et al., 1983; Schiebout et al., 1988). Detailed legend and grain size for the composite section is given in Fig. 2. For the
magnetostratigraphic data, gray shading corresponds to unstudied intervals. For the biostratigraphic data, bone symbols correspond to the stratigraphic position of fossil localities. The
associated letters and numbers correspond to the interpreted North America Land Mammal Ages (NALMA) (Woodburne, 2004). La = Lancian; To = Torrejonian; Ti = Tiffanian;
Wa = Wasachtian. The Lancian sites are from the “Sauropod Hills” site and are approximately anchored to the composite section based on lithostratigraphic correlation (Appendix A,
Fig. A2) (Lehman and Busbey, 2007; Lehman and Coulson, 2002) B. and C. Carbon and oxygen isotope measurements at west Tornillo Flat from paleosol micrite, with selected age tie-points
1 through 10. Tie points 3 and 3′, 4 and 4′ correspond to age model options 1 and 2, respectively. Red line corresponds to a centered 5 point moving average. D. Carbon isotope measure-
ments from benthic foraminifera from ODP site 1209 (Westerhold et al., 2011) and ODP site 1262 (Littler et al., 2014) with selected age tie-points, E. Timescale with geologic epochs and
ages, global magnetic polarity zones, and North American Land Mammal Ages.

the interval deposited between 30 m and 95 m hampers the application and Busbey, 2007). The top of this interval has yielded a fossil assem-
of carbon isotope chemostratigraphy within this interval. If the white, blage from sites TM41377 and C-Con, at 205 m and 205 m to 210 m
ridge-forming sandstone topping the Javelina Fm is effectively Lancian within the stratigraphic section, respectively, which is interpreted as
in age (Lehman and Busbey, 2007) then this interval spans a maximum late Torrejonian to early Tiffanian (Schiebout, 1974). The magnetic po-
age range from ~67.5 Ma to ~66 Ma. However, the Lancian age assigned larity zone beginning at 185 m in the section has been assigned to
to this sandstone is based on fossils collected in float and it is possible C26r by Rapp et al. (1983). Elsewhere in the park, the stratigraphic in-
that this sandstone is younger in age and represents a major erosional terval below the “log jam” sandstone has yielded few vertebrate fossils.
unconformity. The only known vertebrate fossil sites include TMM43380 at Grapevine
Hills, Dogie site and TMM42327 at Dogie Mountains and TMM41400 at
6.1.2. Lower Black Peaks Formation Dawson Creek (Lehman and Busbey, 2007; Standhardt, 1995). At three
of those sites only Torrejonian fauna were recovered; at Dogie site
6.1.2.1. Existing age model. No biostratigraphic and magnetostratigraphic Torrejonian fauna coexisted with purported late Puercan faunal ele-
data were available between 95 m and 185 m in the stratigraphic col- ments. Lehman and Busbey (2007) used those biostratigraphic data
umn (Fig. 5) because the work of Rapp et al. (1983) and Schiebout and correlation between sections to place the K/Pg boundary at ~30 m
(1974) did not include most of the lower Black Peaks Fm (Lehman above the base of the Black Peaks Fm at west Tornillo Flat.
 C.P. Bataille et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 277–289 285

6.1.2.2. Refined age model. The δ13C values from pedogenic carbonates of Wasatchian. TMM40143 (level 377 m to 382 m) has yielded significant
the lower Black Peaks Fm are on average lower than those observed for assemblages of Wasatchian fauna including Hyracotherium, Coryphodon
the Javelina Fm. If the lower 30 m of the Black Peaks Fm was Creta- and Lambdotherium. The magnetostratigraphic data from this interval
ceous in age as indicated by Lehman and Busbey (2007), δ13C values show two thin intervals of normal polarity, interpreted as C26n and
similar to those of the Javelina Fm should be present because marine C25n, with the rest of the interval having a reversed magnetic polarity
records suggest that ocean/atmosphere δ13C values remained high (Schiebout et al., 1987).
during the late Maastrichtian (Cramer et al., 2009). Instead, the low
δ13 C values of this stratigraphic interval correlate better with the
low δ13C values from benthic foraminifera for the early Paleocene. 6.1.3.2. Refined age model. The chemostratigraphic data from pedogenic
Based on biostratigraphy (absence of Torrejonian elements above carbonates support the age model proposed by Schiebout et al. (1987)
the 210 m level), magnetostratigraphy (reversed polarity) and for this stratigraphic interval. The δ13C values from pedogenic carbon-
chemostratigraphy (continuous increase in δ13C), we correlate the ates are higher on average in this interval than for the lower Black
210 m level of the section to the Torrejonian-Tiffanian boundary Peaks Fm and increase up section, matching the δ13C pattern of the
at 60.2 Ma (tie point 5 at 210 m). Below 185 m, the absence of Paleocene Carbon Isotope Maximum from the benthic foraminiferal re-
constraints other than chemostratigraphy makes the age cord. Between 210 m and 251 m, δ13C values from pedogenic carbonates
assessment challenging. We suggest two possible options using display a two-step increase and reach a maximum of ~−9‰ at 255 m,
chemostratigraphic correlations (Fig. 5): just above the top of C26n. This two-step increase correlates well with
the δ13C values from benthic foraminifera, with the maximum δ13C
- Option 1: the δ13C values observed between 95 m and 155 m could
value reached at 58.7 Ma just above the top of C26n. Between 251 m
correspond to the low marine δ13C values following the K/Pg bound-
and 290 m, levels which correspond to the top of C26n (58.8 Ma) and
ary between 66 Ma (tie point 3 at 95 m) and 62.8 Ma (tie point 4 at
the top of C25n (57.05 Ma), respectively, δ13C values from pedogenic
155 m), with the double-peaked excursion observed ~115 m corre-
carbonates remain constant around −9‰, matching the 2 million year
sponding to the Dan-C2 hyperthermal (Coccioni et al., 2010;
long “body” of the Paleocene Carbon Isotope Maximum. We selected
Quillévéré et al., 2008). The δ13C values observed between 155 m
two tie points within this interval, including the top of C26n at
and 210 m could correspond to the low δ13C values between
58.8 Ma (tie point 6 at 251 m), and the top of C25n at 57.05 Ma (tie
62.8 Ma and 60.2 Ma (tie point 5 at 210 m) with the minimum ob-
point 7 at 290 m).
