Download full-text PDF

Antarctic ice sheet discharge driven by atmosphere-ocean feedbacks at the Last Glacial Termination

Article · January 2017with248 Reads
DOI: 10.1038/srep39979
Abstract
Reconstructing the dynamic response of the Antarctic ice sheets to warming during the Last Glacial Termination (LGT; 18,000–11,650 yrs ago) allows us to disentangle ice-climate feedbacks that are key to improving future projections. Whilst the sequence of events during this period is reasonably well-known, relatively poor chronological control has precluded precise alignment of ice, atmospheric and marine records, making it difficult to assess relationships between Antarctic ice-sheet (AIS) dynamics, climate change and sea level. Here we present results from a highly-resolved ‘horizontal ice core’ from the Weddell Sea Embayment, which records millennial-scale AIS dynamics across this extensive region. Counterintuitively, we find AIS mass-loss across the full duration of the Antarctic Cold Reversal (ACR; 14,600–12,700 yrs ago), with stabilisation during the subsequent millennia of atmospheric warming. Earth-system and ice-sheet modelling suggests these contrasting trends were likely Antarctic-wide, sustained by feedbacks amplified by the delivery of Circumpolar Deep Water onto the continental shelf. Given the anti-phase relationship between inter-hemispheric climate trends across the LGT our findings demonstrate that Southern Ocean-AIS feedbacks were controlled by global atmospheric teleconnections. With increasing stratification of the Southern Ocean and intensification of mid-latitude westerly winds today, such teleconnections could amplify AIS mass loss and accelerate global sea-level rise.
1
Scientific RepoRts | 7:39979 | DOI: 10.1038/srep39979
www.nature.com/scientificreports
Antarctic ice sheet discharge driven
by atmosphere-ocean feedbacks at
the Last Glacial Termination
C.J. Fogwill1,2, C.S.M. Turney1,2, N.R. Golledge3,4, D. M. Etheridge5, M. Rubino5,6,
D.P. Thornton5, A. Baker1, J. Woodward7, K. Winter7, T.D. van Ommen8,9, A.D. Moy8,9,
M.A.J. Curran8,9, S.M. Davies10, M.E. Weber11,12, M.I. Bird13, N.C. Munksgaard13,14,
L. Menviel1,2, C.M. Rootes15, B. Ellis16, H. Millman2, J. Vohra1,2, A. Rivera17,18 & A. Cooper19
Reconstructing the dynamic response of the Antarctic ice sheets to warming during the Last Glacial
Termination (LGT; 18,000–11,650 yrs ago) allows us to disentangle ice-climate feedbacks that are key
to improving future projections. Whilst the sequence of events during this period is reasonably well-
known, relatively poor chronological control has precluded precise alignment of ice, atmospheric and
marine records, making it dicult to assess relationships between Antarctic ice-sheet (AIS) dynamics,
climate change and sea level. Here we present results from a highly-resolved ‘horizontal ice core’ from
the Weddell Sea Embayment, which records millennial-scale AIS dynamics across this extensive region.
Counterintuitively, we nd AIS mass-loss across the full duration of the Antarctic Cold Reversal (ACR;
14,600–12,700 yrs ago), with stabilisation during the subsequent millennia of atmospheric warming.
Earth-system and ice-sheet modelling suggests these contrasting trends were likely Antarctic-wide,
sustained by feedbacks amplied by the delivery of Circumpolar Deep Water onto the continental
shelf. Given the anti-phase relationship between inter-hemispheric climate trends across the LGT our
ndings demonstrate that Southern Ocean-AIS feedbacks were controlled by global atmospheric
teleconnections. With increasing stratication of the Southern Ocean and intensication of mid-latitude
westerly winds today, such teleconnections could amplify AIS mass loss and accelerate global sea-level
rise.
Understanding centennial to millennial-scale variability of the Earths ice sheets is key to gaining insights into
ice sheet-climate feedbacks1,2, and quantifying their contribution to past and future environmental change3,4.
is is important, as despite mounting evidence of signicant changes in AIS dynamics5, Southern Ocean6, and
atmospheric circulation7, current projections of global mean sea level (GMSL) imply only moderate increases by
1PANGEA Research Centre, University of New South Wales, 2052, Australia. 2Climate Change Research Centre,
School of Biological Earth and Environmental Sciences, University of New South Wales, 2052, Australia. 3Antarctic
Research Centre, Victoria University of Wellington, Wellington 6140, New Zealand. 4GNS Science, Avalon, Lower
Hutt, New Zealand. 5CSIRO Climate Science Centre, Oceans and Atmosphere, Aspendale, Victoria, 3195 Australia.
6Dipartimento di Matematica e Fisica, Università della Campania - Luigi Vanvitelli, viale Lincoln, 5-81100 Caserta,
Italy. 7Department of Geography, Faculty of Engineering and Environment, Northumbria University, Newcastle
upon Tyne, NE1 8ST, United Kingdom. 8Australian Antarctic Division, 203 Channel Highway, Kingston, Tasmania
7050, Australia. 9Antarctic Climate & Ecosystems Cooperative Research Centre, University of Tasmania, Private
Bag 80, Hobart, Tasmania 7001, Australia. 10Department of Geography, College of Science, Swansea University,
Swansea, United Kingdom. 11Department of Earth Sciences, University of Cambridge, Drummond Street,
Cambridge, United Kingdom. 12Steinmann Institute, University of Bonn, Poppelsdorfer Schloss, Bonn, Germany.
