ArticlePDF Available

Abstract and Figures

A compilation of foraminiferal stable isotope measurements from southern high latitude (SHL) deep-sea sites provides a novel perspective important for understanding Earth's paleotemperature and paleoceanographic changes across the rise and fall of the Cretaceous Hot Greenhouse climate and the subsequent Paleogene climatic optimum. Both new and previously published results are placed within an improved chronostratigraphic framework for southern South Atlantic and southern Indian Ocean sites. Sites studied were located between 58° and 65°S paleolatitude and were deposited at middle to upper bathyal paleodepths. Oxygen isotope records suggest similar trends in both bottom and surface water temperatures in the southern sectors of the South Atlantic and in the Indian Ocean basins. Warm conditions were present throughout the Albian, extreme warmth existed during the Cretaceous Thermal Maximum (early-mid-Turonian) through late Santonian, and long-term cooling began in the Campanian and culminated in Cretaceous temperature minima during the Maastrichtian. Gradients between surface and seafloor δ 18 O and δ 13 C values were unusually high throughout the 11.5 m.y. of extreme warmth during the Turonian-early Campanian, but these vertical gradients nearly disappeared by the early Maastrichtian. In absolute terms, paleotemperature estimates that use standard assumptions for pre-glacial seawater suggest sub-Antarctic bottom waters were ≥21 °C and sub-Antarctic surface waters were ≥27 °C during the Turonian, values warmer than published climate models support. Alternatively, estimated temperatures can be reduced to the upper limits of model results through freshening of high latitude waters but only if there were enhanced precipitation of water with quite low δ 18 O values. Regardless, Turonian planktonic δ 18 O values are ~1.5‰ lower than minimum values reported for the Paleocene-Eocene Thermal Maximum (PETM) from the same region , a difference which corresponds to Turonian surface temperatures ~6 °C warmer than peak PETM temperatures if Turonian and Paleocene temperatures are estimated using the same assumptions. It is likely that warm oceans surrounding and penetrating interior Antarctica (given higher relative sea level) prevented growth of Antarctic ice sheets at all but the highest elevations from the late Aptian through late Campanian; however, Maastrichtian temperatures may have been cool enough to allow growth of small, ephemeral ice sheets. The standard explanation for the sustained warmth during Cretaceous Hot Greenhouse climate invokes higher atmospheric CO 2 levels from volcanic outgassing, but correlation among temperature estimates, proxy estimates of pCO 2 , and intervals of high fluxes of both mafic and silicic volcanism are mostly poor. This comparison demonstrates that the relative timing between events and their putative consequences need to be better constrained to test and more fully understand relationships among volcanism, pCO 2 , temperature ocean circulation, Earth's biota and the carbon cycle.
Content may be subject to copyright.
Contents lists available at ScienceDirect
Global and Planetary Change
journal homepage: www.elsevier.com/locate/gloplacha
The rise and fall of the Cretaceous Hot Greenhouse climate
Brian T. Huber
a,
, Kenneth G. MacLeod
b
, David K. Watkins
c
, Millard F. Con
d,e,f
a
Department of Paleobiology, MRC-121, Smithsonian Institution, Washington, DC 20013, United States
b
Department of Geological Sciences, University of Missouri-Columbia, Columbia, MO 65211, United States
c
Department of Earth & Atmospheric Sciences, University of Nebraska, Lincoln, NE 68588, United States
d
Institute for Marine and Antarctic Studies, University of Tasmania, Private Bag 129, Hobart, TAS 7001, Australia
e
School of Earth and Climate Sciences, University of Maine, Orono, ME 04469-5790, United States
f
Woods Hole Oceanographic Institution, Woods Hole, MA 02543, United States
ARTICLE INFO
Keywords:
Cretaceous Hot Greenhouse
Foraminiferal stable isotopes
Volcanic outgassing
pCO
2
proxies
Greenhouse glacier hypothesis
Southern high latitudes
ABSTRACT
A compilation of foraminiferal stable isotope measurements from southern high latitude (SHL) deep-sea sites
provides a novel perspective important for understanding Earth's paleotemperature and paleoceanographic
changes across the rise and fall of the Cretaceous Hot Greenhouse climate and the subsequent Paleogene climatic
optimum. Both new and previously published results are placed within an improved chronostratigraphic fra-
mework for southern South Atlantic and southern Indian Ocean sites. Sites studied were located between 58° and
65°S paleolatitude and were deposited at middle to upper bathyal paleodepths. Oxygen isotope records suggest
similar trends in both bottom and surface water temperatures in the southern sectors of the South Atlantic and in
the Indian Ocean basins. Warm conditions were present throughout the Albian, extreme warmth existed during
the Cretaceous Thermal Maximum (early-mid-Turonian) through late Santonian, and long-term cooling began in
the Campanian and culminated in Cretaceous temperature minima during the Maastrichtian. Gradients between
surface and seaoor δ
18
O and δ
13
C values were unusually high throughout the 11.5 m.y. of extreme warmth
during the Turonian-early Campanian, but these vertical gradients nearly disappeared by the early
Maastrichtian.
In absolute terms, paleotemperature estimates that use standard assumptions for pre-glacial seawater suggest
sub-Antarctic bottom waters were 21 °C and sub-Antarctic surface waters were 27 °C during the Turonian,
values warmer than published climate models support. Alternatively, estimated temperatures can be reduced to
the upper limits of model results through freshening of high latitude waters but only if there were enhanced
precipitation of water with quite low δ
18
O values. Regardless, Turonian planktonic δ
18
O values are ~1.5
lower than minimum values reported for the Paleocene-Eocene Thermal Maximum (PETM) from the same re-
gion, a dierence which corresponds to Turonian surface temperatures ~6 °C warmer than peak PETM tem-
peratures if Turonian and Paleocene temperatures are estimated using the same assumptions. It is likely that
warm oceans surrounding and penetrating interior Antarctica (given higher relative sea level) prevented growth
of Antarctic ice sheets at all but the highest elevations from the late Aptian through late Campanian; however,
Maastrichtian temperatures may have been cool enough to allow growth of small, ephemeral ice sheets. The
standard explanation for the sustained warmth during Cretaceous Hot Greenhouse climate invokes higher at-
mospheric CO
2
levels from volcanic outgassing, but correlation among temperature estimates, proxy estimates of
pCO
2
, and intervals of high uxes of both mac and silicic volcanism are mostly poor. This comparison de-
monstrates that the relative timing between events and their putative consequences need to be better constrained
to test and more fully understand relationships among volcanism, pCO
2
, temperature ocean circulation, Earth's
biota and the carbon cycle.
1. Introduction
Evidence for warm polar regions during the Cretaceous is well
documented. Terrestrial plant assemblages and dinosaurs from polar
latitudes of both hemispheres indicate mean annual and winter
minimum temperatures much warmer-than-modern (Brouwers et al.,
1987;Case, 1988;Case et al., 2000;Herman and Spicer, 1997;Olivero
et al., 1991;Parrish et al., 1989;Parrish and Spicer, 1988;Rich et al.,
https://doi.org/10.1016/j.gloplacha.2018.04.004
Received 18 December 2017; Received in revised form 27 April 2018; Accepted 27 April 2018
Corresponding author.
E-mail addresses: huberb@si.edu (B.T. Huber), MacLeodK@missouri.edu (K.G. MacLeod), dwatkins1@unl.edu (D.K. Watkins), Mike.Con@utas.edu.au (M.F. Con).
Global and Planetary Change 167 (2018) 1–23
0921-8181/ Published by Elsevier B.V.
T
2002). Presence of champsosaur (a crocodile-like reptile) remains at
~72°N during the Turonian-Coniacian is consistent with mean annual
temperatures of > 14 °C (Tarduno et al., 1998;Vandermark et al.,
2007). Geochemical conrmation and renement of temporal trends
are provided by paleotemperatures inferred from foraminiferal oxygen
isotope compilations (e.g., Cramer et al., 2009;Friedrich et al., 2012),
sh teeth (Pucéat et al., 2007;Martin et al., 2014) and TEX
86
mea-
surements derived from the composition of membrane lipids of marine
Archaea (O'Brien et al., 2017).
The compilations of deep-sea benthic foraminiferal and bulk car-
bonate δ
18
O data reveal that warming of the world's oceans to the mid-
Cretaceous peak was gradual and spanned the early Albian through
middle Cenomanian. Extremely warm temperatures (seaoor tem-
peratures of > 20 °C at mid-bathyal depths) persisted from the late
Cenomanian through Santonian, and then gradually returned to cooler
values (~68 °C at midbathyal depths; ~15002500 m) during the
Maastrichtian (Clarke and Jenkyns, 1999;Huber et al., 1995;Huber
et al., 2011;Huber et al., 2002;Cramer et al., 2009;Friedrich et al.,
2012). Oxygen isotopic records from surface dwelling planktonic for-
aminifera in the southern South Atlantic have been shown generally to
parallel the deep-sea benthic δ
18
O trends with highest planktonic δ
18
O
values during the late Aptian-early Albian and Maastrichtian cool
greenhouseand lowest values during the Turonian-Santonian hot
greenhouse(Huber et al., 1995,2002;Barrera and Savin, 1999;Fassell
and Bralower, 1999).
Extreme warmth during the mid-Cretaceous has been ascribed to
high pCO
2
maintained by high uxes from volcanic activity including
increased oceanic crust formation, eruption of large igneous province
(LIP), and high rates of subduction-related arc volcanism (Arthur et al.,
1985;Barron and Washington, 1985;Crowley, 1991;Larson, 1991a;
Con and Eldholm, 1994;Crowley and Berner, 2001;Van Der Meer
et al., 2014). The buildup of atmospheric CO
2
, palaeogeography, and
warm Cretaceous oceans created conditions near the threshold for
widespread deposition of organic-rich sediments under anoxic condi-
tions, as demonstrated by the repeated occurrence of Oceanic Anoxic
Events (OAEs) and their associated prominent carbon-isotope excur-
sions documented from marine and terrestrial sequences (Ando et al.,
2002;Arthur et al., 1988;Gröcke et al., 1999;Jahren et al., 2001;Jarvis
et al., 2006;Jenkyns, 1980, 2010;Jenkyns et al., 2017). Globally ex-
pressed Cretaceous OAEs occurred during the early Aptian (OAE 1a;
~120 Ma) and at the Cenomanian/Turonian boundary (OAE 2;
~94 Ma) while regionally recognized events occurred during the early
Albian (OAE 1b; ~111 Ma), late Albian (OAE 1d; ~100 Ma) and pos-
sibly late Coniacian-early Santonian (OAE 3; ~86 Ma). Two end-
member models, one invoking reduced ocean circulation (stagnant
ocean basins) and another invoking high plankton productivity and
expansion of the oxygen minimum zone (productivity model), have
been proposed to explain OAEs. Where on this spectrum the answer
remains debated due to discrepancies between model predictions and
empirical observations, and forcing mechanisms for OAEs remain
poorly understood (Jenkyns, 2010). Regardless of triggering mechan-
isms, the burial of large amounts of organic carbon should have reduced
pCO
2
and lowered temperatures globally.
At Falkland Plateau DSDP Site 511 (Fig. 1), planktonic foraminiferal
δ
18
O values from the Turonian are especially low, ranging at times to
between 4.2 and 4.7. These values suggest upper ocean at ~60 °S
paleolatitude was as warm as 2932 °C assuming the standard ice-free
values of 1
SMOW
applies for regional seawater δ
18
O(Bice et al.,
2003). The low planktonic δ
18
O values (and resulting high paleo-
temperature estimates) have deed straightforward heat budget ex-
planations. In the Bice et al. (2003) study, temperatures of ~30 °C were
incompatible with model results unless the models use either im-
probably high pCO
2
levels (6500 ppm) or proscribed at least a 50%
increase in oceanic poleward heat transport relative to earlier mid-
Cretaceous conditions. Reducing calculated temperatures by invoking
moisture balance changes that lowered seawater δ
18
O was considered
impossible as the modeled
18
O-depleted seawater was too brackish for
the diverse, open ocean microfossil assemblages present. Foraminiferal
preservation in the Turonian at Site 511 is excellent arguing against
diagenetic overprints of > 2that are required to match even
2500 ppmv CO
2
models.
In spite of widespread evidence for extreme global warmth during
the CenomanianTuronian, some authors have argued for the episodic
existence of continental ice based on presumed isochrony of sequence
stratigraphic boundaries and correlation with oxygen isotope shifts
(Gale et al., 2002;Miller et al., 2003,2005;Stoll and Schrag, 2000;
Bornemann et al., 2008). The proposed ice sheet events during the
Cretaceous hot greenhouse have been countered by more recent studies
that showed oxygen isotopic records are inconsistent with waxing and
Fig. 1. Paleogeographic reconstructions for the earliest Turonian (94 Ma) and
late Eocene (40 Ma) showing location of deep-sea sites discussed in the text.
Reconstructions based on Müller et al. (2016).
