ArticlePDF Available

Asynchronous Antarctic and Greenland ice-volume contributions to the last interglacial sea-level highstand


Abstract and Figures

The last interglacial (LIG; ~130 to ~118 thousand years ago, ka) was the last time global sea level rose well above the present level. Greenland Ice Sheet (GrIS) contributions were insufficient to explain the highstand, so that substantial Antarctic Ice Sheet (AIS) reduction is implied. However, the nature and drivers of GrIS and AIS reductions remain enigmatic, even though they may be critical for understanding future sea-level rise. Here we complement existing records with new data, and reveal that the LIG contained an AIS-derived highstand from ~129.5 to ~125 ka, a lowstand centred on 125–124 ka, and joint AIS + GrIS contributions from ~123.5 to ~118 ka. Moreover, a dual substructure within the first highstand suggests temporal variability in the AIS contributions. Implied rates of sea-level rise are high (up to several meters per century; m c⁻¹), and lend credibility to high rates inferred by ice modelling under certain ice-shelf instability parameterisations.
ariability in Last Interglacial sea-level time-series. Yellow bar: time-interval of Heinrich Stadial 11 (HS11) 19 . Orange bar: approximate interval of temporary sea-level drop in various records. Dashed line: end of main LIG highstand set to 118.5 ka (cross-bar indicates 95% confidence limits of ±1.2 ka), based on compilations in b and the speleothem sea-level "ceiling" (c). a GrIS contributions to sea level from a model-based assessment of Greenland icecore data (blue) 9 , and changes in surface sea-water δ 18 O at Eirik Drift (black; this study) with uncertainties (2σ) determined from underpinning δ 18 O and Mg/Ca measurement uncertainties and Mg/Ca calibration uncertainties. b Ninety-five per cent probability interval for coral sea-level markers above 0 m 11 (brown), and LIG duration from a previous compilation (black) 26 . c Red Sea RSL stack (red, including KL23) with 1σ error bars. Smoothings are shown to highlight general trends only, and represent simple polynomial regressions with 68% and 95% confidence limits (orange shading and black dashes, respectively). Purple line indicates the sea-level "ceiling" indicated by subaerial speleothem growth (Yucatan) 24 . d Probability maximum (PM, lines) and its 95% confidence interval for Antarctic temperature changes (red) 68 , and proxy for eastern Atlantic water temperature (ODP976, grey) 69 . Blue crosses: composite record of atmospheric CO 2 concentrations from Antarctic ice cores 19 . e Individual records for Red Sea cores KL11 (blue, dots) and KL23 (red, plusses), with 300-year moving Gaussian smoothings (as used in ref. 1 ). Also shown is a replication exercise to validate the single-sample earliest-LIG peak in KL23 (grey, filled squares) with 1 standard error intervals (bars, σ/√{N}, based on N = 5, 5, 4, 4, and 5 replications, from youngest to oldest sample, respectively). Separate blue cross indicates typical uncertainties (1σ) in individual KL11 datapoints prior to probabilistic analysis of the record. f Probabilistic analysis of the KL11 Red Sea RSL record, taking into account the strict stratigraphic coherence of this record. Results are reported for the median (50th percentile, dashed yellow), PM (modal value, black), the 95% probability interval of the PM (dark grey shading), and both the 68% and 95% probability intervals for individual datapoints (intermediate and light grey shading, respectively)
dentification of Greenland Ice Sheet and Antarctic Ice Sheet contributions to Last Interglacial sea-level variations. a Global Mean Sea Level (GMSL) approximation based on the probabilistically assessed KL11 PM (black line) and its 95% probability interval (grey). This record is shown in terms of RSL in Fig. 2f, but here includes the glacio-isostatic correction and its propagated uncertainty. Black triangles identify limits between which sea-level rises R1, R2, and R3 were measured. Rates of rise with 95% bounds: R1 = 2.8 (1.2-3.7) m c −1 ; R2 = 2.3 (0.9-3.5) m c −1 ; R3 = 0.6 (0.1-1.3) m c −1 . b Blue: GrIS sea-level contribution from the model-data assimilation of ref. 9 (shading represents the 95% probability interval). Grey: GrIS contribution based on Eirik Drift δ 18 O sw . Uncertainties as in Fig. 2a. Orange: AIS contribution from subtraction of the blue GrIS reconstruction from the record in a. Green: AIS contribution found by subtracting the grey GrIS reconstruction from the record in a. Orange and green AIS reconstructions are shown as medians (lines) and 95% confidence intervals (shading). Reconstructed AIS contributions cross downward through a fine dashed when they fall below -10 m, which indicates a rough maximum AIS growth limit in terms of sea-level lowering (AIS growth is limited by Antarctic continental shelf edges). When the green/orange curves fall below these limits, North American and/or Eurasian ice-sheet growth is likely implied. The key result from the present study lies in identification of GrIS and AIS sea-level contributions above 0 m. c Southern Ocean ODP (Ocean Drilling Program) Site 1094 authigenic uranium mass accumulation rates, on its original, Antarctic Ice Core Chronology (AICC2012) tuned, age model. Dashed lines indicate potential offsets (within uncertainties) between the ODP 1094 AICC2012-based chronology 36 and our LIG chronology (see refs. 10,19 and this study)
This content is subject to copyright. Terms and conditions apply.