served at ~170 m corresponding to the Latest Danian Event (LDE)
δ13C values from pedogenic carbonates begin to decrease after
(Bornemann et al., 2009). In this scenario the last sandstone of the
290 m, reaching ~−10‰ at 310 m. The δ13C values from pedogenic car-
Javelina Fm corresponds approximately to the K/Pg boundary and
bonates display a series of three ~ 1–2‰ carbon isotope excursions
no major erosional unconformity is present in the early Paleocene.
(CIEs) similar to those observed in the benthic foraminifera record at
- Option 2: the δ13C values observed between 95 m and 155 m could
56.9 Ma, 56.6 Ma and 56.3 Ma. This interval had previously been
correspond to the low marine δ13C values between 62.8 Ma (tie
assigned to the PETM (White and Schiebout, 2004), but as in that
point 3′ at 95 m) and 61.4 Ma (tie point 4′ at 155 m), with the
study our data show a pattern of highly variable carbon isotope fluctu-
double-peaked excursion at ~115 m corresponding to the LDE. The
ations that better match the period of enhanced δ13C variability preced-
δ13C values observed between 155 m and 210 m could correspond
ing the PETM in the marine record. We interpret the third CIE, observed
to the transient drop in marine δ13C values recorded between
at 310 m, as correlative with a minor CIE in the marine record at 56.3 Ma
61.4 Ma and 60.2 Ma. In this scenario, a major unconformity of at
(tie point 8 at 310 m). The last 12 m of this sequence represent the black
least 3 million of years is present within the section. No evidence
“G” layer from White and Schiebout (2004) and did not yield pedogenic
for a major erosional surface was observed in the first 30 m of the
carbonate nodules. We interpret the thick sandstone strata deposited
Black Peaks Fm, but the unconformity might be associated with a hi-
between 322 m and 328.5 m as representing the Paleocene-Eocene
atus in sedimentation in the early Paleocene. Another possibility is
transition at 55.7 Ma. Between 328.5 m and 377 m, δ13C values from
that the last sandstone of the Javelina Fm is Paleocene in age and
pedogenic carbonates show a progressive decrease from ~−9.5‰ to ~
eroded the Puercan and a portion of the Lancian, which would ex-
−11.5‰. This decreasing trend correlates well with the δ13C decrease
plain the Lancian fauna recovered from this bed. In this scenario,
observed in the marine record in the earliest Eocene.
the K/Pg boundary might not be present in this section.

6.1.4. Hannold Hill Formation

In both scenarios, it is also possible that the erosional surfaces
observed at the base of the sand body at level ~ 165 m and at the 6.1.4.1. Existing age model. Vertebrate localities within the Hannold Hill
base of the “log-jam” sandstone at level ~ 185 m correspond to signif- Fm at west Tornillo Flat include TMM41221 (Exhibit sandstone) and
icant time unconformities. The low resolution of carbon isotope TT-Jacks (second conglomeratic sandstone of the Hannold Hill Fm)
chemostratigraphy for this interval is not sufficient to fully resolve and indicate a late Wasatchian age. Unconformably overlying the
the age model of the lower Black Peaks Fm. Magnetostratigraphic Hannold Hill Fm is the Canoe Fm, which contains Bridgerian fauna
work assessing the presence of C29n, C28n and C27n could resolve (Wilson, 1967). Rapp et al. (1983) identified one interval of normal po-
the age uncertainty for the lower Black Peaks Fm. larity just above the Exhibit sandstone, which they interpreted as C24n.

6.1.3. Upper Black Peaks Formation 6.1.4.2. Refined age model. Between the base of C24n at 386.5 m (tie point
9, 54.1 Ma) and the top of the Hannold Hill Fm at 440 m in the compos-
6.1.3.1. Existing age model. Fossil localities TMM40536 and TMM41365 ite column, δ13C values from pedogenic carbonates vary between ~
from Schiebout (1974) were anchored at 240 m and 271 m–277 m − 11‰ and − 12‰. These values are low relative to the rest of the
respectively. These sites have yielded small mammals including Tornillo Group and match well with the low δ13C values observed for
plesiadapoids indicating early Tiffanian ages. TMM41364, at 290 m, has the Early Eocene Climate Optimum, between 54.1 Ma and 50.3 Ma, in
yielded a Psittacotherium making the site late Tiffanian. TMM41221 the benthic record (Zachos et al., 2001). The Hannold Hill Fm thickness
(level 322 m to 328 m) has yielded two mammals identified as varies greatly in other outcrops of the park (e.g. Canoe Valley) and
Hyracotherium and Coryphodon, both from sandstone float (potentially additional work should be conducted to assess the presence of uncon-
eroded from the Exhibit sandstone), making this site Clarkforkian to formities within this Formation.