B.T. Huber et al. Global and Planetary Change 167 (2018) 1–23
2
waning of continental ice sheets at these times (Ando et al., 2009;
MacLeod et al., 2013;Moriya et al., 2007). Signicant uncertainty in
the timing and extent of sea level falls and the inuence of dynamic
topography on local tectonics and regional climatic conditions have
further weakened support for ice volume variation as a global forcing
mechanism for sea level change during the mid-Cretaceous (Haq and
Huber, 2017). While an ice-free Turonian now seems widely accepted,
the existence of ephemeral ice sheets during cool greenhouse intervals
of the AptianAlbian and Maastrichtian continues to be debated (Alley
and Frakes, 2003;Bowman et al., 2013,2014).
High latitudes respond more strongly than low latitudes to climate
forcing, and the Cretaceous southern high latitudes (SHL) are critical in
examining disagreements among studies (e.g., large dierences be-
tween model and proxy temperature estimates exemplied by the Site
511 data, apparent contradictions between inferred glacio-eustatic
sealevel changes and lack of glacial evidence in cores recovered around
Antarctica). To test proposed patterns and improve knowledge of
temperature evolution at SHL, we have compiled published high lati-
tude carbon and oxygen isotopic data from foraminifera with very good
to excellent preservation and generated new data to augment and ex-
pand existing records. All data are plotted using age models revised
herein to facilitate comparisons with CO
2
, volcanism, tectonism, and
proxies for Cretaceous atmospheric CO
2
composition.
Our new compilation spans 80 m.y. of subpolar history and shows
similar oxygen isotopic trends among multiple benthic and planktonic
foraminifer species that reect changes in Earth's sea surface tem-
peratures (SSTs) and sea oor temperatures across the rise and fall of
the Cretaceous hot greenhouse climate between 58° and 65 °S paleola-
titude. Comparison of maximum and minimum Cretaceous tempera-
tures with those of the Paleogene provides a context for identifying
times during the Cretaceous when the existence of polar ice sheets were
most likely. Incorporation of recent estimates of volcanic activity, tec-
tonism, and pCO
2
history into our compilation allows assessment of
relationships between estimated temperatures and these potential
controlling variables and illuminates areas where better constrained
data are needed to advance understanding.
2. Methods and studied sections
New data presented in this study were obtained from single species
separates of well-preserved planktonic and benthic foraminifera picked
from DSDP Sites 257, 258 and 511. Stable isotope results are reported
in δ-notation on the Vienna Pee Dee Belemnite (VPDB) scale. Samples
analyzed were picked from washed residues. Analyses used 130+
individuals for each measurement targeting sample weights of between
50 and 100 μg. Measurements were made on a Kiel III carbonate device
linked to a Delta Plus isotope ratio mass spectrometer. Analytical pre-
cision is < 0.03and 0.06(1 s.d.) for δ
13
C and δ
18
O, respectively,
based on long term tracking of uncorrected results for NBS-19. One split
of NBS-19 is run every 7 to 8 samples, and day-to-day variability is
corrected for by adjusting results for that day by the dierence between
the average value of NBS-19 samples measured within that run and
nominal values of 1.95for δ
13
C and 2.20δ
18
O. Planktonic and
benthic paleotemperatures were calculated from δ
18
O values using the
equation of Kim and O'Neil (1997) reformulated for synthetic calcite by
(Bemis et al. (1998) assuming a δ
18
O
sw
value of 1.27Vienna
standard mean ocean water (VSMOW) for Cretaceous seawater in a
non-glacial world (Shackleton and Kennett, 1975) adjusted by 0.27
to accurately combine input values on the VPDB and VSMOW scales for
calcite and seawater, respectively. The temperature estimates have not
been adjusted for possible regional variations in δ
18
O
sw
, but assump-
tions regarding δ
18
O
sw
are explored in Section 4.1 below. See Pearson
(2012) for thorough discussion of assumptions about δ
18
O
sw
in an as-
sumed ice-free world, planktonic foraminiferal vital eects, inter-in-
dividual planktonic foraminiferal variability in δ
18
O, and variability in
individual depth habitats.
Below is a summary of the geologic setting, age control, and for-
aminiferal preservation for each of these deep-sea sites. The primary
resource for the planktonic foraminiferal taxonomy used in this study is
from the pforams@mikrotaxdatabase (http://www.mikrotax.org/
pforams/index.html;Huber et al., 2016). Calcareous nannofossil age
assignments follow the CC scheme of (Perch-Nielsen, 1985). A more
thorough biostratigraphic and taxonomic treatment of Sites 257 and
258 will be presented in a separate work. Ages used for the age models
are based on the 2016 Geologic Time Scale (Ogg et al., 2016) and a
2017 data extract from TimeScale Creator 2017 (https://engineering.
purdue.edu/Stratigraphy/tscreator/index/index.php).
2.1. Southern Indian Ocean, Wharton Basin: DSDP Site 257
At DSDP Site 257, Albian claystone yielding extraordinarily well
preserved calcareous plankton was recovered in an interval from 199.7
to 248.9 mbsf. The base of this interval is 9.2 m above the basaltic
basement. The site was drilled in 5278 m water depth and is located in
the Perth Abyssal Plain in the southeastern Wharton Basin (30°59.16S,
108°20.99E, SE Indian Ocean; Davies et al., 1974). The paleolatitude of
the site was ~62°S at 110 Ma according to the plate tectonic re-
construction of Müller et al. (2016) and the Albian sequence was
probably deposited at a bathyal paleodepth (Davies et al., 1974).
2.1.1. Biostratigraphy
In most samples analyzed from cores 2578 and 9 foraminiferal
abundance relative to sediment volume is low and planktonic for-
aminifera are less common (typically < 15% of the population) than
benthic foraminifera. The preservation of the calcareous tests, though,
is good to excellent (Table 1). Most shells show no inlling and have
translucent (glassy) walls with no evidence of secondary shell alteration
(see images in Fig. 2). An early Albian age is assigned to cores from 257-
7-CC to -9-2, 4042 cm based on identication of Microhedbergella
praeplanispira, M. renilaevis and M. rischi. This assemblage is correlated
with the Ticinella madecassiana Zone, which spans from 112.4 to
111.8 Ma (Ogg et al., 2016). Planktonic foraminifera are absent from
Sample 257-9-2, 110112 cm.
The interval from Sample 2578-1, 9294 cm through 2579-1,
130132 cm contains abundant to common, moderately-preserved cal-
careous nannofossils. The assemblages include few specimens of
Prediscosphaera columnata with coccolith diameters < 5 μm as well as
rare Hayesites albiensis. There are no specimens of Cribrosphaerella eh-
renbergii or Tranolithus orionatus. This association indicates the sample
comes from the lower, but not the lowest, portion of Subzone CC8a
(=Subzone NC8b) of early Albian age. Specically, this interval is
correlated by (Gradstein et al., 2016) to an age of 111.3112.65 Ma,
with a mean age of 112.0 Ma ± 0.7 Ma uncertainty. The common pre-
sence of Seribiscutum primitivum indicates a high-latitude (austral) a-
nity for the surface waters at this site and it corroborates the age pla-
cement. As reported by Thierstein (1974) in his initial examination of
this interval, samples below Section 257-9-1 exhibit increasingly poor
preservation, and in the oldest samples only robust species are present
and identiable.
2.1.2. Stable isotope data
We measured stable isotopic values of two benthic and one plank-
tonic species from six sample levels between 238.42 and 248.90 mbsf at
Site 257 in sediments assigned to the upper lower Albian (Table 2;
Fig. 3). Insucient numbers of individuals of planktonic foraminifera
were present in samples below 240.22 mbsf to analyze accurately.
2.2. Southern Indian Ocean, Mentelle Basin: DSDP Sites 258, 258A
A 525 m sequence of marine sediments was discontinuously cored at
DSDP Site 258 (33°47.69S, 112°28.42E) in the western Mentelle Basin,
which borders Naturaliste Plateau to the west (Davies et al., 1974).
B.T. Huber et al. Global and Planetary Change 167 (2018) 1–23
3
According to plate reconstructions by Müller et al. (2016) this site was
located at about 62°S during the mid-Cretaceous (Fig. 1). The upper
114 m of the cored interval was assigned to the upper MioceneRecent
and the underlying 411 m was assigned to the mid- to Late Cretaceous.
The sequence was spot-cored with about ~50% recovery, such that in
total only 22% was recovered (Fig. 3). A second hole, Hole 258A, was
drilled to 123.5 meters below seaoor (mbsf) with the goal of im-
proving recovery of the upper portion of the sequence. Cretaceous se-
diments discussed in this study consist of nannofossil ooze (258A-8
through -9), nannofossil chalk (cores 258-5 through -13) and organic-
rich claystone (258-14) with a distinctive black shale bed at 258-14-1,
4757 cm.
2.2.1. Biostratigraphy
Foraminifera are abundant and show frosty preservation (presence
of micron scale overgrowths) in the ooze and chalk cores from cores 8-9
at Hole 258A and cores 5-10 at Site 258 (Fig. 2ab). Most samples from
Cores 258-11, -12, and -13 are weakly indurated and yield foraminifera
that are both strongly recrystallized and inlled with sparry calcite.
Claystone samples from Section 258-14-1 yield well-preserved for-
aminifera, some of which show glassy preservation (shells translucent
with no inlling or overgrowths) (Fig. 2cd). The low abundance of
benthic foraminifera (generally < 3%) and the dominance in the
benthic assemblage of gavelinellid and gyroidinoidid species suggest
that deposition occurred at a middle bathyal depth.
As observed in Herb (1974), biostratigraphic subdivision of the
Cretaceous sequence at Holes 258 and 258A is hampered by the low
diversity of the planktonic foraminiferal assemblage and absence of
standard tropical-subtropical zonal marker species. Nonetheless, sev-
eral observations in the present study provide helpful constraints on the
age of the sequence and have led us to revise the original chronos-
tratigraphic framework for the holes. Revisions include assignment of:
(1) Cores 258A-8 through -9 to the early Campanian Globotruncanita
elevata Zone (previously unzoned but assigned to the Santonian)
based on the presence of Globotruncana neotricarinata (Petrizzo
et al., 2011), which rst occurs in Sample 258A-9-4, 9194
(119.91 mbsf)
(2) Cores 258-5 through -9 to the Coniacian- Santonian Dicarinella
marginata Zone (=Marginotruncana marginata Zone sensu (Petrizzo,
2001) based on the presence of the nominate taxon, several species
of Marginotruncana, and the large biserial species Planoheterohelix
reussi (previously unzoned but assigned to the Coniacian)
(3) Cores 258-11 through -14 to the Turonian Whiteinella baltica Zone
(sensu (Huber, 1992) based on the presence of the nominate spe-
cies, Dicarinella hagni,Praeglobotruncana stephani, and Whiteinella
aprica (previously unzoned but assigned to the Cenomanian)
The calcareous nannofossil biostratigraphy from Site 258 was ori-
ginally documented by Thierstein (1974) using a set of informal zones
that were not subsequently adopted by others. The site was re-examined
two decades later as part of an early attempt to derive an Austral Upper
Cretaceous zonation (Watkins et al., 1996). Those data provide means
to assign the Site 258 material to the CC Zonation as follows:
1) Sections 5.1 to 5.2: Santonian (Zone CC16) based on the presence of
Eprolithus oralis and Lithastrinus grillii without Lithastrinus septe-
narius
2) Section 5.3 to sample 9-1, 40 cm: middle to upper Coniacian (Zone
CC15) based on co-occurrence of Lithastrinus septenarius and
Reinhardtites anthophorus
3) Sample 9-1, 140 cm to -11-1, 90 cm: upper Turonian to lower
Coniacian (Zone CC13) based on the presence of Marthasterites fur-
catus without Micula staurophora, an association that is characteristic
of the upper Turonian (the co-occurrence of the observed FAD of
Micula staurophora and Reinhardtites anthophorus in Sample 9-1,
Table 1
Distribution of Albian foraminifera at DSDP Site 257. Preservation ratings are: E = excellent (glassy shell wall), G = good (frosty shell walls but no shell inlling), and M= moderate (signicant shell recrystallization
and/or shell inlling). Abundance ratings are: C = common, F = few, R = rare. Samples where xis shown in place of relative abundance values denote presence of the species.
Site 257 Depth Age P. F. Zone Preservation Total Foram Abundance % Planktics Mi. renilaevis Mi. praeplanispira Mi. rischi Muricohedbergella sp. Ticinella cf. madecassiana
7, CC 207.08 m. Albian T. madecassiana GC xx xx x
8-1, 9294 238.42 m. Albian T. madecassiana EC 4%9 12 4
8-1, 134136 238.84 m. Albian T. madecassiana E C 52% 49 35 22 2
8-2, 5860 239.58 m. Albian T. madecassiana EC 1%
8-2, 122124 240.22 m. Albian T. madecassiana EC 6%2 2 1 2
91, 105109 248.05 m. Albian T. madecassiana GC xx x x
9-1, 130132 248.30 m. Albian T. madecassiana E C 12% 10 7 3
9-2, 4042 248.90 m. Albian T. madecassiana E F 14% 11 6 2 1
9-2, 110112 249.60 ? ? G R 0%
10-1, 110112 257.60 ? ? M R 0%
B.T. Huber et al. Global and Planetary Change 167 (2018) 1–23
4
3940 cm means Zone CC14 is not identied in this sequence)
4) Sample 11-2, 40 cm to 12-3, 40 cm: middle to late Turonian (Zone
CC12) based on the presence of Eiellithus eximius without
Marthasterites furcatus
5) Sample 12-3, 140 cm to 13-4, 70 cm: lower Turonian (Zone CC11)
based on the presence of Quadrum gartnerii and Eprolithus moratus
and the absence of Eiellithus eximius
6) Core 14: uppermost Cenomanian (Subzone CC10a) based on the
presence of Lithraphidites acutus,Axopodorhabdus biramiculatus (=A.
albianus of some authors), and Helenea chiastia. The black shale
within this core correlates biostratigraphically with OAE 2.