13Centre for Tropical Environmental and Sustainability Science, College of Science and Engineering, James Cook
University, Cairns, Australia. 14Research Institute for the Environment and Livelihoods, Charles Darwin University,
Australia. 15Department of Geography, University of Sheffield, United Kingdom. 16Research School of Earth
Sciences, Australian National University, Canberra, Australia. 17Glaciology and Climate Change Laboratory, Centro
de Estudios Cientcos, Valdivia, Arturo Prat 514, Chile. 18Department of Geography, University of Chile, Santiago,
Chile. 19Australian Centre for Ancient DNA, University of Adelaide, 5005, Australia. Correspondence and requests for
materials should be addressed to C.J.F. (email: c.fogwill@unsw.edu.au)
Received: 19 October 2016
Accepted: 29 November 2016
Published: 05 January 2017
OPEN
www.nature.com/scientificreports/
2
Scientific RepoRts | 7:39979 | DOI: 10.1038/srep39979
the end of the twenty-rst century3. ese projections, however, do not fully include ice-sheet-ocean dynamic
feedbacks which are believed to have triggered rapid continental ice-sheet retreat and driven periods of abrupt
sea-level rise during the geological past2,8. e LGT oers a potential process analogue for future climate trends,
characterised by multi-millennial global warming, poleward migrating and strengthening westerly winds9, and
increasing atmospheric carbon dioxide levels10, similar in magnitude to future projections4.
During the LGT, long-term warming was interrupted by the ~2,000-year duration cold event across the mid
to high latitude Southern Hemisphere, known as the Antarctic Cold Reversal11,12, which was associated with a
~35 m GMSL rise. Within the ACR, Meltwater Pulse 1A (MWP-1A) forms a prominent abrupt rise in sea level
of ~16 m (14,700–14,300 years or 14.7–14.3 ka) that has been a major focus of previous studies, and which was
coincident with a period of enhanced iceberg ux in the Southern Ocean2. However, the actual contribution
of the AIS during this period remains unclear13,14 due to the paucity of geological records capable of resolving
ice-sheet volume changes in such a dynamic contemporary ice sheet setting15, and the diculties in precisely
aligning the chronologies of marine and terrestrial sequences2,11. While the contribution of AIS to GMSL rise
during MWP-1A range from ‘high-end’ scenarios (> 10 m contributing over half of the total GMSL rise), to
‘low-end’ (scenarios with little to no contribution), the AIS input (if any) during the ACR and the subsequent
period of sustained Southern Hemisphere warming remains debated14,16,17. Crucially, the role of the AIS in global
climate-ocean dynamics during the LGT remain uncertain11,18. An improved understanding of the links between
AIS stability and ice-ocean-climate feedbacks throughout the LGT (i.e. not just MWP-1A), and its relationship to
Northern Hemisphere changes, is therefore critical for improving projections of sea-level rise3,4 and understand-
ing ice-sheet-climate feedbacks2,17 in detail.
Here we take a novel approach that investigates a new 800 m long ‘horizontal ice core’ that captures a unique
record of ice-sheet dynamics and climate across the Weddell Sea Embayment (WSE)19, a region which today
drains more than one-h of the ice-mass of continental Antarctica, including sectors of the East and West
Antarctic ice sheets and the Antarctic Peninsula (Fig.1A).
Results
We report results from an exposed ancient blue ice area (BIA) outcropping alongside Patriot Hills in Horseshoe
Valley (Fig.1A), a locally sourced compound glacier that is buttressed by, but ultimately coalesces with the
Institute Ice Stream close to the contemporary grounding line of the AIS20, making the site highly sensitive to ele-
vation changes across the broader WSE region21. Geochemically identied volcanic (tephra) horizons along with
multiple trace gas species (CO2, CH4 and N2O) provide key age tie points across the prole (Fig.2; Methods and
Supplementary Information), and demonstrate that the prole spans from ~2.5 to 50 ka, with two unconformities
(discontinuities D1 and D2; Fig.1B) that mark the build up to, and deglaciation from, the Last Glacial Maximum
(see Methods)20. e conformable BIA layers or ‘isochrons’ between these two unconformities span ~11 to ~23
ka, capturing a unique highly-resolved record of ice-sheet dynamics across the LGT in an area of exceptionally
slow-moving ice20 (Fig.1B).
e water stable isotope deuterium (δ D) from the Patriot Hills BIA identies a two-stepped change in val-
ues during the LGT, with a ~39‰ increase recorded across the ACR between ~14.7–12.7 ka, followed by a
millennial-duration isotopic plateau (~12.7–11.7 ka) (Fig.3G). Using a regionally applied δ D–temperature rela-
tionship of 6.4 ± 1.3‰ per °C22, the Patriot Hills record implies an increase in annual temperature of ~6 °C across
the ACR with little to no subsequent change up to ~11.7 ka; comparable trends are also recorded in the δ18O pro-
le (SI FigureS6). e deuterium excess values demonstrate there is no signicant regime shi across the prole
during the LGT, suggesting no change in precipitation source, or the sign of the isotope-temperature relationship
(Fig.2). ese apparent temperature changes are in marked contrast to regional climate records across continen-
tal Antarctica23 and the Antarctic Peninsula22 that show a clear plateau/reversal in the warming trend during the
ACR (Fig.3). Given the isolated nature of Horseshoe Valley both during contemporary times and at the LGT19,20,
and the buttressing eect of the AIS on ice ow from the valley (Fig.1), we interpret the isotopic trend captured
in the Patriot Hills BIA as the result of ice-sheet elevation changes due to mass loss across the broader WSE21
(see Methods). us, increasing δ D (and δ
18O; Supplementary Information) water isotope values across the ACR
and the apparent local warming can only reect regional ice-sheet draw down. e isotopic prole that followed
(~12.7–11.7 ka) appears to reect a stabilisation of ice-sheet elevation for approximately a millennium.
