V O L . 1 2 0 N O . 5 / 6 M AY / J U N E 2 0 0 8
S UE •
IN
H
S
I
S
I
• IN
E
U
T
SS
T H IS IS S U
China
n
r
e
t
s
We axes
rel
j
a
tarctic
n
A
n
i
fire
Ice on
j
umed
h
x
e
n
sca
Franci Mt. Diablo
in
E
• IN T HIS I
ISSN 0016-7606
Eocene climate record of a high southern latitude continental shelf:
Seymour Island, Antarctica
Linda C. Ivany†
Department of Earth Sciences, Syracuse University, Syracuse, New York 13244, USA
Kyger C. Lohmann
Franciszek Hasiuk
Department of Geological Sciences, University of Michigan, Ann Arbor, Michigan 48109, USA
Daniel B. Blake
Alexander Glass*
Department of Geology, University of Illinois, Urbana, Illinois 61801, USA
Richard B. Aronson
Ryan M. Moody
Dauphin Island Sea Lab, Dauphin Island, Alabama 36528, USA
ABSTRACT
A high-resolution record of Eocene paleotemperature variation is preserved within
the high southern latitude, marine shelf
succession of the La Meseta Formation on
Seymour Island, off the NE side of the Antarctic Peninsula. 87Sr/86Sr ratios of bivalve
shell carbonate indicate that the La Meseta
Formation spans virtually the entire Eocene,
and suggest the presence of an early middle
Eocene unconformity. Paleoclimatic and
paleoceanographic inferences are based on
the stable oxygen and carbon isotope values
of two genera of bivalves collected with a
high degree of stratigraphic resolution within
the formation, and with multiple replicate
samples from each horizon. δ18O data indicate roughly 10 °C of cooling from the early
Eocene climatic optimum (~15 °C) through
the end of the Eocene (minimum ~5 °C),
much of which took place in two comparatively short intervals at ca. 52 and ca. 41 Ma.
Many features of the isotope curves generated
from this Eocene shelf section are apparent
in δ18O and δ13C data from the Southern and
global oceans, including warm intervals that
likely correspond to the early and middle
Eocene climatic optima (EECO and MECO).
†
E-mail: [email protected]
*Present address: Division of Earth and Ocean
Sciences, Duke University, Durham, North Carolina
27708, USA
A rapid middle Eocene shift to much more
positive values is the most significant in the
section and reflects a drop to universally
cooler temperatures in the late middle and
late Eocene that might also be associated
with a short-lived glacial advance. However,
even using a somewhat depleted value for
δ18O of seawater in the Antarctic peninsular
region, average Seymour Island shelf-water
paleotemperatures did not reach freezing
before the end of the Eocene. δ13C data similarly reflect the documented middle Eocene
surface ocean enrichment followed by more
negative values, but depletion is much more
pronounced on Seymour Island and persists
for the remainder of the Eocene, suggesting a
combination of upwelling, metabolic effects,
and/or atypical carbon cycling on the shelf
in this region. Isotope data capture information about changes in the paleoenvironment
that also had consequences for the biota, as
published paleontological records document
marked change in the nature of terrestrial
and marine biota at this time. The fact that
middle Eocene cooling and biotic turnover
in the Peninsular region correspond well in
time to the proposed initial opening of Drake
Passage suggests that the formation of gateways, in addition to changes in pCO2, had
significant consequences for the Earth’s climate system during the Paleogene.
Keywords: stable isotope, Eocene, climate, bivalve, Antarctica, strontium.
INTRODUCTION
The Eocene Epoch encompasses the climate
transition from the warmest interval of the Cenozoic through progressive cooling that culminated
in the initiation of Antarctic glaciation in the
earliest Oligocene. This fundamental shift led to
the establishment of present-day oceanographic
and atmospheric circulation patterns, and set the
stage for evolutionary turnovers that shaped the
modern marine biota. The global climate record
of this interval is reasonably well constrained by
studies of open ocean sequences (e.g., Zachos
et al., 2001, and many others), but it remains
unclear how representative these records are of
conditions on continental margins, from which
most of the macrofaunal marine fossil record is
derived. This is particularly true at high latitudes,
where exposed stratigraphic sections are rare,
and where climate change was most severe.
In this paper, we present a record of high-latitude shelf paleotemperatures from the Eocene
La Meseta Formation on Seymour Island, Antarctica. The Seymour Island section is the only
place known in all of Antarctica where Eocene
shelf sediments are exposed in outcrop and in
stratigraphic continuity. Because ongoing coring efforts on the shelf (Shallow Drilling on the
Antarctic Continental Margin [SHALDRIL]
and Antarctic Geological Drilling [ANDRILL])
have yet to successfully recover Eocene sediments due to the difficulties of drilling through
ice and penetrating the overlying mantle of
glacial debris, this section remains critical for
GSA Bulletin; May/June 2008; v. 120; no. 5/6; p. 659–678; doi: 10.1130/B26269.1; 11 figures; 2 tables.
For permission to copy, contact [email protected]
© 2008 Geological Society of America
659
Ivany et al.
low-water settings is needed. Carbonate macrofossils such as mollusks provide high-quality
records of skeletal δ18O and the shelf environment in which they lived because sampling can
be done within cross sections of shells, taking
care to avoid any material that may be secondary or diagenetically altered—a luxury that is
not possible with microfossils. Despite complications arising from variation in the δ18O of nearshore waters and depositional hiatuses imposed
by relative changes in base level, these data can
provide accurate estimates of shelf paleotemperatures that offer excellent complements to open
ocean records and thereby help to complete the
puzzle of climate history during this most interesting interval of Earth’s history.
studies of high-latitude climate in the Paleogene
greenhouse world. Our paleotemperature record
is derived from the stable isotope values of
bivalve carbonate collected throughout the section. Samples are well constrained stratigraphically and taxonomically, and include multiple
replicate samples from each horizon, and thus
provide for a significantly more detailed record
than earlier work. 87Sr/86Sr ratios of mollusk
carbonate allow samples to be positioned on
the Eocene seawater Sr-evolution curve in order
to augment existing biostratigraphically based
age control. In addition, because the section is
fossiliferous, these data allow for an integrated
examination of the response of the Antarctic
shelf biota to Eocene climate change, which
we hope will facilitate more detailed studies of
paleontological trends within the formation.
A record of Eocene shelf paleotemperatures
from the high southern latitudes will help to
refine climate models and thereby further understanding of the mechanisms responsible for
global climate cooling. High-latitude records are
particularly necessary to constrain pole-to-equator thermal gradients and how they evolved over
the course of the Eocene. In addition, the potential for δ18O values of planktonic foraminifera to
underestimate surface-water paleotemperatures
due to precipitation of secondary calcite in cold
deep waters or pore waters (Zachos et al., 1994;
Pearson et al., 2007) suggests that examining
independent types of records from Eocene shal-
km
Seymour Island is located roughly 100 km
east of the Antarctic Peninsula near its northern
end (Fig. 1). Fossiliferous marine outcrops on the
NE third of the island span much of the Eocene,
and thus provide an ideal opportunity to study
the effects of cooling on high-latitude, marine
shelf environments and their associated faunas.
The Eocene La Meseta Formation (Elliot and
Trautman, 1982) is a shallow marine succession
composed of sandstones, mudstones, and shell
banks that accumulated in a variety of inner
shelf environments (Sadler, 1988; Porębski,
1995, 2000; Marenssi et al., 1998, 2002). The
100
N
0
GEOLOGIC SETTING
Seymour
Island
South
Shetland
Islands
James Ross
Island
a
nsul
Peni
rctic
Anta
Weddell
Sea
Figure 1. The location of Seymour Island in the James Ross
Basin off the east side of the Antarctic Peninsula.
660
unit is confined within a fault-bounded NW-SE–
trending basin on the east side of the Antarctic
Peninsula. Underlying Cretaceous and Paleocene shelf sediments of the James Ross Basin
were tilted gently to the SE by uplift associated with arc volcanism along the west side of
the Peninsula, and dissected prior to the onset
of La Meseta Formation deposition (Elliot,
1988; Macellari, 1988). The formation has been
described as a compound, incised–valley-fill
system (Porębski, 2000). Internal architecture is
complex, with stacked and interfingering sediment lenses and occasional slumps of variable
thickness. Consequently, and because the floor
of the basin at the channel’s axis is not exposed,
formation thickness has been difficult to measure, with estimates ranging from just over
300 m to an expanded composite section estimate of over 700 m (Elliot, 1988; Marenssi et
al., 1998). Petrographic study of the La Meseta
Formation (Marenssi et al., 2002) and diagenetic
analysis of underlying units on the nearby island
of Vega (Pirrie et al., 1994) demonstrate that the
section has not been buried more than 1 km and
has undergone minimal diagenetic alteration.
Stratigraphy and Depositional Environment
The original three-part lithologic division of
the formation by Elliot and Trautman (1982)
was expanded to seven distinct, mappable,
lithologic units, or Telms (“Tertiary Eocene La
Meseta”) by Sadler (1988). More recently, several authors have attempted to recognize genetically related, unconformity-bound, depositional
sequences reflecting the interplay of fault-controlled subsidence and eustatic change in sea
level. Porębski (1995, 2000) defined three large
sequences, the boundaries of which he believes
to be dominantly eustatically controlled. Smaller
scale transgressive packages within, however, are
more likely to be associated with tectonic processes, since the basin is several hundred meters
deep yet facies are dominantly shallow marine
throughout (Porębski, 1995). The six allomembers of Marenssi et al. (1998) are consistent with
those of Porębski, but more finely subdivided.
He sees eustasy as a more important control on
deposition, and calls for fluvial erosion at the
base of each sequence (Marenssi, 2006). A common feature to all, however, is the recognition
of a significant depositional hiatus at the base
of Telm 4, a surface associated with ravinement,
perhaps following incision, and the generation of
a glauconite- and phosphate-rich lag.
Different authors have variously stressed
estuarine, deltaic, and channel or valley fill
interpretations for the depositional setting of
different parts of the La Meseta Formation
(e.g., Elliot and Trautman, 1982; Sadler, 1988;
Geological Society of America Bulletin, May/June 2008
Eocene climate change on the Antarctic shelf
Wiedman and Feldman, 1988; Stilwell and Zinsmeister, 1992; Porębski, 1995, 2000; Marenssi
et al., 1998, 2002). The shallow-water, tidally
influenced nature of component facies, the evidence for channelization, and the fault-bounded
nature of the entire unit have led recent workers
to favor the estuarine model (Porębski, 1995,
2000; Marenssi et al., 1998). While such interpretations imply (but do not necessarily require)
an influx of fresh water from an associated river,
the diverse nature of the invertebrate fauna preserved throughout the unit suggests accumulation under relatively normal marine conditions.
An exploration of the variability in stable isotope
values, both spatially and temporally, will help
to resolve uncertainties in the interpretation of
depositional setting. Note though that, because
geochemical data are derived from carbonate
macrofossils, they say nothing about the conditions that prevailed during deposition of intervening more barren intervals. While we made
a concerted effort to sample every fossiliferous
horizon we encountered in vertical sections, it is
possible that erosional truncation along channel
bases removed portions of the section that might
be preserved elsewhere along strike. More extensive discussions of sedimentology, stratigraphic
relations, and inferred depositional environments
of stratigraphic subunits are found in Elliot and
Trautman (1982), Sadler (1988), Stilwell and
Zinsmeister (1992), Porębski (1995, 2000), and
Marenssi et al. (1998, 2002), as well as below in
the context of isotope geochemistry.
Age
Age control within the La Meseta Formation
has been based primarily on biostratigraphy
and suggests that deposition spanned much of
the Eocene, but uncertainty remains about the
age of particular units within the formation.
Wrenn and Hart (1988) used dinoflagellate
assemblages to identify a late early Eocene and
a late-middle to late Eocene interval separated
by the discontinuity at the base of Telm 4. Askin
et al. (1991), Cocozza and Clarke (1992), and
Askin (1997) support these bounds on the base
and top of the unit with terrestrial and marine
palynomorphs, although they place most of the
formation within the middle Eocene. Diatoms
in the lower part of the formation also yield
an early Eocene age estimate (Harwood, 1985,
1988). Mollusk faunas support an Eocene age,
with a struthiolariid gastropod from Telms 4 and
5 being indicative of the late Eocene (Stilwell
and Zinsmeister, 1992). Hall (1977) used dinoflagellate assemblages to suggest a late Eocene
to early Oligocene age for the top of the formation, and Fordyce (1989) concurred based on the
presence of a cetacean skeleton with mysticete
affinities (Mitchell, 1989).
More recently, workers have used strontium
isotope stratigraphy to augment the existing biostratigraphy. Dutton et al. (2002) present strontium isotope ratios from a limited number of
mollusk shells from Telms 2, 5, and 7 that suggest
early, middle, and late Eocene ages, respectively.
Marenssi (2006) concurs with data from the base
of Telm 2 that yield an early Eocene age. Dingle
and Lavelle (1998) and Ivany et al. (2006) report
strontium isotope ratios from shells at the top
of the formation that are consistent with ages
at or just below the Eocene-Oligocene boundary. Here, we report additional strontium isotope
ratios from mollusks collected in stratigraphic
order throughout the section in the hopes that
they will help resolve some of the inconsistencies in age estimates within the formation.
Previous Paleoclimate Interpretations
Paleontological studies of La Meseta floral
and faunal assemblages suggest cool to warmtemperate conditions during deposition of the
unit (Askin and Fleming, 1982, 1997; Case,
1988; Doktor et al., 1996; Francis and Poole,
2002; Stilwell and Zinsmeister, 1992; Wiedman
et al., 1988). The presence of fossil driftwood
through much of the section (Fig. 2) demonstrates that conditions were mild enough to support the growth of significant temperate rain forests on the peninsula (Case, 1988; Francis and
Poole, 2002; Poole and Cantrill, 2006). Dingle
et al. (1998) used clay mineral assemblages to
recognize a shift from warm, wet conditions
to cooler and drier at the top of the formation.
Paleotemperatures have been estimated using
δ18O values of early marine cements (from the
high-Mg calcite linings of Teredolites borings;
Pirrie et al., 1998) and an assortment of invertebrate macrofossils (Gaździcki et al., 1992; Ditchfield et al., 1994) collected within the formation.
All yield temperature estimates ranging between
5 and 15 °C and generally show cooling where
there is stratigraphic control. Recently, Dutton
and colleagues (2002) constructed a temperature
record based on δ18O values of shell carbonate
from individuals of the bivalve genus Cucullaea,
in which they infer an overall pattern of Eocene
cooling from ~15 °C in Telms 2–5 to ~10.5 °C in
Telms 6 and 7. Shells were pooled from localities resolved only to the level of Telm, however, and hence yielded only six paleotemperature estimates for the Telm 2–7 section. Data
Figure 2. Fossil wood collected from
low in Telm 5 (locality 01–93), La
Meseta Formation. This section of a
large trunk was washed out to sea
and its outside surface colonized by
shipworms (wood-boring bivalves)
while still floating, producing the
trace fossil Teredolites. The wood
ultimately sank and was preserved
with a marine benthic fauna in a
nearshore environment. Such material is common in the formation up
through Telm 5, indicating the presence of forests on land. Telm—Tertiary Eocene La Meseta.