286 C.P. Bataille et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 277–289

6.2. Sediment accumulation rates and stratigraphy Wheeler and Lehman, 2009). Based on our proposed age model, chang-
es in sediment accumulation rates and properties for the Tornillo Group
Sediment accumulation rates of the Tornillo Group varied between can be compared with known global temperature changes of the late
~ 15 m/Myrs and 45 m/Myrs. Those values are an order of magnitude Cretaceous and early Paleogene. Sediment accumulation rates were
lower than those of other Laramide intermountane basins in North high during the hot greenhouse periods of the early Eocene (55.7 Ma
America during the early Paleogene: ~ 600 m/Myrs for the Hoback to 54.1 Ma) and low during the cooler periods of the late Cretaceous
Basin, 210 m/Myrs for the Bighorn Basin and 100 m/Myrs for the San (69 Ma to 67.5 Ma) and Paleocene Carbon Isotope Maximum (58.8 Ma
Juan Basin for Paleocene sediments (Sloan, 1987). The sediment accu- to 57.05 Ma) (Fig. 6). This correlation between global temperature and
mulation rates of the Tornillo Group fluctuated throughout the early Pa- sediment accumulation rates in the Tornillo Group could be explained
leogene and correlated with distinct stratigraphic patterns (Fig. 6). The by two main hypotheses: 1) temperature-driven changes in local
relatively low sedimentation rates observed from 69 Ma to 67.5 Ma, rainfall intensity led to changes in sediment supply rates and/or
58.8 Ma to 57.05 Ma and 57.05 Ma to 56.3 Ma, are associated with a 2) temperature-driven eustatic sea-level changes controlled accommo-
high proportion of fine-grained deposits and a high degree of paleosol dation and deposition regimes in the basin.
maturity (Fig. 2). The relatively high sedimentation rates observed The first hypothesis is consistent with global climate model simula-
from 60.4 Ma to 58.8 Ma are associated with a high proportion of tions of intensified hydrological cycling in subtropical regions during
sand-prone facies, weakly pedogenically modified mudstones and a hotter periods of the early Paleogene (Huber and Goldner, 2012). In
lack of carbonate nodules (Figs. 2 and 6). Depending on the age model modern sub-tropical basins, sediment load is controlled by the intensity
selected for the lower Black Peaks Fm, the sediment accumulation and seasonality of rainfall during the wet season (Cecil et al., 2003).
rates are low (option 1) or high (option 2) relative to the rest of the Twentieth century data from India suggests that increases in global
Black Peaks Fm (Fig. 6). The stratigraphy of the lower Black Peaks Fm temperature can be associated with increased seasonality and intensity
consists primarily of strongly pedogenically modified fine-grained of monsoon rainfall, leading to more flooding events in the fluvial sys-
overbank deposits with large pedogenic carbonate nodules (Fig. 2), tem (Goswami et al., 2006). In the early Paleogene, the Tornillo Basin
which is more consistent with low sediment accumulation rates as cal- was likely located within a hydroclimatic region dominated by the
culated using option 1 (Fig. 6). The character of the upper Black Peaks North American monsoon and received more than 70% of its total rain-
Fm, deposited during the early Eocene (above the last black paleosols), fall during the wet season (Carmichael et al., 2016; Huber and Goldner,
is different from the rest of the Formation, with thick red paleosols and 2012). Influence of changing monsoon regimes may be expressed in the
rare pedogenic carbonate nodules (Figs. 2 and 3E). The sediment contrasting properties of different interval of the Black Peaks Fm. During
accumulation rates for this interval are high relative to the rest of the the relatively cool Paleocene Carbon Isotope Maximum, the Black Peaks
Black Peaks Fm, but the change in paleosol properties suggests the Formation is characterized by low sediment accumulation rates, high
possibility of a significant change in the conditions of and controls on proportions of fine-grained deposits, well-developed meter-scale pur-
sedimentation in the Tornillo Basin. The poor resolution of the age ple paleosols with large carbonate nodules and well-defined horizons,
model for the Hannold Hill Fm does not allow for the accurate calcula- and common occurrence of black paleosols (Figs. 2 and 3D). This de-
tion of sediment accumulation rates. scription contrasts with the Black Peaks strata deposited during the
Early Eocene greenhouse, which are characterized by higher sediment
6.3. Climate influence on sedimentation: 106-year timescale? accumulation rate, coarser-grained deposits, and several meter-thick
red paleosols with rare and small carbonate nodules (Figs. 2 and 3E).
The overall low sediment accumulation rate in the Tornillo Basin are These differences could reflect a change in rates of sediment supply
consistent with the low runoff expected from the warm semiarid condi- associated with an increase in monsoon strength during the warm
tions occurring in the basin during the early Paleogene (Lehman, 1990; Early Eocene.