e marked change in δ D captured in the record across the ACR implies an ice-sheet surface elevation decrease
of ~615 m across Horseshoe Valley (at a rate of ~0.4 m/a), assuming an atmospheric lapse rate of 10 °C/km24.
is rate of change across the WSE is similar to that inferred from terrestrial cosmogenic isotope studies of
mid-Holocene glacier thinning in the WSE25, and lower than recorded in the Amundsen Sea sector of the West
Antarctic in recent decades5. Importantly, the projected elevation changes represent absolute minima. If the
eects of regional ACR cooling and potential glacial isostatic rebound are included this value would exceed
~800 m. e period of mass loss we identify in the WSE parallels a period of enhanced iceberg-raed debris
ux as recorded in marine sediments from the Scotia Sea (Fig.3D)2, strongly indicating enhanced AIS iceberg
discharge was sustained across the ACR (that included but was not limited to MWP-1A). Importantly, during
the subsequent period, modelled ice-sheet outputs (see Methods) suggest a marked reduction in AIS drawdown
between 12.7 and 11.7 ka (Fig.3E and Supplementary Information), consistent within the uncertainties of the
Patriot Hills BIA chronology (Supplementary Information) of invariant stable isotope values following the ACR
(Fig.3G).
Discussion
e regional climate and sea-level ngerprint of the ACR across the mid- to high-latitude Southern Hemisphere
is dicult to reconcile. Recent modelling studies have demonstrated that it is possible to reconstruct the spatial
pattern without substantial fresh water forcing in the Southern Ocean11,14. However, our modelling, together
www.nature.com/scientificreports/
3
Scientific RepoRts | 7:39979 | DOI: 10.1038/srep39979
with the results of previous studies2,13,18, suggests that a signicant fresh water input into the Southern Ocean
provides a potential trigger for the ACR signal, a hypothesis supported by our eld data. Our inferred decou-
pling of ice-sheet elevation from air temperature across the LGT implies ocean forcing was a primary driver of
Antarctic-wide ice-sheet dynamics through this period. Independent ice-sheet modelling experiments, driven
by transient Earth System Model (LOVECLIM) outputs that include fresh water hosing in the Ross and Weddell
seas17 (see Supplementary Information), predict similar changes in ice-sheet geometry and ice-ow dynamics
Figure 1. (A) Location map of Weddell Sea Embayment (WSE)36 and the major ice streams, with the location
of Patriot Hills in the Ellsworth Mountains. Lower le. Inset map of Antarctica, with locations of the Patriot
Hills (PH), WAIS Divide (WDC), Byrd, James Ross Island (JRI) and EPICA Dome C (EDC) ice cores, and the
East Antarctic Ice Sheet (EAIS) and West Antarctic Ice Sheet (WAIS). (B) (i). Moderate Resolution Imaging
Spectroradiometer (MODIS) mosaic36 showing inferred ice ow path from the head of Horseshoe Valley
to Patriot Hills, where discontinuities D1 and D2 formed as a result of Blue Ice Area wind scour in front of
Liberty and Marble Hills respectively20, (ii) schematic stratigraphic succession, indicating ice accumulation
punctuated by two periods of erosion (D1 and D2; red lines). e uppermost panel of ii represents the observed
stratigraphic sequence at the Patriot Hills BIA as seen in (iii), the full GPR stratigraphic sequence at Patriot Hills
BIA, where red lines indicate erosional events D1 and D2. Adapted from Winter et al.20.
www.nature.com/scientificreports/
4
Scientific RepoRts | 7:39979 | DOI: 10.1038/srep39979
Figure 2. (A) Full δ D isotopic prole from the Patriot Hills BIA with chronological age ties (red triangles)
highlighted. e red bars highlight the location of volcanic horizons at 17.8 ka, 18.2 ka and 36.4 ka, as recorded
in other Antarctic ice core records (Supplementary Information). e grey bars highlight the area of the prole
between the unconformities at 247 m (D1) and 360 m (D2) between which the GPR survey demonstrates clear
dipping reectors or isochrons across the prole (see Fig.1B). (B) δ D-excess across prole; dashed horizontal
lines denote potential regime shis across the prole at 99% condence37. (C) CH4 concentrations from ice
extracted from the Patriot Hills prole (lled red circles) plotted against EPICA Dome C (EDC)34,35 (open white
circles), with the approximate timings of the unconformities outlined by the hatched areas and 1σ uncertainty.