1 cm
Geological Society of America Bulletin, May/June 2008
661
Ivany et al.
’W
40
Cape Wiman
’S
13
Telm 4
o
56
o
64 o
15
64 o
14
64
Bahia Lopez de Bertodano
’S
(Paleocene)
’S
ey
Cross Vall
presented here reinforce this overall trend toward
cooler temperatures but, because of much higher
stratigraphic resolution, we are able to identify
structure in the data and offer a more detailed
record of climate change during this important
period of Earth history. The reader is referred
to Dutton et al. (2002) for additional information regarding the preservation and chemistry of
individual shells from this unit.
Telm 5
Telm 3
Telm 2
o
56
Lan
ding
MATERIALS AND METHODS
Strip
Base
Marambio
Field Collection and Specimen Preparation
662
W
Larsen Cove
(Paleocene)
Macrofossils were collected in situ from
shelly facies within the La Meseta Formation on Seymour Island and their locations
recorded using global positioning system (GPS)
coordinates and/or positioned on the U.S.
Geological Survey (USGS) topographic map
64056-T5-TM-010 produced in 1995. Relative
stratigraphic position and correlation with other
samples were determined in as much detail as
possible. Localities are mainly confined to the
north and west of the meseta that gives the
formation its name (Fig. 3). Wherever possible, multiple localities were collected within
a single stratigraphic horizon in order to assess
heterogeneity that may reflect differences in
time averaging or spatial variability in environmental conditions. Sampled horizons for which
there are replicates within Telms 4, 5, most of
6, and 7 could be physically traced along strike.
Two vertical sections were sampled within Telm
3 (Fig. 3, one to the west and one to the east)
between which precise correlation could not be
ascertained; their stratigraphic placement in the
Results section is based on the similarity of the
isotope trajectories between them. One locality
within Telm 6 was stratigraphically ambiguous
but its relative position inferred based on equivalent elevation with another site between which
the beds dipped slightly, requiring that one be
younger than the other.
The most abundant and/or consistently present macrofossil groups were collected for isotope analysis, including the arcid bivalve Cucullaea (C. raea and C. donaldi) and the venerid
bivalve Eurhomalea (E. antarctica, E. newtoni,
and E. florentinoi), both shallow infaunal suspension feeders, the naticid gastropod Polinices
subtenuis, a semi-infaunal predator of other mollusks, and the terebratellid brachiopod Bouchardia antarctica, an epifaunal suspension feeder
(Fig. 4). Species within genera are similar morphologically and occur in the same facies, and
thus are presumed to have very similar ecologies
and physiologies. Fossils were collected from
54 localities distributed within 27 stratigraphic
horizons throughout the formation.
’
36
Telm 7
(Shoreline
deposits, alluvium)
Telm 6
Penguin Bay
Telm 4
Alluvium, slump
Telm 3
Telm 7
Telm 2
Telm 6
Telm 1
Telm 5
isotope sample locality
0
1000m
Figure 3. Geologic map of the NE portion of Seymour Island (based on Sadler, 1988), showing units of the Eocene La Meseta Formation and locations of sampling localities for this
study. The meseta that gives rise to the formation’s name is capped by alluvium overlying
Telm 7 and is occupied by Argentina’s Base Marambio. Telm—Tertiary Eocene La Meseta.
Individuals were sectioned along the maximum growth axis (or as much as could be
revealed in a flat plane through the body whorl
for naticids) and polished to reveal growth
banding and internal architecture. Sections
were visually evaluated for quality of preservation as indicated by presence of original crystal
structure and absence of cracks and diagenetic
cement. Selected sections were scanned using
cathodoluminescent microscopy, and X-ray
diffraction was performed on mollusk subsamples to ensure primary aragonite mineralogy.
Cracks, discolored regions, or other apparent
shell defects that may reflect diagenetic alteration were avoided. Only the best preserved
specimens or portions of specimens were used
for isotope analysis.
Sampling
All samples for geochemical analyses were
collected from polished cross sections, not
external shell surfaces. Initially, we sampled
three to seven individuals of each taxon from
each of two different collecting localities in
order to assess sampling methodology and
differences in stable isotope values among
taxa. The two bivalve genera (Cucullaea and
Eurhomalea) were selected for the remainder of
the study because they were most consistently
present through the section and their internal
growth structure was the most straightforward
for microsampling. For all remaining localities
for which there is good stratigraphic control,
three to five individuals from each bivalve genus
(where possible) were prepared and sampled for
stable isotope geochemistry.
Estimates of average annual δ18O and δ13C
for each sampled shell were attained by drilling a path of sampled carbonate across a number of growth bands so as to obtain an average “bulk” composition representing several
years of growth. Three to five bulk samples
were collected from each individual in this
manner. Consecutive growth-band–parallel
samples were also collected across a number
of growth increments from two individuals of
each bivalve taxon from the same locality in
order to reveal differences in seasonal (intraannual) growth patterns. Shells were sampled
on the high-precision Merchantek (New Wave)
MicroMill computer-controlled milling systems at Syracuse University and at the University of Michigan. Resulting powdered samples
were analyzed directly or split for paired stable
isotopic and Sr-isotopic analysis.
Geological Society of America Bulletin, May/June 2008
Eocene climate change on the Antarctic shelf
B
A
C
D
Figure 4. Taxa used in this study for geochemical analysis. (A) the arcid bivalve Cucullaea raea; (B) the terebratellid
brachiopod Bouchardia antarctica; (C) the naticid gastropod Polinices subtenuis; (D) the venerid bivalve Eurhomalea
newtoni. Scale bars are 1 cm; same scale for B, C, and D.
Strontium Isotope Analysis
In an effort to improve temporal control
within the La Meseta Formation, 87Sr/86Sr ratios
of skeletal aragonite were analyzed from 43
bivalves distributed throughout the section
(Telms 2–7). Samples for Sr isotopic analysis
were dissolved in HNO3 and eluted through a
quartz cation exchange column using Sr-specific
resin (Eichrom). Between 50 and 100 ng of Sr
were dried down, re-dissolved in 1 μL HCl, and
loaded onto tungsten single filaments with Ta2O5
powder and 1 μL 0.3 M H3PO4. Analyses were
performed on the Finnigan MAT 262 thermal
ionization mass spectrometer at the University
of Michigan’s Radiogenic Isotope Geochemistry Laboratory. Sample splits were retained for
stable isotope analysis as well.
While the marine 87Sr/86Sr curve (McArthur
et al., 2001) shows only limited (and undulatory) variation throughout much of the Eocene,
multiple skeletal values distributed throughout the section combined with stratigraphic
constraints on their relative ages allow for the
construction of a coherent chronostratigraphy
into which to place stable isotope values and
interpret climate change. Age assignments were
made by fitting Sr-isotopic data to the 08/2004
version of LOWESS 3, the Phanerozoic Srisotope lookup table for global marine seawater
(Howarth and McArthur, 1997; McArthur et al.,
2001). Differences in inferred age for a given
87
Sr/86Sr ratio between this study and Dutton et
al. (2002) result from the difference in timescale
to which the global marine curve is calibrated;
our data are plotted relative to the Gradstein
et al. (2004) timescale. Best inferred ages are
constrained by both 87Sr/86Sr ratios and relative
stratigraphic position. When samples are plotted in stratigraphic order on the marine curve,
best estimates of their ages are determined by
adjusting age such that, wherever possible, values fall directly on the curve or their analytical
errors intersect it. In cases where multiple ratios
are derived from the same horizon, they are plotted such that they straddle the curve. Relative
stratigraphic position is maintained in all cases.
Error bars on age estimates derived solely from
the LOWESS (locally weighted regression scatterplot smoother) fit, generally ~±2 m.y., are
not plotted here, as stratigraphic position places
more stringent constraints on age. Note though
that poor knowledge of the marine curve during the early Eocene in particular introduces
the greatest amount of error on these estimates.
New data from Hodell et al. (2007) may help to
resolve some of this uncertainty.
Stable Isotope Analysis
remove volatile contaminants and water. They
were then reacted with anhydrous H3PO4 at 76
± 2 °C in a Kiel automated carbonate preparation
device coupled directly to the inlet of a Finnigan
MAT 251 or 253 mass spectrometer at the University of Michigan’s Stable Isotope Laboratory.
17
O-corrected data are adjusted for acid fractionation and source mixing by calibration to a bestfit regression line defined by National Bureau
of Standards (NBS)-18 and NBS-19 standards.
Data are reported in ‰-variation relative to the
Vienna PeeDee Belemnite (VPDB) standard.
Measured precision is better than 0.1‰ for
both δ13C and δ18O. The bulk-sample data set
represents 907 analyses from 187 shells. An
additional 288 samples were collected during
high-resolution analysis of two specimens of
Cucullaea and two of Eurhomalea.
Paleotemperature Determinations
Paleotemperatures are calculated using the
empirically determined paleotemperature equation of Grossman and Ku (1986) for biogenic
aragonite, adjusted to correct for the difference
between “average marine water” and standard
mean ocean water (SMOW) (e.g., Kobashi et
al., 2003):
Samples for stable isotopic analysis were
roasted in a vacuum for one hour at 200 °C to
Geological Society of America Bulletin, May/June 2008
T(°C) = 20.6 – 4.34 [δ18OARAGONITE
– (δ18OSeawater −0.2)]
663
Ivany et al.
Average global δ18OSW values are generally
given as −1‰ for an ice-free Eocene world
(Shackleton and Kennett, 1975; Zachos et al.,
1994), and estimates of δ18OSW from paired
δ18OCALCITE and Mg/Ca ratios of benthic foraminiferans also vary around −1‰ (Lear et al., 2004).
The ocean is heterogeneous, however, and here
we use a value of −2‰ for δ18OSW based on estimates from Huber et al. (2003) specific to the
Eocene Antarctic Peninsula region derived from
a coupled ocean-atmosphere climate model.
Values may have been as low as −3‰ (Huber
et al., 2003), and if we use this value, calculated
temperatures are cooler by ~4.5 °C. If more
positive δ18OSW values are employed to account
for proposed accumulations of Eocene ice (e.g.,
Lear et al., 2000; Tripati et al., 2005), estimated
temperatures are somewhat warmer. If local
shelf water δ18O values were further depleted
from mixing with fresh water, calculated temperatures will be too warm. It is unlikely that
evaporative enrichment is a significant concern
in this high-latitude setting.
RESULTS
Inter-Taxon Comparisons
Stable isotope ratios of the four different taxa
(two bivalves, one gastropod, and one brachiopod) from the same collections at two different
localities were compared to evaluate potential
5
differences among taxa (Fig. 5). The naticid
Polinices showed the greatest spread in values,
with samples from the same location exhibiting
at least 2‰ of variation in both carbon and oxygen isotope values. Their semi-infaunal habit
may introduce more variation in δ13C due to
differences in δ13C of dissolved inorganic carbon (DIC) between the water column and pore
fluids, which might include more negative carbon derived from the decomposition of organic
material. In addition, and perhaps more likely,
their predatory trophic strategy and resulting
difference in metabolic rate from suspensionfeeding taxa (Huebner and Edwards, 1981) may
result in the incorporation of variable amounts
of isotopically negative, respired CO2 into shell
carbonate. Difference in δ13C of tissue for prey
taxa of different trophic levels introduces yet
another variable. The δ13C values of Polinices
overlap the negative end of the range in variation exemplified by the other taxa, suggesting
that one or more of these variables may play a
role in its composition. It is unclear, however,
why δ18O should show such a range, since the
reservoir and its incorporation into shell material is not affected by biological processes. One
possibility is growth rate—if Polinices were
growing so fast that individual bulk samples
did not in fact represent an average of multiple
years, and/or if individuals stopped growing
for variably long intervals of each year, and/
or if growth rates were at times fast enough
Polinices
Bouchardia
Cucullaea
Eurhomalea
4
to introduce kinetic disequilibrium effects, the
δ18O values of our samples would show similar
high variability. Rapid, seasonally interrupted,
and prey-dependent growth is characteristic
of modern Polinices from the coast of Massachusetts, USA (Edwards and Huebner, 1977),
and may be true of Eocene taxa as well. This
hypothesis requires additional high-resolution
sampling to evaluate and, even if substantiated, it would rule out Polinices as a reliable
and consistent recorder of annual average temperatures. Although they present an interesting
paleoecological puzzle, for this reason they
were avoided in construction of the long-term
climate record despite their ubiquity in collections throughout the unit.
The brachiopod Bouchardia shows a tighter
clustering of δ18O data, but examination by
cathodoluminescence revealed the common
presence of secondary luminescent cement filling the punctae of shells. Hinge regions appear
to be relatively free of such cement, suggesting
that sampling difficulties could be overcome
if focused on these regions, but even a small
potential for contamination by isotopically
depleted diagenetic cement makes them a less
than ideal candidate taxon for long-term temperature reconstruction. In addition, the inconsistent relationship between data from brachiopods and the two bivalve taxa is difficult
to explain, and so they were not investigated
further for this particular study.
1
3
δ13C
δ13C
0
2
-1
1
-2
0
B
A
-2
-1
0
1
2
-3
-2
-1
δ O
18
0
1
2
δ O
18
Figure 5. Differences in isotopic composition of bulk samples from different taxa at two localities within Telm 5. Polinices is a
naticid gastropod, Bouchardia is an articulate brachiopod, Cucullaea is an arcid bivalve, and Eurhomalea is a venerid bivalve.
Bouchardia has a calcite shell, the others are aragonite. Telm—Tertiary Eocene La Meseta.
664
Geological Society of America Bulletin, May/June 2008
Eocene climate change on the Antarctic shelf
The bivalves Cucullea and Eurhomalea show
roughly the same spread in δ13C values, and
δ18O values are offset but to a similar degree
in the two test localities. The straightforward
nature of sampling with respect to the growth
trajectory (a single cross section captures the
whole life history, as opposed to naticids’ spiral growth), the lack of complicating diagenetic
cement (as with the brachiopods), and the consistency of isotope results made these two genera the best taxa of choice for reconstruction of
the long-term paleoclimate record. Cucullaea
has a thick, easy to sample shell with clear
growth banding, and has been used already
for paleoclimate reconstruction (Dutton et al.,
2002); therefore, these results will be directly
comparable to this earlier work. Eurhomalea is
a venerid bivalve, and venerids have been used
often and with consistent success for stable isotope reconstruction of environmental conditions
(e.g., Arthur et al., 1983; Jones and Quitmyer,
1996; Surge and Walker, 2006). The combination of the two taxa should therefore produce
robust and reproducible results.