The second hypothesis was tested by Atchley et al. (2004), who
showed that sediment accumulation rate and lithostratigraphy of the
Tornillo Group at the Dawson Creek locality, ~40 km to the west of Tor-
nillo Flat, is potentially controlled by 3rd order sea-level fluctuations. At
million-year timescales, global temperature is an important driver of
sea-level fluctuations (Haq et al., 1987) and would explain a correlation
between global temperature and local sedimentation in the Tornillo
Basin. At the Dawson Creek locality, transgressive systems track equiv-
alents coincided with a rapid increase in fluvial aggradational cycle
thickness and sand-prone deposition, whereas regression and high
stand systems track equivalents were dominated by thin fluvial aggra-
dational cycles and strongly developed paleosols. Similar cyclical strati-
graphic patterns appear to be present at west Tornillo Flat. Further work
to decipher hectometer-scale fluvial aggradational cycles and correlate
them with global sea-level change could help to test the possibility of
a sea-level-sedimentation linkage within the west Tornillo Flat section.
Tectonic controls described by Lehman (1991) likely also played a
critical role in explaining some trends in the sediment accumulation
rates and lithostratigraphy of the Tornillo Group (Lehman, 1991).
Lehman (1991) showed that enhanced sediment accumulation rate in
the Tornillo Basin in the late Cretaceous and Eocene broadly coincided
with two Laramide tectonic pulses. Based on our age model, the deposi-
tion of the Hannold Hill Formation coincides with the second and possi-
bly most pronounced pulse in the region (Chapin and Cather, 1983). In
Fig. 6. Centered 5 point moving average for carbon and oxygen isotope record from
pedogenic carbonates and calculated sediment accumulation rates at west Tornillo Flat
other sections of the Tornillo Group, undeformed middle Eocene strata
plotted against time. Dark and gray lines are the isotope values for the age model option of the Canoe Formation unconformably overlie Laramide folds in the
1 and 2 respectively. Black Peaks, suggesting syndepositional deformation sometime
 C.P. Bataille et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 277–289 287

between the late Paleocene to middle Eocene (Lehman and Busbey, Further evidence for a coupling between stratigraphy and climate at
2007). The presence of limestone clasts from intrabasinal formations west Tornillo Flat arises from the positive correlation observed between
in conglomeratic sandstones of the Hannold Hill also suggests a tectonic the δ13C and δ18O values of pedogenic carbonates throughout the strat-
influence during early Eocene time (Lehman, 1991). Interestingly the igraphic section. In the geologic record, low δ13C periods or events, asso-
Hannold Hill deposition also coincides with the most extreme global ciated with net transfers of organic carbon from rocks to the ocean/
warming period of the Cenozoic: the early Eocene Climatic Optimum atmosphere system, are usually associated with high δ18O in pedogenic
(54 Ma to 50 Ma). Elsewhere in the continental United States, this carbonate, reflecting warming conditions (Kump and Arthur, 1999). The
greenhouse period is correlated with extremely high denudation rates, positive correlation observed between δ13C and δ18O values of pedogen-
possibly due to enhanced chemical weathering rates associated ic carbonates throughout the section is thus reverse to what would be
with higher CO2 levels, a decrease in vegetation cover and/or an in- expected if temperature was the primary driver of δ18O in Tornillo
crease in storm intensity (Elliot Smith et al., 2008). Changes in precipi- Group pedogenic carbonates. Variations in δ18O of soil water are anoth-
tation and tectonic uplift both generate characteristic patterns of grain er potential control of δ18O variations in pedogenic carbonates. δ18O in
size and stratigraphic geometry (Armitage et al., 2011). Detailed soil water is controlled by the isotopic composition of the rainfall, by
lithostratigraphic study of Tornillo Group sandstones combined with the interactions between isotopically distinct groundwater and surface
detrital zircon analysis could help untangle the influence of tectonics water, and by isotopic fractionation occurring during evaporation.
and climate change throughout the section. Two principal hypotheses are proposed to explain the positive corre-
lation between δ13C and δ18O values: 1) more intense precipitation dur-
ing greenhouse periods would coincide with lower δ18O in pedogenic
6.4. Climate influence on sedimentation: 105-year timescale? carbonates due to the increased isotopic fractionation associated with
rainout and, 2) better soil drainage in cooler periods would coincide
The Paleocene Eocene Thermal Maximum (PETM) is a well- with higher δ18O in pedogenic carbonates due to the increased isotopic
recognized hyperthermal event, characterized by a diagnostic 3–6‰ fractionation associated with evaporation. The current δ18O data do not
negative CIE that has been used globally as a chemostratigraphic marker provide enough evidence to support one hypothesis or the other. Inde-
in both marine and terrestrial sections (McInerney and Wing, 2011). pendent constraints on evaporation and/or precipitation rates or
White and Schiebout (2004) identified several negative shifts in the clumped isotope thermometry that disentangles temperature varia-
δ13C values of pedogenic carbonates above and below the black paleosol tions from δ18O variations in soil water could resolve the correct inter-
layer located at level 312 m in our composite section and argued that pretation of the pattern revealed here.