(D) Age-depth model based upon chronological control ties between ~2.5 ka (2,540 years) and ~52 ka (52,170
years) from volcanic ‘tephra’ horizons and most-likely age as derived from multiple trace gas comparison to
published records (CH4, CO2, N2O; see Methods and Supplementary Information). Note: the timing of the onset
of ice accumulation aer D2 in Patriot Hills is a conservative estimate and with future trace gas dating may be
younger than that presented here.
www.nature.com/scientificreports/
5
Scientific RepoRts | 7:39979 | DOI: 10.1038/srep39979
Figure 3. Inter-comparison of deglacial elevation changes from Patriot Hills BIA with modelled and
empirical records. (A) δ
18O from NGRIP (GICC05 chronology)38. (B) Cariaco Basin grey scale, a measure of
latitudinal changes in the trade winds associated with the ITCZ12. (C) Southern Ocean opal ux9. (D) Iceberg-
raed debris ux (IRD; 100-year average) relative to Holocene average from the Scotia Sea2. (E) Modelled
sector-wide AIS mass loss17. (F) Byrd δ
18O (blue) (synchronised to GISP2 chronology) isotopic record39 and
WAIS Divide Core δ
18O (WD2014 chronology) (black)23 correlated with the volcanic horizons at 18 ka and
18.2 ka. (G) δ D isotope prole (black dashed line), with 2-point moving average (green solid line). e red
arrow highlights the apparent 600 m ice-sheet surface elevation change across the WSE estimated from the δ
D isotopic changes recorded during the ACR from the Patriot Hills BIA. Vertical boxes identify the periods
dened by the Antarctic Cold Reversal (ACR) (blue) and the Younger Dryas chronozone (YD) (green). e
black triangles represent the age tie points (derived from geochemically identied volcanic horizons and trace
gases) in this section of the Patriot Hills BIA.
www.nature.com/scientificreports/
6
Scientific RepoRts | 7:39979 | DOI: 10.1038/srep39979
(Figs3 and 4 and Supplementary Information). e modelled increase in freshwater ux strongly suggests the
drawdown of the AIS during the ACR was sustained by a positive feedback operating within the Southern Ocean.
Crucially, we nd a freshening of surface waters leads to a weakening of Southern Ocean overturning, resulting
in reduced Antarctic Bottom Water (AABW) formation, enhanced stratication and sea-ice expansion17 (Fig.4).
e increased delivery of relatively warm Circumpolar Deep Water (CDW)26 onto the continental shelf close
to the grounding line of the AIS thermally erodes marine-based ice, maintaining a positive ice-ocean feedback
(Fig.4)17. High resolution ice sheet modelling suggests that this mechanism predicts increases in ice mass loss
across the AIS during the ACR in excess of 800 Gt/a, with an average of ~400 Gt/a, making a GMSL contribu-
tion of ~0.3 to 0.1 m per century17, importantly this rate almost doubles during the period dened by MWP-1A
(Fig.3E). Modelling of the following millennium implies a marked reduction in mass loss from all sectors of the
Figure 4. LOVECLIM transient model simulations of Southern Ocean fresh water forcing showing
temperature anomalies (°C) for the ACR (14 ka minus 15 ka; le-hand panels)17,18, and subsequent surface
warming during the Younger Dryas chronozone (12 ka minus 14 ka right-hand panels), with sea surface
temperatures and 0.1 m sea ice contour (A,B), ocean temperature anomalies at depth (C,D, averaged over
484–694 m), and ocean temperature anomalies across the Weddell Sea (60°W to 15°W) (E,F) (constructed using
ferret http://ferret.pmel.noaa.gov/Ferret/).
www.nature.com/scientificreports/
7
Scientific RepoRts | 7:39979 | DOI: 10.1038/srep39979
AIS including the Weddell Sea (Fig.3E), reecting reduced Southern Ocean stratication and resumption of
AABW formation post ACR, in agreement with our observations from the Patriot Hills BIA.
e coincidence between changes in AIS elevation from the Patriot Hills, enhanced iceberg ux2, atmospheric
temperature trends22,23, and Southern Hemisphere mid-latitude westerly airow9 through the LGT (Fig.3B,C)
implies a tight coupling between the ice-ocean-atmosphere system. Recent work using absolutely-dated tree
ring chronologies has identied an abrupt increase in the inter-hemispheric radiocarbon gradient as a result
of increased upwelling of 14CO2–depleted abyssal waters from 12.7 ka12, coincident with the maximum south-
erly extent of the Intertropical Convergence Zone (ITCZ) and strengthening Southern Hemisphere Westerlies
(SHW)9. Our results are consistent with these ndings, suggesting that weaker SHW during the ACR enhanced
Southern Ocean stratication and maintained a positive ice-sheet-ocean feedback that drove substantial draw-
down of the AIS (Fig.4). is positive feedback appears to have only been disrupted by the re-expansion of the
tropical belt and Hadley circulation during subsequent Northern Hemisphere cooling, and anti-phase southern
warming aer 12.7 ka (Fig.3), suggesting AIS dynamics are highly sensitive to global atmospheric circulation.
e Patriot Hills preserves a record of signicant AIS ice-sheet drawdown, mass loss and meltwater discharge
during the ACR and across the LGT, contrasting markedly with previous interpretations of the conguration
in the Weddell Sea sector of the AIS, which predict limited ice sheet drawdown since the local Last Glacial
Maximum (LGM)15. Previous terrestrial reconstructions, based upon cosmogenic isotope analysis, predict a
maximum thinning of ~480 m since the LGM, that occurred predominately during the mid-Holocene, suggest-
ing that the WSE only made a minor contribution to GMSL rise since the LGM25. ese estimates contrast with
model-based reconstructions from far-eld sites16, recent ice-sheet modelling studies8, reconstructions of IRD in
the Scotia Sea2, and, crucially, our estimate of ~600 m of ice sheet surface elevation change across the ACR and
MWP-1A (Fig.3). Whilst we cannot dene an upper altitudinal limit of the pre-ACR ice sheet across the WSE,
the results reported here are inconsistent with estimates based upon terrestrial cosmogenic reconstructions25.
