You are on page 1of 15

ARTICLE IN PRESS

Deep-Sea Research II 54 (2007) 22932307


www.elsevier.com/locate/dsr2

Onset of Cenozoic Antarctic glaciation


Peter F. Barkera,, Bernhard Diekmannb, Carlota Escutiac
a

Threshers Barn, Whitcott Keysett, Clun, Shropshire SY7 8QE, UK


Alfred Wegener Institute for Polar and Marine Research (Research Unit Potsdam), Telegrafenberg A43, 14473 Potsdam, Germany
c
Instituto Andaluz de Ciencias de la Tierra (IACT), CSIC-University de Granada, Campus de Fuentenueva s/n, 18002 Granada, Spain
b

Accepted 24 July 2007


Available online 22 October 2007

Abstract
This paper considers the wide range of evidence, both direct and indirect, for the onset of Cenozoic Antarctic glaciation.
It distinguishes two useful phases of Antarctic glacial onset: an initial phase of mountain glaciation, from which ice streams
occasionally and in isolated locations reached sea level, and a subsequent phase of full glaciation, with an ice sheet as
large as todays, extending everywhere to sea level. According to direct evidence, generally proximal, from the continent
itself or surrounding Southern Ocean, the rst of these occurred probably during the late Eocene, while the second
developed at the EoceneOligocene boundary. Indirect evidence, mainly involving proxy measurements from DSDP and
ODP sites remote from the Southern Ocean, suggests that middle and late Eocene glaciations may have been full also (ice
sheets possibly even larger than todays) but short-lived, and that the E/O boundary onset differed from these mainly in
producing a stable ice sheet. In pursuing the notion of glacial onset, we examined the direct record separately for East
Antarctica, West Antarctica, and the Antarctic Peninsula, the different sub-ice topography and geographic positions of
which suggest that their glacial histories could have been different. The direct record for an initial, middle or late Eocene
phase is very sparse: only the rare occurrence of IRD at Southern Ocean DSDP and ODP sites suggests the possibility of
early ice, and all three regions include mountains that could have hosted such ice. Although the indirect record and climate
modelling in combination suggest that full glaciation of each region was probably synchronous, we nd differences in
the available direct evidence. There is abundant evidence that East Antarctica became fully glaciated in the earliest
Oligocene, but certain evidence of glaciation of a similar age extending to sea level is sparse for the Antarctic Peninsula,
and is not found until the late Oligocene for West Antarctica. High-resolution direct evidence is required to resolve
uncertainties in glacial history.
r 2007 Elsevier Ltd. All rights reserved.
Keywords: Antarctic; Glacial onset; Palaeoclimate

1. Introduction
The high latitudes receive little of their heat
directly from the sun. The greater part is received
indirectly, by atmospheric and oceanic transport
Corresponding author.

E-mail address: pfbarker@tiscali.co.uk (P.F. Barker).


0967-0645/$ - see front matter r 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.dsr2.2007.07.027

from lower latitudes. Thus, the high latitudes


experience the coolest climates and, provided that
moisture is available, are the most likely to
experience glaciation. Within the Cenozoic, in
which cooling was near-global in extent, it has been
the Southern Hemisphere, with land over the
geographic pole, that appears to have experienced
glaciation earlier than the Northern Hemisphere

ARTICLE IN PRESS
2294

P.F. Barker et al. / Deep-Sea Research II 54 (2007) 22932307

(although evidence for early Cenozoic bipolar


glaciation is growing) and for which glaciation has
been more persistent. Nevertheless, Cenozoic Antarctic continental glaciation has been a complicated
process: the continent went through different phases
of glaciation, different regions of Antarctica have
had different glacial histories, and periods of
deglaciation and re-glaciation may have occurred.
In the face of this complexity, it is necessary to focus
on an important aspect of glaciation, that has the
potential to provide a greater understanding of its
causes and effects and perhaps the reasons for a
generally lesser expression of bipolar glaciation.
Here we consider glacial onset; the evidence for
subsequent variation is both sparse and controversial, and less obviously related to cause.
We aim to distinguish between an early phase of
glacial onset, in which ice generated at high
elevation may occasionally and in a very few places
drain at sea level, and a later-stage full glaciation,
with an ice sheet almost as large as todays. We are
encouraged that these phases can be distinguished
by the external, essentially distal and indirect,
evidence of ocean-oor geochemistry (from benthic
foraminifera: e.g., oxygen isotopes, Zachos et al.,
2001 and Mg/Ca ratios, Lear et al., 2000) and
numerical modelling (e.g., Huybrechts, 1993; De
Conto and Pollard, 2003), showing that full,
stable Antarctic glaciation probably developed
to an advanced state quite rapidly, around the
Eocene-Oligocene boundary (E/O) interval.
The possible causes of Antarctic glaciation are
poorly understood. The thermal isolation of Antarctica within a growing Southern Ocean, with the
development of a strong, deep-reaching Antarctic
Circumpolar Current (ACC) (e.g., Kennett, 1977)
has long been a preferred explanation. Thus, as an
additional contribution to an understanding of
cause, evidence for the time of onset of the ACC
is also considered elsewhere in this volume. However, numerical modelling has led more recently to
the suggestion (De Conto and Pollard, 2003) that a
steady global decline in atmospheric pCO2 could
have triggered a rapid glaciation even without a
contribution from changing ocean circulation. The
non-linear ice volume response found by De Conto
and Pollard is a feature of Antarctic ice sheet
growth (concerning surface elevation or albedo
perhaps) rather than of CO2 involvement, as it
gures also in a purely temperature-dependent ice
sheet model (e.g., Huybrechts, 1993; Barker et al.,
1999a; Fig. 1). Also, therefore, although not ruled

Fig. 1. Ice sheet volume plotted against the rise in mean annual
temperature at sea level around Antarctica, compared with
modern temperatures (Huybrechts, 1993). A non-linear relationship between temperature and ice volume is clearly seen.

out, an event, such as the onset of an ACC, is not


a requirement. Other suggested causes for global
cooling are changes in ocean circulation (e.g.,
Lawver and Gahagan, 2003) and tectonic uplift
(e.g., Raymo and Ruddiman, 1992) elsewhere in the
world. Here, we discuss the kinds of evidence that
bear on glacial onset, and assess the present state of
knowledge for Antarctica.
2. The present ice sheet
The modern Antarctic ice sheet covers the entire
continent, with only 0.3% of the land area ice free
(e.g., Lythe et al., 2001 and Fig. 2). It reaches a
thickness exceeding 4500 m in places and has a total
volume of 25.4 million km3 (equivalent to a sea-level
rise of 57 m). It acquires volume by snowfall and,
being cold, loses volume largely (up to 90%
Paterson, 1994) by calving ice streams. The mean
residence time of ice within the ice sheet exceeds 0.5
Myr in places close to the centre, but is less than
20 kyr near the edge (Paterson, 1994), where
snowfall is both greater and warmer (leading to
greater basal melting and faster ow). The snow
of which the ice sheet is formed is isotopically
lighter than the oceans, because of fractionation

ARTICLE IN PRESS
P.F. Barker et al. / Deep-Sea Research II 54 (2007) 22932307

2295

Fig. 2. Surface of the present ice sheet (modied from Drewry, 1983). Ice ow directions are drawn perpendicular to surface slopes, and do
not show ice stream locations. Thin dashed lines show ice divides and thick dashed lines divide East from West Antarctica and the
Antarctic Peninsula. Numbered black locations are existing DSDP and ODP drill sites close to the continent. Western Ross Sea drill sites
CR Cape Roberts, MS MSSTS, CI CIROS 1. Labelled red locations are ice core sites V Vostok, B Byrd, C Dome C, S Siple, T Taylor
Dome. GaM Gamburtsev Subglacial Mountains.

during evaporation and precipitation, but by an


amount that may have varied with time. It is
important to appreciate that the relationship
between ice volume and ice temperature is uncertain: cold ice is stiffer than warm ice, so ows less
easily, and there is less melting; but colder air carries
less moisture, suggesting lower snowfall. Thus, a
colder ice sheet may not be larger. The present
regime is interglacial, in which the grounding
line of the Antarctic ice-sheet is well away from
the continental shelf edge, mainly because there
are no major Northern Hemisphere ice sheets,
that lower sea level during glacial maxima so as
to permit grounding line advance (but, most
probably, only a minor increase in Antarctic icesheet volume, as the continental shelves are largely
oored by low-viscosity till, permitting rapid ice
ow). A signicantly larger ice sheet than this in the
past is difcult to imagine. Many workers distinguish between a cold and a temperate ice sheet,
and describe todays as cold. An Oligocene ice sheet
is generally considered to have been warmer than
todays.