this entire ~30 m interval represents the PETM. Based on the sediment
accumulation rates estimated here, 30 m appears far too thick to repre-
7. Conclusion
sent this ~ 200 kyr-long event (McInerney and Wing, 2011). Our field
work included thorough collection of this stratigraphic interval to re-
Carbon isotope chemostratigraphy, in combination with magnetic
cover carbonate nodules and further resolve the diagnostic CIE of the
and biostratigraphic data, is a powerful tool to date fluvial stratigraphic
PETM. Although several minor 1–2‰ fluctuations in δ13C values are
sequences over multi-million year timescale. The relative variations of
present, a clear, coherent CIE that could be assigned to the PETM was
carbon isotope ratios of Tornillo Group pedogenic carbonates strongly
not identified. As mentioned previously, nodules collected in black
resemble those of the marine record, allowing correlations which signif-
paleosols have anomalously low δ13C and δ18O values relative to pedo-
icantly enhance the resolution and robustness of age models for these
genic carbonate nodules collected in other soil types (Fig. 4). White
deposits. At a million-year timescale, the sediment accumulation rates
and Schiebout probably included values from these nodules in their
and lithostratigraphy of the Tornillo Group correlate with global tem-
chemostratigraphic results, which might in part explain the large and
perature trends of the early Paleogene. We interpreted this correlation
multiple CIEs presented in their record around the black paleosols.
as reflecting a coupling between temperature-driven changes in North
Based on our refined age model, the approximate location of the
American monsoon strength and sediment supply in the fluvial system.
PETM coincides with a multistory channel sandstone located between
At 105-year timescales, early Eocene hyperthermal events coincide with
322 m and 328.5 m in the composite section. The approximate strati-
the deposition of thick sand bodies which could also reflect dramatic
graphic positions of subsequent Eocene hyperthermals, including the
changes in rainfall occurring during these events. Tectonics and possibly
Eocene Thermal Maximum 2 and 3, also appear to coincide with
eustatic sea-level changes also play a critical role in determining the dy-
massive, conglomeratic sandstones within the Hannold Hill Fm. These
namics of the fluvial system in the Tornillo Basin. This system presents
sandstones of the upper Black Peaks and Hannold Hill Formation
numerous opportunities for further study to investigate and clarify the
could reflect dramatic changes in rainfall occurring during these
interplay of climate, tectonic and sea-level changes in controlling fluvial
hyperthermal events (Huber and Goldner, 2012; Winguth et al.,
regimes over multiple timescales.
2010). In other basins, these hyperthermal events often coincide with
Supplementary data to this article can be found online at http://dx.
changes in fluvial system architecture, including influx of coarser-
doi.org/10.1016/j.palaeo.2016.06.023.
grained sediments, increased frequency of avulsion, or deposition of
thick sand bodies (Foreman, 2014; Foreman et al., 2012; Kraus et al.,
2015; Schmitz and Pujalte, 2007). In the Bighorn and Piceance Creek Ba- Acknowledgments
sins these coarser sediments are thought to reflect rapid increase in sed-
iment flux and discharge instigated by changes in vegetation cover and This study was supported by ACS-PRF 2012 number 52222-ND2 to
monsoon strength (Foreman, 2014; Foreman et al., 2012). While the Gabriel Bowen and the 2012 Ed Picou Fellowship awarded to Clement
change in overall grain size, sediment provenance, sediment accumula- Bataille. Access to outcrops in Big Bend National Park was provided by
tion rate and lithostratigraphy between the Hannold Hill Fm and Black the United States Department of the Interior via permit # BIBE-2014-
Peaks Fm are probably controlled by a pulse of Laramide uplift SCI-0015. We thank Don Corrick for his help. We thank Thomas Lehman
(Lehman, 1991), individual distinctive conglomeratic sandstones of who provided numerous maps, cross-sections and suggestions to facili-
this formation might reflect basin-scale sediment flushing events asso- tate our field work. We thank Alex Gonzalez, Amy Steimke and Galen
ciated with hydroclimatic change during hyperthermals. Additional Gorski for their help in the field. We thank the handing editor, Isabel
chemostratigraphic and stratigraphic work in the upper Black Peaks Montanez, and three reviewers, Thomas Lehman, Blaine Cecil and
and Hannold Hill Formations would be required further test this idea. Sebastien Blanchard, for their suggestions, which greatly improved
Hydrological and carbon cycle connections at 106-year timescale? this manuscript.
288 C.P. Bataille et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 277–289

References Kump, L.R., Arthur, M.A., 1999. Interpreting carbon-isotope excursions: carbonates and
organic matter. Chem. Geol. 161 (1–3), 181–198.
Abels, H.A., Kraus, M.J., Gingerich, P.D., 2013. Precession-scale cyclicity in the fluvial lower Lawson, D.A., 1976. J. Paleontol. 50, 158.
Eocene Willwood Formation of the Bighorn Basin, Wyoming (USA). Sedimentology Lehman, T.M., 1986. Late cretaceous sedimentation in trans-pecos Texas. In: Pause, P.H.,
60 (6), 1467–1483. Spears, R.G. (Eds.), Geology of the Big Bend Area and Solitario DomeFieldtrip
Abels, H.A., et al., 2012. Terrestrial carbon isotope excursions and biotic change during Guidebook. West Texas Geological Society, Texas, pp. 105–110.