We suggest these contrasts may reect two factors: rstly, there is a question over the eectiveness of terres-
trial cosmogenic isotope studies to truly reect the former elevation of the LGM ice-sheet surface in areas of cold
based non-erosive polythermalice sheet settings27,28. Secondly, it is possible under a scenario of dynamic deglaci-
ation during the LGT that rapid regional bedrock isostatic variations may have eectively masked rapid ice-sheet
elevation changes that have occurred during deglaciation and subsequently during the Holocene (Supplementary
Information). erefore, terrestrial cosmogenic isotope reconstructions from mountains and nunataks across the
WSE are only likely to robustly reconstruct ice-sheet surface elevation changes during Holocene deglaciation25,29,30.
is is an issue which requires future detailed analysis, with multiple lines of evidence pointing towards a dynamic
history of ice-sheet change across the WSE during the Holocene29–31, with signicant implications for dening the
pre-Holocene history of this sector of the AIS. Innovative reconstructions, such as that provided by the Patriot Hills
BIA, are urgently required to dene in detail dynamic Antarctic ice-climate feedbacks and better constrain the ice
sheets contribution to global sea level rise during periods of rapid climate transition such as the LGT.
Supported by marine geological evidence of enhanced iceberg calving2, and independent ice-sheet and
Earth system modelling experiments17, the Patriot Hills BIA provides the rst direct terrestrial evidence that
the Antarctic ice sheet was highly responsive to global ice-ocean-atmosphere feedbacks during the LGT2,17.
Modelling suggests this pattern could be Antarctic wide, sustained by ice-ocean feedbacks amplied by the deliv-
ery of CDW onto the Antarctic Continental Shelf. e counterintuitive nding of sustained ice-sheet mass loss
across this sector of the AIS during a period of atmospheric cooling suggests that Southern Ocean AIS feedbacks
were likely modulated by global atmospheric teleconnections during a period of asynchronous hemispheric cli-
mate change. Dening the details of these atmosphere-ocean-ice feedbacks is crucial to reducing uncertainty
in sea level projections4,32, and understanding the implications of observed high-latitude Southern Hemisphere
environmental changes today6,7, which may conspire to amplify future Antarctic ice mass loss.
Methods
Description of the Patriot Hills BIA. e characteristics of the Patriot Hills BIA are rare in Antarctic
terms, with Horseshoe Valley being a slow flowing (< 5 m/a) compound glacier system situated within an
over-deepened catchment, that coalesces with the Institute Ice Stream at the periphery of the WSE (Fig.1), mak-
ing it highly sensitive to grounding line changes across the WSE33; this contrasts with the relatively insensitive
inland ice domes, the sites of traditional ice cores23 (Supplementary Information; FigureS8). In the lee of the
Patriot Hills – a small mountain chain at the end of Horseshoe Valley – strong local katabatic winds descend
into the valley from the polar plateau, ablating the ice sheet surface, drawing up ancient ice from depth within
Horseshoe Valley, and forming the extensive (>800 m) Patriot Hills BIA19,20.
High-resolution analysis of the Patriot Hills BIA using ground-penetrating radar (GPR) demonstrates a
remarkably consistent pattern of layering along the 800 m transect out from Patriot Hills with two distinct uncon-
formities at 247 m (D1) and 360 m (D2) along the prole (Fig.1). ese unconformities are interpreted as periods
of BIA formation within Horseshoe Valley20, occurring in the lee of mountains in the upper part of Horseshoe
Valley during normal ice ow in the build-up to the LGM and at some point during the LGT. is interpretation
is further supported by high-resolution ice-sheet modelling and GPR analysis, which concludes that there was no
major regional ow direction change into Horseshoe Valley during the buildup of the AIS at the LGM20. Together,
these lines of evidence conrm that the ice that accumulated between the unconformities at 247 m (D1) and
360 m (D2) is formed within the valley, thus providing a faithful record of environmental change in the catchment
of Horseshoe Valley in response to broader changes across the WSE19–21.