Asynchronous Antarctic and Greenland ice-volume
contributions to the last interglacial sea-level
Eelco J. Rohling 1,2,7*, Fiona D. Hibbert 1,7*, Katharine M. Grant1, Eirik V. Galaasen 3, Nil Irvalı3,
Helga F. Kleiven 3, Gianluca Marino1,4, Ulysses Ninnemann3, Andrew P. Roberts1, Yair Rosenthal5,
Hartmut Schulz6, Felicity H. Williams 1& Jimin Yu 1
The last interglacial (LIG; ~130 to ~118 thousand years ago, ka) was the last time global sea
level rose well above the present level. Greenland Ice Sheet (GrIS) contributions were
insufcient to explain the highstand, so that substantial Antarctic Ice Sheet (AIS) reduction is
implied. However, the nature and drivers of GrIS and AIS reductions remain enigmatic, even
though they may be critical for understanding future sea-level rise. Here we complement
existing records with new data, and reveal that the LIG contained an AIS-derived highstand
from ~129.5 to ~125 ka, a lowstand centred on 125124 ka, and joint AIS +GrIS contributions
from ~123.5 to ~118 ka. Moreover, a dual substructure within the rst highstand suggests
temporal variability in the AIS contributions. Implied rates of sea-level rise are high (up to
several meters per century; m c1), and lend credibility to high rates inferred by ice modelling
under certain ice-shelf instability parameterisations. OPEN
1Research School of Earth Sciences, The Australian National University, Canberra, ACT 2601, Australia. 2Ocean and Earth Science, University of
Southampton, National Oceanography Centre, Southampton SO14 3ZH, UK. 3Department of Earth Science and Bjerknes Centre for Climate Research,
University of Bergen, Allegaten 41, 5007 Bergen, Norway. 4Department of Marine Geosciences and Territorial Planning, University of Vigo, 36310 Vigo,
Spain. 5Institute of Marine and Coastal Sciences, Rutgers University, New Brunswick, NJ 08903, USA. 6Department of Geology and Paleontology,
University of Tuebingen, Sigwartstrasse 10, D-7400 Tuebingen, Germany.
These authors contributed equally: Eelco J. Rohling, Fiona D. Hibbert.
NATURE COMMUNICATIONS | (2019) 10:5040 | | 1
Content courtesy of Springer Nature, terms of use apply. Rights reserved
The magnitudes and rates of mass reductions in todays
remaining ice sheets (GrIS and AIS) in response to (past or
future) warming beyond pre-industrial levels remain
poorly understood. With sea levels reaching a highstand of +6to
13, or up to 2 m higher4, relative to the present (hereafter
0 m), the last interglacial (LIG) is a critical test-bed for improving
this understanding. Thermosteric and mountain glacier con-
tributions fell within 0.4 ± 0.3 m and at most 0.3 ± 0.1 m,
respectively5,6, and also Greenland Ice Sheet (GrIS) contributions
were insufcient to explain the LIG highstand79. Hence, sub-
stantial Antarctic Ice Sheet (AIS) reduction is implied13.
Determining AIS and GrIS sea-level contributions during the LIG
in more detail requires detailed records with tightly constrained
chronologies, along with statistical and model-driven assessments
(e.g., see refs. 13,915; Supplementary Note 1). To date, however,
chronological (both absolute and relative) and/or vertical uncer-
tainties in LIG sea-level data have obscured details of the timings,
rates, and origins of change.