3. Nature and validity of evidence of onset of


glaciation
3.1. Indirect evidence
Some of the distal or indirect evidence has already
been mentioned. Oxygen isotopic measurements on
benthic foraminifera (Zachos et al., 2001, provided
a useful recent compilation) are ambiguous, as
changes in isotopic ratio may result from changes of
seawater temperature or of mean seawater composition (which is affected mainly by the volume, but
also by the isotopic chemistry, of sequestered ice:
e.g., Mix and Ruddiman, 1984). Mg/Ca measurements, also on benthic foraminifera, (e.g., Lear
et al., 2000) are insensitive to ice volume/chemistry,
so have the ability to resolve the isotopic ambiguity
between ice volume and water temperature (Fig. 3):
the data base of reliable measurements is growing
but some interpretational uncertainties remain, as
the ratio may be sensitive to other factors. Global
sea-level change is sensitive to grounded ice volume
(Browning et al., 1996; Pekar et al., 2005), but is not

ARTICLE IN PRESS
2296

P.F. Barker et al. / Deep-Sea Research II 54 (2007) 22932307

Fig. 3. Zachos et al (2001) compilation of benthic oxygen isotopic data (raw data blue, 0.2 Ma RMS red) for the Cenozoic (modied from
Barker and Camerlenghi, 2002), showing oxygen isotopic events Mi-1 and Oi-1 and hypothesised short-lived Eocene glaciations, and
comparison of Zachos et al. (2001) data with Lear et al. (2000) measurements of Mg/Ca ratio in benthic foraminifera from DSDP Site 522
(Walvis Ridge, SE Atlantic) across the E/O interval.

widely used as a determinant at present because of


difculties in separating global and regional effects
(but see Pekar et al., 2002; Miller et al., 2005 for
useful recent comparisons that suggest greater
future use; Miller et al., 2005, propose intermittent
substantial glaciation throughout the past 100 Ma,
to account for global sea-level changes). The
indirect data are mainly volume effects, so (at the
present level of understanding) can indicate only
major changes, such as the inferred dramatic growth
in ice-sheet volume in the E/O interval. The
interpretation of other oxygen isotopic changes of
similar magnitude, as within the late Oligocene and
early Miocene, is less certain. Perhaps signicant
also, in view of speculation that short-lived (unstable?) Eocene glaciations and the initial earliest
Oligocene Oi-1 isotopic peak infer ice volumes
larger than the present Antarctic ice volume, is the
suggestion of Toggweiler and Bjornsson (2000) that
one effect of an ACC would have been to enhance
the climatic difference between Northern and
Southern Hemispheres. Before ACC onset therefore, climate in high latitudes of the two hemispheres could have been more equal, with signicant
ice volumes developing in the north.
An additional problem is distinguishing between
orbital forcing of ice-sheet volume and the longerterm changes that are of interest here. Orbital
variation in ice sheet volume is very evident from all
high-resolution oxygen isotopic studies, and in
several (e.g., Paul et al., 2000) it has been suggested
that orbital variation triggered major ice-sheet
volume change. However, it is necessary to separate
the two time scales, as our main concern here is with

the longer-term, irreversible changes in earth


response. The model of De Conto and Pollard
(2003) suggested that ice-sheet volume variation
over an orbital cycle could be quite large, provided
that a basal sliding mechanism (such as a renewable
basal till) was always and everywhere available.
Barker et al. (1999a) speculated that orbital variation in ice-sheet volume could be large in the initial
phases of glaciation, before ice-sheet extent had
become constrained by the continental shelf edge
(rather like the unconstrained Pleistocene Northern
Hemisphere ice sheets). However, uctuations in
oxygen isotopic measurements at single sites (e.g.,
Kennett and Stott, 1990; Zachos et al., 1996) place a
limit on such volume changes (although Wade and
Palike, 2004, argue for larger amplitude orbital
variation, from measurements at equatorial Site
1218), suggesting that, at the lower time resolution
we can employ (i.e. limited by onshore geology), we
should be able to identify an advanced full glacial
onset without confusion from glacialinterglacial
variation. An additional consideration, not addressed by numerical models, concerns ice-stream
drainage, which could have distorted the simple
picture produced by models by providing evidence
of glaciation at the coast well in advance of the main
ice-sheet margin.
Measurements on ODP samples far from the
Southern Ocean have pointed to additional complexity, for which more proximal and direct
evidence is required. One such circumstance, already mentioned, is the combination of the benthic
oxygen isotopic compilation of Zachos et al. (2001)
and measurements of Mg/Ca ratio in benthic

ARTICLE IN PRESS
P.F. Barker et al. / Deep-Sea Research II 54 (2007) 22932307

foraminifera from Site 522 in the SE Atlantic


(Fig. 3), which is used to argue that the entire
oxygen isotopic shift at the E/O interval is caused by
ice volume change, with no change in deep and
bottom-water temperature. This is of concern, as it
implies that another process, now perhaps inoperative, caused pre-glacial high-latitude creation,
through cooling, of deep and bottom water, and
that the (glacially induced) processes that now
operate, namely brine production beneath forming
(fresh-water) sea ice followed by supercooling
beneath cold ice shelves (Foster and Middleton,
1980; Foldvik and Gammelsrd, 1988) have been
less important in the past. An alternative, perhaps,
is to investigate the possibility of additional
inuences, not as yet determined for the Site 522
data, on measured Mg/Ca ratios (e.g., Lear et al.,
2004), or to question the value of compilations of
isotopic data from a range of sites, lest there be
inter-site differences in bottom-water source or in
dating. Closely related are two additional questions,
of the large amplitude of the initial (early Oligocene)
Oi-1 oxygen isotopic excursion (implying an ice
volume larger than that of the present day, if
conned to the Antarctic and having the same
average isotopic composition) and its rapid development, and of the proposed occurrence of earlier
(middle and late Eocene) glaciations, short-lived
(ca. 500 ka) but again larger in amplitude than the
present Antarctic ice volume and considered in
part to represent Northern Hemisphere glaciation
(Tripati et al., 2005). In both cases, the data set is
diverse, comprising oxygen isotopic and Mg/Ca
data and carbonate concentrations used to determine variations in carbonate compensation depth
(CCD), mainly from the equatorial Pacic (ODP
Sites 1218 and 1219). Support for Northern Hemisphere glaciation (which does not concern us here) is
said to be Oligocene NADW formation and the
presence of IRD within the Eocene of the only
Arctic Ocean section extant, from the Lomonosov
Ridge (Shipboard Scientic Party, 2005). Southern
Hemisphere support is cited as Eocene glacial
sediments from McMurdo Sound (Ehrmann,
1998), and uctuations in Antarctic-derived clay
minerals (kaolinite, smectite, chlorite and illite) on
Maud Rise and the Kerguelen Plateau (Ehrmann
and Mackensen, 1992; Robert and Kennett, 1992).
In all the instances described above, the assumption is made that atmospheric pCO2 is the dominant
inuence on glaciation. It should be pointed out
also that, according to the model of De Conto and

2297

Pollard (2003), ice volume changes in the earliest


phase of development could be rapid (though not
quite as rapid as required by Coxall et al., 2005), so
that short-lived ice sheets are mechanically feasible.
3.2. Direct evidence
A wealth and a wide range of direct evidence
bears on the history of East Antarctic glaciation,
much less on West Antarctic and Antarctic Peninsular glaciation but sufcient to suggest that the
three regions should be treated separately. The East
Antarctic ice-sheet is today by far the largest, with
the greatest scope for minor variation (although the
West Antarctic ice sheet is considered the least
stable): changes inferred from indirect evidence
cannot be assumed to reect changes in any
particular region of Antarctica alone. Also, it would
be unwise to rely overmuch on numerical models:
among the problems with many of the most
interesting numerical models of ice-sheet development currently available are that they do not treat
of ice grounded well below present sea level (i.e.
much of the West Antarctic ice sheet), of ice-sheet
drainage by ice streams (affecting all regions) or of
sea ice capable of moving on the ocean surface
(perhaps relevant to the Antarctic Peninsula).
Ice-rafted debris (IRD) is an important direct
indicator of glaciation. It may be released from
melting ice grounded below sea level both proximally (e.g., in the continental shelf facies residual
glacial marine of Anderson et al., 1980) and
distally, after transport in icebergs, within oceanoor sediments. There is a limit to the extent to
which variations in IRD at a particular site may be
used simply to infer variations in glaciation, since
melting is also affected by storminess and water
temperature along iceberg paths, by changing ocean
circulation, and by the location of debris within the
iceberg. In particular, debris may fall onto the
surface of a valley glacier, and wind-blown terrigenous material (usually therefore of a more restricted
size range) may accumulate on sea ice that then
melts or moves. Both of these are typical of the early
shallow phases of glaciation, while thicker-icesheet regimes are dominated by basally transported
debris within and beneath ice streams. The largest
ice streams, such as those within the Filchner
Ronne ice shelf today, carry no IRD to the iceberg
calving zone because of prior basal melting. In
summary, IRD onset may be used to infer the start,
or existence, of a phase of glaciation in which