Samples from the base of Section 258A-8-6 and Core 258A-9 contain
abundant, well-preserved calcareous nannofossils including
Arkhangelskiella cymbiformis,Marthasterites furcatus,Lithastrinus grillii,
and Reinharditites biperforatus. This assemblage, in the absence of
Broinsonia parca parca and B. parca constricta, indicates the middle to
upper part of Zone CC17 of earliest Campanian age. The presence of
abundant Kamptnerius magnicus, common Seribiscutum primitivum, and
frequent Biscutum dissimilis implies a strongly austral high-latitude af-
nity for these assemblages.
Ages and depths of calcareous nannofossil and planktonic for-
aminiferal datum events identied at Holes 258 and 258A are presented
in Table 3 and Figs. 4, 6 and 7.
2.2.2. Stable isotope data
We measured stable isotopic values of ve benthic and eight
planktonic species from 19 sample levels between 124.53 and
264.26 mbsf at Holes 258 and 258A in sediments ranging from the
middle Cenomanian-Santonian age (Tables 4, 5;Fig. 3). Previously
published δ
13
C and δ
18
O data from Site 258 from Huber et al. (1995)
are also presented. Species belonging to the same genus or family group
were combined to maximize stratigraphic continuity of the data plots.
2.3. Southern South Atlantic, Falkland Plateau: DSDP Site 327, Hole 327A
DSDP Site 327 was drilled on the Falkland Plateau Basin (50°52S,
46°47W), in 2410 m water depth. The upper Campanianlower
Maastrichtian interval included in this study was spot-cored between 90
and 140 mbsf with < 50% core recovery. Cores 327A-10 through 13
yielded calcareous marl with abundant, well preserved foraminiferal
assemblages that are dominated by (> 90%) planktonic foraminifera.
Fig. 2. Preservation of planktonic foraminifera from
DSDP Sites 258 and 257. (1a, b) Santonian specimen
of Globigerinelloides bollii from Sample 258A-5-1,103-
106 cm showing minor shell recrystallization in
broken section of shell wall; (2a, b) early Turonian
specimen of Whiteinella brittonensis from uppermost
Cenomanian Sample 258-14-1,126-129 cm showing
pristine preservation in broken section of shell wall;
(3) Gyroidinoides globosus from lower Albian Sample
257-9-1, 130132 cm; (4) Osangularia schloenbachi
from lower Albian Sample 257-9-1, 130132 cm; (5)
Microhedbergella renilaevis from lower Albian Sample
257-8-1, 134136 cm.
Table 2
Oxygen and carbon isotope measurements from DSDP Site 257 foraminifera. Benthic species include Gyroidinoides globosus and Osangularia schloenbachi; mixed
planktonic samples all belong to the genus Microhedbergella.
Site 257 Depth Gyroid. globosus Os. schloenbachi Microhedberg. spp.
(mbsf) δ
13
Cδ
18
Oδ
13
Cδ
18
Oδ
13
Cδ
18
O
8-1, 9294 cm 238.42 1.89 0.53 1.75 0.33 3.01 1.46
8-1, 9294 cm 238.42 1.76 0.37 1.84 0.28
8-1, 134136 cm 238.84 1.63 0.60 1.78 0.34 3.12 1.62
8-1, 134136 cm 238.84 1.71 0.51 2.92 1.60
8-2, 5860 cm 239.58 1.74 0.69 1.99 0.53
8-2, 122124 cm 240.22 1.70 1.27 1.62 0.74 3.08 1.89
8-2, 122124 cm 240.22 1.59 1.14 1.49 0.55 3.09 1.80
8-2, 122124 cm 240.22 1.92 1.18
9-1, 130132 cm 248.30 1.65 0.68 1.64 0.64
9-2, 40cm 248.90 1.93 0.44 1.85 0.57
Average 1.75 0.74 1.75 0.50 3.05 1.67
Std. Dev. 0.12 0.33 0.16 0.16 0.08 0.17
B.T. Huber et al. Global and Planetary Change 167 (2018) 1–23
5
2.3.1. Biostratigraphy
The planktonic foraminiferal biostratigraphy of the site was dis-
cussed in Huber et al. (1995) and is revised in the present study based
on additional observations of planktonic foraminifera (datum events
listed in Table 6) that were used to construct a new age model from
which we estimate sedimentation rates for that 50 m interval (Fig. 4).
2.3.2. Stable isotope data
Previously published Hole 327A δ
13
C and δ
18
O data from Huber
et al. (1995) are presented in Figs. 6 and 7.
2.4. Southern South Atlantic, Falkland Plateau: DSDP Site 511
DSDP Site 511 is located in the Falkland Plateau Basin at 51°00S
and 47°58W at a water depth of 2589 m, about 10 km southeast of
DSDP Site 327. Paleogeographic reconstructions indicate that this site
was at ~61°S during the early Albian and migrated to ~56°S by the end
of the Campanian (Fig. 1). Core recovery, lithologic descriptions,
magnetostratigraphy, and initial biostratigraphic interpretations for
this site are presented in Ludwig et al. (1983). The upper Aptian-lower
Maastrichtian sediments in the present study were cored between 195
and 499 mbsf (cores 2356) with an average core recovery of 70%.
Material at Site 511 was deposited at outer shelf to upper slopes depths
during the late Aptian and deepened to middle bathyal depths by the
late Albian (Basov and Krasheninnikov, 1983). Lithologies include
muddy nannofossil chalk and black shale from 480 to 499 mbsf (cores
5556), reddish brown muddy nannofossil chalk from 413 to 480 mbsf
(cores 4854), zeolitic claystone from 209 to 412 mbsf, (cores 2547),
and calcareous ooze between 195 and 212 mbsf (cores 2324). The
percent carbonate in the cores is also variable, ranging from 65% in the
calcareous ooze to 1% in the zeolitic clay. Dilution by terrigenous clays
probably accounts for most of the variability in carbonate content, but
shallowing of the foraminiferal lysocline has been proposed to explain
loss of the calcareous microfossil record during the late Cenomanian
and middle Campanian (Wise Jr, 1983;Basov and Krasheninnikov,
1983).
2.4.1. Biostratigraphy
The more carbonate-rich intervals in the Cretaceous sequence yield
abundant foraminiferal assemblages that are dominated by planktonic
specimens (usually > 80%), but the claystone sediments yield fewer
foraminifera and more variable planktonic:benthic foraminiferal ratios
(Huber et al., 1995). Foraminifera show glassy preservation in nearly
all samples from the upper Aptian-Campanian interval. In the upper
Campanian interval, foraminifera are well preserved but exhibit small-
scale shell recrystallization (Krasheninnikov and Basov, 1983;Basov
and Krasheninnikov, 1983;Huber et al., 1995;Bice et al., 2003;Fassell
and Bralower, 1999).
The planktonic foraminiferal biostratigraphy applied in Huber et al.
(1995) and Bice et al. (2003) has been modied in the following ways:
(1) Core 511-23 is reassigned from the lower Maastrichtian to the
upper Campanian based on the absence of Rugotruncana circumno-
difer and Globotruncana subcircumnodifer, which are consistently
present in Maastrichtian sediments at Maud Rise ODP Sites 689 and
690 (Huber, 1990)
(2) Sections 511-49-5 and -6 are reassigned from the Cenomanian to
Depth (mbsf)
(c) DSDP Site 511
Depth (mbsf)
200
250
300
350
400
450
500
75 80 85 90 95 100 105 110 11 5
Forams
Nannos
tPr
bM
r
bAa
bEt
bMa
bKm
bEe
bMd
bM
f
tLs
bAp
tGi
7.5 m/m.y.
24.1 m/m.y.
10.4 m/m.y.
6.23 m/m.y.
(b) DSDP Site 327
A
Depth (mbsf)
Age (Ma)
95
100
105
110
115
120
125
130
135
140
73 74 75
Forams
tGi
bGh
bGa
12.6 m/m.y.
14.2 m/m.y.
(a) DSDP Sites 258, 258
A
4.1 m/m.y.
11.8 m/m.y.
25.4 m/m.y.
120
140
160
180
200
220
240
260
280
83 84 85 86 87 88 89 90 91 92 93 94 95
Forams
Nannos
bQg
bEe
bRa
tLs
bGn
bAc
20.9 m/m.y.
Fig. 3. Age-depth models showing changes in sedimentation rates at DSDP Sites
258 and 258A on Mentelle Basin and DSDP Sites 327A and 511 on Falkland
Plateau. Datum abbreviations are listed in Tables 3, 6, 7 and 10.
Table 3
Ages and depths for planktonic foraminiferal (F) and calcareous nannofossil (N)
species from DSDP Sites 258 and 258A used to dene the line of correlation in
Fig. 4a. Genus spellings are provided in the text. Plot code refers to genus-
species abbreviations; FAD = rst appearance datum.
Group Event Plot code Age (m.y.) Top depth
(mbsf)
Bottom depth
(mbsf)
N FAD A.
cymbiformis
bAc 83.20 120.94 127.58
F FAD G.
neotricarinata
bGn 83.64 118.30 119.95
N LAD L. septenarius tLs 85.60 126.39 127.39
N FAD R.
anthophorus
bRa 88.14 180.89 181.89
N FAD E. eximius bEe 92.99 237.91 238.91
N FAD Q. gartneri bQg 93.80 258.71 263.19
N FAD L. acutus/
alatus
bLa 96.16 264.19 264.19
N FAD H. chiastia bHc 93.90 264.19 264.19
B.T. Huber et al. Global and Planetary Change 167 (2018) 1–23
6
the late Albian-early Cenomanian based on the presence of
Muricohedbergella astrepta and Pseudothalmaninnella ticinensis
(3) the Aptian/Albian boundary is placed at 485.46 mbsf (Sample 511-
55-4, 60 cm) using the lowest occurrence (LO) of Microhedbergella
renilaevis, which is the datum now used to dene the base of the
Albian Stage (Kennedy et al., 2014, 2017). The extinction of several
Aptian planktonic foraminifer species, which was the previous cri-
terion for placement of the Aptian/Albian boundary (Huber and
Leckie, 2011;Huber et al., 2011), occurs at 486.22 mbsf (511-55-5,
5152 cm)
Calcareous nannofossils in samples from Sections 511-49-4 through
511-49-6 were investigated to better resolve the previous late Albian-
Cenomanian age assignment (Wise, 1983). The samples contain
common to few, moderate to poorly-preserved calcareous nannofossils
in assemblages dominated by Watznaueria barnesiae and Seribiscutum
primitivum. The high relative abundance of the latter species indicates
the strong austral anity of samples in this interval precluding direct
correlation with low latitude zonal schemes. The presence of (rare)
Gartnerago segmentatum indicates that the assemblages are no older than
early (but not earliest) Cenomanian. Observed Gartnerago segmentatum
is probably the form reported as Gartnerago sp. cf. G.confossus by Wise
Jr (1983). A maximum early Cenomanian age is corroborated by the
presence of Eiellithus turrisseifellii and Eiellithus casulus without their
(older) late Albian sister taxa (e.g., Eiellithus praestigium;(Watkins and
Bergen, 2003)). Determination of the minimum age is more problematic
as it is based on the absence of taxa. The absence of Quadrum gartneri
and Kamptnerius magnicus, which are common components of other
austral Turonian assemblages, suggests that this interval is no younger
than Cenomanian.
Ages and depths of calcareous nannofossil and planktonic for-
aminiferal datum events identied at Site 511 are presented in Table 7
240
250
CC16
e. Camp.
CC11 CC12 CC13 CC15
G.elevata Unzoned
Marginotruncana marginata
Whiteinella baltica
Coniacian
Turonian
Cen. Sa.
Siite 258 Siite 258A
CC10a CC17
e. Albian
Ti. mad.
Siite 257
CC8a
-1.0 0.0 1.0 2.0 3.0 4.0
δ
13
C
VPDB
-4.0-3.0-2.0-1.00.0
δ
18
O
VPDB
Angulogavelinella sp.
Gavelinella sp.
Stensioina sp.
Len culina sp.
Gyroidinoides globosus
Benthics
Globigerinell. spp.
Pl. reussi
Pl. globulosa
Globotruncanids
Ar. bosquensis
Muricohed.
Whiteinella sp.
Microhed. sp.
Planktics
120
130
140
150
160
170
180
190
200
210
220
230
240
8
9
10
250
260
5
6
7
11
12
13
14
110
120
9
8
8
9
8
8
Fig. 4. Oxygen and carbon isotope measurements
from planktonic and benthic foraminifera from
DSDP Sites 257, 258 and 258A. Columns left to right
refer to: DSDP Site, Age, calcareous nannofossil
biozone, planktonic foraminiferal biozone, meters
below seaoor, and core number. Core ll patterns
include black for recovered core, white for un-
recovered core, and diamond crosses for uncored
intervals. Illustrated specimens on far left are
benthics including Gyroidinoides globosus and
Gavelinella cf. ahuvae; in middle and right from
bottom to top: Microhedbergella rischi;Whiteinella
aprica;Globigerinelloides yaucoensis;Planoheterohelix
reussi;Globotruncana linneiana;Archaeoglobigerina
bosquensis;Globotruncana neotricarinata.