Age control is most precise for radiometrically dated coral-based
sea-level data, but stratigraphically discontinuous LIG coverage of
these complex three-dimensional systems, and species- or region-
specic habitat-depth uncertainties affect the inferred sea-level
estimates11. Stratigraphic coherence and, therefore, relative age
relationships among samples are stronger in the sediment-core-
based Red Sea relative sea-level (RSL) record1,10,1618 (Methods),
but its LIG signals initially lacked replication and sufcient age
control1,17. Chronological alignment of the Red Sea record with
radiometrically dated speleothem records has since settled its age
for the LIG-onset 10,18,19, but the LIG-end remains poorly con-
strained (Methods). Also, the Red Sea record has since 2008 (ref. 1)
been a statistical stack of several records without the tight sample-
to-sample stratigraphy of contiguous sampling through a single
core, and this has obscured details that are essential for studying
centennial-scale changes10,1719. Advances in understanding LIG
sea-level contributions therefore relied on statistical deconvolu-
tions based on multiple datasets and associated evaluations with
ambiguous combining of chronologies2,12,13,20,orconsideredonly
mean LIG contributions21. Some of these studies suggest that AIS
contributions likely preceded GrIS contributions, and that there
were intra-LIG sea-level uctuations, with kilo-year averaged rates
of at most 1.1 m per century (and likely smaller)13,thoughthis
does not discount higher values for centennial-scale averages
(e.g., ref. 1).
To quantify centennial-scale average sea-level-rate estimates
that may reveal rapid events and processes of relevance to the
future, and robustly distinguish AIS from GrIS contributions, we
present an approach that integrates precise event-dating from
coral/reef and speleothem records3,2224 with stratigraphically
tightly constrained Red Sea sea-level records and a broad suite of
palaeoceanographic evidence. Results indicate that the LIG con-
tained an early AIS-derived highstand, followed by a drop centred
on 125124 ka, and then joint AIS +GrIS contributions for the
remainder of the LIG. We also infer high rates of sea-level change
(up to several metres per century; m c1), that likely reect
complex interactions between oceanic warming, dynamic ice-
mass loss, and glacio-isostatic responses.
Overview of LIG sea-level evidence. The nature of LIG sea-level
variability remains strongly debated, with emphasis on two issues.
First, near-eld sites (close to the ice sheets) in NW Europe
suggest LIG sea-level stability, although resolution and age con-
trol remain limited and other N European sites might support
sea-level uctuations25. Second, there is a wealth of global sites
(mostly in the far eld relative to the ice sheets) that implies LIG
sea-level variability (Fig. 1), but which also reveals a striking
divergence between site-specic signals with respect to both
timing and amplitude of variability (Supplementary Note 1). This
suggests that individual sites are overprinted by considerable site-
specicinuencese.g., prevailing isostatic, tectonic, physical,
biological, biophysical, and biochemical characteristicsrather
than reecting only global sea-level changes. Regardless, a more
coherent pattern seems to be emerging from the more densely
dated and stratigraphically well-constrained sites, which include
the Seychelles, Bahamas, and also Western Australia (Supple-
mentary Note 1, synthesis). The Seychelles coral data are radio-
metrically precisely dated, avoid glacio-isostatic offsets among
sites, and include stratigraphic relationships that unambiguously
reveal relative event timings3,22. The Bahamas data comprise
stratigraphically well-documented and dated evidence of different
reef-growth phases23. Nevertheless, the overall coral-based lit-
erature suggests at least two plausible types of LIG history (early
vs. late highstand solutions) that remain to be reconciled (Sup-
plementary Note 1, synthesis).
Updated Red Sea age model. Regarding the Red Sea RSL record,
we improve its LIG-end age control10,18 by comparing the entire
dataset (the stack) with radiometrically dated coral-data compi-
lations11,26 and Yucatan cave-deposits that indicate when sea
level dropped below the cave (i.e., a ceilingfor sea level)24. This
comparison reveals that the 95% probability limit of the Red Sea
stack on its latest chronology10,19 dropped too early (123 ka; see
Methods and Supplementary Note 2) relative to the well-dated
archives (119118 ka; Fig. 2b, c; Supplementary Figs. 2 and 3).
We, therefore, adjust this point to 118.5 ± 1.2 ka (95% uncertainty
bounds) (Fig. 2, Supplementary Figs. 2 and 3), and accordingly
revise all interpolated LIG ages with fully propagated uncertain-
ties (Supplementary Fig. 2).