ARTICLE IN PRESS
2298

P.F. Barker et al. / Deep-Sea Research II 54 (2007) 22932307

grounded ice reaches sea level at the continental


margin, or sea ice formed there, but more-detailed
interpretation, involving changes in IRD concentration, is more difcult.
Observations of cold surface water and sea ice
around Antarctica in the past, determined using
microfossil assemblages for example, provide an
inference of conditions onshore, but not a clear
indication of the phase of glaciation. Because of
uncertainties over the existence of inter-ocean
connections at a particular time, such an inference
can be identied with one particular region of
Antarctica only if the ocean off other regions is
markedly different.
Clay mineralogy (e.g., Ehrmann and Mackensen,
1992; Robert and Chamley, 1992) at offshore sites
has been used to distinguish a cold climate on a
neighbouring landmass (with dominantly physical
weathering of continental rock exposure, producing
more chlorite or well-crystallised illite) from a warm
wet climate (dominantly chemical weathering, producing more kaolinite and smectite). It is necessary
to know the geographic origin of the clays, and to
beware changes in the composition of the eroding
rocks, as might easily occur with continued erosion.
For example, physical re-erosion of older, chemically weathered sediment or older marine sediment
is a possibility. Also, an authigenic component may
mislead (e.g., Diekmann et al., 2004), as may a
change in ocean circulation if samples are from a
distal site.
Rock exposure onshore or on the continental
shelf is a potential source of climatic information,
provided it is representative. We should perhaps
bear in mind that, given the erosion common at the
base of grounded ice and, particularly, glacial
overdeepening at the continental margin, onshore
preservation of a geological record of an intervening
warmer period is unlikely. In fact, continental rock
outcrop relating to Cenozoic climate is sparse. Some
does occur, and there are indications of continental
geology also in glacial erratics exposed onshore, in
deep-sea sediments off drainage basins, and in drill
holes. Useful indicators include marine macro- and
microfossils, pollen and spores. The former reect
conditions inshore in shallow water, the latter
onshore vegetation, therefore non-glacial conditions, and constraints on glaciation are provided
by their declining diversity or geographic distribution, and disappearance. In circumstances of rapid
and perhaps reversing climate change, as within a
well-developed orbital cycle, the preferential pre-

servation of representatives of one part of a change


can lead to misunderstanding.
There is a particular interest in the ora, in the
context of glacialinterglacial variation. Southern
Hemisphere (including Oligocene Antarctic) ora
usually include Nothofagus, for which it is generally
accepted that even moderate seaways present a
barrier to dispersal (Hill and Dettmann, 1996). We
therefore may assume perhaps, despite the greater
Oligocene proximity of East Antarctica and Australia, or the Antarctic Peninsula and South
America, that re-colonisation after Oligocene glacial
maxima had to be from Antarctic refugia rather
than more distant sources. This would limit the
areal coverage of a maximal Oligocene ice sheet, in
contrast to conclusions based on indirect data.
Antarctic continental shelves are commonly deep
and inward-sloping, having been eroded by moving,
grounded ice. Many shelves have been extended by
prograding wedges of diamictunsorted terrigenous material deposited by or from ice (most
commonly by the rapidly moving ice streams,
grounded across much or all of the present
continental shelf around glacial maxima). The
wedges display a range of geometries, and are
difcult to recover by drilling, owing to the lack of
sorting and poor consolidation of the sediments,
exposure of the drill ship to ocean swell etc. (see
Hayes et al., 1975; Barker et al., 1999b; Cooper and
OBrien, 2004 for results of drilling the wedges).
Dates from the wedges probably indicate the
existence of an ice sheet of signicant size at that
part of the continental margin at that time. It has
been argued, on the basis of inverse modelling of
exure and sediment compaction, that glacial overdeepening of the shelves is preceded by a phase of
lesser glaciation, with shallow, outward-sloping
continental shelves (e.g., De Santis et al., 1999;
Camerlenghi et al., 2002). A proposal to combine
drill data from different parts of the continental
margin, where the prograded wedge and derived
upper rise drift reect different phases of Antarctic
glaciation in different regions, was generated by
ANTOSTRAT, an alliance of Antarctic marine
geoscientists formalised within the Scientic Committee for Antarctic Research (SCARsee Barker
et al., 1998). Directed by a model of glacial
evolution adduced from numerical modelling
(Huybrechts, 1993; Barker et al., 1999a) it was only
partly carried out. Recovery, though poor in places,
was sufcient to provide useful information about
the onset and recent (last 9 myr) phases of glaciation

ARTICLE IN PRESS
P.F. Barker et al. / Deep-Sea Research II 54 (2007) 22932307

(Barker and Camerlenghi, 2002; Cooper and


OBrien, 2004), and an extant proposal for drilling
the shelf and upper rise off Wilkes Land (Escutia
et al., 1997, 2005) seems capable of elucidating the
intervening phases.
Along with the classical model of a glaciation
caused by thermal isolation of the continent by an
ocean current, goes the glacial production of cold,
saline bottom water (at present by sinking of a brine
concentrated by persistent sea ice formation,
followed by supercooling beneath oating ice
shelves), invigorating ocean thermohaline circulation. Under this scheme, which operates today,
evidence of invigorated thermohaline circulation
could be used as an indicator of the existence of
Antarctic glaciation. However, the validity of this
indicator is now in question: attribution of virtually
all of the E/O interval isotopic shift to ice sheet
growth (Lear et al., 2000) implies that intermediate
and bottom waters of the world oceans were already
cooling in the absence of an Antarctic ice sheet, and
did not cool dramatically when an ice sheet was
formed.
4. Regional pattern of glaciation
We choose to treat separately East and West
Antarctica and the Antarctic Peninsula, partly on
account of the existing evidence, but also because of
clear differences in sub-ice topography and position
between them. For example, the South Pole has
been located within East Antarctica since the
Early Cretaceous (DiVenere et al., 1994); the other
two have lain farther north. East Antarctic ice is
mostly grounded above (pre-ice-sheet) sea level,
whereas most West Antarctic ice, including the
thickest parts, is grounded below sea level. The
Antarctic Peninsula is the farthest north, but
has a narrow axis at high elevation, which would
have acted as a nucleus for early ice formation.
The existence at onset times of an ACC or similar,
to reduce sea-surface temperature variation
around the Antarctic margin or to extend the
implications of marine-derived observations to
other regions of Antarctica, is uncertain (see Barker
et al., 2007).
5. East Antarctica
The isostatically compensated sub-ice topography
of East Antarctica is mostly above sea level, and the
comparatively large size of the present East

2299

Antarctic Ice Sheet (52 m sea-level rise on melting


Lythe et al., 2001) is generally taken to imply that
it was the earliest to form. This impression is
strongly supported by numerical modelling (e.g.,
Huybrechts, 1993; De Conto and Pollard, 2003).
The models suggest that an ice sheet developed rst
on mountain ranges (the Sr Rondane and Gamburtsev Mountains in particular, and the Transantarctic Mountains if they were then elevated), then
spread relatively rapidly (i.e. non-linearly in terms
of response to external variables) to form an ice
sheet extending everywhere to sea level. It is possible
to identify Wilkes Land (in particular the Wilkes
and Aurora Subglacial basins) as the part of East
Antarctica likely to have become glaciated last.
However, as already mentioned, such numerical
models do not include ice-stream drainage, which
could have transported potential IRD material to
the coast well in advance of the ice-sheet margin.
A comparison of benthic oxygen isotopic and Mg/
Ca ratio data (Zachos et al., 2001; Lear et al., 2000)
also supports the view that a large ice sheet (at least
as large as todays) developed rapidly within the
E/O interval.
Abundant fossiliferous ice-transported fragments
found onshore in the southern McMurdo Sound
area of the Ross Sea reect a cool temperate middle
to late Eocene coastal climate (Harwood and Levy,
2000), adjacent to the rising Transantarctic Mountains. No Eocene glacial facies were found, suggesting there was no ice at sea level, whereas Oligocene
diamictites were found.
Drilling in the Ross Sea region, mostly close to
East Antarctica and the Transantarctic Mountains,
has been a source of information about glacial
history. Early Oligocene ice-rafting may be seen at
DSDP Site 274 (Hayes et al., 1975), but other
DSDP Leg 28 sites (including Sites 267 and 268 on
the abyssal plain and continental rise off Wilkes
Land) show only early Miocene or later ice rafting.
Late Oligocene (27 Ma) glacial sediments were
described from the MSSTS-1 borehole, and earliest
Oligocene diamicton and IRD were seen in the
CIROS-1 core (Barrett, 1986, 1989). The age of the
lowest part of the CIROS-1 core is uncertain (early
Oligocene to middle Eocene), because of the long
ranges of the microfossils described, but diamicton
persists to the base. Drilling at Cape Roberts (Cape
Roberts Science Team, 2000, 2001) recovered a long
section of early Oligocene (and possibly
latest Eocene) age, its base estimated as 33.5
35.0 Ma. Low-diversity fauna in its upper part,