B.T. Huber et al. Global and Planetary Change 167 (2018) 1–23
7
Table 4
Oxygen and carbon isotope measurements from DSDP Site 258 foraminifera. Benthic taxa include species of Angulogavellinella,Gavelinella,Stensioina,Lenticulina, and Nuttallides and the remaining taxa are planktonic
species. Genus abbreviations include: Pl.=Planoheterohelix;Gt.=Globotruncana;Ar.=Archaeoglobigerina and Muricohed.=Muricohedbergella.
Site 258 Depth Age Stage Nanno Angulogavellinella sp. Gavelinella sp. Stensioina sp. Lenticulina sp. Nuttallides sp.
(mbsf) (m.y.) Zone δ
13
Cδ
18
Οδ
13
Cδ
18
Οδ
13
Cδ
18
Οδ
13
Cδ
18
Οδ
13
Cδ
18
Ο
5-1, 103106 124.53 83.68 Santonian CC16 1.42 1.17 0.05 1.25 1.40 0.88
5-1, 103106 124.53 83.68 Santonian CC16 0.03 1.45 1.65 0.65
5-1, 103106 124.53 83.68 Santonian CC16 0.00 1.23
5-2, 108110 126.08 83.80 Santonian CC16 0.48 1.39
5-2, 108110 126.08 83.80 Santonian CC16 0.41 1.45
5-3, 108111 127.58 83.91 Santonian CC16 0.28 1.20
5-3, 108111 127.58 83.91 Santonian CC16 1.37 0.79
5-3, 108111 127.58 83.91 Santonian CC16
6-2, 110113 145.10 85.25 Coniacian CC15 0.63 1.68
6-3, 110113 146.60 85.36 Coniacian CC15
6-3, 110113 146.60 85.36 Coniacian CC15
6-3, 110113 146.60 85.36 Coniacian CC15
6-4, 110113 148.10 85.48 Coniacian CC15 1.34 1.16
6-4, 110113 148.10 85.48 Coniacian CC15 1.42 1.12
6-5, 112115 149.62 85.60 Coniacian CC15
7-1, 106109 153.06 85.86 Coniacian CC15 1.39 1.16
7-1, 106109 153.06 85.86 Coniacian CC15 1.48 1.13 0.08 1.40
7-2, 110112 154.60 85.98 Coniacian CC15
7-2, 110112 154.60 85.98 Coniacian CC15
7-2, 110112 154.60 85.98 Coniacian CC15
7-2, 110112 154.60 85.98 Coniacian CC15
7-3, 110113 156.10 86.09 Coniacian CC15
9-1, 3841 180.88 87.99 Coniacian CC15 0.88 1.09
9-1, 3841 180.88 87.99 Coniacian CC15
9-1, 7982 181.29 88.02 Coniacian CC13 0.83 1.10
9-1, 7982 181.29 88.02 Coniacian CC13 0.92 1.12
9-1, 7982 181.29 88.02 Coniacian CC13 0.77 1.03
10-1, 142145 200.92 89.52 Coniacian CC13 1.15 1.15
10-1, 142145 200.92 89.52 Coniacian CC13
10-2, 5354 201.53 89.57 Coniacian CC13
10-2, 5354 201.53 89.57 Coniacian CC13
10-2, 5354 201.53 89.57 Coniacian CC13
11-3, 7680 219.26 90.92 Turonian CC12 1.08 2.25
12R-1, 5355 235.05 92.13 Turonian CC12 0.26 2.77
13R-1, 116118 254.68 93.63 Turonian CC11 2.07 1.85 0.28 2.06
13R-1, 116118 254.68 93.63 Turonian CC11 1.98 1.74 0.47 2.11
132, 54.555.5 255.57 93.70 Turonian CC11 2.24 1.62
(continued on next page)
B.T. Huber et al. Global and Planetary Change 167 (2018) 1–23
8
Table 4 (continued)
Site 258 Depth Age Stage Nanno Angulogavellinella sp. Gavelinella sp. Stensioina sp. Lenticulina sp. Nuttallides sp.
(mbsf) (m.y.) Zone δ
13
Cδ
18
Οδ
13
Cδ
18
Οδ
13
Cδ
18
Οδ
13
Cδ
18
Οδ
13
Cδ
18
Ο
132, 54.555.5 255.57 93.70 Turonian CC11 2.32 1.86
141, 8386 263.83 94.33 Cenomanian CC10a
141, 8386 263.83 94.33 Cenomanian CC10a
141, 126129 264.26 94.37 Cenomanian CC10a 1.51 0.70 0.18 1.13
141, 126129 264.26 94.37 Cenomanian CC10a 1.58 0.79
Site 258 Gl. prairiehill. Globigerinell. sp. Pl. planata Pl. globulosa Gt. linneiana Ar. bosquensis Whiteinella spp. Muricohed. sp.
δ
13
Cδ
18
Οδ
13
Cδ
18
Οδ
13
Cδ
18
Οδ
13
Cδ
18
Οδ
13
Cδ
18
Οδ
13
Cδ
18
Οδ
13
Cδ
18
Οδ
13
Cδ
18
Ο
5-1, 103106 2.86 1.68 2.89 1.24 3.25 1.97
5-1, 103106 3.01 1.09
5-1, 103106
5-2, 108110 3.35 1.03 3.12 1.20 2.73 1.62
5-2, 108110
5-3, 108111 3.22 0.97
5-3, 108111 3.58 1.75
5-3, 108111
6-2, 110113 2.53 2.08 2.90 2.12
6-3, 110113 2.93 2.26
6-3, 110113
6-3, 110113 2.62 2.08
6-4, 110113 2.86 2.57 2.54 2.12
6-4, 110113 2.45 2.17
6-5, 112115
7-1, 106109 2.17 1.89
7-1, 106109 2.76 1.81 2.40 1.93
7-2, 110112 2.75 1.51 3.2 2.24
7-2, 110112
7-2, 110112
7-2, 110112 2.82 1.99
7-3, 110113
9-1, 3841 2.91 2.01
9-1, 3841 3.10 2.33
9-1, 7982 2.58 2.02 2.89 1.76 2.97 2.48 3.23 2.88
9-1, 7982 2.70 2.49
9-1, 7982
10-1, 142145 2.61 2.48 2.98 2.40
10-1, 142145 2.73 2.39
10-2, 5354 2.89 2.82
10-2, 5354
10-2, 5354 2.74 2.54
11-3, 7680 3.15 2.40 3.10 2.65
12R-1, 5355 2.72 3.10
13R-1, 116118 3.65 3.76
13R-1, 116118
132, 54.555.5 3.90 3.20 3.35 2.35
132, 54.555.5
141, 8386
141, 8386 3.38 1.85
141, 126129 2.61 1.81 3.42 2.35 2.30 1.74
141, 126129
B.T. Huber et al. Global and Planetary Change 167 (2018) 1–23
9
and estimates of sedimentation rates are shown in Fig. 4. The age model
indicates that the stable isotope record covers a range from 113.8 to
74.3 Ma and is interrupted by a 10.4 m.y. hiatus spanning from the late
Albian (102.2 Ma) through early Turonian (91.8 Ma).
2.4.2. Stable isotope data
New δ
13
C and δ
18
O data from Site 511 presented in Table 8 and
Fig. 6 were generated mostly from the uppermost Aptian through upper
Albian interval. Additional late Aptian-late Campanian data were
compiled from several sources (Huber et al., 1995;Fassell and
Bralower, 1999;Bice et al., 2003).
2.5. Southern South Atlantic, Maud Rise: ODP Sites 689 and 690
Ocean Drilling Program (ODP) Sites 689 (64°31S, 03°06E) and 690
(65°10S, 1°12E) were drilled on Maud Rise in the southern South
Atlantic at 2080 and 2900 water depth, respectively (Barker and
Kennett, 1988). The sites occupied nearly the same latitude during the
Late Cretaceous-Paleogene (Müller et al., 2016), and benthic for-
aminiferal assemblages indicate they were deposited at upper abyssal to
lower bathyal paleodepths (Thomas, 1990). Core recovery, lithologic
descriptions, magnetostratigraphy, and biostratigraphic interpretations
for the sites are presented in Barker and Kennett (1988).
2.5.1. Biostratigraphy
Age models for Sites 689 and 690 (Tables 9, 10;Fig. 5) are primarily
based on the high quality magnetostratigraphic records reported in
Hamilton (1990) and Spiess (1990) with additional control provided by
the level of the Cretaceous/Paleogene boundary and several planktonic
foraminiferal datum events recorded by Stott and Kennett (1990). Both
records show continuous deposition except for a brief hiatus identied
in the upper lower Eocene (Thomas et al., 1990).
2.5.2. Stable isotopes
Oxygen and carbon isotope data from ODP Sites 689 and 690 in
Fig. 6 are compiled from Barrera and Huber (1990),Stott and Kennett
(1990),Barrera and Savin (1999),Huber et al. (2002), and Friedrich
et al. (2006). They are plotted using the revised age models discussed
above.
3. Stable isotope results
New stable isotope results from Holes 257, 258, 258A, and 511 are
presented in Tables 2, 4, 5, and 8, respectively. Temperature estimates
shown parenthetically assume seawater δ
18
Oof1
SMOW
and use the
paleotemperature equation of (Kim and O'Neil, 1997) reformulated for
synthetic calcite by Bemis et al. (1998). The summaries below integrate
both new and published data.
3.1. Wharton Basin
In the early Albian samples from Hole 257 (Fig. 3), the δ
18
O com-
position of the two benthic species, Osangularia schloenbachi and Gy-
roidinoides globosus, average 0.5(13.5 °C) and 0.8(14.7 °C),
respectively, with G. globosus δ
18
O values showing greater variability.
Combined planktonic species of Microhedbergella rischi and M. praepla-
nispira yield average δ
18
O values of 1.7(~18 °C) and show little
variability. The δ
13
C analyses of the two benthic species both average
1.8, with G. globosus showing greater variability than O. schloenbachi.
The δ
13
C values from Microhedbergella spp. average 3.1and show
little variability. The average vertical (planktonic to benthic) δ
18
O
gradient for the lower Albian is 0.9 to 1.2(~3.3 to 4.5 °C), and the
average vertical δ
13
C gradient is 1.3.
Table 5
Oxygen and carbon isotope measurements from DSDP Hole 258A foraminifera. Benthic taxa include species of Gavelinella,Angulogavellinella, and Lenticulina all remaining species are planktonic. See caption of Table 4 for
explanation of genus abbreviations.
Site 258A Depth Stage Nanno Gavelinella sp. Lenticulina sp. Angulagavelinella sp. Gl. prairiehillensis Pl. planata Gt. linneiana Ar. bosquensis
(mbsf) Zone δ
13
Cδ
18
Οδ
13
Cδ
18
Οδ
13
Cδ
18
Οδ
13
Cδ
18
Οδ
13
Cδ
18
Οδ
13
Cδ
18
Οδ
13
Cδ
18
Ο
8-6, 141144 113.41 82.83 lwr. Camp. CC17 2.81 1.12 2.29 0.79 2.02 0.77 2.49 1.72
8-6, 141144 113.41 82.83 lwr. Camp. CC17 1.84 0.69
9-1, 141143 115.41 82.98 lwr. Camp. CC17 2.90 0.72 1.02 0.75 2.86 0.68 2.78 0.57 2.53 0.92
9-1, 141143 115.41 82.98 lwr. Camp. CC17 0.23 1.11 2.77 0.62
9-1, 141143 115.41 82.98 lwr. Camp. CC17 0.07 1.18
9-2, 105107 116.55 83.07 lwr. Camp. CC17 0.80 2.30 1.22 2.54 2.08 0.54 2.41 1.11
9-2, 105107 116.55 83.07 lwr. Camp. CC17 0.97 2.26 1.05 0.80 1.18 2.29 2.40 0.84 1.86 0.21 2.58 1.85
9-3, 130133 118.30 83.20 lwr. Camp. CC17 1.33 0.46 2.37 0.57 2.59 1.45 3.10 1.61
9-3, 130133 118.30 83.20 lwr. Camp. CC17
9-3, 130133 118.30 83.20 lwr. Camp. CC17
9-3, 130133 118.30 83.20 lwr. Camp. CC17
9-4, 9194 119.41 83.28 lwr. Camp. CC17 1.41 0.87 0.55 1.21 2.96 0.66 2.52 0.68 2.49 1.02
9-4, 9194 119.41 83.28 lwr. Camp. CC17 0.25 1.24 2.52 0.76 2.90 1.65
9-5, 9597 119.45 83.29 lwr. Camp. CC17 1.45 0.18 0.20 1.09 2.87 0.76 2.74 0.52 2.84 1.41
9-5, 9597 119.45 83.29 lwr. Camp. CC17 1.41 0.37 0.71 1.00 2.65 0.49
9-6, 9496 120.94 83.40 lwr. Camp. CC17 0.28 0.98 1.45 0.33 2.94 0.71
9-6, 9496 120.94 83.40 lwr. Camp. CC17 1.51 0.37 2.40 0.95
9-6, 9496 120.94 83.40 lwr. Camp. CC17 1.72 0.21 2.43 0.66
B.T. Huber et al. Global and Planetary Change 167 (2018) 1–23
10
3.2. Mentelle Basin
Early Turonian-early Campanian foraminifera from DSDP Holes 258
and 258A yield oxygen isotope ratios that suggest warmest tempera-
tures during the early Turonian and coolest temperatures during the
early Campanian (Fig. 3). In Core 258-13 benthic foraminiferal values
range from 1.7 to 2.0(~1819 °C), and surface mixed layer
planktonic foraminiferal values range from 3.2 to 3.8
(2528 °C). These values are ~1.0lower than the underlying up-
permost Cenomanian δ
18
O values, which were obtained from just below
the inferred OAE 2 black shale bed in Core 14. Benthic and planktonic
values are relatively constant from the middle Turonian through late
Coniacian and then increase by ~0.5near the Coniacian/Santonian
boundary. The highest δ
18
O values occur in the lower Campanian at
Hole 258A with benthic values near 0.5(~13 °C) and mixed layer
planktonics averaging 1.4(~17 °C).