An examination of bulk sample data from
additional localities confirms that mean values
from Cucullaea are ~0.4‰ more positive than
those from Eurhomalea (Fig. 6), and hence will
yield paleotemperatures cooler by ~2 °C. The
origin of this taxon-specific offset was investigated through high-resolution microsampling
within and across growth bands of two speci-
mens of each bivalve from a single locality in
Telm 5 (Fig. 7). This approach reveals that Cucullaea grows mainly during the Austral winter
(Buick and Ivany, 2004), while Eurhomalea
appears to capture more of the Austral summer.
The average (bulk) composition of Eurhomalea
shell carbonate is therefore more isotopically
negative, yielding warmer average annual temperatures. Dark (translucent) growth bands, indicating slower growth due to physiological stress
(Pannella and MacClintock, 1968), in Cucullaea
correspond to more negative δ18O values and
warm temperatures (Figs. 7A and 7B; Buick and
Ivany, 2004), while in Eurhomalea they fall at
δ18O maxima (cool temperatures, Fig. 7D). This
is consistent with the truncation of the annual
temperature cycle in the summer for Cucullaea
but in the winter for Eurhomalea, giving rise to
their different bulk isotopic compositions.
The thin outermost layer of the shell in both
taxa was often abraded or not preserved, and here
the potential was greater for encountering diagenetic cements in surface cracks or altered shell
in contact with matrix material. As a result, all
samples for this study come from internal cross
sections rather than the exposed shell surface.
While most studies of shell geochemistry in
Recent specimens have used material from the
outer shell layer, Surge and Walker (2006) show
that δ18O values from the middle shell layer are
indistinguishable from those in the outer and
hence provide a robust paleoclimate archive for
ancient settings. In this study, bivalve carbonate was sampled from both the middle and inner
shell layers. Compositions of microsampled carbonate from the middle and inner shell layers of
the same specimen of Cucullaea (Figs. 7B and
7C) exhibit several differences. First, because
the growth bands are much closer together in the
inner shell layer, it is impossible to achieve here
the temporal resolution attainable in the middle
layer. Each individual sample is therefore more
time-averaged, and hence the amplitude of the
seasonal cycle is much reduced in the record from
the inner shell. Secondly, the mean δ18O value
for each region is identical, but δ13C is markedly
more negative in the inner shell layer (Fig. 7C).
This offset is likely due to the preferential incorporation of isotopically negative metabolic CO2
into the inner shell layer, where the mantle is in
constant contact with the shell and where the
seawater reservoir is more distant. While this
poses no problem for the δ18O-derived paleoclimate record, it may introduce an added source of
variation for the long-term δ13C record.
Intra-Horizon Comparisons
While some variation exists within and
between shells from a given locality that reflects
a combination of seasonal and interannual temperature variation during each time-averaged
interval of deposition, the range of values
exhibited by shells at different localities from
Cucullaea
Eurhomalea
2
*
1
δ18O
*
0
*
*
*
*
*
-1
-2
00
-1
* *
00
Figure 6. Comparison of δ18O values from co-occurring Cucullaea
and Eurhomalea bivalves. With
very few exceptions, Cucullaea is
more positive than Eurhomalea.
On average, the difference is
0.42 ± 0.1‰. Localities are not
arranged in stratigraphic order.
-11 0 -17 1- 04 1-87 1-12 1-29 1-30 1- 45 1- 46 1-53 1-54 - 004 - 005 - 600
0
0
0
0
0
0
0
0
0
0
03 03 03
Locality
Geological Society of America Bulletin, May/June 2008
665
Ivany et al.
Isotope Value (VPDB)
3.0
A
Cucullaea 3
2.5
2.0
1.5
1.0
0.5
0.0
-0.5
0
2
4
6
8
10
12
14
Distance (mm)
B
2.5
18
δ O
Cucullaea 1
2.0
Isotope Value (VPDB)
Isotope Value (VPDB)
13
δ C
3.0
1.5
1.0
0.5
0.0
-0.5
0
1
2
3
4
5
6
7
C
Cucullaea 1
inner shell layer
1.5
1.0
0.5
0.0
-0.5
1
8
5
3.0
3.0
D
Eurhomalea 4
2.0
1.5
1.0
0.5
0.0
40
25
50
60
Eurhomalea 5
0.5
0.0
-1.0
30
21
1.0
-1.0
20
17
1.5
-0.5
10
13
2.0
-0.5
0
E
2.5
Isotope Value (VPDB)
2.5
9
Sample Number
Distance (mm)
Isotope Value (VPDB)
16
0
5
Sample Number
10
15
20
25
30
35
Sample Number
Figure 7. Microsampled isotope data from two specimens of Cucullaea (A–C) and two of Eurhomalea (D and E) from the same locality in
Telm 5. Dashed horizontal lines at 1.5‰, 0.5‰, and −0.5‰ are merely to aid in visual comparison; note that the vertical axes for Eurhomalea
(D and E) extend to more negative values than those for Cucullaea. Vertical gray bars in A, B, and D denote dark growth bands. Isotope
values from both Cucullaea specimens are very similar (A and B), as are values from both Eurhomalea specimens (D and E). However, δ18O
values of Eurhomalea extend to more negative (warmer) values than Cucullaea and do not record the more positive (cooler) values. Growth
slow-downs or cessations marked by growth bands occur in the summer for Cucullaea but the winter in Eurhomalea. Inner shell-layer carbonate from Cucullaea 1 (C) shows a smaller range of variation than the middle shell layer (B) due to closer spacing of growth increments
and lower temporal resolution of samples. Mean value of δ18O is similar between transects, but δ13C is markedly more negative in the inner
shell layer, suggesting vital effects associated with incorporation of respired CO2. Telm—Tertiary Eocene La Meseta.
666
Geological Society of America Bulletin, May/June 2008
Eocene climate change on the Antarctic shelf
the same stratigraphic horizon is nearly always
statistically indistinguishable (Table 1). This is
true even for shells collected from transgressive
lag deposits that may represent significant intervals of time. Telm 4, equivalent to the base of
the Cucullaea I Allomember of Marenssi et al.
(1998), is one such horizon (“4.5” in Table 1).
Another is a laterally extensive horizon marked
by the abundance of small to moderate-sized
Cucullaea bivalves found near the top of Telm
5 (“5.9” in Table 1). Unlike most other horizons, these two are easily recognized, and can
be traced throughout much of the outcrop area.
Their sedimentology makes them suboptimal for
reconstructing climate information due to their
potentially time-averaged nature, and therefore
they were not heavily sampled for geochemical
analysis. Nevertheless, the fact that shells from
different localities yield similar overall ranges
of values suggests that the amount of time represented by accumulated shell is not significant
and/or does not encompass much secular variation in temperature. These results suggest that
conditions across the depositional surface at a
given time were reasonably consistent from
place to place, and therefore that any temporal
change in the average δ18O-derived paleotemperature is likely to represent a meaningful shift
in Eocene peninsular climate rather than simply
the incomplete sampling of a heterogeneous
environment at each stratigraphic level.
Stratigraphic and Strontium Isotopic
Constraints on Age
Before presenting stable isotope data in a
stratigraphic context, we first explore the potential of strontium isotope ratios to provide additional age control within the formation. Secular
variation in the 87Sr/86Sr ratio of marine waters
is reflected in the chemistry of carbonate precipitates through time. In portions of the record
where that ratio is well constrained and changing rapidly, the 87Sr/86Sr ratio of carbonate precipitated in fully marine seawater is an extremely
useful tool for age control (e.g., Howarth and
McArthur, 1997; McArthur et al., 2001). The
Eocene Epoch, however, presents difficulties in
that the marine ratio undulates around a value of
~0.70775; thus there is not a unique age solution
for any single measured ratio. In addition, data
constraining the marine curve for the early and
early middle Eocene are very sparse (McArthur
et al., 2001; but see new data from Hodell et al.,
2007), and therefore assigned ages in this part of
the section must be considered hypotheses to be
tested with additional approaches.
While unequivocally resolving Eocene ages
using strontium isotope stratigraphy is a daunting task under the best of circumstances, the
TABLE 1. ANALYSIS OF VARIANCE RESULTS FOR COMPARISONS OF MEAN
δ18O VALUES BETWEEN CORRELATIVE LOCALITIES WITHIN A STRATIGRAPHIC
HORIZON
Stratigraphic
Taxon
Number of
Number of
F-statistic
p-value
horizon
localities
analyses
7.7
Eurhomalea
2
40
1.06
0.310
7.5
Cucullaea
2
39
0.79
0.380
7.5
Eurhomalea
2
45
7.03
0.011
6.3
Eurhomalea
3
29
2.66
0.089
6.1
Eurhomalea
5
56
1.03
0.399
5.9
Cucullaea
2
45
0.97
0.331
5.7
Cucullaea
3
36
12.83
0.000
5.7
Eurhomalea
3
46
3.00
0.060
5.3
Cucullaea
2
25
0.06
0.814
4.5
Cucullaea
3
14
0.78
0.481
3.5
Cucullaea
2
21
0.05
0.824
Note: Horizons are coded as Telm (“Tertiary Eocene La Meseta”) number with
relative position indicated by the decimal. Only those horizons for which there were
multiple localities where the same taxon was analyzed are included. Tests indicating
significant difference between localities within horizons are marked by an asterisk (*).
TABLE 2. 87Sr/86Sr RATIOS AND ANALYTICAL ERROR
FROM LA MESETA FORMATION BIVALVES
87
Telm
Stratigraphic
Locality
Sample
Sr/86Sr
Error
Best inferred age
horizon
for horizon (m.y.)
7
7.7
01-82
C1
0.707780
0.000009
34.0
7.6
01-55
E2
0.707726
0.000014
37.3
7.4
01-99
C1
0.707741
0.000011
37.5
7.2
01-50
E2
0.707731
0.000011
39.1
6.4
01-10
E1
0.707748
0.000011
6
6.4
01-10
E3
0.707738
0.000020
41.0
6.2
01-63
E2
0.707799
0.000010
6.2
01-63
E3
0.707852
0.000010
6.1
01-68
E4
0.708158
0.000010
6.1
01-8
E2
0.707840
0.000013
5.9
01-73
C5
0.708243
0.000010
48.8
5
5.9
01-70
C1
0.707704
0.000016
5.8
01-35
E1
0.707732
0.000010
49.0
5.8
01-35
C3
0.707715
0.000010
5.7
01-29
C1
0.707781
0.000013
5.7
01-29
E4
0.708187
0.000013
5.7
01-30
C1
0.707874
0.000010
5.5
01-46
C1
0.707712
0.000010
49.4
5.5
01-46
E4
0.707733
0.000013
5.4
01-45
C2
0.707710
0.000012
50.8
5.4
01-45
E1
0.707708
0.000011
5.3
01-43b
C1
0.707827
0.000011
51.0
4
4.8
01-42
C20
0.707769
0.000011
4.5
01-27
C20
0.707723
0.000011
52.5
3.8
00-17
C2
0.707771
0.000011
52.8
3
3.8
00-17
E1
0.707740
0.000012
3.5
01-12a
E3
0.707717
0.000010
53.0
3.5
01-12
C1
0.707848
0.000012
3.5
01-12
E3
0.707762
0.000012
3.2
01-22
E1
0.707712
0.000010
53.2
3.1
01-23
E3
0.707698
0.000009
53.5
3.1
01-23
E1
0.707736
0.000011
3.1
01-23
E4
0.707791
0.000013
2
2.5
01-04
C3
0.707744
0.000011
54.0
2.5
01-04
C2
0.707738
0.000012
Note: Samples are arranged in stratigraphic order by locality. Telm—“Tertiary Eocene La
Meseta”; C—Cucullaea; E—Eurhomalea. Best inferred age is determined by both 87Sr/86Sr and
relative stratigraphic position. Error is ±2σ. See text for discussion.
nearshore facies that characterize the La Meseta
Formation pose additional concerns. Mixing
of marine waters with freshwater carrying an
87
Sr/86Sr signature that differs from the marine
value may be difficult to recognize given the
undulatory nature of the curve during the
Eocene. Without estimates of the 87Sr/86Sr ratio
of Peninsula basement rocks and the concentration of Sr in Eocene runoff, it is impossible to
determine how salinity variation would affect
the 87Sr/86Sr ratio of nearshore carbonates (e.g.,
Bryant et al., 1995). Continental runoff is generally more radiogenic while Eocene marine
87
Sr/86Sr values are particularly nonradiogenic;
therefore, the influence of freshwater influx
could be significant. Nevertheless, while there
is some variation, most La Meseta Formation
samples produce isotope ratios that fall within
the range for typical Eocene seawater (Table 2,
Fig. 8A), suggesting that any salinity variation
Geological Society of America Bulletin, May/June 2008
667
Ivany et al.
Telm 7
Telm 6
Telms 4, 3, 2
Telm 5
0.70785
0.70780
87
Sr
86
Sr
(unconformity)
0.70775
Eurhomalea
Cucullaea
0.70770
Ivany et al. 2006
Dutton et al. 2002
A
35
40
45
Age (my)
50
55
7
0.70785
Sr
4
86
y = -0.17x + 13.95
r 2 = 0.78
5
y = -1.24x + 69.39
r 2 = 0.94
3
2
Sr
6
87
Relative Strat. Position
0.70790
y = -0.23x + 17.09
r 2 = 0.85
B
0.70780
0.70775
0.70770
C
0.70765
35
40
45
50
Age (my)
is relatively minor. In addition, freshwater influx
would be expected to affect not only the 87Sr/86Sr
ratio, but also the stable isotope values of carbonate. There is, however, no consistent relationship between the 87Sr/86Sr ratio and the mean
δ18O value of shell carbonate (Fig. 8C). Furthermore, the regular and low-amplitude variation in
δ18O exhibited in microsampled bivalves is more
consistent with seasonal temperature variation
than with fluctuating salinity. Lastly, Dutton et
al. (2002) demonstrate that multiple 87Sr/86Sr
measurements on the same Cucullaea shell are
statistically indistinguishable, suggesting little
if any variation in salinity occurred over timescales of an individual animal’s lifetime. As this
taxon can live for over a century (Buick and
Ivany, 2004), the within-shell consistency of Srisotope values suggests sediment accumulation
668
55
-1.5
r 2 = 0.0042
-0.5
0.5
18
δ OVPDB
in a marine environment of consistently normal
salinity. These geochemical data therefore support interpretations based on faunal composition
(Wiedman and Feldmann, 1988; McKinney et
al., 1988; Blake and Zinsmeister, 1988; Stilwell
and Zinsmeister, 1992; Filkhorn, 1994; Baumiller and Gaździcki, 1996; Blake and Aronson,
1998) and published geochemical data (Dutton
et al., 2002) that, while shallow and nearshore,
the La Meseta Formation accumulated under
basically normal marine salinities.