ARTICLE IN PRESS
2300

P.F. Barker et al. / Deep-Sea Research II 54 (2007) 22932307

and occasional striated clasts within all but the


lowest few tens of metres of this section, attest to its
glacial nature, despite the dominant conglomerate/
debris-ow lithology that reects rapid subsidence
and proximity to a rising Transantarctic Mountains.
Strontium- and oxygen-isotopic measurements on
marine bivalves from the upper part of the section
suggest early Oligocene shelf bottom-water temperatures (depending on the ice volume and salinity
models) of 7.6 and 5.2 1C, with a seasonal range of
between 1.2 and 4.9 1C. Temperatures were signicantly warmer and had a much greater range than
those of today. Unfortunately, an E/O transition
was not found conclusively.
It should be noted that palynomorphs and plant
fragments occur in sediment interbedded with
Oligocene diamictites in the CIROS-1 core, suggesting interglacial expansion of vegetation from refugia
(Barrett, 1989), and palynomorphs at Cape Roberts
have been described from lower Oligocene sediments (Raine and Askin, 2001). A very different,
more diverse palynomorph fauna was found at the
base of the hole, suggesting an Eocene age and a
warmer environment.
The southern Indian Ocean has been well studied,
with drilling in Prydz Bay at the East Antarctic
continental margin (ODP Legs 119, 188) and farther
offshore on the Kerguelen Plateau (ODP Legs 119,
120, 183) opposite Prydz Bay. On the continental
shelf in Prydz Bay, diamictite occurs at the base of
holes at Sites 739 and 742 (ODP Leg 119: Barron
et al., 1991), but dating (early Oligocene and
possibly older) is uncertain. From grounding line
locations, the Oligocene ice sheet is inferred to have
been slightly larger than todays. ODP Leg 188
drilling (Site 1166: OBrien et al., 2001; Cooper and
OBrien, 2004) recovered upper-middle Eocene
carbonaceous sands in Prydz Bay, with palynomorphs and with sand-grain surface textures that
suggested mountain glaciation to the south. Around
the E/O interval, these sands were overlain by sandclay layers and, above an erosion surface, by
glaciomarine diamictites containing dinoagellates,
diatoms and lonestones. Leg 188 did not recover
well-dated lower Oligocene sediments in Prydz Bay
(Cooper and OBrien, 2004). Evidence of Antarctic
glacial development is provided also by drilling on
the Kerguelen Plateau (ODP Legs 119, 120, 183).
Possible IRD was seen in middle Eocene sediments
at Sites 738 and 744 (Leg 119: Barron et al., 1991)
on the southern Plateau, and a change in clay
mineralogy (a decrease in smectite and increase in

kaolinite from glacial weathering on Antarctica) in


the late Eocene was attributed to glacial onset. This
was followed in the early Oligocene by unequivocal IRD at these sites and farther north at Site
748 (Leg 120: Wise et al., 1992), by a dramatic
decrease in smectite and increase in illite concentration within clays, reecting a change to physical
weathering onshore, and by coeval biogenic changes
(increases in opal and decreases in calcium carbonate, changes to radiolarian, nannofossil and
foraminiferal assemblages) suggesting cooling of
surface and deep waters. The principal objectives of
ODP Leg 183 (Frey et al., 2003) concerned aspects
of the Mesozoic volcanic evolution of the Kerguelen
Plateau, but it found that cooling waters from
middle Eocene time onward affected radiolarian
and foraminifer preservation, nannofossil assemblages, discoaster and sphenolith abundance and
diversity.
ODP Leg 113 drilled in the Weddell Sea. At Site
693, on the East Antarctic margin, middle lower
Oligocene sediments contained sparse gravel- and
pebble-sized IRD (Grobe et al., 1990). IRD was not
seen in sediments of the same age on Maud Rise,
but a siliceous biofacies was rst seen there in the
latest Eoceneearliest Oligocene.
Pre-glacial sediments from the Wilkes Land
margin have been recovered by dredging the eroded
anks of the Mertz-Ninnis Glacial Trough.
Dredged rocks included in situ Paleogene palynomorphs (Escutia et al., 2005).
6. Antarctic Peninsula
The Antarctic Peninsula is a long, narrow
dissected plateau, giving the impression of a
peneplain (rising from 900 m at its northern end to
17502000 m in the south). The plateau is broader in
the south, and total width (between Weddell Sea
and Southeast Pacic shelf edges) ranges from
200 km in the north to 600 km in the south. During
glacial maxima, an ice sheet probably extended to
the continental shelf edge. Snowfall is heavy, but the
present Antarctic Peninsular ice sheet is a few
hundred metres thick at most, and its climatic
regime has been characterised as small-volume,
high-throughput, rapid-response (Barker and
Camerlenghi, 1999).
There is a discrepancy between onshore and
offshore information on glacial history: offshore
ocean drilling during ODP Leg 178 (Barker and
Camerlenghi, 2002) detected a late Miocene (9 Ma)

ARTICLE IN PRESS
P.F. Barker et al. / Deep-Sea Research II 54 (2007) 22932307

to present ice sheet (grounded regularly to the shelf


edge), extending back speculatively to about 15 Ma
using related seismic reection studies (e.g., Rebesco
et al., 1997, 2002). Ice-sheet volume was found to be
insensitive to climate change, over the past 9 myr.
Earlier drilling (DSDP Leg 35Hollister et al.,
1976) was ambiguous because coring was discontinuous, but very sparse dropstones were reported
from 15 to 16 Ma-aged sediment at two sites.
Onshore exposures of shallow glacial marine sediments on the South Shetland Is (separated from the
northern Antarctic Peninsula by the Plio-Pleistocene opening of Branseld Strait), described by
Troedson and Smellie (2002) and Troedson and
Riding (2002), have been dated by Sr-isotopic
measurements (Dingle and Lavelle, 1998) on included bivalves and brachiopods, and 40Ar/39Ar
determinations (Troedson and Smellie, 2002) for
interbedded lavas, as middle and late Oligocene,
and earliest Miocene. These ages supersede older
KAr dates reported by Birkenmajer et al. (1987),
which had led to Eocene and other glaciations that
were difcult to understand. The sediments contained clasts of rocks now exposed only in the
Transantarctic and Ellsworth mountains, which was
explained in terms of a glaciation in those regions,
draining into a Weddell Sea that was experiencing a
clockwise circulation, as today. Intervening nonglacial sediments were recorded, suggesting a
glacial/interglacial cyclicity operating at the margins
of glaciation. The sparseness of the surviving record
did not permit of a precise age for Oligocene glacial
onset, but Barker and Camerlenghi (2002) inferred
an Oligocene glaciation to sea level along the
Antarctic Peninsula, assisted by the clockwise
circulation of sea ice and icebergs within the
Weddell Sea, followed (perhaps) by deglaciation,
and renewed glaciation at sea level in the middle
Miocene. More recently, Birkenmajer et al. (2005)
have reported evidence of mountain glaciation on
King George Island that they propose, on the basis
of scattered KAr dates, is of middle Eocene age.
Both the scatter and the vulnerability to other
effects of KAr dates make this occurrence uncertain evidence of early glaciation.
Eocene sediments (of the shallow-water La
Meseta Formation) have been examined on Seymour Island, on the Weddell Sea continental shelf at
the northern end of the Antarctic Peninsula, and
numerous climate-related studies undertaken. The
sedimentary section extends up into the latest
Eocene. Most recently, Dingle et al. (1998) de-

2301

scribed cooling effects on clay mineralogy and


sediment maturity, in the late middle and late
Eocene, and Dutton et al. (2002) interpreted
measured oxygen, carbon and strontium isotopic
ratios on bivalves, to show cooling of surface waters
from ca. 14 to 10 1C in the middle Eocene (assuming
ice-free conditions), remaining at that temperature until ca. 34 Ma. Other studies (e.g. on
palynomorphs, Askin, 1997; bryozoa, Hara, 1997;
mollusca, Stillwell and Zinsmeister, 1992; land
fauna, Reguero et al., 2002) show a matching
decrease in diversity. No late Eocene glacial
sediments have been identied, but Ivany et al.
(2006) have reported an E/O interval occurrence of
glaciation at sea level from directly above the La
Meseta Fm., with dating based on dinoagellates
and clast lithologies, and on Sr87/Sr86 ratios in
bivalves. Younger sediments than these (i.e. Oligocene and younger) could be accessible offshore.
Eocene sediments have been sampled farther east
on the South Orkney microcontinental block by
drilling, and within the Scotia Sea. ODP Site 696, on
the former, provided a shallow-water late middle to
late Eocene and possibly early Oligocene section
beneath a (presumed break-up) unconformity.
Mohr (1990) suggested a slightly warmer environment than on Seymour Island on the basis of
palynomorphs, and for ferns suggested mean annual
temperatures of 911 1C for the late Eocene and
57 1C for the early Oligocene (Mohr, 2001).
A dredge site in the Scotia Sea yielded a middle
Eocene (4546 Ma) age (Toker et al., 1991) and fern
palynomorphs indicated mean annual temperatures
onshore of 1620 1C (Mohr, 2001). These data
are incompatible with prolonged pre-Oligocene
glaciation.
7. West Antarctica
West Antarctica comprises the elevated and
rugged provinces of Marie Byrd Land and the
Jones Mountains along the Pacic margin and the
Ellsworth-Whitmore Mountains closer to East
Antarctica, and an inner, much lower province that
is partly smooth (close to the Ross Sea) but has
elsewhere considerable sub-ice topographic relief,
including the Byrd Subglacial Basin and Bentley
Subglacial Trough, which are at present up to
2500 m below sea level (Lythe et al., 2001). The deep
regions now contain the thickest ice, but there is
little doubt that West Antarctic glacial onset would
have involved ice-sheet nucleation on the more