Carbon isotope ratios from planktonic and benthic foraminifera
(excluding Lenticulina) also show relatively small dierences between
the highest values in the Turonian and the lowest values in the lower
Campanian. Lenticulina is not considered for comparison because the
highly variable δ
13
C are characteristic of the genus perhaps related to
an unusually exible ecology with individuals able to shift from epi-
faunal to deep infaunal habitats (see Wendler et al., 2013 and refer-
ences therein). The vertical δ
18
O gradient is largest in the Turonian
(~1.9) and smallest in the early Campanian (< 1.0), whereas the
vertical δ
13
C gradient (excluding Lenticulina) shows little change from
the Turonian through lower Campanian with dierences mostly ranging
between 1.0 and 1.8.
3.3. Falkland Plateau
Oxygen isotope ratios for planktonic foraminifera from samples at
Hole 327A and Site 511 are notable for their remarkably low values of
4.2to 4.7that suggest temperatures of 2932 °C during the
Turonian, as discussed above (Fig. 6). Prior to this interval, a gap spans
the entire Cenomanian, and, below that, Albian planktonics exhibit
values that are 23higher (cooler) than seen in the Turonian sam-
ples. Planktonic δ
18
O values also increase (cool) following the Turonian
warm interval. Maximum δ
18
O values of 0.6(13 °C) for surface
dwelling species are reached in the late Campanian, the youngest
Cretaceous recovered on Falkland Plateau. Benthic taxa show similar
patterns but with lower amplitude shifts such that δ
18
O values are
0.5in the Albian (8 °C), decrease to 1.9(19 °C) during the
Turonian, and then gradually increase back to 0.5(8 °C) from the
Coniacian through the Campanian.
Carbon isotope ratios also show a major negative excursion across
the Aptian/Albian boundary interval (AABI), with pre-excursion
benthic values at 113.8 Ma ranging between 1.4 and 2.2, decreasing
to minimum values of 3.3at 113.0 Ma, and increasing to between
0.0 and 1.4at 112.8 Ma. The latest Aptian planktonic data from
within the excursion interval 30 kyr below the boundary level
(113.0 Ma) mostly range from 0to 1.1with minimum and
maximum values of 2.2and 1.4, respectively. Benthic and
planktonic δ
13
C values show no obvious trends from the early through
the late Albian. Benthic values mostly range between 0.7 and 0.0,
and planktonic values range between 1.4 and 2.8. Above the
Cenomanian gap, benthic δ
13
C values shift subtly lower whereas
planktonic δ
13
C values are largely unchanged until a shift to higher
values in Santonian and early Campanian samples. During the
Campanian, benthic and planktonic values converge with δ
13
C values of
0.5and 2in these two groups, respectively, in the youngest
samples.
The trends summarized above indicate large changes in surface to
seaoor isotopic gradients. A vertical δ
18
O and δ
13
C gradient is essen-
tially absent within the AABI excursion interval as the extremes of
benthic and planktonic values and the wide range of values (particu-
larly benthic) in both isotope systems is remarkable and result in
overlap between sea surface and seaoor values. The vertical δ
18
O
gradient in the Albian sequence mostly ranges from 1.0to 1.5;it
reaches 3.5during the Turonian and remains high through the early
Campanian, and decreases to 1.5in the late Campanian. The vertical
δ
13
C gradient is also eectively non-existent during the AABI (although
Table 6
Ages and depths determined for planktonic foraminifera from DSDP Site 327A used to constrain the line of correlation in Fig. 4b. Plot code refers to genus-species
abbreviations; FAD = rst appearance datum.
Event Plot code Core Sample Top Depth Bott. Depth Mean Depth Datum Age
a
FAD Globotruncana arca bGa 327A-10-3, 2230 94.50 92.74 93.62 72.95
FAD Globtruncanella havanensis bGh 327A-12-1, 6972 110.80 109.20 110.00 74.25
FAD Globigerinelloides impensus tGi 327A-13-1, 143145 113.20 138.43 125.82 75.36
a
Mean ages calculated from Maud Rise datum events; see Tables 9, 10
Table 7
Ages and depths for planktonic foraminiferal (F) and calcareous nannofossil (N) species from DSDP Site 511 used to constrain the line of correlation in Fig. 4c. Plot
code refers to genus-species abbreviations; FAD = rst appearance datum; LAD = last appearance datum.
Group Event Plot code Age Top depth Bottom depth Mean depth
F FAD Globigerinelloides impensus tGi 75.36 195.38 204.70 200.04
N FAD Aspidolithus parcus parcus bAp 81.43 245.00 246.50 245.75
N LAD Lithastrinus septenarius tLs 86.38 366.00 366.00 366.00
N FAD Micula furcatus bMf 90.24 399.00 404.40 401.70
N FAD Micula decussata bMd 89.77 399.09 404.40 401.75
N FAD Eifellithus eximius bEe 92.99 413.90 427.90 420.90
N FAD Kamptnerius magnicus bKm 92.99 413.90 427.90 420.90
F FAD Muricohedbergella astrepta bMa 102.43 429.65 430.96 430.31
N FAD E. turriseieli bEt 103.13 434.00 434.00 434.00
F FAD Microhedbergella renilaevis bMr 113.00 485.11 485.46 485.29
F LAD Paraticinella rohri tPr 113.00 486.12 486.14 486.13
B.T. Huber et al. Global and Planetary Change 167 (2018) 1–23
11
Table 8
Oxygen and carbon isotope measurements from DSDP Site 511 foraminifera. Benthic taxa include species of Osangularia, Berthelina, and Lenticulina, and the re-
maining taxa are planktonic species. Genus abbreviations include: Os. = Osangularia;Gy. = Gyroidinoides, Paraticin. = Paraticinella,Hd. = Hedbergella,M. =
Muricohedbergella, and Ar. = Archaeoglobigerina.
Age Os. schoenb. Gy. globosus Berthelina sp. Lenticulina sp. Paraticin. rohri Hd. infracretacea M. praeplanispira Ar. bosquensis
δ
13
Cδ
18
Oδ
13
Cδ
18
Oδ
13
Cδ
18
Oδ
13
Cδ
18
Oδ
13
Cδ
18
Oδ
13
Cδ
18
Oδ
13
Cδ
18
Oδ
13
Cδ
18
O
86.54 2.58 1.83
102.82 2.11 1.12
102.82 1.93 0.86
103.09 1.81 1.26
103.72 2.33 0.68
103.72 2.35 0.67
104.57 2.22 0.72
104.74 2.42 0.84
107.25 1.62 0.79
107.25 1.70 0.80
108.88 2.31 1.27
110.24 2.46 0.93
110.30 1.41 0.05 2.11 0.09 0.26 0.46 2.18 0.79
110.30 1.06 0.46 2.31 1.05
110.54 1.21 0.13 1.74 0.16 0.25 0.70
110.54 1.20 0.10 1.72 0.08 0.47 0.69
110.78 1.24 0.26 1.82 0.09 0.92 0.60 2.29 0.92
110.99 1.19 0.19 1.72 0.12 0.76 0.72 2.20 1.55
110.99 0.97 0.11
111.17 1.44 0.17
111.65 2.05 0.42
111.92 2.58 0.18 3.38 0.93
111.92 2.50 0.27 3.30 1.10
112.16 1.32 0.04 1.70 0.07
112.16 1.44 0.02 1.73 0.01
112.39 1.65 0.47 2.96 1.22
112.39 1.65 0.17 2.00 0.11 0.63 0.60 2.69 1.35
112.39 1.55 0.09 1.77 0.01 0.43 0.83
112.64 2.68 0.64
112.65 1.60 0.13 2.18 0.01 0.15 0.56
112.65 1.65 0.12 1.08 0.28 1.65 0.22
112.65 1.94 0.07 1.07 0.24
112.73 0.83 0.62 1.09 0.19 1.49 0.15 0.09 0.71
112.73 1.00 0.26 1.35 0.32
112.85 0.68 1.54 0.79 1.69 2.05 0.42
112.85 0.71 1.35 0.79 1.68 1.31 0.75
112.91 0.79 2.97 1.85 0.80
112.91 0.94 3.65 0.32 1.71
112.92 0.55 3.01 0.04 1.97
112.92 1.22 0.53 0.80 2.26
112.94 0.37 2.75 0.70 3.42 0.42 2.75
112.94 0.81 1.12 1.52 4.47 0.59 2.97
113.00 1.59 1.40
113.00 3.31 1.92 5.19 2.84
113.01 0.02 1.04 0.57 0.93
113.01 1.11 1.69
113.02 0.38 0.52 1.72 0.28
113.02 0.21 0.18 0.32 0.73
113.02 1.42 0.59 0.13 1.13 1.31 3.28 1.40 2.61 0.66 4.68 1.08 2.87
113.02 0.57 1.48 1.09 2.65 0.98 0.91 1.25 2.66 0.13 3.74 2.16 4.61
113.03 0.16 2.07 1.06 3.83 1.50 0.46 1.37 3.86
113.03 1.25 4.04 1.53 0.51 1.06 3.43
113.03 1.72 4.05
113.03 0.45 0.86
113.05 0.42 1.79 1.16 3.71 0.27 1.96
113.06 0.00 2.36 0.54 2.06
113.06 0.21 2.39 0.26 2.17
113.06 0.10 2.29 0.85 3.63 0.56 2.22
113.06 0.68 2.73
113.07 1.23 3.84 1.33 1.11
113.07 0.19 2.37
113.08 0.19 2.65 1.31 2.96
113.09 1.37 0.58 0.62 3.37 0.20 2.46 2.52 3.12
113.09 0.94 3.85 0.53 3.03
113.11 1.84 2.99
113.13 0.11 2.42 0.86 3.27 1.30 4.08 2.79 4.84
113.13 1.58 4.07 0.96 3.66 0.55 0.77
113.30 1.28 3.43
113.42 1.79 4.53 2.30 5.17 0.73 1.29 1.00 5.19 1.72 5.64
113.42 1.10 3.75 1.80 4.29
(continued on next page)
B.T. Huber et al. Global and Planetary Change 167 (2018) 1–23
12
variability is high), it is up to 1.2during the Albian, increases to
2.5during the middle Turonian, and it diminishes to 0.8in the late
Campanian.
3.4. Maud Rise
Upper Campanian-Maastrichtian samples from Maud Rise (Fig. 6)
continue the pattern of change seen in Santonian-upper Campanian
samples on Falkland Plateau. The Campanian δ
18
O measurements from
benthic foraminifera for the Maud Rise sites match values from Hole
327A at 0.5(8 °C), and they continue to generally increase through
the Maastrichtian to δ
18
O values of 0.7 to 1.3(4 to 7 °C) in samples
between 68.3 and 66.5 Ma. A brief 1negative shift between 66.5 and
66.2 Ma is followed by a return to higher values just below the Cre-
taceous/Paleogene boundary. The planktonic δ
18
O data from both
Maud Rise sites increase from 1(14 °C) in the late Campanian to
0.5(8 °C) between 66.8 and 66.5 Ma and show a short, 1negative
excursion between 66.5 Ma and the Cretaceous/Paleogene boundary.
Planktonic δ
18
O values are ~0.5higher in samples from Site 689
than in correlative Campanian samples from the same species analyzed
from Sites 690 and Hole 327A, but there is no obvious dierence in
values in the mid- to late Maastrichtian from the two Maud Rise sites.
Benthic δ
13
C values in the Maud Rise sites increase gradually from
0.5to 1.5whereas planktonic values vary between 2and 3.
Late Campanian values in both sites match closely values in correlative
samples from Falkland Plateau.
Maud Rise trends also continue the pattern of decreasing vertical
δ
18
O and δ
13
C gradients seen at Falkland Plateau. Maud Rise gradients
are highest during the late Campanian at 1to 1.5for δ
18
O and
2.5to 3for δ
13
C, decrease to a mid-Maastrichtian minimum of
~0for δ
18
O and 1.5for δ
13
C. The δ
18
O increases to 0.5in the
late Maastrichtian.