The chronology of the Patriot Hills BIA. Chronological control across the profile is provided by
Antarctic-wide volcanic tephra horizons at 282 m (~17.8 ka), 279 m (18.2 ka) and 190 m (36.4 ka) (Supplementary
Information), and a comprehensive suite of trace gas samples – carbon dioxide (CO2), methane (CH4) and nitrous
www.nature.com/scientificreports/
8
Scientific RepoRts | 7:39979 | DOI: 10.1038/srep39979
oxides (N2O) – taken from depth (> 3 m) along the BIA transect (Supplementary Information). e trace gases
were extracted and measured at CSIRO’s Ocean and Atmosphere ICELAB facility in Melbourne, calibrated
to internationally-recognised standards, and aligned to published values reported from EPICA Dome C34,35
(Supplementary Information), providing a conservative range of possible age solutions, which together with the
absolute constraints provided by the tephra horizons, allows the development of a robust chronological frame-
work (Supplementary Information) that can be tied directly to the isotopic prole through high-resolution GPR
survey19,20. e integrity of the extracted air was further checked using sulfur hexauoride (SF6) as an indicator of
contamination by modern air. e average concentration of 8 samples analysed for SF6 was about 5% of modern
day atmospheric concentrations and less than 2% for two of the samples selected to develop the chronology. e
available constraints indicate the complete 800-m long Patriot Hills BIA transect spans ~50 to ~2.3 ka. Here we
focus on the central section of the record, captured between the unconformities at 247 and 360 m, which, with
multiple age ties together with the presence of the volcanic horizons dated to 17.8 ka and 18.2 ka, provides a
robust chronology across the Patriot Hills BIA sequence of uninterrupted isochrons (Fig.1). Further details are
available in the Supplementary Information.
Isotopic analysis. δ D isotopic measurements at 1 m resolution were performed across the Patriot Hills BIA
transect at the Australian Antarctic Division (AAD). ese results were conrmed and augmented by δ D and
δ
18O isotopic measurements at 5 m resolution at James Cook University (JCU), and the University of New South
Wales (UNSW) ICELAB. At the AAD an on-line chromium reduction method on a EuroVector EuroPyrOH-HT
system interfaced in continuous ow mode to an Isoprime isotope ratio mass spectrometer. Analytical precision
is < 0.5‰ and δ D values are expressed relative to the Vienna Standard Mean Ocean Water (VSMOW) scale.
To conrm δ D values, particularly the rapid transitions across the periods dened by the LGT, δ D and δ
18O
were measured independently at JCU using Diusion Sampling - Cavity Ring-down Spectrometry (DS-CRDS).
is system continuously converts liquid water into water vapour for real-time stable isotope analysis by laser
spectroscopy (Picarro L2120-i, Sunnyvale, CA, USA). Each analytical run consisted of 12 standards interspersed
with 44 unknown samples. Data processing was performed using a customised Excel template and included
correction for between-sample memory, instrumental dri and normalisation to the VSMOW scale. Further
details are available in the Supplementary Information. Finally, to ensure reproducibility a sub set of samples
were rerun at UNSW ICELAB for δ D and δ
18O using a Las Gatos Research Liquid Water Isotope Analyzer 24d
(International Atomic Energy WICO Lab ID. 16117). Reported overall analytical precision on long term ice core
standards are < 0.32‰ for δ D, and < 0.13‰ for δ
18O values are expressed relative to the (VSMOW Scale).
References
1. Masson-Delmotte, V. et al. In Climate Change 2013: The Physical Science Basis. Contribution of Woring Group I to the Fifth
Assessment eport of the Intergovernmental Panel on Climate Change (eds T.F. Stocer et al.) Ch. 5, 383–464 (Cambridge University
Press, 2013).
2. Weber, M. E. et al. Millennial-scale variability in Antarctic ice-sheet discharge during the last deglaciation. Nature 510, 134–138, doi:
10.1038/nature13397 (2014).
3. Church, J. A. et al. In Climate Change 2013: e Physical Science Basis. Contribution of Woring Group I to the Fih Assessment eport
of the Intergovernmental Panel on Climate Change (eds T.F. Stocer et al.) Ch. 13, 1137–1216 (Cambridge University Press, 2013).
4. Collins, M. et al. In Climate Change 2013: e Physical Science Basis. Contribution of Woring Group I to the Fih Assessment eport
of the Intergovernmental Panel on Climate Change (eds T.F. Stocer et al.) Ch. 12, 1029–1136 (Cambridge University Press, 2013).
5. Joughin, I., Smith, B. E. & Medley, B. Marine ice sheet collapse potentially under way for the waites Glacier Basin, West Antarctica.
Science 344, 735–738 (2014).
6. van Wij, E. M. & intoul, S. . Freshening drives contraction of Antarctic Bottom Water in the Australian Antarctic Basin.
Geophysical esearch Letters 41, 2013GL058921, doi:10.1002/2013gl058921 (2014).
7. Jones, J. M. et al. Assessing recent trends in high-latitude Southern Hemisphere surface climate. Nature Clim. Change 6, 917–926,
doi: 10.1038/nclimate3103 (2016).
8. Golledge, N. . et al. e multi-millennial Antarctic commitment to future sea-level rise. Nature 526, 421–425, doi: 10.1038/
nature15706 (2015).
9. Anderson, . F. et al. Wind-Driven Upwelling in the Southern Ocean and the Deglacial ise in Atmospheric CO2. Science 323,
1443–1448, (2009).
10. Marcott, S. A. et al. Centennial-scale changes in the global carbon cycle during the last deglaciation. Nature 514, 616–619 (2014).
11. Pedro, J. B. et al. e spatial extent and dynamics of the Antarctic Cold eversal. Nature Geosci 9, 51–55 (2015).
12. Hogg, A. et al. Punctuated shutdown of Atlantic Meridional Overturning Circulation during the Greenland Stadial 1. Scientic
eports 6 (2016).