ARTICLE IN PRESS
2302

P.F. Barker et al. / Deep-Sea Research II 54 (2007) 22932307

elevated regions. The West Antarctic ice sheet is


grounded largely below sea level and is therefore
considered by some to be unstable: melting would
produce a sea-level rise of about 5 m, one-tenth of
the result of East Antarctic ice-sheet melting. The
model of Huybrechts (1993) suggests that onset of a
West Antarctic ice sheet (i.e. nucleation on elevations) would have occurred at a lower mean annual
SST (around the margin, compared with todays)
than for East Antarctica.
The glacial history of Marie Byrd Land has been
reviewed recently by LeMasurier and Rocchi (2005).
The record of glaciation is partial and ambiguous,
occurring mainly within hyaloclastites accompanying volcanic eruptions, usually at considerable
present-day elevation (e.g. 2927 Ma 40Ar/39Ar ages
for the base of section at 2700 m on Mt. Petras
Wilch and McIntosh, 2000). A Marie Byrd Land
Dome is inferred from progressive uplift and block
faulting, over the past 34 myr or more, of a Late
Cretaceous (West Antarctic-New Zealand separation) West Antarctic Erosion Surface (most probably, originally near sea level), that accompanied
volcanic activity. Included marine microfossils,
theoretically capable not only of dating these events
but also of demonstrating that glaciation occurred
at sea level, are rare: in their absence, the palaeoelevation of the site is in question. The detection of
open-ocean forms among these microfossils is used
to infer episodes of deglaciation, but entanglement
with the Pliocene deglaciation controversy (Webb
and Harwood, 1991; Sugden et al., 1993; Stroeven,
1997; Gersonde et al., 1997), which concerns the
origins of microfossils in glacial sediment, renders
these interpretations uncertain.
Glacial history may be inferred from the nature of
erosion, under the assumption that it is sensitive to
both elevation and climatic change. The anomalous
unroong of a 34 Ma (40Ar/39ArRocchi et al.,
2006) gabbro is here used to imply an early, warmer
phase of glaciation, and the contrast of cirque
erosion of an early Miocene volcano with minimal
dissection since ca. 14 Ma is taken to mark the
transition from a warmer to a colder, less erosive
regime, perhaps in the period 1715 Ma. These may
translate into West Antarctic glacial onset in the
Oligocene, and a signicant change at 17-15 Ma,
but there is no unequivocal indication of the phase
of glacial onset that each represents, or of any long
intervening non-glacial period. The likely contribution of uplift to glaciation in Marie Byrd Land adds
uncertainty.

During DSDP Leg 28, drilling at Site 270 in the


central part of the Ross Sea recovered 1 m of preglacial Oligocene (26 Ma K/Ar age) glauconitic
sand, directly overlain by Oligocene to early
Miocene glacial marine silty claystone with sparse
erratics distributed throughout (Barrett, 1974a, b).
No source region was identied, but (perhaps
because of the telling absence of Trans-Antarctic
Mountains clasts except for one brief interval in the
middle Miocene) a source within the lower-relief
part of West Antarctica may be inferred. This is a
clear indication of West Antarctic glaciation,
possibly glacial onset, at or close to sea level,
shortly after 26 Ma.
Piston cores are a useful supplement to drilling in
the Southern Ocean, when they sample older
sediments. Wei (1992) has rened the age of Eltanin
piston cores from the South Pacic, reported by
Margolis and Kennett (1971) to contain IRD
(quartz grains), as early and middle Eocene,
suggesting that some ice occurred at sea level
around Antarctica at that time. The closest region
is West Antarctica, and the presence of IRD at more
northerly locations is justied by the likelihood of
different iceberg paths at a time when there were no
inter-ocean pathways. A contribution from other
regions (the Antarctic Peninsula, or East Antarctica
via the Ross Sea?) is possible, but perhaps unlikely.

8. Discussion and conclusions


We have described above the direct and indirect
published evidence of development of the Antarctic
ice sheet. In general, it is important to have the
benet of both kinds of evidence before drawing
rm conclusions. For example, use of the combination of benthic oxygen isotopic and Mg/Ca data
(assuming that all such data are valid) to reveal ice
volume should not be pushed too far. The isotopic
effect of an ice sheet is a combination of volume and
isotopic ratio of the sequestered water, which
depends partly on such factors as the prior history
of precipitation and the source of the accumulated
snow. These may have changed signicantly between the Oligocene and now (particularly if the
oceans were less well connected then), leading
perhaps to a misplaced appreciation (from proxy
data) of the nature of the substantial Oligocene
glaciation. Also, if the Oligocene ice sheet was
warmer than todays, the balance between snow
input and output (involving softer ice and perhaps

ARTICLE IN PRESS
P.F. Barker et al. / Deep-Sea Research II 54 (2007) 22932307

additional processes such as ablation), would have


been different.
We consider that glaciation included the Antarctic Peninsula and West Antarctica to some extent, as
well as East Antarctica. Despite the lack of direct
evidence, might it also have included a West
Antarctic and Antarctic Peninsular glaciation as
substantial as todays? There is direct evidence for
late Oligocene and subsequent West Antarctic
glaciation, but to infer an earlier glaciation (i.e.
E/O interval onset) at sea level in West Antarctica
without unequivocal direct evidence is to go too far.
Also, it is premature to interpret the oxygen isotopic
peak Oi-1 as an entirely Antarctic glacial volume.
More generally, the involvement of an orbital
variation of insolation in ice-sheet volume changes
is certain, in the Oligocene as in the Pleistocene,
perhaps also in the Eocene, but it is difcult to draw
rm conclusions about its function as a trigger for
any specic phase of onset, or the nature of any
short-lived glaciation: we have not attempted such
studies here. The possible existence of refugia for
plant species would limit the extent of Oligocene
glacial maxima.
One problem concerning some of the direct
evidence for glacial onset within the E/O interval,
particularly onshore and in the older published
record, and unavoidable at times, is the use for
dating of low-latitude microfossil assemblages that
have an age range extending into both the early
Oligocene and the late Eocene. It is clear that, in or
close to Antarctica, late Eocene and early Oligocene
environments were, or could have been, dramatically different, so that high-latitude faunal assemblages could have been different too. At lower
latitudes, however, such differences across the E/O
interval could have been muted, or absent entirely.
We hesitate to insist on such an explanation for the
common disagreements, particularly in earlier work,
over the age of the base of glacial sections during
this period, but consider that it has played a part.
Direct evidence for an early phase of glacial onset
has been published, mainly (but not entirely) in the
initial occurrence of IRD (striated quartz, minor
gravel, etc.), and in changes in clay mineralogy, in
deep-water sediments offshore. Based on Northern
Hemisphere occurrences, it is difcult to distinguish
between material from a dry, cold but non-glacial
environment, wind-blown onto sea ice, and material
within mountain glaciers that owed long distances
to sea level and beyond, unless the grain size is
too large for wind transport. Given a cooling

2303

environment, either occurrence is likely to be the


preserved indication of the earliest (mountain)
phase of glaciation.
Such indications are seen in the late and middle
Eocene in Prydz Bay, on the southern Kerguelen
Plateau and in the South Pacic. At the same time,
waters around the Antarctic margin (Prydz Bay, the
Ross Sea, Seymour Island and the Scotia Sea),
though cooling, appear too warm in most cases for
even seasonal sea ice to form. Thus, we support
conclusions that an early phase, with mountain
glaciers reaching the sea at intervals and in a few
places, developed through the middle and late
Eocene. We cannot see trends (although the late
Eocene was colder than the middle Eocene) or
specify locations. Although neither numerical models nor proximal evidence show it, the Pacic Ocean
piston cores suggest that an early ice sheet may have
developed on the coastal mountains of West
Antarctica or the southern Antarctic Peninsula, as
well as within East Antarctica, possibly because of
abundant snowfall. Few data points are available
and inter-ocean connections are unlikely.
A second phase, the development of a largevolume ice sheet (as large as, if not as cold as,
todays), in the E/O interval, is better documented.
Evidence of ice at sea level at that time includes
diamictite in the Ross Sea and Prydz Bay, indicating
grounding, and abundant IRD on the Ross Sea
slope, the Kerguelen Plateau (including its northern
part), and the Weddell Sea margin of East
Antarctica. Interbedded palynomorph-bearing sediments at proximal sites may be taken to indicate
warmer, presumably interglacial conditions, with
recolonisation from refugia. Clay mineralogy on the
Kerguelen Plateau provides clear evidence of a
change in Antarctic weathering mode, and biofacies
changes occur widely. This evidence applies only to
East Antarctica: the existence of a deep-water
connection between the Pacic and Atlantic oceans
at this time cannot be guaranteed (although a deepwater IndianPacic connection seems likely, and a
shallow PacicAtlantic connection cannot be ruled
out), so even the more distal evidence cannot
certainly be taken to apply to other parts.
For West Antarctica and the Antarctic Peninsula,
unequivocal direct evidence is sparse, but that which
exists suggests a later time of full glacial onset on
the former than for East Antarctica. Central Ross
Sea drilling suggests an onset of glaciation at sea
level for West Antarctica shortly after 26 Ma, and
onshore exposures on Seymour Island indicates