4. Subantarctic temperature changes across 80 m.y
An 80 m.y. compilation of benthic and planktonic δ
18
O changes for
the late Aptian through late Eocene is based on foraminiferal data
plotted against the revised age models for all SHL sites (Fig. 7). High
latitude Paleogene isotopic trends have been discussed in many well
cited studies (e.g., Stott and Kennett, 1990;Kennett and Stott, 1990;
Bohaty and Zachos, 2003;Mackensen and Ehrmann, 1992;Stott et al.,
1990;Thomas and Shackleton, 1996); we include those data for SHL
here to facilitate comparison of the two most recent and arguably best
studied greenhouse intervalsthe Late Cretaceous and the early
Table 8 (continued)
Age Os. schoenb. Gy. globosus Berthelina sp. Lenticulina sp. Paraticin. rohri Hd. infracretacea M. praeplanispira Ar. bosquensis
δ
13
Cδ
18
Oδ
13
Cδ
18
Oδ
13
Cδ
18
Oδ
13
Cδ
18
Oδ
13
Cδ
18
Oδ
13
Cδ
18
Oδ
13
Cδ
18
Oδ
13
Cδ
18
O
113.54 1.57 3.07 1.23 3.45 0.36 1.60 2.75 5.37
113.54 1.85 3.40
113.65 1.04 0.04
113.76 1.88 0.25 0.58 0.38
113.77 1.44 0.18
113.77 2.13 0.51 2.35 0.38 1.15 0.17
113.77 2.16 0.45
Table 9
Oxygen and carbon isotope measurements from DSDP Site 511 foraminifera.
Group Event Plot code Age Depth
M Base Chron 8N-2 bC8n.2n 25.99 75.97
M Top Chron 9N bC8r 26.42 79.46
M Base Chron I0N bC10n.2n 28.28 91.93
M Base Chron11N bC11n.2n 29.97 103.38
M Base Chron 12N bC12n 31.03 106.88
M Base Chron 15R bC15r 35.71 135.77
M Base Chron 16N bC16r 36.97 145.02
M Base Chron 17N-3 bC17n.3n 38.33 152.73
M Top Chron 18N b17r 38.62 153.70
M Base Chron 19N bC19n 41.39 163.16
M Top Chron 20N bC19r 42.30 165.55
F FAD Planomalina australiformis bPa 55.50 207.92
F FAD Parasubbotina inconstans bPi 62.90 229.84
F FAD Parasubbotina pseudobulloides bPp 65.70 231.89
K/Pg K/Pg K/Pg 66.00 233.42
M C30N/C30R bC30N 68.20 246.60
M C30R/C31N bC30R 68.37 248.08
M C31N/C31R bC31N 69.27 252.92
M C31R/C32N bC31R 71.45 259.97
M C32R.2N/C32n.2R b32R.2N 73.65 272.33
M 32R.2R/C33N bC32R.2R 74.31 277.32
Table 10
Ages and depths for planktonic foraminiferal (F) and calcareous nannofossil (N)
species and base magnetic polarity chron boundaries from ODP Site 690 used to
constrain the line of correlation in Fig. 6b. Plot code refers to genus-species
abbreviations; FAD = rst appearance datum.
Group Event Plot Code Age Depth (mbsf)
M Top Chron 8N bC7r 24.76 53.25
M Base Chron 8N bC8n.2n 25.99 60.11
M Top Chron 9N bC8r 26.42 60.99
M Base Chron 9N bC9N 27.44 68.48
M Base Chron11N bC11n.2n 29.97 73.93
M C11R-2/C12N bC11r 30.59 80.76
M C12N/C12R bC12n 31.03 84.01
M Base C12R b12r 33.16 91.70
M C16n.2n bC16n.2n 36.70 95.70
M Top Chron 17N bC16r 36.97 96.59
M bC19n bC19n 41.39 105.65
M ~ Top Chron 20N bC19r 42.30 106.27
M ~Top Chron 2IN bC20r 45.72 118.74
M Base Chron 21N bC21n 47.35 123.63
M bC21r bC21r 48.57 130.48
M Base Chron 22N bC22n 49.34 132.23
M C23N-2/C23R-2 bC23n.2 51.83 133.18
M bC23r bC23r 52.62 137.33
M C24R-1/C24N-2 bC24n.1r 53.20 144.42
M Base Chron 24N C24n.3n 53.98 154.62
M Top Chron 25N bC24r 57.10 185.48
M C25N/C25R bC25n 57.66 195.94
M C25R/C26N bC25r 58.96 210.20
M C26N/C26R bC26n 59.24 213.06
M C29N/C29R bC29n 65.69 247.55
K/Pg K/Pg K/Pg 66.04 247.81
M C29R/C30N bC29R/C30N 66.40 252.28
M C31N/C31R bC31N/C31R 69.27 272.25
M C31R/C32N bC31R/C32N 71.45 283.39
M C32R.2N/C32n.2R bC32R.2N/C32n.2R 73.65 302.78
M 32R.2R/C33N b32R.2R/C33N 74.31 308.02
B.T. Huber et al. Global and Planetary Change 167 (2018) 1–23
13
Eocene.
Long-term trends as well as short-term excursions are apparent in
the compilation. The most signicant oxygen isotope events resolved
are the Aptian/Albian boundary interval event (AABI; Huber et al.,
2011), Cretaceous Thermal Maximum (KTM), Deccan basaltic province
eruption event (Deccan; Barrera and Savin, 1999;Wilf et al., 2003),
Paleocene/Eocene Thermal Maximum (PETM; Stott and Kennett, 1990;
Kennett and Stott, 1991;Thomas and Shackleton, 1996), and Middle
Eocene Climatic Optimum (MECO; e.g., (Bohaty et al., 2009). The ex-
tremely low oxygen isotopic values across the KTM and AABI relative to
the other events is readily apparent as are generally lower δ
18
O during
most of the Cretaceous greenhouse interval compared to the Eocene
greenhouse. In contrast, values at the transitions are quite comparable.
That is, whereas the δ
18
O values suggest most of the Cretaceous
greenhouse was warmer than its Cenozoic counterpart, apparent SHL
temperatures during the Albian, Maastrichtian, and late Eocene are
quite comparable. As ice sheets are known to have formed in Antarctica
by the late Eocene (e.g., Ehrmann and Mackensen, 1992;Carter et al.,
2017), determining whether or not ice was present during the Maas-
trichtian and Albian would provide two important, additional test cases
for the conditions when (and if) glacial ice accumulates and decays,
thereby better constraining over what range of conditions the climate
system can tip from an ice house to an ice-free state.
4.1. Cretaceous temperatures and δ
w
of southern high latitude seawater
New benthic and planktonic foraminiferal δ
18
O data from the
southern Indian Ocean (Site 258) corroborate low to extremely low
δ
18
O values previously reported for Falkland Plateau (Sites 511 and
327) and strengthen the argument that the Cretaceous hot greenhouse
was warmer than its Cenozoic counterpart. These new data undermine
arguments that Falkland Plateau results are anomalous due to local
oceanographic and/or sedimentological conditions. Diagenetic artifacts
also seem an unlikely explanation for the low values. Excellent pre-
servation at Site 511 is one critical counter argument (Bice et al., 2003).
The very good, but frosty, preservation at 258 is a second. During early
burial in near seaoor conditions, overgrowths (Fig. 2) would have
shifted values toward results for co-occurring benthics and imparted a
positive bias in the 258 planktonic results. That is, Turonian planktonic
and benthic δ
18
O values < 3.7and 1.8, respectively, are
observationally well supported based on results from two widely se-
parated sites in which the foraminifera were preserved in dierent
lithologies with dierent burial histories. If regional seawater had δ
18
O
values close to the canonical ice-free value of 1
SMOW
,δ
18
O values
of < 3.7and 1.8imply surface water temperatures > 27 °C
and bottom water temperatures > 19 °C.
Surface water temperatures in excess of 27 °C at 5860 °S are pro-
blematically warm based on all model results of which we are aware. To
achieve ~30 °C surface temperatures during the Turonian, modeling
experiments required 65007500 ppm CO
2
and a 50% increase in
poleward heat transport (Bice et al., 2003). SpecicCO
2
levels reported
are likely a function of the General Circulation Model used by Bice et al.
(2003), as other models with dierent climate sensitivities and assumed
boundary conditions produce dierent temperature simulations at the
same CO
2
levels (e.g., Lunt et al., 2016), but ~30 °C temperatures at
60°S is warm regardless. In a more recent study, at 2800 ppmv CO
2
,
which is above the upper limit of pCO
2
estimated by paleobarometric
proxies for the Late Cretaceous (e.g., Hong and Lee, 2012;Foster et al.,
2017;Barral et al., 2017; see Fig. 6), modeled SHL sea surface tem-
peratures are between 10 and 20 °C (Zhou et al., 2012). Sea surface
temperatures < 20 °C at 60° paleolatitude are also predicted from lati-
tudinal gradients estimated from Cenomanian and Turonian sh teeth
δ
18
O values (Pucéat et al., 2007;Martin et al., 2014).
Higher precipitation rates and a steeper trend in the latitudinal
gradient in the δ
18
O values of precipitation provide a potential re-
solution to this paradox. Bice et al. (2003) discounted a lowering of
seawater δ
18
O values through increased runoand/or a positive pre-
cipitation balance to explain Site 511 data as their calculations resulted
in predicted salinities that would be too brackish for foraminifera. That
study used a δ
18
Oof10
SMOW
for freshwater input based on the
modern precipitation at comparable latitudes. Using an end member
value of 20
SMOW
for precipitation, though, can result in seawater
having a δ
18
O value of 3.2
SMOW
and salinity of 30 psu (within the
range at which modern planktonic foraminifera survive). Foraminiferal
tests with a δ
18
O value of 4
V-PDB
secreted in equilibrium with
3.2
SMOW
water would have formed at 17.3 °C, still warm for high
latitudes, but within the upper range of model predicted temperatures
(Fig. 8). An enhanced hydrologic cycle is commonly invoked for
greenhouse times, and a positive precipitation balance has been pre-
dicted by climate models in the southern South Atlantic during the
Turonian (Donnadieu et al., 2016;Zhou et al., 2008). Using predicted
salinity and/or predicted δ
18
O values of seawater and precipitation
inferred from proxy studies (e.g., Huber et al., 2011;MacLeod et al.,
2017) relative to model results (Zhou et al., 2008) could be a valuable
new validation target in model-data comparisons.
Low salinity surface waters also could explain high surface to sea-
oor isotopic gradients at SHL that are particularly apparent when SHL
trends are compared to long-term trends in other regions (Fig. 9). The
Turonian planktonic to benthic δ
18
O gradient is as large at SHL as it is
in the North Atlantic and larger than in the Indian Ocean. However, if
SHL planktonic foraminiferal δ
18
O values are low due to both low
surface water δ
18
O values and warm temperatures whereas surface
water δ
18
O values contributes little to the planktic to benthic δ
18
O
ODP Site 689
Depth (mbsf)
150
200
250
40 45 50 55 60 65 70 75
Forams
M
K/Pg bC32R.2R
b32R.2N
bC31R
bC31N
bC30R
bC30N
K/Pg
bPp
bPi
bPa
bC19r
bC19n
b17r
bC17n.3n
bC16r
bC15r
ODP Site 690
Depth (mbsf)
Age (Ma)
100
150
200
250
300
40 45 50 55 60 65 70 75
M
K/Pg
b32R.2R
bC32R.2N
bC31R
bC31N
bC29R
K/Pg
bC29n
bC26n
bC25r
bC25n
bC24r
C24n.3n
bC24n.1r
bC23r
bC23n.2
bC22n
bC21r
bC21n
bC20r
bC19r
bC19n
bC16r
bC16n.2n
Fig. 5. Age-depth models showing changes in sedimentation rates at ODP Sites
689 and 690 on Maud Rise.
B.T. Huber et al. Global and Planetary Change 167 (2018) 1–23
14
gradient at other sites, this comparison would be misleading. SHL
benthic δ
18
O values are generally comparable to values in the North
Atlantic, lower than those in the Pacic, and generally higher than
those in the Indian Ocean. SHL planktonic to benthic δ
18
O gradients
decrease markedly after the early Campanian due mainly to a dis-
portionately large increase in SHL planktonic δ
18
O values.
High δ
13
C gradients in SHL data are even more dramatic than the
high δ
18
O gradients. Temporal trends in planktonic and benthic δ
13
C
values for the three other regions plot almost entirely within the yellow
shading bounded by SHL planktonic and benthic δ
13
C curves.
Planktonic δ
13
C values are similar among regions, but SHL benthic δ
13
C
values are 1‰–2lower than Pacic, North Atlantic, and Indian
Ocean benthic values throughout the interval from the Turonian into
the late Campanian (Fig. 9). A well stratied water column with limited
mixing but with high productivity and ecient organic carbon export
from surface waters (that is, an ecient biological pump) would create
such a situation.
The success of combined high temperatures and altered hydrology
at explaining SHL data for much of the Late Cretaceous notwith-
standing, explanations for isotopic patterns across the Aptian/Albian
boundary interval (AABI) at Site 511 seem to require incorporation of
additional variables. Excess rainfall with low δ
18
O values at 20 °C
would result in the low planktonic δ
18
O values, but similarly low
benthic δ
18
O values are puzzling and more dicult to understand be-
cause AABI benthic δ
18
O values span a 5range. The problem is re-
versed for δ
13
C values where planktonic values span a range of 5.