87
Sr/86Sr ratios in the lower part of the section tend to be more ambiguous than those
above. Telm 2 yielded values from four different
specimens (two from this study and two from
Dutton et al., 2002) that are virtually identical and pin the age of this part of the section
at ~54 m.y. 87Sr/86Sr ratios from Telms 3–5,
1.5
Figure 8. (A) 87Sr/86Sr ratios
from the La Meseta Formation
relative to the global marine
seawater curve (gray band
encompasses 95% confidence
limits around curve, LOWESS
Version 4: 08/03; McArthur et
al., 2001) and consistent with
their relative stratigraphic
positions. Error bars on the
y-axis reflect analytical error.
Associated error on age estimates during the Eocene generally approximates ±2 m.y.;
those error bars are not plotted here for clarity, and due to
additional controls on age by
stratigraphic position that are
not considered in the LOWESS
(locally weighted regression
scatterplot smoother) fit. (B)
Age model for the La Meseta
Formation based on the best
estimates of age derived from
A (Table 2). The section was
divided into three intervals and
best-fit lines were determined
for each. Y axis values reflect
relative stratigraphic positions
(Telm numbers), so differences
in slope do not directly relate
to change in sedimentation
rate. (C) Relationship between
87
Sr/86Sr ratio and mean δ18Ο
value for sampled shells. Telm—
Tertiary Eocene La Meseta.
however, are much more variable. While the
majority of values can be placed such that their
analytical error bars intersect the marine curve,
some horizons (e.g., the base of Telm 3) show a
large range of variation in presumably correlative samples. Several samples are more radiogenic than expected, suggesting the potential
for limited freshwater mixing. Acknowledging
this uncertainty, 87Sr/86Sr ratios from Telms 2
through 5 are nearly all low enough to ensure
that they represent the early Eocene minimum
in marine 87Sr/86Sr values, rather than the latemiddle Eocene saddle in the marine 87Sr/86Sr
curve. This suggests that deposition of the lower
part of the formation took place in the early
Eocene, an interpretation consistent with estimates derived from dinoflagellate (Wrenn and
Hart, 1988; Cocozza and Clarke, 1992) and
Geological Society of America Bulletin, May/June 2008
Eocene climate change on the Antarctic shelf
diatom (Harwood, 1985, 1988) biostratigraphy. Telm 5, however, is apparently older than
expected based on molluscan and microfossil
biostratigraphy, only just crossing into the base
of the middle Eocene. While Sadler (1988) and
Stilwell and Zinsmeister (1992) had speculated
that the basal unconformity of the La Meseta
Formation may have been generated by a major
sea-level lowstand at 49.5 Ma (Vail et al., 1980;
Haq et al., 1987), these strontium isotope data
confirm that deposition of La Meseta Formation
sediments began several millions of years earlier. Miller et al. (2005) report an early Eocene
eustatic rise following a relative Paleocene stillstand that may correspond to the initiation of La
Meseta Formation marine deposition, but the
potential competing influence of local tectonism
makes this correlation tentative at best.
A potential explanation for 87Sr/86Sr variability in the lower part of the section is the incorporation of small amounts of siliciclastic material
as contaminants in samples of shell carbonate.
Mollusks can entrain minute amounts of sediment in their shells as they grow, and siliciclastic sediments tend to be quite radiogenic. If a
small amount of this material is leached during
sample preparation, its influence on the 87Sr/86Sr
of the sample could be significant. Mud is more
likely to be incorporated than sand, and because
the lower part of the section is muddier than
above, this may explain the more variable ratios
there. It does not appear that differing degrees
of diagenesis within the section can explain the
higher strontium variability in Telms 3–5. Preservation of shell material is equally good here
as in the upper part of the formation, and shells
tend to be larger and thicker, facilitating greater
ease of microstructural evaluation and selective
sampling.
There has been some discussion, both informally and in the literature, about the stratigraphic and age relationships between Telms
2 and 3 (e.g., Sadler, 1988; Stilwell and Zinsmeister, 1992; Marenssi et al., 1998; Porębski,
1995, 2000). The muds and coarser grained
lenses of Telm 2 may represent a deeper-water
facies that everywhere predates and underlies
Telm 3, and the lack of clear shallow-water
sedimentary structures in Telm 2 supports this
interpretation. Alternatively, the Telm 2 muds
may instead be a lower-energy facies that accumulated along the margins of the trough that
constrains the La Meseta Formation, while
Telm 3 sands are at least in part laterally correlative to the Telm 2 muds and represent the
higher-energy facies in the axis of the trough.
The base of Telm 3 is not exposed in the axis,
thus it is unknown whether there are Telm 2
muds below, but field relations suggest that the
two units are at least in part correlative (Stilwell
and Zinsmeister, 1992; Porębski, 1995). While
the strontium isotope ratios from Telm 2 are
consistent and unambiguous, the variation
in ratios from Telm 3 makes age estimates of
the two units overlap and, therefore, they are
consistent with either scenario. Stable isotope
values from Telm 2 material, however, are distinct from those in the lower part of Telm 3 (see
next section), suggesting that the two units are
not time equivalent, at least in portions of the
island sampled for this study.
Telm 4 clearly represents a ravinement surface
and associated transgressive lag characterized
by abundant phosphate pebbles and glauconite,
and is conformably overlain by sediments of the
ensuing Telm 5 transgression. The fact that fossils from Telm 4 include a significant number
of reworked shells could explain the similarity
of their ratios to values from Telm 3, although
an effort was made to sample in situ (articulated) specimens. Because of the potential for
reworking, this unit was not a focus of attention
for geochemistry, and therefore data are limited,
but immediately overlying shells from Telm 5
produce ratios that suggest an age for the base
of the unit of ca. 51 Ma. Porębski (2000) and
Marenssi (2006) propose that the Telm 4 disconformity may correspond to the previously
mentioned 49.5 Ma sea-level lowstand. This
is a plausible scenario, given the uncertainty
associated with 87Sr/86Sr values in this setting.
However, shells from multiple localities at the
base of Telm 5 consistently yield values that
correspond to the ca. 51 Ma minimum in marine
strontium isotope ratios, and the Gradstein et al.
(2004) timescale to which the marine strontium
isotope curve is calibrated now places the (formerly 49.5 Ma) lowstand closer to 48.5 Ma. If
this is correct, then there is no obvious manifestation of that eustatic lowstand in the La Meseta
section. Pekar et al. (2005) note the presence of
a sequence-bounding unconformity at 50.9 Ma
on the Tasman Rise that likely correlates with
a hiatus in the New Jersey, USA, shelf record
(Miller et al., 1998), suggesting eustasy as a
driver. This may correlate with the Telm 4 hiatus
as constrained here, and if so, implies that this
aspect of the Seymour Island section reflects
eustatic sea-level change rather than regional
patterns of tectonism and base-level change (see
also Porębski, 1995). While the sedimentology
of Telm 4 suggests a hiatus of some duration,
strontium isotope ratios suggest no more than
~2 million years of section could be missing.
Age control based on strontium isotopes
is much better in the upper part of the formation. Specimens from Telm 7 and most of
Telm 6 yield consistent and unambiguous
ratios that establish an age ranging from late
middle Eocene (ca. 41 Ma) up through virtu-
ally the Eocene-Oligocene boundary (Fig. 8A;
Ivany et al., 2006). That values fall consistently
on the marine curve corroborates a normal
marine inference based on sedimentology and
stratigraphic relations (Porębski, 1995, 2000).
Samples from the lowermost part of Telm 6 are
radiogenic, suggesting the potential for freshwater mixing, but sediments appear to be in stratigraphic and temporal continuity with the overlying sediments, and we observed no lithologic
differences sufficient to suggest any significant
difference in depositional setting. This part
of the section will be discussed in more detail
below. A late-middle to late Eocene age for the
upper part of the La Meseta Formation is consistent with existing biostratigraphy.
Stratigraphic constraints on samples and comparison of their 87Sr/86Sr ratios with the global
marine seawater curve suggest the presence
of a middle Eocene (Lutetian) unconformity
between Telms 5 and 6, spanning from roughly
48 to 43 Ma. Sedimentologic evidence for such
a hiatus, however, is not immediately apparent.
Scoured surfaces and concentrated shelly horizons are present throughout most of the section (Sadler, 1988; Porębski, 1995, 2000) but,
aside from the base of Telm 4, none are associated with features that clearly record significant
missing time. Shells in these layers are generally
in good condition and often articulated, indicating much less time averaging than in Telm 4. A
more laterally persistent shell bed with a scoured
base is present near the top of Telm 5 (likely the
base of Marenssi’s Cucullaea II Allomember;
Dingle et al., 1998), suggesting a break in sedimentation, but lithofacies and biofacies as well
as geochemistry clearly place it within Telm
5 rather than the overlying Telm 6 sands. The
87
Sr/86Sr ratios from shells in this bed (horizon
“5.9” in Table 2) suggest that in fact the base
of this layer may represent the aforementioned
sea-level lowstand, now positioned at roughly
48.5 Ma. The short-lived nature of this particular lowstand (Haq et al., 1987) may explain the
lack of a significant temporal hiatus evident in
87
Sr/86Sr ratios from below and above.
The facies change from Telm 5 to Telm 6,
marked by an influx of rust-weathering, sandbearing shell beds, low-angle trough crossbedding, and Ophiomorpha burrow networks,
appears to correspond to the interval of missing time suggested by our strontium isotopic
data. This is the most significant lithologic
and facies shift in the section. Porębski (1995,
2000) notes the presence of an angular unconformity associated with the erosive base of a
channel between Telm 6 strata and the underlying Telm 5, consistent with an interval of time
not represented by sediment, but this is the first
explicit suggestion of an unconformity of some
Geological Society of America Bulletin, May/June 2008
669
Ivany et al.
duration at this level. Miller et al. (2005) infer a
substantial eustatic sea-level fall ca. 42–43 Ma
based on backstripping data from the New
Jersey coastal plain, and this may be responsible for the unconformity noted here. Field
observations, however, suggest that Telm 5
and 6 lithologies may be interbedded in some
areas, implying correlative ages not supported
by 87Sr/86Sr data or the hypothesized unconformity. Porębski (1995) too notes that the unconformable contact between Telms 5 and 6 may
in fact be conformable in the south-southwest
part of the outcrop belt, and correlate to a horizon within what has been mapped as Telm 5.
The transition, however, is poorly exposed due
to slumping and hence it is difficult to rule out
any single interpretation at this time. It is possible that the apparent gap in time recorded by
strontium isotope values is at least in part a
function of incomplete sampling, because shell
beds are not consistently present throughout
the unit and exposure was intermittent. Telm
5 sediments younger than those we recovered
may be preserved in a different part of the section where they were not as severely truncated
by the base of Telm 6.
While Telm 7 is a bit finer grained and the
bedding more even and laterally continuous
than in Telm 6, both units appear to be shoreface
deposits. The somewhat more variable and
nearshore nature of Telm 6 may be associated
with deposition in a bay-mouth bar (Porębski,
1995), while Telm 7 deposits accumulated just
seaward. 87Sr/86Sr ratios give no indication of
a significant time gap at the base of Telm 7, as
suggested by Porębski (2000) from patterns of
onlap and facies change.
In the following section, isotope data are
plotted stratigraphically using a combination of
best age estimates from Table 2 and linear interpolation for intervening horizons from the age
model in Figure 8B.
Long-Term Stable Isotope Record and
Eocene Paleotemperatures
Within the temporal context provided by
strontium isotope ratios and stratigraphic relationships, values of δ18O of shell aragonite indicate warm early Eocene temperatures with a
maximum of ~15 °C near the base of Telm 3
(Fig. 9A). Telm 2 samples below (ca. 54 Ma)
suggest slightly cooler temperatures, particularly if data from Eurhomalea are considered,
although comparatively few samples constrain
this inference (four shells total, only one of
which is from Eurhomalea). Estimated temperatures drop within Telm 3 to ~10–11 °C, and
remain essentially stable at this value through
the end of Telm 5 (earliest middle Eocene).
670
Immediately above the middle Eocene
unconformity, δ18O values decrease to −1.0‰
in the lower part of Telm 6, recording a shortlived return to temperatures more typical of the
base of Telm 3 (~15 °C). Immediately above
this (ca. 41 Ma), δ18O increases rapidly, reflecting a drop in temperature to ~7 °C by the end
of Telm 6. Temperatures continue to cool into
Telm 7 and reach an Eocene minimum value of
~5 °C by ca. 37 Ma. Upper Telm 7 (late Eocene)
temperatures exhibit a modest warming back to
~7–8 °C. For comparison, water temperature
around the Antarctic Peninsula today is ~−2 °C.
With the possible exception of Telm 2, the
overall trend in stable oxygen isotope values
is consistent between the two bivalve genera
examined (Fig. 9A), providing assurance that
the observed secular pattern is robust. Interestingly, the characteristic offset between Cucullaea and Eurhomalea is less apparent in the
upper part of the section. This could indicate that
the taxonomic difference in preferred season of
growth decreases upsection, or that the seasonal
range of temperature variation had decreased to
a point where summer and winter temperatures
were not as distinct as in Telm 5, and therefore
any seasonal bias in growth had only a minimal
effect on the average δ18O value of shells. These
hypotheses need to be evaluated with high-resolution microsampling of shells from Telm 7.
While it may be tempting to interpret the
range of isotope values from shells within any
stratigraphic horizon as an indication of the
seasonal range of temperature variation at that
time, it should be remembered that this range is
also influenced by the degree of time averaging
within the horizon, the amount of interannual
variation in temperature, and the range of temperatures over which the taxon produces shell
material. Because all of these variables might
change through the section, with or without any
accompanying change in mean temperature or
seasonal range in temperature, caution is advised
in making inferences about changes in seasonality through time. Seasonality can only be confidently addressed with records from multiple
microsampled shells distributed throughout the
section, reinforced with complementary independent records based on paleobotanical (e.g.,
Francis and Poole, 2002; Gandolfo et al., 1998a,
1998b, in Case, 2007) or clay mineralogical
(e.g., Dingle et al., 1998) data. Initial results of
microsampling on shells in Telms 3, 5, and 7
suggest a significant decrease in seasonality in
the coldest parts of Telm 7, associated especially
with a decrease in summer temperatures (Ivany,
2007; Miklus, 2008).
Stable carbon isotope values fall between
−2.0‰ and +2.0‰ from Telm 2 up through
the lower part of Telm 6 (Fig. 9B). A short-
lived excursion to values near −6‰ takes place
within Telm 6 above the negative values of δ18O
and coincident with the shift back to more positive δ18O and cooler temperatures (ca. 41 Ma).