Palaeogene hyperthermals. Nat. Geosci. 5 (5), 326–329. Lehman, T.M., 1988. Stratigraphy of the Cretaceous-Tertiary and Paleocene-Eocene
Abels, H.A., et al., 2016. Environmental impact and magnitude of paleosol carbonate Transition rocks of Big-Bend-National-Park, Texas — discussion. J. Geol. 96 (5),
carbon isotope excursions marking five early Eocene hyperthermals in the Bighorn 627–631.
Basin, Wyoming. Clim. Past 12 (5), 1151–1163. Lehman, T.M., 1990. Paleosols and the Cretaceous Tertiary transition in the Big Bend
Allen, J.P., Fielding, C.R., Gibling, M.R., Rygel, M.C., 2014. Recognizing products of Region of Texas. Geology 18 (4), 362–364.
palaeoclimate fluctuation in the fluvial stratigraphic record: an example from the Lehman, T.M., 1991. Sedimentation and tectonism in the Laramide Tornillo Basin of West
Pennsylvanian to Lower Permian of Cape Breton Island, Nova Scotia. Sedimentology Texas. Sediment. Geol. 75 (1–2), 9–28.
61 (5), 1332–1381. Lehman, T.M., Busbey, A.B., 2007. Big Bend Field Trip Guide. Society of Vertebrate Paleon-
Armitage, J.J., Duller, R.A., Whittaker, A.C., Allen, P.A., 2011. Transformation of tectonic and tology, 67th Annual Meeting, Field Trip Guidebook, p. 69.
climatic signals from source to sedimentary archive. Nat. Geosci. 4 (4), 231–235. Lehman, T.M., Coulson, A.B., 2002. A juvenile specimen of the sauropod dinosaur
Atchley, S.C., Nordt, L.C., Dworkin, S.I., 2004. Eustatic control on alluvial sequence stratig- Alamosaurus sanjuanensis from the Upper Cretaceous of Big Bend National Park,
raphy: a possible example from the cretaceous-tertiary transition of the Tornillo Texas. J. Paleontol. 76 (1), 156–172.
Basin, Big Bend National Park, West Texas, USA. J. Sediment. Res. 74 (3), 391–404. Lehman, T.M., McDowell, F.W., Connelly, J.N., 2006. First isotopic (U-Pb) age for the Late
Blum, M.D., Törnqvist, T.E., 2000. Fluvial responses to climate and sea-level change: a Cretaceous Alamosaurus vertebrate fauna of West Texas, and its significance as a
review and look forward. Sedimentology 47, 2–48. link between two faunal provinces. J. Vertebr. Paleontol. 26 (4), 922–928.
Bornemann, A., et al., 2009. Latest Danian carbon isotope anomaly and associated envi- Littler, K., Röhl, U., Westerhold, T., Zachos, J.C., 2014. A high-resolution benthic stable-
ronmental change in the southern Tethys (Nile Basin, Egypt). J. Geol. Soc. 166 (6), isotope record for the South Atlantic: implications for orbital-scale changes in Late
1135–1142. Paleocene–Early Eocene climate and carbon cycling. Earth Planet. Sci. Lett. 401 (0),
Bowen, G.J., et al., 2015. Two massive, rapid releases of carbon during the onset of the 18–30.
Palaeocene-Eocene thermal maximum. Nat. Geosci. 8 (1), 44–47. Lourens, L.J., et al., 2005. Astronomical pacing of late Palaeocene to early Eocene global
Breecker, D.O., Sharp, Z.D., McFadden, L.D., 2009. Seasonal bias in the formation and stable warming events. Nature 435 (7045), 1083–1087.
isotopic composition of pedogenic carbonate in modern soils from central New Maxwell, R.A., Lonsdale, J.T., Hazzard, R.T., Wilson, J.A., 1967. Geology of Big Bend National
Mexico, USA. Geol. Soc. Am. Bull. 121 (3–4), 630–640. Park, Brewster County, Texas. Bureau of Economic Geology Publication, p. 6711.
Busbey, A.B., Lehman, T.M., Wilson, J.A., Society of Vertebrate Paleontology. Meeting, McInerney, F.A., Wing, S.L., 2011. The Paleocene-Eocene Thermal Maximum: a perturba-
1989. Vertebrate Paleontology, Biostratigraphy, and Depositional Environments, tion of carbon cycle, climate, and biosphere with implications for the future. In:
Latest Cretaceous and Tertiary, Big Bend Area, Texas (s.n., [Texas, 90 p. pp.). Jeanloz, R., Freeman, K.H. (Eds.), Annual Review of Earth and Planetary Sciences Vol
Carmichael, M.J., et al., 2016. A model–model and data–model comparison for the early 39, pp. 489–516.
Eocene hydrological cycle. Clim. Past 12 (2), 455–481. Mintz, J.S., Driese, S.G., Breecker, D.O., Ludvigson, G.A., 2011. Influence of changing hydrol-
Cecil, C.B., et al., 2003. Observations on Climate and Sediment Discharge in Selected Trop- ogy on pedogenic calcite precipitation in vertisols, dance bayou, Brazoria County,
ical Rivers, Indonesia. Society for Sedimentary Geology(Special Publication No. 77), Texas, U.S.A.: implications for estimating Paleoatmospheric PCO2. J. Sediment. Res.
pp. 29–50. 81 (6), 394–400.