Foraminiferal specimens are excellently preserved (glassy) throughout
the nearly 9 m thickness of the AABI interval. Further arguing against a
diagenetic explanation is that no lithologic break coincides with the
transition into or out of the AABI, but lithology varies considerably
across the interval with anomalous isotopic values (interbedded black
shale and muddy reddish and greenish gray nannofossil chalks). As a
working hypothesis, we propose sediments at Site 511 were deposited
within a small, shallow, quite restricted and isolated sub-basin at this
time. Isolation and a relatively small volume of water within the sub-
basin would mean the entire water column might be responsive to
changing local conditions. We acknowledge, though, that this is an ad
hoc explanation and not an explanationsupported by independent evi-
dence.
4.2. Deepwater circulation
In terms of circulation, the very low benthic δ
13
C values indicate
SHL bathyal waters were not a signicant source for intermediate or
Fig. 6. Compiled Cretaceous oxygen and carbon isotope data for benthic and planktonic foraminifera from Sites 257 and 258 in the southern Indian Ocean and Sites
327, 511, 689 and 690 in the southern South Atlantic (this study) shown relative to (left to right): (1) proxy estimates for Cretaceous pCO
2
including the following:
blue squares with crosses: Frenelopolis conifer estimates with ± 1σaround the mean pCO
2
level (Barral et al., 2017); green triangles: liverwort δ
13
C, red circles:
pedogenic carbonate δ
13
C, and crosses: leaf stomata shown with LOESS best t line through all but conifer data (from Foster et al., 2017 compilation); (2); (3) Sr
isotope seawater curve (McArthur et al., 2012); (4) regional and global Oceanic Anoxic Events (Takashima et al., 2006); (5) global large igneous province (LIP)
magma ux estimated by Con et al. (2006); and (6) global mid-ocean ridge magma ux estimated by Müller et al. (2016). Global subduction zone length estimated
by van der Meer et al. (2014). Strongly negative δ
18
O and δ
13
C values across the Aptian-Albian boundary interval are considered an artifact of a more restricted and
shallower depositional basin compared to later periods (see text). (For interpretation of the references to colour in this gure legend, the reader is referred to the web
version of this article.)
B.T. Huber et al. Global and Planetary Change 167 (2018) 1–23
15
deep waters in the Pacic, Indian, or North Atlantic at least until the
Maastrichtian (Fig. 9). SHL benthic δ
13
C values are higher by 12
than in the other regions from the Turonian through the Campanian
suggesting either SHL deep waters were largely isolated from deeper
water in other oceans or that any ow between regions would be to-
ward SHL. Circulation patterns inferred from neodymium isotopes (ε
Nd
)
support the idea that the SHL was not a source region for intermediate
or deep waters for most of the Cretaceous. Whereas ε
Nd
trends in North
Atlantic, Pacic and Tethyan sites indicate changing Late Cretaceous
circulation patterns among Atlantic, Pacic, Tethyan, and Boreal waters
(Dameron et al., 2017;Frank et al., 2005;Jiménez Berrocoso et al.,
2010;MacLeod et al., 2008;Martin et al., 2012;Pucéat et al., 2005;
Zheng et al., 2013), analogous studies of SHL sites nd highly variable
ε
Nd
values and trends that were attributed to local inputs and internal
circulation in the region at least until the Campanian (Moiroud et al.,
2016;Murphy and Thomas, 2012;Robinson et al., 2010;Robinson and
Vance, 2012;Thomas et al., 2014;Voigt et al., 2013).
During the late Campanian to Maastrichtian, planktonic and benthic
δ
13
C values converge, benthic δ
13
C increase to values similar to or
above values in other basins, and planktonic δ
18
O values increase. At
the same time, SHL ε
Nd
values converge. Together, these observations
suggest deep circulation in SHL became better connected both intern-
ally and with other basins coincident with cooling temperatures and the
transition away from the Late Cretaceous greenhouse interval (e.g.,
MacLeod and Huber, 1996;Robinson et al., 2010;Voigt et al., 2013).
Finally, relative to other regions, SHL δ
18
O trends (Fig. 9) seem to
best approximate the traditional characterization of Late Cretaceous
climate evolution with warming from the Albian through the Cen-
omanian followed by a hot greenhouse interval that spans the Turonian
to Santonian before a gradual and progressive cooling during the
Campanian and Maastrichtian (e.g., Douglas and Savin, 1975;Savin,
1977;Huber et al., 1995,2002;Cramer et al., 2009;Friedrich et al.,
2012;O'Brien et al., 2017). In other regions, patterns vary in the timing
and pattern of cooling. Stable isotope results from Exmouth Plateau
(Falzoni et al., 2016; Indian Ocean trend on Fig. 9) indicate that cooling
began in the Coniacian and eectively stabilized by the middle of the
Campanian. Long-term Pacic planktonic trends are not available, but
Pacic benthic δ
18
O values also show no long-term trends from the late
Campanian onwards. Big gaps from the early through middle Campa-
nian record at Blake Nose (Huber et al., 2002) complicate its use as
representative of the North Atlantic as do potential diagenetic concerns
(MacLeod et al., 2005), but these problems are mitigated by the fact
that TEX
86
trends from the U.S. Gulf Coast (Linnert et al., 2014) and
Demerara Rise (Forster et al., 2007) support and ll in gaps in the Blake
Nose record. Together, data from these sites indicate that North Atlantic
surface cooling eectively ended by the middle Campanian.
Fig. 7. Cretaceous and Paleogene oxygen isotope
data compiled for southern high latitude deep-sea
sites, shown relative to changes in global climatic
states. Columns are (left to right): (1) geologic age
and magnetic reversal chron (exported from
TimeScale Creator v. 7.1, https://engineering.
purdue.edu/Stratigraphy/tscreator/index/index.
php); (2) age; (3) foraminiferal oxygen isotope va-
lues for southern high latitude deep-sea drill sites
(see Fig. 5 for symbol explanation); (4) climatic
events, including AABI = Aptian/Albian Boundary
Interval, KTM = Cretaceous Thermal Maximum;
Deccan = Deccan volcanism event, PETM = Paleo-
cene/Eocene Thermal Maximum, and
MECO = Middle Eocene Climatic Optimum; and (5)
Earth's greenhouse climatic state. Dashed vertical
line at 1.8is threshold value used by Miller et al.
(1987) for middle bathyal benthic foraminiferal δ
18
O
values to mark Paleogene transition from ice-free to
ice sheet conditions in Antarctica.
B.T. Huber et al. Global and Planetary Change 167 (2018) 1–23
16
4.3. Threshold for ice sheet growth
The existence of ice sheets in Antarctica during the Cretaceous has
been debated for many years. On one hand, faunal and oral evidence
for high latitude warmth has led many to consider the Late Cretaceous
as too warm and equable to support ice sheets. Therefore, calculations
of Cretaceous oxygen isotope paleotemperatures commonly assume ice-
free conditions (Barron et al., 1981;Barron and Washington, 1985;
Savin et al., 1975;Shackleton and Kennett, 1975). On the other hand,
sequence stratigraphic studies have suggested that globally correlatable
rapid sea level falls of at least 50 m during the Cretaceous must have
been caused by the buildup of ice on Antarctica (Vail et al., 1977;Haq
et al., 1987;Miller et al., 2005;Müller et al., 2016).
The only direct evidence for glacial ice during the Cretaceous is a
possibly Valanginian (Early Cretaceous) tillite discovered in the Cadna-
owie Formation in the northern Flinders Range of South Australia,
(Alley and Frakes, 2003), which was located at 75 °S paleolatitude.
Slightly younger striated boulders with chatter marks and polished
surfaces found in association with the Bulldog Shale Member in the NE
Flinders Range have been interpreted as having been deposited during
glacial events in the mid- to late Aptian and early to middle Albian
(Alley et al., 2011). However, the glacial origin of the boulders remains
unproven since they occur adjacent to, rather than within, the Bulldog
Shale outcrop and the Valanginian age is based on palynomorphs ob-
tained from units that overlie rather than occur within the tillite bed.
Coordinated positive shifts in Cretaceous planktonic and benthic
foraminiferal δ
18
O records that occur at levels of sea level fall would
provide secondary evidence for ice sheet growth (Prentice and
Matthews, 1991) and such evidence has been proposed for events in the
mid-Cenomanian (~95 Ma) and the early Turonian (~91 Ma; Miller
et al., 2004,2005;Bornemann et al., 2008). More detailed study,
though, has shown the proposed correlation fails tests at higher spatial
and temporal resolution (Ando et al., 2009;MacLeod et al., 2013;
Moriya et al., 2007) and is not supported by climate modeling studies
(Ladant and Donnadieu, 2016). Further, the accuracy of cited eustatic
curves during the late Cenomanian through Turonian has been ques-
tioned with alternative dating and regional forcing proposed for key
events (Haq and Huber, 2017).
The best candidate for a time with Late Cretaceous ice sheets is the
early Maastrichtian. Like above, proposals that a small ice sheet formed
during the early Maastrichtian are founded on purported synchroneity
between sea level events and positive increases in foraminiferal δ
18
O
values (Barrera, 1994;Barrera and Savin, 1999;Barrera et al., 1997;
Miller et al., 1999). However, uncertainty in age correlations of the sea-
level events, inconsistency in the timing of δ
18
O data trends among
dierent locations, and lack of sedimentological evidence of con-
temporary glaciation cast doubt on the conclusions (MacLeod and
Huber, 2001). Unlike the mid-Cenomanian and early Turonian, though,
bathyal temperatures inferred from early Maastrichtian benthic δ
18
O
are at a Late Cretaceous minimum (Figs. 6, 7), climate model simula-
tions exist that favor perennial ice accumulation (Ladant and
Donnadieu, 2016), and seasonal sea ice has been proposed to have
existed based on sedimentological and paleontological evidence from
the Arctic (Davies et al., 2009;Ladant and Donnadieu, 2016) and
Antarctic (Bowman et al., 2013). Still, no glacial deposits have been
found.
Foraminiferal δ
18
O records alone can place few rm constraints on
when or if ice sheets were present during greenhouse climates, but
knowing what benthic values correspond to times of ice sheet growth
and which correspond to ice-free conditions would help establish
thresholds for when transitions might occur. Ice sheets formed and
grew as the late Eocene climate cooled, and a large East Antarctic ice
sheet eventually advanced to sea level while benthic foraminiferal δ
18
O
values increased. A critical value for middle bathyal benthic for-
aminifera has been proposed as 1.8(Miller et al., 1987,2005).
Benthic foraminiferal δ
18
O measurements from Kerguelen Plateau
(southern Indian Ocean) reach this value by ~35 Ma, which coincides
with a sharp increase in ice-rafted debris (IRD) (Barron et al., 1991;
Wise Jr. et al., 1992;Zachos et al., 1992). Latest Eocene benthic for-
aminiferal δ
18
O values from Maud Rise range between 1.0and 1.6
and increase to 2.6by the time the East Antarctic ice sheet formed in
the early Oligocene (Stott and Kennett, 1990;Mackensen and Ehrmann,
1992;Fig. 7). The oldest Cenozoic IRD was reported from late middle
Eocene (~4544 Ma) sediments on Kerguelen Plateau and Maud Rise
(Ehrmann and Mackensen, 1992). Corresponding δ
18
O values at Maud
Rise for that time range between 1.0 and 0.5, which is comparable
to the range of values at Maud Rise during the Maastrichtian. Note,
though, that dierences in palaeogeography, moisture sources, trans-
port pathways, altitude of ice formation, and CO
2
thresholds all aect
ice sheets (Linnert et al., 2014;Ladant and Donnadieu, 2016). Thus,
while Maastrichtian benthic foraminiferal δ
18
O values approach those
from Paleogene times of known ice build-up, caution should be used in
deciding whether or not this similarity signies ice existed in the Late
Cretaceous.
Cool temperatures inferred from fossils collected in shallow marine
deposits on Seymour Island (northeast Antarctic Peninsula), located at
~62 °S paleolatitude (Müller et al., 2016), continue to fuel speculation
on the presence of Cretaceous greenhouse glaciers. The δ
18
O values of
Maastrichtian mollusks yield a wide range of values within the last
4 m.y. of the Maastrichtian, with coolest mean temperatures ranging
from 4 to 6 °C (Tobin et al., 2012). These values compare well with
coolest temperature estimates from several benthic foraminiferal spe-
cies (Barrera et al., 1987), but they dier in that the foraminiferal re-
cords show no discernable stratigraphic trend while the mollusk data
suggest a slight warming toward the end of the Maastrichtian. Clumped
isotope analyses of a subset of the Tobin et al. (2012) samples, on the
other hand, suggested more dramatic temperature swings than inferred
because of correlated changes in temperatures and local seawater δ
18
O
values that tended to cancel each other out (Petersen et al., 2016).