13. Deschamps, P. et al. Ice-sheet collapse and sea-level rise at the Bølling warming 14,600 years ago. Nature 483, 559–564 (2012).
14. Liu, J., Milne, G. A., opp, . E., Clar, P. U. & Shennan, I. Sea-level constraints on the amplitude and source distribution of
Meltwater Pulse 1A. Nature Geosci 9, 130–134 (2016).
15. Hillenbrand, C.-D. et al. econstruction of changes in the Weddell Sea sector of the Antarctic Ice Sheet since the Last Glacial
Maximum. Quaternary Science eviews 100, 111–136 (2014).
16. Clar, P. U., Mitrovica, J. X., Milne, G. A. & Tamisiea, M. E. Sea-level ngerprinting as a direct test for the source of global meltwater
pulse IA. Science 295, 2438–2441 doi: 10.1126/science.1068797 (2002).
17. Golledge, N. . et al. Antarctic contribution to meltwater pulse 1A from reduced Southern Ocean overturning. Nat Commun 5, doi:
10.1038/ncomms6107 (2014).
18. Menviel, L., Timmermann, A., Elison Timm, O. & Mouchet, A. Deconstructing the Last Glacial Termination: the role of millennial
and orbital-scale forcings. Quaternary Science eviews 30, 1155–1172 (2011).
19. Turney, C. S. M. et al. Late Pleistocene and early Holocene change in the Weddell Sea: a new climate record from the Patriot Hills,
Ellsworth Mountains, West Antarctica. Journal of Quaternary Science 28, 697–704 (2013).
20. Winter, . et al. Assessing the continuity of the blue ice climate record at Patriot Hills, Horseshoe Valley, West Antarctica. Geophys.
es. Lett. 10.1002/2015GL066476 (2016).
21. Hein, A. S. et al. Evidence for the stability of the West Antarctic Ice Sheet divide for 1.4 million years. Nature Communciations 7
(2016).
www.nature.com/scientificreports/
9
Scientific RepoRts | 7:39979 | DOI: 10.1038/srep39979
22. Mulvaney, . et al. ecent Antarctic Peninsula warming relative to Holocene climate and ice-shelf history. Nature 489, 141–144, doi:
10.1038/nature11391 (2012).
23. WAIS Divide Members. Onset of deglacial warming in West Antarctica driven by local orbital forcing. Nature 500, 440–444
(2013).
24. Jouzel, J. et al. Magnitude of isotope/temperature scaling for interpretation of central Antarctic ice cores. Journal of Geophysical
esearch 108 (2003).
25. Bentley, M. J. et al. Deglacial history of the West Antarctic Ice Sheet in the Weddell Sea embayment: Constraints on past ice volume
change. Geology 38, 411–414 (2010).
26. Jacobs, S. S., Jenins, A., Giulivi, C. F. & Dutrieux, P. Stronger ocean circulation and increased melting under Pine Island Glacier ice
shelf. Nature Geosci 4, 519–523 (2011).
27. Fabel D. et al. Landscape preservation under Fennoscandian ice sheets determined from in situ produced Be-10 and Al-26. Earth
and Planetary Science Letters 201, 397–406 (2002).
28. Clar, P. U. Deglacial history of the West Antarctic Ice Sheet in the Weddell Sea embayment: Constraints on past ice volume change:
Comment. Geology 39, 240 (2011).
29. Fogwill, C. J. et al. Drivers of abrupt Holocene shis in West Antarctic ice stream direction from combined ice sheet modelling and
geologic signatures. Antarctic Science. Vol . 26(6), 674 (2014).
30. Hein, A. et al. Mid-Holocene pulse of thinning in the Weddell S ea sector of the West Antarctic ice sheet. Nature Communications 7,
12511 (2016).
31. Bradley, S. L. et al. Low post-glacial rebound rates in the Weddell Sea due to Late Holocene ice-sheet readvance. Earth and Pl anetary
Science Letters 413, 79–89 (2015).
32. Golledge et al. e multi-millennial Antarctic commitment to future sea-level rise. Nature. 526, 421–425 (2015).
33. Fogwill C. J. et al. Do blue-ice moraines in the Heritage ange show the West Antarctic Ice Sheet survived the last interglacial?
Palaeogeography, Palaeoclimatology, Palaeoecology 335–336 (2012).
34. Lüthi, D. et al. High-resolution carbon dioxide concentration record 650,000–800,000 years before present. Nature 453, 379–382
(2008).
35. Monnin E. et al. Evidence for substantial accumulation rate variability in Antarctica during the Holocene, through synchronization
of CO2 in the Taylor Dome, Dome C and DML ice cores. Earth and Planetary Science Letters 224, 45–54 (2004).
36. Haran, T., Bohlander, J., Scambos, T., Painter, T. & Fahnestoc, M. MODIS Mosaic Image of Antarctica. National Snow and Ice Data
Center, Boulder, Colo., Digital media (2006).
37. odionov, S. N. A sequential algorithm for testing climate regime shis. Geophys. es. Lett. 31 (2004).
38. asmussen, S. O. et al. A new Greenland ice core chronology for the last glacial termination. J. Geophys. es. 111, D06102, doi:
10.1029/2005jd006079 (2006).