Above this, δ13C values from Eurhomalea
remain on the low side (~−3 to −4‰), while
those from Cucullaea return to near zero by
the top of the section. Variation in δ13C within
a horizon is considerable, yet the seasonal range
of δ13C in microsampled shells from Telm 5
is generally less than 2‰ (Fig. 7). The large
spread of values within horizons could reflect a
combination of real variation in the composition
of DIC at the scale of time averaging and variation introduced by sampling from both the inner
and middle shell layers of bivalves. Despite the
variability within the section, the trend toward
more negative values in Telms 6–7 is highly
robust (p <.0001).
PALEOENVIRONMENTAL
IMPLICATIONS
Data on stable isotopic values and strontium
isotope ratios provide constraints on the depositional environment of the La Meseta Formation. Most importantly, the dominantly marine
values of shell carbonate offer little evidence in
support of a traditional deltaic or river-mouth
estuary setting. The consistent range of δ18O
variation and complete lack of any pronounced
negative oxygen isotope values argues strongly
against such an interpretation. In addition, no
covariation exists between δ18O and δ13C values (see also Dutton et al., 2002), as would be
expected with seawater-freshwater mixing (e.g.,
Ingram et al., 1996). Even at the base of Telm 6,
where the potential for freshwater influx seems
highest based on the negative δ18O values and
radiogenic strontium ratios, mean δ18O values
(constrained by 114 analyses from 29 clams at
10 different localities) are no more variable than
elsewhere in the section. This is unlike what
would be expected in a brackish-water setting,
particularly at high latitude where precipitation
can be expected to be more distilled and hence
isotopically negative. Instead, these data in combination with the coast-normal, fault-bounded
nature of the basin suggest accumulation in a
relatively small, tectonically generated graben
with minimal freshwater flow reflecting a limited catchment area. The general compositional
maturity of sediments (Marenssi et al., 2002;
field observations) suggests to us that sediments
were likely delivered not by an associated river
but instead primarily by longshore transport
along the coast combined with mass wasting and
localized runoff from valley margins. Sediments
accumulated here due to subsidence associated
with downdropping of the graben rather than
Geological Society of America Bulletin, May/June 2008
Eocene climate change on the Antarctic shelf
Ice-Free Temperature (°C)
5
10
15
Figure 9. Stable oxygen (A) and carbon (B) isotope values for all individual analyses from every bivalve shell from all horizons, plotted by
taxon. Intensity of color reflects stacking of multiple data points. Age is plotted according to strontium isotope stratigraphy presented in
Figure 8 and Table 2.
proximity to the source of sediment delivery. The
predominantly landward paleoflow directions in
at least parts of the formation combined with the
“conspicuous absence” of the expected fluvial
deposits associated with lowstands (Porębski,
1995) and the strong tidal influence throughout
the unit support this interpretation.
The great challenge of working in shelf settings is determining which of the observed
changes seen in the section can be attributed to
local facies shifts and which reflect more significant trends driven by climatic and/or oceanographic change (even assuming that changes
due to variable diagenetic alteration can be ruled
out). Both can be manifest as changes in lithofacies, isotope values, and faunal composition, but
the first might simply represent a shift in depositional setting along the shelf while the second
indicates a wholesale change in conditions that
affects all (or most) shelf environments simultaneously. Without widely distributed, correlative exposures both along strike and updip and
downdip where appropriate, it can be difficult to
distinguish the effects of these two possibilities.
This is a consideration for the Seymour Island
section, since the area of outcrop is limited and
there is limited potential for finding correlative sections exposed elsewhere in the region
(but see Askin et al., 1991). Nonetheless, we
have sampled the same horizons along strike
wherever possible so as to better understand
the variation within the environment at the time
of deposition. By comparing geochemical data
with lithologic and faunal properties, we can
begin to resolve differences between faciesimposed and climate-imposed trends.
The shift in δ18O values from more negative
toward more positive in Telm 3 likely reflects
real climate change, as there is effectively no
change in lithofacies associated with it. Both
taxa exhibit a consistent change of a similar
magnitude through time, the trend is maintained
Geological Society of America Bulletin, May/June 2008
671
Ivany et al.
through four different sampling horizons, and
there is no difference in variability among horizons. This is more consistent with a trend from
warm to cooler temperatures. Likewise, the secular variation in δ18O seen in the upper part of
Telm 6 and all of Telm 7 is likely a real climate
signal, because these samples all yield strontium isotope values that consistently fall on the
marine curve and, for the most part, samples are
from similar lithofacies. As discussed below, it
is possible that some of this δ18O variation could
be due to changes in ice volume, but it is not
likely to be associated simply with facies shift
on the shelf.
The unusual isotopic variation in the lower
part of Telm 6, however, is more difficult to
resolve. There is a marked facies change to
cross-bedded sand associated with the Telm
5–6 contact. The depleted oxygen isotope values and radiogenic 87Sr/86Sr ratios at the base
of Telm 6 in association with this facies change
suggest that all these phenomena could be due
to a change in depositional setting to a locus
more proximal to a freshwater source. However,
the lack of evidence for riverine discharge elsewhere in the unit (see above) makes it difficult
to argue for here. The Telm 6 δ18O depletion
more likely reflects a real climate signal than a
facies change. The fact that Cucullaea is virtually absent at sample sites during this interval
might also be an indication of warmer (rather
than fresher) water, since this taxon grows preferentially during the winter months and therefore might prefer cooler conditions.
In general then, we maintain that most of the
oxygen isotope variation exhibited in these data
can be attributed to climate change rather than
significant environmental and facies shifts along
the shelf. Because there is little convincing evidence for significant freshwater influx to the
depositional setting, relative changes in water
depth or distance from open water are unlikely
to result in changes in δ18O that could be misinterpreted as temperature change.
PALEOCLIMATIC AND
PALEOCEANOGRAPHIC
IMPLICATIONS
Antarctic Shelf Paleotemperatures
These data provide the first taxonomically
consistent record of paleotemperature variation with this level of stratigraphic resolution
throughout the Eocene in an Antarctic marine
shelf section. The overall trend in the La Meseta
Formation is broadly consistent with the pattern of increasing δ18O and cooling temperature reported by Gaździcki et al. (1992) and
Dutton et al. (2002), but these data reveal a
672
much greater level of complexity in δ18O and
temperature through the Eocene section than
anticipated from the earlier, lower resolution
work. [Note that our inferred temperatures are
cooler than Dutton et al. (2002) for equivalent
carbonate isotope values due to a difference in
assumed δ18O of seawater (−2‰ versus −1‰);
see Fig. 11 for the effect on temperature of different presumed δ18OSW values.] Absolute paleotemperature estimates from these La Meseta
Formation bivalves are in good agreement with
data from high southern latitude surface waters
summarized by Zachos et al. (1994).
Several features of the global (Zachos et al.,
2001) and Southern Ocean (Stott et al., 1990;
Bohaty and Zachos, 2003) marine records of
δ18O-based paleotemperatures are evident in
this high-resolution, shelf data set. First, given
the uncertainty in strontium isotope age estimates, it is likely that maximum temperatures
near the base of Telm 3 reflect the early Eocene
climatic optimum (EECO). Strontium isotope
ratios from Telm 2 consistently yield age estimates of ca. 54 Ma, indicating that the EECO
(roughly 51–53 Ma; Zachos et al., 2001) should
be encompassed within the La Meseta Formation section. Maximum inferred temperatures in
Telm 3 are ~15 °C, at least 4–5 °C warmer than
earlier or later samples. The degree of warmth
slightly exceeds that seen in the global deepocean record (Zachos et al., 2001). Inferred
paleotemperatures at Seymour Island during the
EECO are comparable to those inferred in the
Arctic Ocean at the time based on TEX86 (~18 °C,
Sluijs et al., 2006), but less than those suggested
at the Paleocene-Eocene thermal maximum
(PETM). Cooling on the Antarctic continent following the EECO at ca. 52 Ma is also inferred
based on changes in the clay mineral assemblage of cores from the Southern Ocean (Robert and Kennett, 1992). Significantly, the age
of Telm 2 samples, just postdating the PETM,
suggests as well that this interesting event might
also be recorded in Seymour Island sediments,
perhaps manifest within Telm 1 and its unusual
assemblage of calcitic fossils and deeply ironstained sediments [see Sadler (1988), Stilwell
and Zinsmeister (1992), and Hara (2001) for
descriptions]. The underlying Cross Valley Formation is upper Paleocene in age (Askin, 1988;
Harwood, 1988). Inferred temperatures for the
remainder of Telm 3, Telm 4, and Telm 5 are
reasonably consistent, hovering around 10 °C
for the rest of the early Eocene.
The extreme but short-lived excursion toward
negative δ18O values at the base of Telm 6 might
also have a parallel in the open ocean, making
the alternative explanation of a facies change
to brackish water even less likely. Taken at face
value, the δ18O shift indicates warming of shelf
waters to temperatures comparable to those
reached during the EECO, followed rapidly by
~7 °C of cooling. 87Sr/86Sr ratios from bivalves
immediately above the negative excursion
yield ages of 40.4 to 41.3 Ma, indicating that
this warming could coincide with the middle
Eocene climatic optimum (MECO, 41–42 Ma)
observed in sediments deposited in the Southern
Ocean on the Maud Rise and the Kerguelen Plateau (Bohaty and Zachos, 2003) (Figs. 10A and
10B). The amount of warming during this event
on the Seymour Island shelf is ~5 °C, slightly
more than that noted in Southern Ocean surface
waters (Bohaty and Zachos, 2003). While the
duration of warmth on Seymour Island is poorly
constrained, samples that encompass it span a
limited stratigraphic thickness (on the order of
a few meters) and show no obvious evidence
for disconformity within. The MECO as recognized in the Southern Ocean persists for less
than a million years, but it is not inconceivable
that such an event could be recognized in a shelf
section. The pattern of variation in δ13C around
the excursion is also consistent with, albeit more
extreme than, that seen in the Southern Ocean
(Figs. 10C and 10D) (Bohaty and Zachos,
2003), where a drop in δ13C occurs after peak
MECO warmth is reached in the surface waters
and coincident with the shift to cooler temperatures (see below). Whether this negative oxygen
isotope excursion reflects warming temperatures, a flush of fresh water, or a combination of
both, 87Sr/86Sr age estimates from immediately
upsection indicate that the circumstances associated with the base of Telm 6 likely correlate
with the MECO. Warmer temperatures might
reasonably be expected to lead to an increase in
precipitation and hence runoff, and so the negative δ18O values and shift to sandier sediments
could represent a combination of both influences. Refinements in age estimates either in the
La Meseta Formation or in the deep-sea record
may call for an adjustment of this suggested correlation. However, there is no other clear interval
of warm temperatures during the middle Eocene
trend toward cooler temperatures represented in
data thus far available from the Southern Ocean;
therefore, it is likely that the excursion noted on
Seymour Island correlates with the MECO, if
indeed the negative Telm 6 δ18O values primarily reflect temperature.
The ensuing shift toward substantially more
positive values beginning ca. 41 Ma, hereafter
the Middle Eocene Cooling, is the most significant interval of cooling in the Seymour Island
section. Interestingly, Birkenmajer et al. (2005)
report evidence for localized mountain glaciers
in the South Shetland Islands that, while difficult to constrain, yield ages consistent with
cooling at this time. In addition, Dallai et al.
Geological Society of America Bulletin, May/June 2008
Geological Society of America Bulletin, May/June 2008
55
50
40
35
1.0
δ18O
VPDB
0.0
-1.0
Eurhomalea
Cucullaea
A
2.0
1.5
0.0 -0.5
VPDB
0.5
δ18O
1.0
-2.0
0.0
VPDB
δ13C
2.0
0.5
2.0
VPDB
1.0 1.5
δ13C
Data from
Bohaty and Zachos (2003)
2.5
(Holes 738B, 748B)
Benthic (Holes 689B,
690B, 738B/C, 748B)
Data from
Bohaty and Zachos (2003)
Fine fraction
D
(Holes 738B, 748B)
Benthic (Holes 689B,
690B, 738B/C, 748B)
-6.0 -4.0
C
Fine fraction
B
Ice-Free Temperature (°C)
4
6
8 10 12
MECO
2
Figure 10. Stable oxygen and carbon isotope values of La Meseta Formation bivalves (plotted by genus) compared to data from the Southern Ocean (from Bohaty and Zachos,
2003). Pale symbols are mean values for each shell; bold symbols are average values and standard deviations for each horizon by taxon. (Α) δ18O values and calculated paleotemperatures using δ18OSW = −2‰; gray line is fitted through data from Eurhomalea, because values from that taxon more closely approximate mean annual temperature;
(B) δ18O data from the Southern Ocean, showing middle Eocene climatic optimum (MECO); temperatures calculated using δ18OSW = −1‰; C); δ13C data from the La Meseta
Formation, showing potential for divergence between taxa in Telm 7; (D) δ13C data from the Southern Ocean. Note the difference in the X axes between shelf and oceanic data,
and that δ18O and δ13C axes increase in opposite directions. Telm—Tertiary Eocene La Meseta.
Age (my)
Late
EOCENE
Middle
Early
Ice-Free Temperature (°C)
5
10
15
Eocene climate change on the Antarctic shelf
673
Ivany et al.
35
-3
-2
δ18OSW
-1
Late
-4
50
(freezing)
EOCENE
Middle
45
(missing section)
Early
Age (my)
40
-4
-3
0
5
-2
Carbon Cycling on the Shelf
-1
10
15
Temperature (°C)
20
Figure 11. Calculated paleotemperatures for a range of
seawater compositions (−1‰ to −4‰). Mean temperatures do not reach freezing unless δ18OSW value exceeds
−3‰. Winter temperatures presumably would be colder,
however, and may support the existence of seasonal sea
ice in the late middle and late Eocene.
(2001) use δ18O and δD values of granitoid
minerals altered by exchange with meteoric
fluids in hydrothermal systems to suggest an
interval of climate cooling in the Ross Sea of
West Antarctica between 42 Ma and 38 Ma.
Robert and Chamley (1991) and Robert and
Kennett (1992) document an increase in illite
at the expense of kaolinite and palygorskite in
Southern Ocean cores, consistent with cooling
during the middle Eocene (ca. 41 Ma). Thomas
(2004) also suggests a link with high-latitude
climate cooling to explain εNd data suggesting
the reinstatement of Southern Ocean deepwater formation around 40 Ma.