ARTICLE IN PRESS
2304

P.F. Barker et al. / Deep-Sea Research II 54 (2007) 22932307

grounded ice to sea level in the early Oligocene on


the Antarctic Peninsula. While in the former case,
there is no evidence of subsequent deglaciation, the
preservation of South Shetland Is. and Seymour I.
evidence suggests that glaciation there did not
persist. Also, the long time gap between mountain
and sea-level glaciation in West Antarctica may
reect the comparative difculty of submarine
grounding of most of the present West Antarctic
ice sheet.
Direct evidence for short-lived Eocene glaciations, as large as or larger than todays, is sparse.
Reporters of the indirect, distal proxy evidence cite
changes in clay mineralogy on Maud Rise and
glacial sediments in the Ross Sea in support, but
such evidence is ambiguous and poorly dated. It is
possible also that other direct evidence cited here in
support of Eocene mountain glaciation was in fact
produced by the few fuller but short-lived glaciations inferred from the indirect evidence. We should
consider what is required to test such hypotheses
conclusively.
9. Future directions
It is clear that the main uncertainties concerning
Antarctic glacial onset are two: its causes, and the
reality of the short-lived late and middle Eocene
glaciations hypothesised by such as Tripati et al.
(2005) on the basis of CCD and stable isotope
studies on samples from an Equatorial ODP site,
and from sea-level variations (e.g., Miller et al.,
2005; Pekar et al., 2005). Of more regional
signicance is the age of full glacial onset in West
Antarctica.
The question of cause reduces to the relative
importance of ocean circulation (in particular,
development of the ACC) and variations in atmospheric pCO2. It is difcult to determine past pCO2,
and consideration of the causes of its variation and
their response times is in its infancy. Onset of the
ACC is considered in detail in another paper
(Barker et al., 2007). In essence, if it can be shown
that ACC onset was not coeval with E/O interval
glacial development, a dominant role for ocean
circulation would be difcult to sustain. A wide
range of times of onset are postulated, on the basis
of tectonics or marine geology, and a program of
(IODP) drilling close to what is considered to be a
strong candidate for the nal barrier in a circumpolar deep-water path has been proposed, and
appears to be the next logical step in resolving the

relative importance of ocean circulation and atmospheric greenhouse gases.


Consideration of possible pre-Oligocene, shortlived Antarctic glaciations is delayed by lack of
high-resolution, direct data from within or close to
the continent itself. Grounded ice is erosional,
usually at a level signicantly below wave base, so
preservation of continental records of glaciation is
uncommon. Continuous records capable of yielding
high-resolution information must be sought offshore (in deep water of the Southern Ocean, close to
the continent for an unambiguous signal), on
continental shelves or within interior basins deeper
than the palaeo-ice base. Consideration of access
requires that the overburden should be small.
Assuming that such early glaciations were largevolume (as the indirect measurements suggest), the
most likely locations are perhaps:

1. The proximal Southern Ocean. IRD indicates


glaciation, but not glacial extent (i.e. mountain
glaciers vs. an ice sheet). The key would be
demonstration of coeval short-lived glaciation at
several sites.
2. Around the East Antarctic continental margin:
suitable sediments might be found off Wilkes
Land (Escutia et al., 2005), in Prydz Bay (ODP
Legs 119 and 188: Barron et al., 1991; Cooper
and OBrien, 2004), and along the East Antarctic
margin of the Weddell Sea. Tills or proximal
glacial marine sediments, IRD, would be likely
indicators. The most easily accessible La Meseta
Fm is exposed onshore on Seymour I., Antarctic
Peninsula, and appears to be continuous to the
E/O interval. Ivany et al. (2006) have shown that
the E/O interval full glaciation included
grounded ice at sea level here, but a short-lived
glaciation might not.
3. Interior Basins. The Byrd Subglacial Basin and
Bentley Subglacial Trough are ruled out by thick
present ice cover and uncertain sedimentary
section, but are the major intra-Antarctic submarine basins. Much of the Ross Sea basin
subsided during the Oligocene, and is lled by
thick sediments: basal pre-Oligocene sediments
are most probably inaccessible. Harwood and
Levy (2000) report that the potentially informative source rocks of the Eocene McMurdo
erratics lie beneath the Ross Ice Shelf, in an area
that could be sampled post-2008 by ANDRILL
(Harwood et al., 2005).

ARTICLE IN PRESS
P.F. Barker et al. / Deep-Sea Research II 54 (2007) 22932307

Sampling sites in and around West Antarctica,


and possibly also the Antarctic Peninsula, are
perhaps less likely to provide certain evidence of
early, short-lived glaciations than sites in and
around East Antarctica, although sites around West
Antarctica might provide an age for subsequent full
glacial onset. And, of course, in all cases sampling
would have to be preceded by detailed reconnaissance.
Acknowledgement
We are grateful to the convenors of the JOIUSSAC Southern Oceans Workshop, from which
this contribution derives. Two anonymous reviewers
made very helpful suggestions which have improved
the manuscript.

References
Anderson, J.B., Kurtz, D.D., Domack, E.W., Balshaw, K.M.,
1980. Glacial and glacial marine sediments of the Antarctic
continental shelves. Journal of Geology 88, 399414.
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, pp. 993996.
Barker, P.F., Camerlenghi, A., 1999. An approach to Antarctic
glacial history: the aims of Leg 178. In: Barker, P.F.,
Camerlenghi, A., Acton, G.D., et al., 1999. Proceedings of
ODP, Initial Reports., 178, 144 [CD-ROM]. Available from:
Ocean Drilling Program, College Station, TX 77845-9547, USA.
Barker, P.F., Camerlenghi, A., 2002. Glacial history of the
Antarctic Peninsula from Pacic margin sediments. In:
Barker, P.F., Camerlenghi, A., Acton, G.D. and Ramsay,
A.T.S. (Eds.). Proceedings of ODP, Scientic Results, 178:
College Station TX (Ocean Drilling Program), pp. 140.
Barker, P.F., Barrett, P.J., Camerlenghi, A., Cooper, A.K.,
Davey, F.J., Domack, E.W., Escutia, C., Kristoffersen, Y.,
OBrien, P.E., 1998. Ice sheet history from Antarctic
continental margin sediments: the ANTOSTRAT approach.
Terra Antartica 5, 737760.
Barker, P.F., Barrett, P.J., Cooper, A.K., Huybrechts, K., 1999a.
Antarctic glacial history from numerical models and continental margin sediments. Palaeogeography, Palaeoclimatology, Palaeoecology 150, 247267.
Barker, P.F., Camerlenghi, A., Acton, G.D., et al., 1999b.
Proceedings of ODP, Initial Reports, 178 [CD-ROM].
Available from: Ocean Drilling Program, College Station,
TX 77845-9547, USA.
Barker, P.F., Filippelli, G.M., Florindo, F., Martin, E.E., Scher,
H.D., 2007. Onset and Role of the Antarctic Circumpolar
Current. Deep-Sea Research II, in press, doi:10.1016/j.dsr2.
2007.07.028.
Barrett, P.J., 1974a. Textural characteristics of Cenozoic preglacial and glacial sediments at Site 270, Ross Sea, Antarctica.
Initial Reports of the Deep Sea Drilling Project 28, 757767.