39. Blunier, T. et al. Asynchrony of Antarctic and Greenland climate change during the last glacial period. Nature 394, 739–743
(1998).
Acknowledgements
C.J.F., C.S.M.T. and N.R.G. are supported by their respective Australian Research Council (ARC) and Royal
Society of N.Z. fellowships. Fieldwork was undertaken under ARC Linkage Project (LP120200724), supported
by Linkage Partner Antarctic Logistics and Expeditions. J.W. and K.W. undertook GPR survey of the Patriot
Hills record under NERC grant NE/I027576/1 with logistical eld support from the British Antarctic Survey. We
thank Dr Chris Hayward for electron microprobe assistance, Dr Nelia Dunbar for providing the Siple Dome data
and Kathryn Lacey and Gareth James for help with preparing the tephra samples, CSIRO GASLAB personnel
for support of gas analysis, and Prof. Bill Sturges and Dr Sam Allin of the Centre for Ocean and Atmospheric
Sciences, University of East Anglia for performing the SF6 analyses. CSIROs contribution was supported in
part by the Australian Climate Change Science Program (ACCSP), an Australian Government Initiative. SMD
acknowledges nancial support from Coleg Cymraeg Cenedlaethol and the European Research Council (ERC
grant agreement no. 25923), LM acknowledges funding from the ARC (DE150100107). We thank A/Prof.
Andrew Mackintosh for detailed discussions over the implications of our data and acknowledge the eorts of two
anonymous reviewers whose detailed reviews strengthened the manuscript. e data reported in this paper are
archived on the NOAA Paleoclimatology website. e author(s) wish to acknowledge use of the Ferret program
for analysis and graphics of the LOVECLIM outputs presented in this paper. Ferret is a product of NOAAs
Pacic Marine Environmental Laboratory. (Information is available at http://ferret.pmel.noaa.gov/Ferret/).
Author Contributions
C.J.F., C.S.M.T. and N.R.G. conceived this work under their respective A.R.C. and Royal Society of N.Z.
fellowships. Fieldwork was undertaken under A.R.C. Linkage Project LP120200724, supported by Antarctic
Logistics and Expeditions. D.E., M.R., D.P.T. undertook extraction, measurement and interpretation of the
trace gases. J.W. and K.W. undertook G.P.R. survey of the Patriot Hills record under N.E.R.C. grant NE/
I027576/1. TDvO, A.D.M., M.A.J.C., M.B., N.C.M., C.J.F. and A.B. undertook isotopic analysis. S.M.D.
undertook analysis of the volcanic tephra. C.M.R. and H.M., helped in field work and sampling. B.E.,
A.C., J.V. and H.M. undertook sample preparation, eld sampling and analyses. N.R.G. and L.M. designed
and undertook the LOVECLIM and PISM Earth system and ice sheet model simulations. A.R. advised on
the Patriot Hill BIA and M.E.W. advised on the integration with existing marine records and terrestrial
reconstructions. All authors contributed to the development of ideas and writing of the manuscript.
Additional Information
Supplementary information accompanies this paper at http://www.nature.com/srep
Competing nancial interests: e authors declare no competing nancial interests.
How to cite this article: Fogwill, C.J. et al. Antarctic ice sheet discharge driven by atmosphere-ocean feedbacks
at the Last Glacial Termination. Sci. Rep. 7, 39979; doi: 10.1038/srep39979 (2017).
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional aliations.
www.nature.com/scientificreports/
10
Scientific RepoRts | 7:39979 | DOI: 10.1038/srep39979
is work is licensed under a Creative Commons Attribution 4.0 International License. e images
or other third party material in this article are included in the article’s Creative Commons license,
unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license,
users will need to obtain permission from the license holder to reproduce the material. To view a copy of this
license, visit http://creativecommons.org/licenses/by/4.0/
© e Author(s) 2017
Project
Existe a escala global un creciente interés en la disponibilidad de los recursos hídricos en respuesta al crecimiento sostenido de la población mundial y a la mayor presión que las sociedades moder…" [more]
Project
Recently published findings indicate that the Antarctic Ice Sheet has been much more dynamic during the last deglaciation than previously anticipated. In this project we like to collect and discuss…" [more]
Conference Paper
March 2016
    During the last deglaciation (21,000 to 7,000years ago) global sea level rise was punctuated by several abrupt meltwater spikes triggered by the retreat of ice sheets and glaciers worldwide. However, the debate regarding the relative timing, geographical source and the physical mechanisms driving these rapid increases in sea level has catalyzed debate critical to predicting future sea level... [Show full abstract]
    Conference Paper
    April 2015
      During the last deglaciation (21,000 to 7,000 years ago) global sea level rise was punctuated by several abruptmeltwater spikes triggered by the retreat of ice sheets and glaciers world-wide. However, questions regarding therelative timing, geographical source and the physical mechanisms driving these rapid increases in sea level havecatalysed significant debate that is critical to predicting... [Show full abstract]
      Conference Paper
      December 2014
        The transition from the late Pleistocene to the Holocene (30 000-5000 years ago) was a period of considerable climate variability, which has been associated with changes in deep water formation and the intensity of the Meridional Overturning Circulation. Although numerous records exist across the North Atlantic region, few Antarctic ice core records have been obtained from the south. Here we... [Show full abstract]
        Discover more