It is possible that the Middle Eocene Cooling inferred from Telm 6 oxygen isotope enrichment reflects not only falling temperatures but
also limited growth of ice on the Antarctic continent at this time. Mg/Ca ratios from the Southern Ocean suggest the potential for ice growth
on Antarctica starting ca. 40 Ma (Billups and
Schrag, 2003). Tripati et al. (2005) have suggested a short-lived middle Eocene glaciation
ca. 41 Ma based on benthic δ18O data from the
equatorial Pacific Ocean. Edgar et al. (2007)
note a correlative pattern in the equatorial Atlantic, but demonstrate that enrichment (≤0.6‰) is
not enough to warrant the hypothesis (Tripati et
al., 2005) of an ice sheet at both poles. Billups
and Schrag (2003) similarly estimate that the ice
volume contribution to the global δ18O signal at
674
is equivocal. The slight trend back to more negative δ18O values in bivalves preserved in the
upper part of Telm 7 may reflect the interval of
late Eocene warmth, or at least the interruption
of cooling (Poag et al., 2003), recognized in
both the Southern Ocean (Bohaty and Zachos,
2003) and the global benthic record (Zachos et
al., 2001) just prior to the onset of glaciation in
the early Oligocene.
this time is ~0.4‰. A short-lived deepening of
the carbonate compensation depth (CCD) noted
in carbonate accumulation rates in the equatorial Pacific (Lyle et al., 2005) as well suggests
that at least some ice growth occurred in concert
with the middle Eocene positive δ18O shift, as
glacioeustatic sea-level fall led to a shift in the
locus of carbonate deposition from the shelves to
the deep sea. However, even given that roughly
0.5‰ of the observed shift in δ18O on Seymour
Island might be due to the growth of glacier ice
in the interior of Antarctica, late middle and late
Eocene mean temperatures are still universally
cooler than in the earlier part of the record.
As in the Southern Ocean record from
Bohaty and Zachos (2003), minimum temperatures (most positive δ18O values) are attained
ca. 37 Ma. This is roughly coincident with the
earliest hypothesized evidence for glaciation in
Greenland based on the presence of ice-rafted
debris (Eldrett et al., 2007). Mean temperatures
at this time approach 5 °C assuming a seawater δ18O value of −2‰, but would nearly reach
freezing if seawater were 1‰ more depleted
(Fig. 11). Since Huber et al. (2003) predict
values as low as −3‰ east of the Peninsula
from their early Eocene climate model, this is
not out of the question. If so, conditions would
likely be cool enough to support the existence
of sea ice at least in the winter. The evidence
for this in the sedimentologic record, however,
The pattern of variation in δ13C through time
on the shelf is less straightforward. As in the
Southern Ocean (Bohaty and Zachos, 2003)
and recently documented in the Italian Apennines (Jovane et al., 2007), δ13C values in the La
Meseta Formation are generally positive during
what we suggest to be the MECO (Figs. 10C and
10D), perhaps associated with increased burial
of organic carbon during warm conditions. And
like both of these records, values become more
negative in concert with the subsequent cooling.
However, the magnitude of this negative shift
in δ13C on the Antarctic shelf is much larger
and persists for much longer, making it clear
that perturbations to the carbon cycle imposed
by cooling were more pronounced or magnified in this shallow-water setting. Gaździcki et
al. (1992) also recognize this significant shift
toward more negative δ13C in the upper part of
the La Meseta Formation, reporting values as
low as −5‰ from the shells of both Cucullaea
and Eurhomalea.
Factors responsible for the large, late-middle
Eocene negative shift in δ13C in the Seymour
Island section are unclear, and more likely
reflect aspects of the carbon cycle restricted to
the shelf in this region. The long-term change
to more negative values in the Southern Ocean
following the MECO takes place synchronously
in both surface and deep waters, indicating that
it is likely due to a decrease in the proportion of
organic carbon stored in the global sedimentary
reservoir. This shift in the δ13C of exogenic carbon contributes to the pattern on the shelf, but
only accounts for ~0.8‰ of the at least 6‰ shift
in the Seymour Island section. The timing of the
most precipitous drop in δ13C in both settings
appears to correlate with a shift toward more
radiogenic εNd values in the Atlantic Ocean,
interpreted by Scher and Martin (2006) to reflect
the initial opening of the Drake Passage and
mixing of Pacific and Atlantic Ocean waters.
That change in circulation might have led to
episodes of high productivity associated with
upwelling of nutrient-rich bottom water in some
regions (Anderson and Delaney, 2005; DiesterHaass and Zahn, 1996). Although negative δ13C
values are consistent with increased upwelling,
Geological Society of America Bulletin, May/June 2008
Eocene climate change on the Antarctic shelf
a phenomenon also not unexpected in association with cooling, values in the La Meseta Formation become significantly more negative than
those recorded from benthic foraminifera in the
deep ocean, indicating that upwelling of isotopically depleted deep water onto the shelf cannot
account for them alone.
An influx of isotopically negative dissolved
inorganic carbon (DIC) from the terrestrial
realm is a possible explanation, but one would
expect such a pulse to correlate with more negative δ18O values associated with freshwater runoff, and this is not the case. Another potential
explanation involves the formation of sea ice
that limited air-sea exchange, allowing depleted
carbon generated by decomposition to accumulate in the water column. However, if there
was enough water movement on the shelf to
transport and deposit the sand that characterizes much of Telm 6, then it is also likely that
circulation was vigorous enough to mix and
ventilate the waters. Another alternative relies
on the ecology of the bivalves themselves. It is
possible that, as temperatures cooled rapidly,
physiological stress led to incorporation of a
fraction of isotopically depleted metabolic CO2
in the shell. This would likely be a short-lived
response to rapidly changing conditions, and
natural selection should have produced adaptations to the cooler conditions relatively quickly,
making it difficult to explain why values remain
comparatively negative upsection. The fact that
δ13C values for the two genera diverge further
upsection in Telm 7 (Fig. 10C) does suggest that
they did not incorporate carbon into their shells
in the same way, potentially supporting this, but
the hypothesis has yet to be evaluated.
Implications for Meridional Temperature
Gradients
In addition to providing high-latitude shelf
paleotemperatures through the Eocene for climate reconstruction, these data can be compared
to those from tropical settings to quantify the
evolution of the pole-to-equator thermal gradient on shelves during the course of global cooling. This gradient is important because it reflects
the processes associated with heat distribution
on the surface of the planet, and therefore has
significant implications for our understanding of
the climate system (Zachos et al., 1994). Pearson et al. (2001, 2007) report tropical Eocene
shelf temperatures that hover around 30–33 °C
for the duration of the epoch, warmer and surprisingly stable in comparison to those based
on deep-sea cores (e.g., Zachos et al., 1994).
Based on this, paleotemperatures derived from
La Meseta Formation bivalves suggest that the
meridional temperature gradient at the EECO
approximates 15 °C, and increases to near
25 °C at the coolest part of the late Eocene.
These values are consistent with those reported
by Pearson et al. (2001), and somewhat exceed
those of Zachos et al. (1994) due to underestimates of tropical sea surface temperatures. The
steepening of this gradient has consequences for
the biogeographic distribution of marine shelf
organisms (e.g., Crame, 1997). These data combined with those of Pearson et al. (2007) predict
that thermally induced biogeographic differentiation peaked in the early Eocene shortly after
the EECO, and again in association with cooling in the middle Eocene. This pattern might not
entirely hold true for the northern hemisphere,
however, because similar data from the length
of the Atlantic Ocean suggest that development
of cooler conditions in the middle Eocene was
a phenomenon restricted to high southern latitudes (Robert and Chamley, 1991).
IMPLICATIONS FOR THE BIOTA
One of our underlying objectives in researching the climate record of the La Meseta Formation is to better understand the conditions experienced by Antarctic marine macrofauna during
the course of Eocene cooling. Presumably, temperature is an important variable driving evolutionary and ecological change at high latitudes,
but relationships during this time interval have
not been clearly demonstrated. Here, we show
that the most pronounced temperature change in
the section is the middle Eocene cooling, occurring low in Telm 6 following a brief excursion
to warmer conditions. A review of the published
paleontological literature suggests that faunal
and floral change within the La Meseta Formation is also highest at the Telm 5–6 transition. In
the marine realm, molluscan data from Stilwell
and Zinsmeister (1992) indicate a drop in taxonomic diversity as high as 50% at this time.
Feldmann et al. (2003) similarly see a marked
shift in the composition of the decapod fauna.
Among marine vertebrates, Long (1992) notes
the virtual elimination of sharks above Telm 5
(see also Case, 1992), and Case (2007) reports a
significant increase in the abundance of penguin
and cetacean material in the upper part of the
formation. The transition appears to have had
ecological as well as taxonomic consequences.
Aronson et al. (1997, 2007) and Aronson and
Blake (2001) report a change in the nature of
predator-prey relationships at roughly the Telm
5–6 contact due to the appearance of dense populations of predation-sensitive, epifaunal suspension feeders. Werner et al. (2004) also relate
morphological data from mollusks suggesting a
drop in the importance of shell-crushing predators and a significant decline in the frequency of
predatory drilling by naticid gastropods. Change
evidently extended to the terrestrial flora and
fauna as well, based on specimens washed in
and preserved in marine sediments, confirming that changes in the marine assemblage are
not simply related to facies change on the shelf
but reflect a wholesale change in environmental
conditions in the Peninsular region. Case (2007)
documents a substantial drop in vertebrate
diversity going into Telms 6 and 7, with changes
particularly noted among the mammal and bird
faunas. We also note based on field observations
that fossil wood is not seen anywhere in the
section above Telm 5, suggesting that climate
change could mark the elimination of significant forest cover from the Peninsula. Although
we recognize an unconformity at the Telm 5–6
contact based on 87Sr/86Sr values, and there is a
coincident facies change to sandier sediments
that might also have affected taphonomic conditions (e.g., Feldmann et al., 2003), the composition and character of the biota is sufficiently distinct on either side of this transition that we do
not believe the difference can be explained by
the simple accumulation of evolutionary change
during the time missing from the stratigraphic
section or by a shift in the local depositional and
preservational environment. We hypothesize
that the middle Eocene cooling of roughly 10 °C
imposed metabolic constraints upon organisms
with which they had not been faced for several
tens of millions years. Whether temperature
itself was the proximal cause, or associated
changes in productivity on the shelf (Aronson
et al., 2007) were more important (e.g., Dayton
and Oliver, 1977), cooling initiated substantial
ecosystem reorganization associated with the
new physiological challenges of dealing with
cold temperatures. The excellent fossil record of
the La Meseta Formation allows for a detailed
examination of the specifics of this relationship.
We anticipate that future work will test aspects
of this hypothesis and clarify the mechanisms
through which climate change imposed such
consequences for the Antarctic biota.
CONCLUSIONS
The agreement of our δ18O data from nearshore peninsular Antarctic habitats with those
from the Southern Ocean, in both general pattern as well as detail, affirms the use of shallow
shelf carbonate fossils as useful archives for
paleoclimate reconstruction. In addition, the La
Meseta Formation record demonstrates that climate perturbations recorded as far away as the
Kerguelan Plateau (Bohaty and Zachos, 2003)
were ubiquitous, high southern latitude phenomena, experienced on the shelf as well as in
the open ocean. Likewise, aspects of the global
Geological Society of America Bulletin, May/June 2008
675
Ivany et al.
Eocene climate record are also evident in this
high-latitude shelf section. Acknowledging the
limitations of strontium isotopic age control, the
most prominent events in the Seymour Island
climate record include the EECO and subsequent cooling at ca. 52 Ma, the middle Eocene
warming that we tentatively correlate with the
MECO, and most prominently, Middle Eocene
Cooling ca. 41 Ma.
The cause(s) of the middle Eocene shift to
warmer conditions and the subsequent reversal
into Middle Eocene Cooling remains a matter
for discussion. Bohaty and Zachos (2003) rule
out methane release as the proximal cause for
MECO warming because their transient δ13C
excursion does not happen until after warming has already begun. They instead suggest
a tectonic trigger, calling for CO2 release and
greenhouse warming associated with ridge or
arc volcanism. Scher and Martin (2006) suggest
that cooling immediately following the MECO
is coincident with the opening of the nascent
Drake Passage, and that associated changes in
circulation and productivity led to a drawdown
of CO2 that may have accelerated cooling and
brought about the proposed transient middle
Eocene glaciation. It is possible that tectonism
around Antarctica led to both these phenomena, first releasing CO2 to cause warming, and
then, once a threshold is breached to allow for
circum-Antarctic circulation, facilitating CO2
drawdown and cooling by subsequent upwelling
and phytoplankton production. While decreasing CO2 is now favored as the trigger for rapid
ice growth near the Eocene-Oligocene boundary
(DeConto and Pollard, 2003), the correlation
between the Middle Eocene Cooling recognized
here in the Peninsular region and the opening of
a shallow-water connection through the Drake
Passage suggests that gateways nevertheless
might have played an important role in global
Paleogene cooling.
ACKNOWLEDGMENTS
We thank John Evans and Kurt Burmeister for
help in the field, and Devin Buick, Nicole Miklus,
Allison Cattani, and Cristina Story for microsampling bivalves in the lab. John Schue and Devin
Buick evaluated growth banding in Eurhomalea and
Cucullaea, respectively. Lora Wingate conducted all
stable isotope analyses. Devin Buick and Jim Brower
helped with specimen photography. Jim Zachos and
Steve Bohaty offered thoughtful comments early on
and made their data available. Bruce Wilkinson and
Ellen De Man read drafts of the manuscript. Steve
Bohaty and Greg Ludvigson reviewed the paper and
made many helpful suggestions, and Hope Jahren
handled the paper for Geological Society of America
Bulletin. This research was supported by grants from
the National Science Foundation (NSF)’s Office of
Polar Programs to Ivany and Lohmann (OPP-0125409
and OPP-0125589), and a Research Experience for
Undergraduate (REU) Program supplement to Ivany.
676
Fieldwork in the winters of 2000–2001, 2001–2002,
and 2003–2004 was supported by NSF grants to
Aronson and Blake (OPP-9908828, OPP-9908856,
and ANT-0245563). Data from this study are available from the Antarctic Master Directory, under
Paleoclimate–Land Records–Isotopes.
REFERENCES CITED
Anderson, L.D., and Delaney, M.L., 2005, Middle Eocene
to early Oligocene paleoceanography from Agulhas
Ridge, Southern Ocean (Ocean Drilling Program Leg
177, Site 1090): Paleoceanography, v. 20, p. PA 1013,
doi:10.1029/2004PA001043.
Aronson, R.B., and Blake, D.B., 2001, Global climate change
and the origin of modern benthic communities in Antarctica: American Zoologist, v. 41, p. 27–39, doi: 10.166
8/0003-1569(2001)041[0027:GCCATO]2.0.CO;2.
Aronson, R.B., Blake, D.B., and Oji, T., 1997, Retrograde
community structure in the late Eocene of Antarctica:
Geology, v. 25, p. 903–906, doi: 10.1130/0091-7613(1
997)025<0903:RCSITL>2.3.CO;2.
Aronson, R.B., Thatje, S., Clarke, A., Peck, L.S., Blake,
D.B., Wilga, C.D., and Seibel, B.A., 2007, Climate
change and invasibility of the Antarctic benthos:
Annual Review of Ecology, Evolution, and Systematics, v. 38, p. 129–154, doi: 10.1146/annurev.ecolsys.38.091206.095525.