Cerling, T.E., 1984. The stable isotopic composition of modern soil carbonate and its Nordt, L., Atchley, S., Dworkin, S., 2003. Terrestrial evidence for two greenhouse events in
relationship to climate. Earth Planet. Sci. Lett. 71 (2), 229–240. the Latest Cretaceous. GSA Today 13 (12), 4–9.
Cerling, T.E., Quade, J., 1989. Carbon isotopes in paleosol carbonates as Paleoecologic indi- Olsen, T., Steel, R., Hogseth, K., Skar, T., Roe, S.L., 1995. Sequential architecture in a fluvial
cators. Am. J. Phys. Anthropol. 78 (2), 203. succession — sequence stratigraphy in the Upper Cretaceous Mesaverde Group, Price
Cerling, T.E., Wright, V.P., Vanstone, S.D., 1992. Further comments on using carbon iso- Canyon, Utah. J. Sediment. Res. 65 (2), 265-&.
topes in paleosols to estimate the CO2 content of the palaeo-atmosphere. J. Geol. Paul Wright, V., Marriott, S.B., 1993. The sequence stratigraphy of fluvial depositional
Soc. 149, 673–676. systems: the role of floodplain sediment storage. Sediment. Geol. 86 (3–4),
Chapin, C.E., Cather, S.M., 1983. Eocene Tectonics and Sedimentation in the Colorado 203–210.
Plateau-Rocky Mountain Area. Rocky Mountain Foreland Basins and Uplifts. Rocky Quillévéré, F., Norris, R.D., Kroon, D., Wilson, P.A., 2008. Transient ocean warming and
Mountains Association of Geologists, Denver, Colorado. shifts in carbon reservoirs during the early Danian. Earth Planet. Sci. Lett. 265
Coccioni, R., et al., 2010. The Dan-C2 hyperthermal event at Gubbio (Italy): global impli- (3–4), 600–615.
cations, environmental effects, and cause(s). Earth Planet. Sci. Lett. 297 (1–2), Rapp, S.D., Macfadden, B.J., Schiebout, J.A., 1983. Magnetic polarity stratigraphy of the
298–305. Early Tertiary Black Peaks Formation, Big-Bend-National-Park, Texas. J. Geol. 91 (5),
Cramer, B.S., Toggweiler, J.R., Wright, J.D., Katz, M.E., Miller, K.G., 2009. Ocean overturning 555–572.
since the Late Cretaceous: inferences from a new benthic foraminiferal isotope com- Runkel, A.C., 1988. Stratigraphy, Sedimentology and vertebrate paleontology of Eocene
pilation. Paleoceanography 24 (4), PA4216. rocks, Big Bend region, Texas. Unpublished Ph.D dissertation, University of Texas,
Elliot Smith, M., Carroll, A.R., Mueller, E.R., 2008. Elevated weathering rates in the Rocky Austin: 1–310.
Mountains during the Early Eocene Climatic Optimum. Nat. Geosci. 1 (6), 370–374. Sankey, J.T., 2001. Late Cretaceous mammals and magnetostratigraphy, Big Bend, Texas.
Foreman, B.Z., 2014. Climate-driven generation of a fluvial sheet sand body at the Occasional Papers of the Museum of Natural Science 77, p. 16.
Paleocene–Eocene boundary in north-west Wyoming (USA). Basin Res. 26 (2), Sankey, J.T., 2005. Late Cretaceous vertebrate paleoecology of Big Bend National Park,
225–241. Texas. Dinosaur Park Symposium. Royal Tyrrell Museum Special Publication.
Foreman, B.Z., Heller, P.L., Clementz, M.T., 2012. Fluvial response to abrupt global Schiebout, J.A., 1974. Vertebrate paleontology and paleoecology of Paleocene Black Peaks
warming at the Palaeocene/Eocene boundary. Nature. Formation, Texas. Texas Memorial Museum Bulletin (88 p.).
Friedrich, O., Norris, R.D., Erbacher, J., 2012. Evolution of middle to Late Cretaceous Schiebout, J.A., Rigsby, C.A., Rapp, S.D., Hartnell, J.A., Standhardt, B.R., 1987. Stratigraphy of
oceans—a 55 m.Y. Record of earth's temperature and carbon cycle. Geology 40 (2), the Cretaceous-Tertiary and Paleocene-Eocene Transition Rocks of Big-Bend-
107–110. National-Park, Texas. J. Geol. 95 (3), 359–375.
Goswami, B.N., Venugopal, V., Sengupta, D., Madhusoodanan, M.S., Xavier, P.K., 2006. Schiebout, J.A., Rigsby, C.A., Standhardt, B.R., Rapp, S.D., Hartnell, J.A., 1988. Stratigraphy of
Increasing trend of extreme rain events over India in a warming environment. the Cretaceous-Tertiary and Paleocene-Eocene Transition rocks of Big-Bend-
Science 314 (5804), 1442–1445. National-Park, Texas — reply. J. Geol. 96 (5), 631–634.
Hajek, E.A., Heller, P.L., Schur, E.L., 2012. Field test of autogenic control on alluvial stratig- Schmitz, B., Pujalte, V., 2007. Abrupt increase in seasonal extreme precipitation at the
raphy (Ferris Formation, Upper Cretaceous–Paleogene, Wyoming). Geol. Soc. Am. Paleocene-Eocene boundary. Geology 35 (3), 215–218.