Kemp et al. (2014) also suggest cool temperatures (8 to 12 °C ± 5 °C)
Fig. 8. Relationships between the oxygen isotopic composition of calcite as a
function of the temperature, salinity, and density of seawater using the model of
(Railsback et al., 1989). Numbers in the gray circles are values for density (in ơ
t
units) for the isopycnals, which are shown as gray lines. Isopleths of δ
18
O
planktonic foraminiferal calcite are calculated in equilibrium (dotted lines)
using the paleotemperature equation of Kim and O'Neil, 1997; as reformulated
by Bemis et al., 1998) assuming (1) that water with a salinity of 34 psu had a
δ
18
O composition of 1
SMOW
(appropriate for an ice-free world), (2) fresh
water inputs had a value of 20
SMOW,
and (3) the evaporative fraction was
0.63/psu (Huber et al., 2011).
B.T. Huber et al. Global and Planetary Change 167 (2018) 1–23
17
for the latest Cretaceous on Seymour Island using the MBT/CBT (me-
thylation index of branched tetraethers/cyclization ratio of branched
tetraethers) paleothermometer. Analyses of growth rings from Seymour
Island fossil wood were tentatively interpreted to indicate late Maas-
trichtian temperatures between 4 and 8 °C (Francis and Poole, 2002).
However, several uncertainties, including taxonomic biases, tree com-
petition, tree maturity, and moisture availability, limit reliability of
temperature estimates using this proxy (Creber and Chaloner, 1984).
Finally, peaks in abundance of a dinoagellate resting cyst, also from
the Maastrichtian sequence on Seymour Island, were interpreted to
represent times of sea ice formation based on weak correlation with the
Tobin et al. (2012) molluscan δ
18
O dataset and speculative paleobio-
logic assumptions (Bowman et al., 2013,2014). Considering sedi-
mentologic evidence for glaciation or sea level change has never been
found in the intensively studied and extensively exposed Seymour Is-
land sequence (e.g., Askin, 1988;Huber, 1988;Macellari, 1988), pro-
posals for ice sheet growth in the Antarctic Peninsula region remain
unsupported.
In the absence of direct evidence for ice sheets at any time during
the mid-Late Cretaceous, the next strongest possible support for the
presence of ice sheets would be well documented synchroneity between
positive shifts in δ
18
O from biogenic carbonate and/or phosphate and
sea level falls across a wide geographic range and dierent pa-
leoenvironments which has not yet been demonstrated. Absence of such
evidence indicates that if Cretaceous ice sheets existed, their size was
too small to have aected at a resolvable level and on a global scale
both the oxygen isotopic composition of the ocean and changes sea
level given current limitations of correlation.
5. Implications for cretaceous climate forcing mechanisms
The importance of atmospheric carbon dioxide concentrations
(pCO
2
) in regulating global temperature variations throughout the
Phanerozoic has been accepted for more than a century (Arrhenius,
1896;Arthur et al., 1985;Budyko and Ronov, 1979;Chamberlin, 1899;
Barron and Washington, 1985;Hay, 2008, 2011). Natural shifts in pCO
2
have primarily been inuenced by changes in the relative rates of
volcanic outgassing, chemical weathering of silicate rocks, and organic
carbon burial. Periods of intensied oceanic crust formation, LIP ac-
tivity, and/or arc production rates lead to a buildup of pCO
2
and in-
creased global temperatures (Con and Eldholm, 1994;Larson, 1991b;
Con et al., 2006), while times of intensied continent-continent plate
collision result in increased chemical weathering rates, a drawdown of
pCO
2
, and cooler global temperatures (e.g., Raymo, 1991;Raymo et al.,
1988). Obtaining accurate records of ancient temperature and pCO
2
changes and linking them to tectonic forcing mechanisms have been a
major challenge for climate scientists because of ambiguities or un-
certainties in the reliability of proxies used to reconstruct past tem-
peratures and pCO
2
concentrations and incomplete information on the
timing and rates of global tectonic events.
Comparison of the oxygen isotope paleotemperature compilation for
the Cretaceous southern high latitudes with estimates of variation in
pCO
2
, crustal production at mid-ocean ridges and (LIPs), and lengths of
continental and island arcs (Fig. 6) demonstrates that links between
Fig. 9. Late Cretaceous planktonic and benthic foraminiferal δ
18
O values (left panel) and δ
13
C values (center panel) plotted against age illustrating regional
dierences in isotopic trends and large changes in the planktonic to benthic isotopic gradient at southern high latitudes (SHL) through time. Compiled data for each
region were plotted on a common age axis constructed by interpolating between reported Age boundaries. A qualitative t was then drawn to present a simplied
depiction of trends through time. The right hand panel shows a representative plot of this approach applied to the SHL δ
13
C data from Fig. 6.Dierent regions are
shown by dierent colors. Black lines depict SHL trends with yellow shading between curves included to emphasize changes in gradients between the sea surface and
the sea oor gradients at SHL; gray band indicates Cenomanian interval where data are eectively absent for SHL. Other regions and data sources are brown- North
Atlantic (Huber et al., 2002;MacLeod et al., 2005;Linnert et al., 2014), blue- subtropical Pacic(Friedrich et al., 2012;Ando et al., 2013), and green- Exmouth
Plateau in the subtropical Indian Ocean (Falzoni et al., 2016). Solid lines represent trends in planktonic δ
18
O values; dashed lines, benthic values; dotted brown line
in δ
18
O plot trend in expected North Atlantic δ
18
O values estimated from TEX
86
-based temperature estimates for surface water (see text). (For interpretation of the
references to colour in this gure legend, the reader is referred to the web version of this article.)
B.T. Huber et al. Global and Planetary Change 167 (2018) 1–23
18
Cretaceous temperature variations and any individual primary forcing
factors are not well established. In particular, the proxies used to esti-
mate Cretaceous pCO
2
show signicant inconsistencies among methods
and variable degrees of correlation to long-term paleotemperature
trends. For example, the hot greenhouse temperatures of the Turonian
correlate with a period of relatively low pCO
2
according to the fossil
conifer leaf and stomatal index proxies, while the cooler temperatures
of the Albian correlate with a time of higher pCO
2
according to the
stomatal and liverwort indices but lower pCO
2
according to the fossil
conifer pCO
2
proxy. Large uncertainties regarding the inuence of
moisture and sunlight availability, on leaf stomata density and meta-
bolics, the δ
13
C composition of the preserved organic matter and esti-
mates of mean land surface temperatures used in pCO
2
estimates from
soil carbonates may explain some of these discrepancies (Royer et al.,
2014;Breecker et al., 2010;Foster et al., 2017).
Long-term variations in
87
Sr/
86
Sr ratios of marine sediments track
the relative importance and absolute rate of continental erosion vs.
oceanic magmatism to the ocean's Sr budget with shifts moderated by
erosion of older sedimentary rocks (Bralower et al., 1997;Jones et al.,
1994;Larson, 1991a,b;McArthur et al., 2012). Higher
87
Sr/
86
Sr ratios
are taken to indicate dominance of continental mountain building and
chemical weathering while lower ratios indicate increased oceanic
magmatism and/or erosion of large basaltic provinces on land. The
mean seawater
87
Sr/
86
Sr curve (Fig. 6) shows generally good agreement
with the paleotemperature compilation, with lowest values occurring
during the hot greenhouse temperatures in the Turonian-Santonian and
ratios that increase minimum values during the Maastrichtian when
temperatures were coolest.
Changes in rates of Cretaceous oceanic crust production (Müller
et al., 2016), in the timing of formation of LIP (Con et al., 2006), and
in lengths of subduction-related volcanic arcs (van der Meer et al.,
2014), correspond to variations in Cretaceous SHL temperatures in
some time intervals, but not in others (Fig. 6). The largest short-term
variation in oceanic crustal production rate in the ~115 to ~66 Ma
interval, a decrease of ~8 km
3
/yr between ~100 and ~98 Ma
(Matthews et al., 2012), occurred during a gap in the SHL δ
18
O record.
However, the decrease in δ
18
O values between ~102 and ~94 Ma ar-
gues for increased temperatures, the opposite of the expected impact of
a decrease in oceanic crustal production and associated lower CO
2
de-
gassing rates. Similarly, the overall trend of increasing oceanic crustal
production rates from ~98 to ~65 Ma corresponds to the long-term
trend of decreasing temperatures from the Turonian to the Maas-
trichtian. In contrast, a peak in LIP production and associated CO
2
degassing between ~100 and ~95 Ma corresponds well with the de-
crease in δ
18
O values between ~102 and ~94 Ma arguing for increased
temperatures. Otherwise, the overall trend of decreasing LIP production
rates from ~95 to ~66 Ma agrees with the overall trend of decreasing
temperatures in the same time interval. Subduction-related volcanic arc
lengths, a proxy for arc crustal production and associated CO
2
degas-
sing, show an overall decrease from ~115 to ~65 Ma, whereas tem-
peratures increase between ~102 and ~94 Ma but decrease from ~95
to ~66 Ma.
Correspondence between the timing of peak volcanism at the
Deccan Traps (southwest India) and global warming provides strong
support for volcanic CO
2
as the trigger for a warming event during the
late Maastrichtian. Magma eruption rates for the Deccan Event were
highest during the late Maastrichtian, ~150300 k.y. before the
Cretaceous/Paleogene boundary (Ravizza et al., 2001;Robinson et al.,
2009;Renne et al., 2015;Font et al., 2016), which coincides with a
2.55 °C temperature increase estimated from δ
18
O measurements from
deep sea benthic foraminifera from multiple equatorial, mid-latitude
and high latitude locations (Barrera, 1994;Barrera and Savin, 1999;
Huber et al., 2002;Westerhold et al., 2011;Barnet et al., 2018) and an
estimated warming of ~3 °C on Seymour Island, Antarctica (Tobin
et al., 2012). This event is well-delineated between 66.5 and 66.2 Ma
with an ~4 °C warming of SHL at ODP Site 690 (Barrera and Savin,
1999), as shown in Figs. 6 and 7.
Inconsistencies between the timing of elevated rates of tectonic
sources of CO
2
and evidence for increased SHL warming (Fig. 6) can be
attributed to multiple factors, including uncertainties in: (1) the rates of
lithosphere production and consumption (van der Meer et al., 2014),
(2) the amount of carbon concentrated in descending slabs and dec-
arbonation eciency at subduction zones (Johnston et al., 2011), (3)
estimates of the timing and volume of magma production during LIP
eruptions (Con et al., 2006), (4) estimates for the amount of CO
2
released from various types of volcanoes (Burton et al., 2013), (5) es-
timates of the length of island arcs (van der Meer et al., 2014), and (6)
estimates of weathering rates of volcanic and non-volcanic rocks
(Berner, 2006). These inconsistencies and uncertainties will diminish as
gaps in the SHL oxygen isotope paleotemperature curve are lled and
constraints on changing pCO
2
that integrate pooled estimates for mantle
derived and oxidized organic carbon CO
2
sources as well as weathering
and organic carbon burial sinks improve.
6. Conclusions
New benthic and planktonic foraminiferal δ
18
O data from the
southern Indian Ocean (Mentelle Basin DSDP Site 258) conrm low to
extremely low δ
18
O values previously reported for the southern South
Atlantic (Falkland Plateau DSDP Sites 511 and 327) and strengthen the
argument that the Cretaceous Hot Greenhouse was considerably
warmer than its Cenozoic counterpart. At Site 258 warmest tempera-
tures occur within the lowermost Turonian and just above a black shale
bed that is correlated with Oceanic Anoxic Event 2. These new data
undermine arguments that Falkland Plateau results are anomalous due
to local oceanographic and/or sedimentological conditions. If regional
seawater had δ
18
O values close to the canonical ice-free value of
1
SMOW
, Turonian foraminiferal δ
18
O values imply surface water
temperatures > 27 °C and bottom water temperatures > 19 °C. Such
extreme warmth at 5860°S seems improbable and an alternative pos-
sibility is that SHL δ
18
O values were at least partially inuenced by
higher precipitation rates and a steeper trend in the seawater latitudinal
δ
18
O gradient.
The δ
18
O trends at SHL seem to best approximate the traditional
characterization of Late Cretaceous climate evolution with warming
from the Albian through the Cenomanian, followed by peak warmth
during the KTM and sustained warmth through the Santonian before a
gradual and progressive cooling during the Campanian and
Maastrichtian. We nd no compelling evidence for ice sheets of sig-
nicant size at any time during this entire time. Compilations of δ
18
O
paleotemperature data from the SHL (this study) and other global δ
18
O
and TEX
86
compilations (e.g., Huber et al., 2002;Pucéat et al., 2005;
Cramer et al., 2009;Friedrich et al., 2012;O'Brien et al., 2017) along
with climate modeling studies (Ladant and Donnadieu, 2016) all in-
dicate that the Cretaceous Hot Greenhouse (Turonian-Santonian) was
too warm to support polar ice sheets except possibly at the highest al-
titudes in Antarctica. Even during the coolest times of the mid-Late
Cretaceous (late Aptian and mid-Maastrichtian), absence of glacial de-
posits, of coordinated positive shifts in Cretaceous planktonic and
benthic foraminiferal δ
18
O records and of corresponding eustatic falls in
sea level cast doubt on recurring proposals for the existence of polar ice
sheets during those times (e.g., Barrera, 1994;Barrera and Savin, 1999;
Barrera et al., 1997;Miller et al., 1999;Bowman et al., 2013,2014).
Throughout the Cretaceous Hot Greenhouse the SHL sites show high