Arthur, M.A., Williams, D.F., and Jones, D.S., 1983, Seasonal temperature-salinity changes and thermocline
development in the mid-Atlantic Bight as recorded by
the isotopic composition of bivalves: Geology, v. 11,
p. 655–659, doi: 10.1130/0091-7613(1983)11<655:ST
CATD>2.0.CO;2.
Askin, R.A., 1988, Campanian to Paleocene palynological
succession of Seymour and adjacent islands, northeast
Antarctic Peninsula, in Feldmann, R.M., and Woodburne, M.O., eds., Geology and paleontology of Seymour
Island, Antarctic Peninsula: Boulder, Colorado, Geological Society of America Memoir 169, p. 131–153.
Askin, R.A., 1997, Eocene–? earliest Oligocene terrestrial
palynology of Seymour Island, Antarctica, in Ricci,
C.A., ed., The Antarctic region: Geological evolution
and processes: Siena, Terra Antarctica, p. 993–996.
Askin, R.A., and Fleming, R.F., 1982, Palynological investigations of Campanian to lower Oligocene sediments on
Seymour Island, Antarctic Peninsula: Antarctic Journal
of the United States, v. 17, p. 70–71.
Askin, R.A., Elliot, D.H., Stilwell, J.D., and Zinsmeister,
W.J., 1991, Stratigraphy and paleontology of Campanian and Eocene sediments, Cockburn Island, Antarctic
Peninsula: Journal of South American Earth Sciences,
v. 4, p. 99–117, doi: 10.1016/0895-9811(91)90021-C.
Baumiller, T.K., and Gaździcki, A., 1996, New crinoids from
the Eocene La Meseta Formation of Seymour Island,
Antarctic Peninsula: Paleontologica Polonica, v. 55,
p. 101–116.
Billups, K., and Schrag, D.P., 2003, Application of benthic
foraminiferal Mg/Ca ratios to questions of Cenozoic climate change: Earth and Planetary Science Letters, v. 209,
p. 181–195, doi: 10.1016/S0012-821X(03)00067-0.
Birkenmajer, K., Gaździcki, A., Krajewski, K.P., Przybycin,
A., Solecki, A., Tatur, A., and Yoon, H.I., 2005, First
Cenozoic glaciers in West Antarctica: Polish Polar
Research, v. 26, no. 1, p. 3–12.
Blake, D.D., and Aronson, R.B., 1998, Stelleroidea of the
late Eocene La Meseta Formation, Seymour Island,
Antarctic Peninsula: Journal of Paleontology, v. 72,
p. 339–352.
Blake, D.B., and Zinsmeister, W.J., 1988, Eocene asteroids
(Echinodermata) from Seymour Island, Antarctic Peninsula, in Feldmann, R.M., and Woodburne, M.O., eds.,
Geology and paleontology of Seymour Island, Antarctic Peninsula: Boulder, Colorado, Geological Society
of America Memoir 169, p. 489–498.
Bohaty, S.M., and Zachos, J.C., 2003, Significant Southern
Ocean warming event in the late middle Eocene: Geology, v. 31, p. 1017–1020, doi: 10.1130/G19800.1.
Bryant, J.D., Jones, D.S., and Mueller, P.A., 1995, Influence
of freshwater flux on 87Sr/86Sr chronostratigraphy in
marginal marine environments and dating of vertebrate
and invertebrate faunas: Journal of Paleontology, v. 69,
p. 1–6.
Buick, D.P., and Ivany, L.C., 2004, 100 years in the dark:
Extreme longevity of Eocene bivalves from Antarctica:
Geology, v. 32, p. 921–924, doi: 10.1130/G20796.1.
Case, J.A., 1988, Paleogene floras from Seymour Island, Antarctic Peninsula, in Feldmann, R.M., and Woodburne,
M.O., eds., Geology and paleontology of Seymour
Island, Antarctic Peninsula: Boulder, Colorado, Geological Society of America Memoir 169, p. 523–530.
Case, J.A., 1992, Evidence from fossil vertebrates for a
rich Eocene, Antarctic marine environment, in Kennett, J., and Warnke, D., eds., Paleoenvironment evolution of Antarctica and the Southern Oceans: American Geophysical Union Antarctic Research Series 56,
p. 119–130.
Case, J.A., 2007, Opening of the Drake Passage: Does this
event correlate to climate change and biotic events from
the Eocene La Meseta Formation, Seymour Island,
Antarctic Peninsula?, in Cooper, A.K., and Raymond,
D.R., et al., eds., Antarctica: A keystone in a changing
World—Online Proceedings of the 10th International
Symposium on Antarctic Earth Sciences: U.S. Geological Survey Open-File Report 2007-1047, Extended
Abstract 117, 3 p.
Cocozza, C.D., and Clarke, C.M., 1992, Eocene microplankton from La Meseta Formation, northern Seymour
Island: Antarctic Science, v. 4, no. 3, p. 355–362, doi:
10.1017/S0954102092000506.
Crame, J.A., 1997, An evolutionary framework for the polar
regions: Journal of Biogeography, v. 24, p. 1–9, doi:
10.1111/j.1365-2699.1997.tb00045.x.
Dallai, L., Ghezzo, C., and Longinelli, A., 2001, Fossil
hydrothermal systems tracking Eocene climate change
in Antarctica: Geology, v. 29, p. 931–934, doi: 10.1130
/0091-7613(2001)029<0931:FHSTEC>2.0.CO;2.
Dayton, P.K., and Oliver, J.S., 1977, Antarctic soft-bottom benthos in oligotrophic and eutrophic environments: Science, v. 197, p. 55–58, doi: 10.1126/science.197.4298.55.
DeConto, R.M., and Pollard, D., 2003, Rapid Cenozoic
glaciation of Antarctica triggered by declining atmospheric CO2: Nature, v. 421, p. 245–249, doi: 10.1038/
nature01290.
Diester-Haass, L., and Zahn, R., 1996, Eocene-Oligocene
transition in the Southern Ocean: History of water
mass circulation and biological productivity: Geology,
v. 24, p. 163–166, doi: 10.1130/0091-7613(1996)024<
0163:EOTITS>2.3.CO;2.
Dingle, R.V., and Lavelle, M., 1998, Antarctic Peninsular cryosphere: Early Oligocene (c. 30 Ma) initiation and a revised glacial chronology: Journal of the
Geological Society, v. 155, p. 433–437, doi: 10.1144/
gsjgs.155.3.0433.
Dingle, R.V., Marenssi, S.A., and Lavelle, M., 1998, High
latitude Eocene climate deterioration: Evidence from
the northern Antarctic Peninsula: Journal of South
American Earth Sciences, v. 11, no. 6, p. 571–579, doi:
10.1016/S0895-9811(98)00035-2.
Ditchfield, P.W., Marshall, J.D., and Pirrie, D., 1994, High latitude paleotemperature variation: New data from the Tithonian to Eocene of James Ross Island, Antarctica: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 107,
p. 79–101, doi: 10.1016/0031-0182(94)90166-X.
Doktor, M., Gaździcki, A., Jerzmanska, A., and Porębski,
S.J., 1996, A plant-and-fish assemblage from the
Eocene La Meseta Formation of Seymour Island (Antarctic Peninsula) and its environmental implications:
Palaeontologica Polonica, v. 55, p. 127–146.
Dutton, A.L., Lohmann, K.C., and Zinsmeister, W.J., 2002,
Stable isotope and minor element proxies for Eocene
climate of Seymour Island, Antarctica: Paleoceanography, v. 17, p. doi: 10.1029/2000PA000593.
Edgar, K.M., Wilson, P.A., Sexton, P.F., and Suganuma, Y.,
2007, No extreme bipolar glaciation during the main
Eocene calcite compensation shift: Nature, v. 448,
p. 908–911, doi: 10.1038/nature06053.
Edwards, D.C., and Huebner, J.D., 1977, Feeding and growth
rates of Polinices duplicatus preying on Mya arenaria
at Barnstable Harbor, Massachusetts: Ecology, v. 58,
p. 1218–1236, doi: 10.2307/1935077.
Eldrett, J.S., Harding, I.C., Wilson, P.A., Butler, E., and Robert, A.P., 2007, Continental ice in Greenland during the
Geological Society of America Bulletin, May/June 2008
Eocene climate change on the Antarctic shelf
Eocene and Oligocene: Nature, v. 446, p. 176–179, doi:
10.1038/nature05591.
Elliot, D.H., 1988, Tectonic setting and evolution of the
James Ross Basin, northern Antarctic Peninsula, in
Feldmann, R.M., and Woodburne, M.O., eds., Geology and paleontology of Seymour Island, Antarctic
Peninsula: Boulder, Colorado, Geological Society of
America Memoir 169, p. 541–555.
Elliot, D.H., and Trautman, T.A., 1982, Lower Tertiary strata
on Seymour Island, Antarctic Peninsula, in Craddock,
C., ed., Antarctic Geosciences: University of Wisconsin Press, p. 287–297.
Feldmann, R.M., Schweitzer, C.E., and Marenssi, S.A., 2003,
Decapod crustaceans from the Eocene La Meseta Formation, Seymour Island, Antarctica: A model for preservation of decapods: Journal of the Geological Society,
v. 160, p. 151–169, doi: 10.1144/0016-764901-136.
Filkhorn, H.F., 1994, Fossil Scleractinian Corals from James
Ross Basin, Antarctica: Washington, D.C., American
Geophysical Union, Antarctic Research Series, v. 65,
96 p.
Fordyce, R.E., 1989, Origins and evolution of Antarctic
marine mammals, in Crame, J.A., ed., Origins and
evolution of the Antarctic biota: London, Geological
Society, p. 269–281.
Francis, J.E., and Poole, I., 2002, Cretaceous and early Tertiary climates of Antarctica: Evidence from fossil wood:
Palaeogeography, Palaeoclimatology, Palaeoecology,
v. 182, p. 47–64, doi: 10.1016/S0031-0182(01)00452-7.
Gandolfo, M.A., Hoc, P., Santillana, S.N., and Marenssi,
S.A., 1998a, Una flor fosil morfologicamente afin a
las Grossulariaceae (Orden Rosales) de la Formacion
La Meseta (Eoceno medio), Isla Marambio, Antartida,
in Casadio, S., ed., Paleogeno de America del Sur y
de la Peninsula Antarctica: Association Paleontologica
Argentina Publicacion Especial, 5, p. 147–154.
Gandolfo, M.A., Marenssi, S.A., and Santillana, S.N.,
1998b, Flora y paleoclima de la Formacion La Meseta
(Eoceno medio), Isla Marambio, Antartida, in Casadio,
S., ed., Paleogeno de America del Sur y de la Peninsula
Antarctica: Association Paleontologica Argentina Publicacion Especial, 5, p. 155–162.
Gaździcki, A.J., Gruszczynski, M., Hofan, A., Malkowski,
K., Marenssi, S.A., Halas, S., and Tatur, A., 1992,
Stable carbon and oxygen isotope record in the Paleogene La Meseta Formation, Seymour Island, Antarctica: Antarctic Science, v. 4, p. 461–468, doi: 10.1017/
S0954102092000671.
Gradstein, F., Ogg, J., and Smith, A., 2004, A geologic time
scale 2004: Cambridge, UK, Cambridge University
Press, 589 p.
Grossman, E.L., and Ku, T.-L., 1986, Oxygen and carbon
isotope fractionation in biogenic aragonite: Temperature effects: Chemical Geology, v. 59, p. 59–74, doi:
10.1016/0009-2541(86)90044-6.
Hall, S.A., 1977, Cretaceous and Tertiary dinoflagellates
from Seymour Island, Antarctica: Nature, v. 267,
p. 239–241, doi: 10.1038/267239a0.
Haq, B.U., Hardenbol, J., and Vail, P.R., 1987, Chronology of
fluctuating sea levels since the Triassic: Science, v. 235,
p. 1156–1167, doi: 10.1126/science.235.4793.1156.
Hara, U., 2001, Bryozoans from the Eocene of Seymour
Island, Antarctic Peninsula, in Gaździcki, A., ed., Palaeontological results of the Polish Antarctic expeditions.
Part III: Palaeontologia Polonica, v. 60, p. 33–156.
Harwood, D.M., 1985, Cretaceous to Eocene Seymour Island
siliceous microfossil biostratigraphy, in Workshop on
Cenozoic Geology of the Southern High Latitudes:
Columbus, Ohio, Sixth Gondwanan Symposium, Ohio
State University, p. 17–18.
Harwood, D.M., 1988, Upper Cretaceous and lower Paleocene diatom and silicoflagellate biostratigraphy of
Seymour Island, eastern Antarctic Peninsula, in Feldmann, R.M., and Woodburne, M.O., eds., Geology and
paleontology of Seymour Island, Antarctic Peninsula:
Boulder, Colorado, Geological Society of America
Memoir 169, p. 55–129.
Hodell, D.A., Kamenov, G.D., Hathorne, E.C., Zachos, J.C.,
Röhl, U., and Westerhold, T., 2007, Variations in the
strontium isotope composition of seawater during the
Paleocene and early Eocene from ODP Leg 208 (Walvis Ridge): Geochemistry Geophysics Geosystems,
v. 8, p. Q09001, doi: 10.1029/2007GC001607
Howarth, R.J., and McArthur, J.M., 1997, Statistics for
strontium isotope stratigraphy: A robust LOWESS fit
to the marine Sr-isotope curve for 0 to 206 Ma, with
look-up table for derivation of numeric age: The Journal of Geology, v. 105, p. 441–456.
Huber, M., Sloan, L., and Shellito, C., 2003, Early Paleogene oceans and climate: A fully coupled modeling
approach using the NCAR CCSM, in Wing, S.L., Gingerich, P.D., Schmitz, B., and Thomas, E., eds., Causes
and consequences of globally warm climates in the
Early Paleogene: Boulder, Colorado, Geological Society of America Special Paper 369, p. 25–47.
Huebner, J.D., and Edwards, D.C., 1981, Energy budget
of the predatory marine gastropod Polinices duplicatus: Marine Biology, v. 61, p. 221–226, doi: 10.1007/
BF00386663.
Ingram, B.L., Ingle, J.C., and Conrad, M.E., 1996, A 2000 yr
record of Sacramento–San Joaquin river inflow to San
Francisco Bay estuary, California: Geology, v. 24,
p. 331–334, doi: 10.1130/0091-7613(1996)024<0331
:AYROSS>2.3.CO;2.
Ivany, L.C., 2007, Contributions to the Eocene climate
record of the Antarctic Peninsula, in Cooper, A.K., and
Raymond, C.R., et al., eds., Antarctica: A keystone in
a changing world—Online Proceedings of the 10th
International Symposium on Antarctic Earth Sciences:
U.S. Geological Survey Open-File Report 2007-1047,
Extended Abstract 068, 4 p.