2305

Barrett, P.J., 1974b. Characteristics of pebbles from Cenozoic


marine glacial sediments in the Ross Sea (Sites 270274) and
the southern Indian Ocean (Site 268). Initial Reports of the
Deep Sea Drilling Project 28, 769784.
Barrett, P.J., (Ed.), 1986. Antarctic Cenozoic History from the
MSSTS-1 Drillhole, McMurdo Sound. DSIR Bulletin (New
Zealand), 237.
Barrett, P.J., (Ed.), 1989. Antarctic Cenozoic History from the
CIROS-1 Drillhole, McMurdo Sound. DSIR Bulletin (New
Zealand), 245.
Barron, J., Larsen, B., et al., 1991. Proceedings of ODP, Scientic
Results, vol 119, College Station, TX (Ocean Drilling
Program).
Birkenmajer, K., Soliani, E., Kawashita, K., 1987. Geochronology of Tertiary glaciations on King George Island, West
Antarctica. Bulletin of the Polish Academy of Sciences
Earth Sciences 37, 2748.
Birkenmajer, K., Gazdzicki, A., Krajewski, K.P., Przybycin,
Solecki, A., Tatur, A., Yoon, H.I., 2005. Polish Polar
Research, 26, 312.
Browning, J.V., Miller, K.G., Pak, D.K., 1996. Global implications of lower to middle Eocene sequence boundaries on the
New Jersey coastal plainthe icehouse cometh. Geology 24,
639642.
Camerlenghi, A., Rebesco, M., De Santis, L., Volpi, V., De
Rossi, A., 2002. The Antarctic Peninsula Pacic margin:
modelling exure and decompaction with constraints from
ODP Leg 178 initial results. In: Gamble, J., Skinner, D.,
Henrys, S. (Eds.), Antarctica at the Close of a Millennium.
Royal Society of New Zealand, Wellington.
Cape Roberts Science Team, 2000. Studies from the Cape
Roberts Project, Ross Sea, Antarctica. Initial Report on
CRP-3. Terra Antartica, 7, 1209.
Cape Roberts Science Team, 2001. Studies from the Cape
Roberts Project, Ross Sea, Antarctica. Scientic Results of
CRP-3. Terra Antartica 8, 123620.
Cooper, A.K., OBrien, P.E., 2004. Leg 188 synthesis: transitions
in the glacial history of the Prydz Bay region, East Antarctica,
from ODP drilling. In: Cooper, A.K., OBrien, P.E., Richter,
C. (Eds.), Proceedings of ODP Scientic Results, vol. 188:
College Station TX (Ocean Drilling Program), 142.
Coxall, H.K., Wilson, P.A., Palike, H., Lear, C.H., Backman, J.,
2005. Rapid stepwise onset of Antarctic glaciation and deeper
calcite compensation in the Pacic Ocean. Nature 433, 5357.
De Conto, R., Pollard, D., 2003. Rapid Cenozoic glaciation of
Antarctica induced by declining atmospheric CO2. Nature
421, 245249.
De Santis, L., Prato, S., Brancolini, G., Lovo, M., Torelli, L.,
1999. The eastern Ross Sea continental shelf during the
Cenozoic: implications for the West Antarctic ice sheet
development. In: Van der Wateren, F.M., Cloetingh, S.A.
(Eds.), Lithosphere dynamics and environmental change of
the Cenozoic West Antarctic Rift System. Global and
Planetary Change 23, 173196.
Diekmann, B., Kuhn, G., Gersonde, R., Mackensen, A., 2004.
Middle Eocene to early Miocene environmental changes in
the sub-Antarctic Southern Ocean; evidence from biogenic
and terrigenous depositional patterns at ODP Site 1090.
Global and Planetary Change 40, 295313.
Dingle, R.V., Lavelle, M., 1998. Antarctic Peninsular cryosphere:
early Oligocene (ca. 30 Ma) initiation and revised glacial
chronology. Journal of Geological Society 155, 433437.

ARTICLE IN PRESS
2306

P.F. Barker et al. / Deep-Sea Research II 54 (2007) 22932307

Dingle, R.V., Marenssi, S.A., Lavelle, M., 1998. High latitude


Eocene climate deterioration: evidence from the northern
Antarctic Peninsula. Journal of South American Earth
Sciences 11, 571579.
DiVenere, V.J., Kent, D.V., Dalziel, I.W.D., 1994. Mid-Cretaceous paleomagnetic results from Marie Byrd Land, West
Antarctica: a test of post-100 Ma relative motion between
East and West Antarctica. Journal of Geophysical Research
99, 1511515139.
Drewry, D.J., 1983. Antarctica: Glaciological and Geophysical
Folio, Scott Polar Res. Inst., Cambridge England, 9 sheets.
Dutton, A.L., Lohmann, K.C., Zinsmeister, W.J., 2002. Stable
isotope and minor element proxies for Eocene climate of
Seymour Island, Antarctica. Paleoceanography, 17, doi:10.
1029/2000PA00593.
Ehrmann, W.U., 1998. Implications of late Eocene to early
Miocene clay mineral assemblages in McMurdo Sound
(Ross Sea, Antarctica) on paleoclimate and ice dynamics.
Palaeogeography, Palaeoclimatology, Palaeoecology 139,
213231.
Ehrmann, W.U., Mackensen, A., 1992. Sedimentological evidence for the formation of an East Antarctic ice sheet in
Eocene/Oligocene time. Palaeogeography Palaeoclimatology
Palaeoecology 93, 85112.
Escutia, C., Eittreim, S.L., Cooper, A.K., 1997. Cenozoic
sedimentation on the Wilkes Land continental rise, Antarctica. In: Ricci, C.A. (Ed.), The Antarctic Region: Geological
Evolution and Processes, pp. 791795.
Escutia, C., De Santis, L., Donda, F., Dunbar, R.B., Brancolini,
G., Eittreim, S.L., Cooper, A.K., 2005. Cenozoic ice sheet
history from east Antarctic Wilkes Land continental margin
sediments. Global and Planetary Change 45 (13), 5181.
Foldvik, A., Gammelsrd, T., 1988. Notes on Southern Ocean
hydrography, sea-ice and bottom-water formation. Palaeogeography Palaeoclimatology Palaeoecology 67, 317.
Foster, T.D., Middleton, J.H., 1980. Bottom water formation in
the western Weddell Sea. Deep-Sea Research 27A, 367381.
Frey, F.A., Cofn, M.F., Wallace, P.J., Quilty, P.G. (Eds.), 2003.
Proceedings of ODP Scientic Results, vol. 183 [CD-ROM].
Available from: Ocean Drilling Program, Texas A&M
University, College Station TX 77845-9547, USA.
Gersonde, R., Kyte, F.T., Bleil, U., Diekmann, B., Flores, J.A.,
Gohl, K., Grahl, G., Hagen, R., Kuhn, G., Sierro, F.J.,
Volker, D., Abelmann, A., Bostwick, J.A., 1997. Geological
record and reconstruction of the late Pliocene impact of the
Eltanin asteroid in the Southern Ocean. Nature 390, 357363.
Grobe, H., Futterer, D.K., Spiess, V., 1990. Oigocene to
Quaternary sedimentation processes on the Antarctic continental margin, ODP Leg 113, Site 693. In: Barker, P.F.,
Kennett, J.P., et al., Proceedings of ODP, Scientic Results,
113: College Station, TX (Ocean Drilling Program),
pp. 121131.
Hara, U., 1997. Bryozoan assemblages from the La Meseta
Formation (Eocene) of Seymour Island, Antarctic Peninsula.
In: Ricci, C.A. (Ed.), The Antarctic Region: Geological
Evolution and Processes, pp. 10011006.
Harwood, D.M., Levy, R.H., 2000. The McMurdo erratics:
introduction and overview. In: Stilwell, J.D., Feldmann, R.M.
(Eds.), Paleobiology and Paleoenvironments of Eocene
Rocks, McMurdo Sound, East Antarctica. Ant. Res. Ser.,
vol. 76, American Geophysical Union, Washington DC,
pp. 118.

Harwood, D., Florindo, F., Fielding, C., Levy, R., Pekar, S.F.,
2005. Southern McMurdo Sound Drilling Prospectus. ANDRILL Program.
Hayes, D.E., Frakes, L.A., et al., 1975. Initial Reports of the
Deep Sea Drilling Project, 28, 1017.
Hill, R.S., Dettmann, M.E., 1996. Origin and diversication of
the genus Nothofagus. In: Veblen, T.T., Hill, R.S., Read, J.
(Eds.), The Ecology and Biogeography of Nothofagus
Forests. Yale University Press, New Haven, CT, pp. 1124.
Hollister, C.D., Craddock, C., et al. (1976). Initial Reports of the
Deep Sea Drilling Project, Vol. 35. US Government Printing
Ofce, Washington, 930pp.
Huybrechts, P., 1993. Glaciological modelling of the Late
Cenozoic East Antarctic ice sheet: stability or dynamism.
Geograska Annaler 75A, 221238.
Ivany, L.C., Van Simaeys, S., Domack, E.W., Samson, S.D.,
2006. Evidence for an earliest Oligocene ice sheet on the
Antarctic Peninsula. Geology 34, 377380.
Kennett, J.P., 1977. Cenozoic evolution of Antarctic glaciation,
the Circum-Antarctic ocean, and their impact on global
paleoceanography. Journal of Geophysical Research 82,
38433860.
Kennett, J.P., Stott, L.D., 1990. Proteus and proto-Oceanus:
ancestral Paleogene oceans as revealed from Antarctic stable
isotopic results; ODP Leg 113. In: Barker, P.F., Kennett, J.P.,
et al., Proceedings of ODP, Scientic Research, vol. 113:
College Station, TX (Ocean Drilling Program), pp. 865880.
Lawver, L.A., Gahagan, L.M., 2003. Evolution of Cenozoic
seaways in the circum-Antarctic region. Palaeogeography
Palaeoclimatology Palaeoecology 198, 1138.
Lear, C.H., Eldereld, H., Wilson, P.A., 2000. Cenozoic deep-sea
temperatures and global ice volumes from Mg/Ca in benthic
foraminiferal calcite. Science 287, 269272.
Lear, C.H., Rosenthal, Y., Coxall, H.K., Wilson, P.A., 2004.
Late Eocene to early Miocene ice sheet dynamics and the
global carbon cycle. Paleoceanography 19; PA4015,
doi:10.1029/2004PA001039.
Le Masurier, W.E., Rocchi, S., 2005. Terrestrial record of postEocene climate history in Marie Byrd Land, West Antarctica.
Geograsker Annalen 87A (1), 5166.
Lythe, M.B., Vaughan, D.V., BEDMAP Consortium, 2001.
BEDMAP: A new ice thickness and subglacial topographic
model of Antarctica. Journal of Geophysical Research 106,
1133511351.
Margolis, S.V., Kennett, J.P., 1971. Cenozoic paleoglacial history
of Antarctica recorded in subantarctic deep-sea cores.
American Journal of Scientic 271, 136.
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., Pekar, S.F., 2005. The Phanerozoic record
of global sea-level change. Science 310, 12931298.
Mix, A.C., Ruddiman, W.F., 1984. Oxygen-isotope analyses and
Pleistocene ice volumes. Quaternary Research 21, 120.
Mohr, B.A.R., 1990. Eocene and Oligocene sporomorphs and
dinoagellate cysts from Leg 113 drill sites, Weddell Sea,
Antarctica. In: Barker, P.F., Kennett, J.P., et al., Proceedings
of ODP, Scientic Results, vol. 113: College Station, TX
(Ocean Drilling Program), pp. 595612.
Mohr, B.A.R., 2001. The development of Antarctic fern oras
during the Tertiary, and palaeoclimatic and palaeobiogeographic implications. Sonder-Abdruck aus Palaeontographica
B 259, 167208.