Bull. 124 (11−12), 1898–1912. Shanley, K.W., McCabe, P.J., 1994. Perspectives on the sequence stratigraphy of continen-
Haq, B.U., Hardenbol, J., Vail, P.R., 1987. Chronology of fluctuating sea levels since the tal strata. AAPG Bull. 78 (4), 544–568.
Triassic. Science 235 (4793), 1156–1167. Sheldon, N.D., Tabor, N.J., 2009. Quantitative paleoenvironmental and paleoclimatic re-
Held, I.M., Soden, B.J., 2006. Robust responses of the hydrological cycle to global warming. construction using paleosols. Earth Sci. Rev. 95 (1–2), 1–52.
J. Clim. 19 (21), 5686–5699. Sloan, R.E., 1987. Paleocene and latest Cretaceous mammal ages, biozones, magnetozones,
Huber, M., Caballero, R., 2011. The early Eocene equable climate problem revisited. Clim. rates of sedimentation, and evolution. Geol. Soc. Am. Spec. Pap. 209, 165–204.
Past 7 (2), 603–633. Standhardt, B.R., 1995. Early Paleocene (Puercan) vertebrates of the Dogie Locality, Big
Huber, M., Goldner, A., 2012. Eocene monsoons. J. Asian Earth Sci. 44, 3–23. Bend National Park, Texas. In: Santucci, V., McClelland, L. (Eds.), National Park
Hunt, R.K., Lehman, T.M., 2008. Attributes of the ceratopsian dinosaur Torosaurus, and Services Paleontological Research. Natural Resources Publication Office, Fort Collins
new material from the Javelina Formation (Maastrichtian) of Texas. J. Paleontol. 82 CO., pp. 46–48.
(6), 1127–1138. Turner, K.J. et al., 2011. Geologic map of Big Bend National Park, Texas: U.S. Geological
Kraus, M.J., 1999. Paleosols in clastic sedimentary rocks: their geologic applications. Earth Survey Scientific Investigations Map 3142. In: 3142, U.S.G.S.S.I.M. (Editor).
Sci. Rev. 47 (1–2), 41–70. Voigt, S., Gale, A.S., Jung, C., Jenkyns, H.C., 2012. Global correlation of Upper Campanian
Kraus, M.J., 2002. Basin-scale changes in floodplain paleosols: implications for – Maastrichtian successions using carbon-isotope stratigraphy: development
interpreting alluvial architecture. J. Sediment. Res. 72 (4), 500–509. of a new Maastrichtian timescale. Newsl. Stratigr. 45 (1), 25–53.
Kraus, M.J., Woody, D.T., Smith, J.J., Dukic, V., 2015. Alluvial response to the Paleocene– Westerhold, T., Röhl, U., Donner, B., McCarren, H.K., Zachos, J.C., 2011. A complete high-
Eocene Thermal Maximum climatic event, Polecat Bench, Wyoming (U.S.A.). resolution Paleocene benthic stable isotope record for the central Pacific (ODP Site
Palaeogeogr. Palaeoclimatol. Palaeoecol. 435, 177–192. 1209). Paleoceanography 26 (2), PA2216.
 C.P. Bataille et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 277–289 289

Wheeler, E.A., 1991. Paleocene dicotyledonous trees from Big Bend National Park, Texas: Wilson, J.A., 1967. Early Tertiary Mammals. In: Maxwell, R.A., Lonsdale, J.T., Hazzard, R.T.,
variability in wood types common in the Late Cretaceous and Early Tertiary, and eco- Wilson, J.A. (Eds.), Geology of Big Bend National Park, Brewster County, Texas, Uni-
logical inferences. Am. J. Bot. 78 (5), 658–671. versity of Texas Bureau of Economic Geology Report of Investigations, p. 671.
Wheeler, E.A., Lehman, T.M., 2009. New Late Cretaceous and Paleocene Dicot woods of Big Winguth, A., Shellito, C., Shields, C., Winguth, C., 2010. Climate response at the Paleocene-
Bend National Park, Texas and review of Cretacous wood characteristics. IAWA J. 30 Eocene Thermal Maximum to greenhouse gas forcing-a model study with CCSM3.
(3), 293–318. J. Clim. 23 (10), 2562–2584.
White, P., Schiebout, J., 2004. The stable isotope stratigraphy and paleosols of North Woodburne, M.O., 2004. Late Cretaceous and Cenozoic Mammals of North America:
America's most southern exposure of Late Paleocene/Early Eocene fossiliferous conti- Biostratigraphy and Geochronology. Columbia University Press.
nental deposits: documenting the initial Eocene Thermal Maximum in Big Bend Zachos, J., Pagani, M., Sloan, L., Thomas, E., Billups, K., 2001. Trends, rhythms, and
National Park Texas. J. Vertebr. Paleontol. 24 (3), 129a-129a. aberrations in global climate 65 Ma to present. Science 292 (5517), 686–693.
White, P.D., Schiebout, J., 2008. Paleogene paleosols and changes in pedogenesis during
the initial Eocene thermal maximum: Big Bend National Park, Texas, USA. Geol. Soc.
Am. Bull. 120 (11/12), 1347–1361.