Ivany, L.C., Van Simaeys, S., Domack, E.W., and Samson,
S.D., 2006, Evidence for an earliest Oligocene ice sheet
on the Antarctic Peninsula: Geology, v. 34, p. 377–380,
doi: 10.1130/G22383.1.
Jones, D.S., and Quitmyer, I.R., 1996, Marking time with
bivalve shells: Oxygen isotopes and season of annual
increment formation: Palaios, v. 11, p. 340–346, doi:
10.2307/3515244.
Jovane, L., Florindo, F., Coccioni, R., Dinarés-Turell, J.,
Marsili, A., Monechi, S., Roberts, A.P., and Sprovieri,
M., 2007, The middle Eocene climatic optimum event
in the Contessa Highway section, Umbiran Apennines,
Italy: Geological Society of America Bulletin, v. 119,
no. 3/4, p. 413–427, doi: 10.1130/B25917.1.
Kobashi, T., Grossman, E.L., Yancey, T.E., and Dockery,
D.T., 2003, The oxygen isotopic record of seasonality
in Conus shells and its application to understanding
late middle Eocene (38 Ma) climate: Paleontological
Research, v. 7, p. 343–355, doi: 10.2517/prpsj.7.343.
Lear, C.H., Elderfield, H., and Wilson, P.A., 2000, Cenozoic
deep-sea temperatures and global ice volumes from
Mg/Ca in benthic foraminiferal calcite: Science, v. 287,
p. 269–272, doi: 10.1126/science.287.5451.269.
Lear, C.H., Rosenthal, Y., Coxall, H.K., and Wilson, P.A.,
2004, Late Eocene to early Miocene ice sheet dynamics and the global carbon cycle: Paleoceanography,
v. 19, p. PA4015, doi: 10.1029/2004PA001039
Long, D.J., 1992, Sharks from the La Meseta Formation
(Eocene), Seymour Island, Antarctic Peninsula: Journal of Vertebrate Paleontology, v. 12, no. 1, p. 11–32.
Lyle, M., Olivarez-Lyle, A., Backman, J., and Tripati, A.,
2005, Biogenic sedimentation in the Eocene equatorial
Pacific—The stuttering greenhouse and Eocene carbonate compensation depth, in Wilson, P.A., Lyle, M.,
and Firth, J.V., eds., Proceedings of the Ocean Drilling
Program, Scientific Results, 199, p. 1–35. Available
online <http://www-odp.tamu.edu/publications/199_
SR/VOLUME/CHAPTERS/219.PDF>
Macellari, C.E., 1988, Stratigraphy, sedimentology, and
paleoecology of Upper Cretaceous/Paleocene shelfdeltaic sediments of Seymour Island, in Feldmann,
R.M., and Woodburne, M.O., eds., Geology and paleontology of Seymour Island, Antarctic Peninsula:
Boulder, Colorado, Geological Society of America
Memoir 169, p. 25–53.
Marenssi, S.A., 2006, Eustatically controlled sedimentation
recorded by Eocene strata of the James Ross Basin,
Antarctica, in Francis, J.E., Pirrie, D., and Crame, J.A.,
eds., Cretaceous-Tertiary high-latitude paleoenvironments, James Ross Basin, Antarctica: London, Geological Society of London, Special Publication 258,
p. 125–133.
Marenssi, S.A., Santillana, S.N., and Rinaldi, C.A., 1998,
Stratigraphy of the La Meseta Formation (Eocene),
Marambio (Seymour) Island, Antarctica: Asociación
Paleontológica Argentina, Publicación Especial, v. 5,
p. 137–146.
Marenssi, S.A., Net, L.I., and Santillana, S.N., 2002, Provenance, environmental and paleogeographic controls
on sandstone composition in an incised-valley system:
The Eocene La Meseta Formation, Seymour Island,
Antarctica: Sedimentary Geology, v. 150, p. 301–321,
doi: 10.1016/S0037-0738(01)00201-9.
McArthur, J.M., Howarth, R.J., and Bailey, T.R., 2001,
Strontium isotope stratigraphy: LOWESS Version 3:
Best fit to the marine Sr-isotope curve for 0–509 Ma
and accompanying look-up table for deriving numerical age: The Journal of Geology, v. 109, p. 155–170,
doi: 10.1086/319243.
McKinney, M.L., McNamara, K.J., and Wiedman, L.A.,
1988, Echinoids from the La Meseta Formation
(Eocene), Seymour Island, Antarctica, in Feldmann,
R.M., and Woodburne, M.O., eds., Geology and paleontology of Seymour Island, Antarctic Peninsula:
Boulder, Colorado, Geological Society of America
Memoir 169, p. 499– 503.
Miklus, N.M., 2008, The high-latitude response of temperature seasonality to global Eocene cooling [M.S. thesis]:
Syracuse, Syracuse University, 62 p.
Miller, K.G., Mountain, G.S., Browning, J.V., Kominz, M.,
Sugarman, P.J., Christie-Blick, N., Katz, M.E., and
Wright, J.D., 1998, Cenozoic global sea level, sequences,
and the New Jersey Transect: Results from coastal plain
and continental slope drilling: Reviews of Geophysics,
v. 36, p. 569–601, doi: 10.1029/98RG01624.
Miller, K.G., Kominz, M.A., Browning, J.V., Wright, J.D.,
Mountain, G.S., Katz, M.E., Sugarman, P.J., Cramer,
B.S., Christie-Blick, N., and Pekar, S.F., 2005, The
Phanerozoic record of global sea-level change: Science, v. 310, p. 1293-1298.
Mitchell, E.D., 1989, A new cetacean from the late Eocene
La Meseta Formation, Seymour Island, Antarctic Peninsula: Canadian Journal of Fisheries and Aquatic Sciences, v. 46, p. 2219–2235.
Pannella, G., and MacClintock, C., 1968, Biological and
environmental rhythms reflected in molluscan shell
growth, in Macurda, D.B.J., ed., Paleobiological
aspects of growth and development, p. 64–81.
Pearson, P.N., Ditchfield, P.W., Singano, J., Harcourt-Brown,
K.G., Nicholas, C.J., Olsson, R.K., Shackleton, N.J.,
and Hall, M.A., 2001, Warm tropical sea surface temperatures in the Late Cretaceous and Eocene epochs:
Nature, v. 413, p. 481–487, doi: 10.1038/35097000.
Pearson, P.N., van Dongen, B.E., Nicholas, C.J., Pancost,
R.D., Schouten, S., Singano, J.M., and Wade, B.S.,
2007, Stable warm tropical climate through the Eocene
Epoch: Geology, v. 35, no. 3, p. 211–214, doi: 10.1130/
G23175A.1.
Pekar, S.F., Hucks, A., Fuller, M., and Li, S., 2005, Glacioeustatic changes in the early and middle Eocene
(51–42 Ma): Shallow-water stratigraphy from ODP
Leg 189 Site 1171 (South Tasman Rise) and deep-sea
δ18O records: Geological Society of America Bulletin,
v. 117, p. 1081–1093, doi: 10.1130/B25486.1.
Pirrie, D., Marshall, J.D., and Crame, J.A., 1998, Marine
high Mg calcite cements in Teredolites-bored fossil
wood: Evidence for cool paleoclimates in the Eocene
La Meseta Formation, Seymour Island, Antarctica:
Palaios, v. 13, p. 276–286, doi: 10.2307/3515450.
Pirrie, D., Dithfield, P.W., and Marshall, J.D., 1994, Burial
diagenesis and pore-fluid evolution in a Mesozoic
back-arc basin: the Marambio Group, Vega Island:
Antarctica: Journal of Sedimentary Research A, v. 64,
p. 541–552.
Poag, C.W., Mankinen, E., and Norris, R.D., 2003, Late
Eocene impacts: Geologic record, correlation, and
paleoenvironmental consequences, in Prothero, D.R.,
Ivany, L.C., and Nesbitt, E., eds., From greenhouse to
icehouse: The Marine Eocene-Oligocene Transition:
New York, Columbia University Press, p. 495–510.
Poole, I., and Cantrill, D.J., 2006, Cretaceous and Cenozoic
vegetation of Antarctica: Integrating the fossil wood
record, in Francis, J.E., Pirrie, D., and Crame, J.A.,
eds., Cretaceous-Tertiary high-latitude paleoenvironments, James Ross Basin, Antarctica: London, Geological Society of London, p. 63–81.
Porębski, S.J., 1995, Facies architecture in a tectonicallycontrolled incised-valley estuary: La Meseta Forma-
Geological Society of America Bulletin, May/June 2008
677
Ivany et al.
tion (Eocene) of Seymour Island, Antarctic Peninsula:
Studia Geologica Polonica, v. 107, p. 7–97.
Porębski, S.J., 2000, Shelf-valley compound fill produced
by fault subsidence and eustatic sea-level changes,
Eocene La Meseta Formation, Seymour Island, Antarctica: Geology, v. 28, p. 147–150, doi: 10.1130/0091
-7613(2000)028<0147:SVCFPB>2.3.CO;2.
Robert, C., and Chamley, H., 1991, Development of early
Eocene warm climates, as inferred from clay mineral
variations in oceanic sediments: Palaeogeography,
Palaeoclimatology, Palaeoecology, v. 89, p. 315–331,
doi: 10.1016/0031-0182(91)90169-R.
Robert, C., and Kennett, J.P., 1992, Paleocene and Eocene
kaolinite distribution in the South Atlantic and Southern Ocean: Antarctic climatic and paleoceanographic
implications: Marine Geology, v. 103, p. 99–110, doi:
10.1016/0025-3227(92)90010-F.
Sadler, P.M., 1988, Geometry and stratification of uppermost
Cretaceous and Paleogene units on Seymour Island,
northern Antarctic Peninsula, in Feldmann, R.M., and
Woodburne, M.O., eds., Geology and paleontology of
Seymour Island, Antarctic Peninsula: Boulder, Colorado,
Geological Society of America Memoir 169, p. 303–320.
Scher, H.D., and Martin, E.E., 2006, Timing and climatic
consequences of the opening of Drake Passage: Science,
v. 312, p. 428–430, doi: 10.1126/science.1120044.
Shackleton, N.J., and Kennett, J.P., 1975, Paleotemperature
history of the Cenozoic and the initiation of Antarctic glaciation: Oxygen and carbon isotope analyses in
DSDP Sites 277, 279, and 281: Initial Reports of the
Deep Sea Drilling Program, v. 74, p. 743–755.
Sluijs, A., Schouten, S., Pagani, M., Woltering, M.,
Brinkhuis, H., Damsté Sinninghe, J.S., Dickens, G.R.,
Huber, M., Reichart, G., Stein, R., Matthiessen, J.,
Lourens, L.J., Pedentchouk, N., Backman, J., Moran,
K., and the Expedition 302 Scientists, 2006, Subtropi-
678
cal Arctic Ocean temperatures during the Paleocene/
Eocene thermal maximum: Nature, v. 441, p. 610–613,
doi: 10.1038/nature04668.
Stilwell, J.D., and Zinsmeister, W.J., 1992, Molluscan systematics and biostratigraphy, Lower Tertiary La Meseta
Formation, Seymour Island, Antarctic Peninsula: Washington D.C., American Geophysical Union, 192 p.
Stott, L.D., Kennett, J.P., Shackelton, N.J., and Corfield,
R.M., 1990, The evolution of Antarctic surface waters
during the Paleogene: Inferences from the stable isotopic composition of planktonic foraminifers, ODP Leg
113, in Barker, P.F., Kennett, J.P., et al., eds., Proceedings of the Ocean Drilling Project, Scientific Results
Volume 113, p. 849–863.
Surge, D., and Walker, K.J., 2006, Geochemical variation
in microstructural shell layers of the southern quahog (Mercenaria campechiensis): Implications for
reconstructing seasonality: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 237, p. 182–190, doi:
10.1016/j.palaeo.2005.11.016.
Thomas, D.J., 2004, Evidence for deep-water production
in the North Pacific Ocean during the early Cenozoic
warm interval: Nature, v. 430, p. 65–68, doi: 10.1038/
nature02639.
Tripati, A., Backman, J., Elderfield, H., and Ferretti, P.,
2005, Eocene bipolar glaciation associated with global
carbon cycle changes: Nature, v. 436, p. 341–346, doi:
10.1038/nature03874.
Vail, P.R., Mitchum, R.M., Shipley, T.H., and Buffler, R.T.,
1980, Unconformities of the North Atlantic: Philosophical Transactions of the Royal Society of London,
v. 294, p. 137–155, doi: 10.1098/rsta.1980.0021.
Werner, J.E., Blake, D.B., and Aronson, R.B., 2004, Effects of
late Eocene cooling on Antarctic marine communities:
GeoSur 2004—International symposium on the geology and geophysics of the southernmost Andes, the
Scotia Arc and the Antarctic Peninsula: Bollettino di
Geofisica, v. 45, supplement 2, p. 262–265.
Wiedman, L.S., and Feldmann, R.M., 1988, Ichnofossils,
tubiform body fossils, and depositional environments
of the La Meseta Formation (Eocene) of Antarctica, in
Feldmann, R.M., and Woodburne, M.O., eds., Geology and paleontology of Seymour Island, Antarctic
Peninsula: Boulder, Colorado, Geological Society of
America Memoir 169, p. 531–539.
Wiedman, L.S., Feldmann, R.M., Lee, D.E., and Zinsmeister, W.J., 1988, Brachiopoda from the La Meseta Formation (Eocene), Seymour Island, Antarctica, in Feldmann, R.M., and Woodburne, M.O., eds., Geology and
paleontology of Seymour Island, Antarctic Peninsula:
Boulder, Colorado, Geological Society of America
Memoir 169, p. 449–457.
Wrenn, J.H., and Hart, G.F., 1988, Paleogene dinoflagellate
cyst biostratigraphy of Seymour Island, Antarctica, in
Feldmann, R.M., and Woodburne, M.O., eds., Geology and paleontology of Seymour Island, Antarctic
Peninsula: Boulder, Colorado, Geological Society of
America Memoir 169, p. 321–447.
Zachos, J.C., Stott, L.D., and Lohmann, K.C., 1994, Evolution
of early Cenozoic marine temperatures: Paleoceanography, v. 9, no. 2, p. 353–387, doi: 10.1029/93PA03266.
Zachos, J.C., Pagani, M., Sloan, L., Thomas, E., and Billups,
K., 2001, Trends, rhythms, and aberrations in global
climate 65 Ma to present: Science, v. 292, p. 686–693,
doi: 10.1126/science.1059412.
MANUSCRIPT RECEIVED 1 JUNE 2007
REVISED MANUSCRIPT RECEIVED 5 DECEMBER 2007
MANUSCRIPT ACCEPTED 16 DECEMBER 2007
Printed in the USA
Geological Society of America Bulletin, May/June 2008