ARTICLE IN PRESS
P.F. Barker et al. / Deep-Sea Research II 54 (2007) 22932307
OBrien, P.E., Cooper, A.K., Richter, C., et al., 2001. Proc ODP
Init. Repts., 188 [CD-ROM]. Available from: Ocean Drilling
Program, Texas A&M University, College Station, TX 778459547, USA.
Paterson, W.S.B., 1994. The Physics of Glaciers, third ed.
Pergamon Press, Oxford.
Paul, H.A., Zachos, J.C., Flower, B.P., Tripati, A., 2000. Orbitally
induced climate and geochemical variability across the
Oligocene/Miocene boundary. Paleoceanography 15, 471485.
Pekar, S.F., Christie-Blick, N., Kominz, M.A., Miller, K.G.,
2002. Calibration between eustatic estimates from backstripping and oxygen isotopic records for the Oligocene.
Geology 30, 903906.
Pekar, S.F., Hucks, A., Fuller, M., Li, S., 2005. Glacioeustatic
changes in the early and middle Eocene (51-42 Ma) greenhouse world based on shallow-water stratigraphy from ODP
Leg 189 Site 1171 and oxygen isotope records. Geological
Society of America Bulletin 117, 10811093.
Raine, J.I., Askin, R.A., 2001. Terrestrial palynology of Cape
Roberts Project Drillhole CRP-3, Victoria Land Basin,
Antarctica. In: Cape Roberts Science Team. Studies from
the Cape Roberts Project, Ross Sea, Antarctica. Scientic
Report of CRP-3, Terra Antartica, 8, 389400.
Raymo, M.E., Ruddiman, W.F., 1992. Tectonic forcing of late
Cenozoic climate. Nature 359, 117122.
Rebesco, M., Larter, R.D., Barker, P.F., Camerlenghi, A.,
Vanneste, L.E., 1997. The history of sedimentation on the
continental rise west of the Antarctic Peninsula. In: Barker,
P.F., Cooper, A.K. (Eds.), Geology and Seismic Stratigraphy
of the Antarctic Margin, Part 2, AGU Antarctic Research
Series vol. 71, pp. 2949.
Rebesco, M., Pudsey, C.J., Canals, M., Camerlenghi, A., Barker,
P.F., Estrada, F., Giorgetti, A., 2002. Case Study 27:
sediment drifts and deep-sea channel systems, Antarctic
Peninsula Pacic margin, mid-Miocene to Present. In: Stow,
D.A.V., Pudsey, C.J., Howe, J., Fauge`res, J.-C., Viana, A.
(Eds.), Deep-water Contourite Systems: Modern Drifts and
Ancient Series, Seismic and Sedimentary Characteristics,
Geological Society, London.
Reguero, M.A., Marenssi, S.A., Santillana, S.N., 2002. Antarctic
Peninsula and South America (Patagonia) Paleogene terrestrial faunas and environments: biogeographic relationships.
Palaeogeography Palaeoclimatology Palaeoecology 179,
189210.
Robert, C., Chamley, H., 1992. Late Eoceneearly Oligocene
evolution of climate and marine circulation: deep-sea clay
mineral evidence. In: Kennett, J.P., Warnke, D.A. (Eds.), The
Antarctic Paleoenvironment: A Perspective on Global
Change, Part One. Antarctic Research Series. American
Geophysical Union, Washington DC, 97117.
Robert, C., Kennett, J.P., 1992. Paleocene and Eocene kaolinite
distribution in the South Atlantic and Southern Ocean:
Antarctic climatic and paleoceanographic implications. Marine Geology 103, 99110.
Rocchi, S., LeMasurier, W.E., Di Vincenzo, G., 2006. Oligocene
to Holocene erosion and glacial history in Marie Byrd Land,

2307

West Antarctica, inferred from exhumation of the Dorrel


Rock intrusive complex and from volcano morphologies.
Bulletin of Geological Society of the America. 118, 9911005.
Shipboard Scientic Party, 2005. Arctic Coring Expedition
(ACEX). Paleoceanographic and tectonic evolution of the
central Arctic Ocean. IODP Prel. Rep. Vol. 302, /http://
www.ecord.org/exp/acex/302PR.pdfS.
Stillwell, J.D., Zinsmeister, W.J., 1992. Molluscan systematics
and biostratigraphy: Lower Tertiary La Meseta Formation,
Seymour Island, Antarctic Peninsula. Antarctic Research
Series. 55, 192.
Stroeven, A.P., 1997. The Sirius Group of Antarctica: age and
environments. In: Ricci, C.A. (Ed.), The Antarctic Region:
Geological Evolution and Processes, pp. 747761.
Sugden, D.E., Marchant, D.R., Denton, G.H., 1993. The Case
for a Stable East Antarctic Ice Sheet. Geograska Annaler
75A (4), 151351.
Toggweiler, J.R., Bjornsson, H., 2000. Drake Passage and
paleoclimate. Journal of Quaternary Science 15, 319328.
Toker, V., Barker, P.F., Wise, S.W., 1991. Middle Eocene
carbonate-bearing sediment off the northern Antarctic
Peninsula. In: Thomson, M.R.A., Crame, J.A., Thomson,
J.W. (Eds.), Geological Evolution of Antarctica. Cambridge
University Press, Cambridge, pp. 639644.
Tripati, A., Backman, J., Eldereld, H., Ferretti, P., 2005. Eocene
bipolar glaciation associated with global carbon cycle
changes. Nature 436, 341346.
Troedson, A.L., Riding, J.B., 2002. Late Oligocene to earliest
Miocene strata of King George Island, South Shetland
Islands, Antarctica: stratigraphy, facies analysis and implications for Antarctic Peninsula glacial history. Journal od
Sedimantary Research 72, 510523.
Troedson, A.L., Smellie, J.L., 2002. The Polonez Cove Formation of King George Island, West Antarctica; stratigraphy,
facies and palaeoenvironmental implications. Sedimentology
49, 277301.
Wade, B.S., Palike, H., 2004. Oligocene climate dynamics.
Paleoceanography, 19, PA4019, doi:10.1029/2004PA001042.
Webb, P.-N., Harwood, D.M., 1991. Late Cenozoic glacial
history of the Ross Embayment, Antarctica. Quaternary
Science Review 10, 215224.
Wei, W., 1992. Calcareous nannofossil stratigraphy and reassessment of the Eocene glacial record in Subantarctic piston cores
of the southeast Pacic. In: Wise Jr., S.W., Schlich, R., et al.,
Proceedings of ODP, Science Results, vol. 120: College
Station, TX (Ocean Drilling Program), pp. 10931104.
Wise Jr., S.W., Schlich, R., et al., 1992. Proceedings of ODP,
Science Results, Part 2, vol. 120: College Station, TX (Ocean
Drilling Program), pp. 4511155.
Zachos, J.C., Quinn, T.M., Salamy, K.A., 1996. High-resolution
(104 years) deep-sea foraminiferal stable isotope records of the
Eocene-Oligocene climate transition. Paleoceanography 11,
251266.
Zachos, J.C., Pagani, M., Sloan, L., Thomas, E., Billups, K.,
2001. Trends, rhythms and aberrations in global climate,
65 Ma to present. Science 292, 686693.

You might also like