Estimates of Greenland mass loss. Next, we compare the Red Sea
sea-level information (Fig. 2b, c, e, f) with estimates of GrIS-
derived LIG sea-level contributions from a model-data-
assimilation of Greenland ice-core data for summer tempera-
ture anomalies, accumulation rates, and elevation changes9
(Fig. 2a). We add independent support for the inferred late GrIS
contribution9, based on a newly extended record of sea-water
oxygen isotope ratios (δ18O
) from a sediment core from Eirik
Drift, off southern Greenland. In this location, δ18O
Greenland meltwater input with a sensitivity of 4 ± 1.2 m global
sea-level rise for the 1.3change seen in the δ18O
from ~128 to ~118 ka (Fig. 2a) (Methods, Supplementary
Note 3). This record suggests (albeit within combined uncer-
tainties) generally lower GrIS contributions than Yau et al.9,
which may agree with results from other modelling studies for
GrIS14,15. Both the modelling and δ18O
approaches indicate a
late GrIS contribution to LIG sea level, which is further supported
by wider N. Atlantic and European palaeoclimate data, which
reveal that contributions started after 127 ka, while GrIS started
to regain net mass from 121 ka27.
AIS and GrIS distinction. Although GrIS did not affect LIG sea-
level change signicantly before 126.5127 ka (Fig. 2a), the Red
Sea and coral data compiled here imply that sea level crossed 0 m
at 130129.5 ka, during a rapid rise to a rst highstand apex that
was reached at ~127 (Fig. 2b, c, e, f). The Seychelles record
indicates specically that sea level reached 5.9 ± 1.7 m by 128.6 ±
0.8 ka3. We infer that both the rst LIG rise above 0 m and the
subsequent rapid rise between 129.5 and 127 ka resulted from
AIS reduction. Similar qualitative inferences about an early-LIG
AIS highstand contribution have been made previously3,9,19,
2NATURE COMMUNICATIONS | (2019) 10:5040 | |
Content courtesy of Springer Nature, terms of use apply. Rights reserved
including attribution to sustained heat advection to Antarctica
during Heinrich Stadial 11 (HS11; 135130 ka)19, when a
northern hemisphere deglaciation pulse (~70 m sea-level rise in
5000 years) caused overturning-circulation shutdown28, a wide-
spread North Atlantic cold event, and southern hemisphere
warming (Fig. 2d). Here we present a quantitative AIS and GrIS
separation with comprehensively evaluated uncertainties.
First, we determine centennial-scale LIG sea-level variability
from the continuous (and contiguous) single-core RSL record of
central Red Sea core KL11 on our new Red Sea LIG age model.
We validate this record with new data for high-accumulation-rate
core KL23 from the northern Red Sea; i.e., from a physically
separate setting than KL11 (Methods) (Fig. 2e). Given this
validation, we continue with KL11 alone because it remains the
most detailed record from the best-constrained (central) location
in the Red Sea RSL quantication method, where δ18O is least
affected by either Gulf of Aden inow effects in the south, or
northern Red Sea convective overturning and Mediterranean-
derived weather systems in the north16,29.
Second, we perform a Monte Carlo (MC)-style probabilistic
analysis of the KL11 record (Fig. 2f), which accounts for all
uncertainties in individual-sample RSL and age estimates (cf. blue
cross in Fig. 2e). This procedure mimics that applied previously to
the Red Sea stack10,18, but now contains an additional criterion of
strict stratigraphic coherence (Methods). The analysis leads to
statistical uncertainty reduction based on datapoint character-
istics, density, and stratigraphy. Remaining RSL uncertainties are
±2.0 to 2.5 m for the 95% probability zone of the probability
maximum (PM, modal value; Fig. 2f; Methods).
Both PM and median reveal an initial RSL rise from ~129.5 to
~127 ka to a highstand apex centred on ~127 ka, followed by a
drop to a lowstand centred on 125124 ka at a few metres below
0 m, and then a small return to a minor peak above 0 m at
~123 ka (Fig. 2f). To quantify AIS contributions, we apply a rst-
order glacio-isostatic correction (with uncertainties) to translate
the record from RSL to global mean sea level (GMSL)
(Supplementary Note 4) (Fig. 3a), and then subtract the GrIS-
contribution records (Figs. 2a and 3b). Our results quantify
signicant asynchrony and amplitude-differences between GrIS
and AIS ice-volume changes during the LIG (Fig. 3b, c). A caveat
applies in intervals where the reconstructed AIS sea-level record
drops below 10 m, because at that stage the maximum AIS
growth limit is approximated (AIS growth is limited by Antarctic
continental shelf edges). Whenever the reconstructed AIS sea-
level record falls below 10 m (notably after ~119 ka), North
American and/or Eurasian ice-sheet growth contributions likely
became important. This timing agrees with a surface-ocean
change south of Iceland from warm to colder conditions27.
Intra-LIG sea-level variability. Red Sea intra-LIG variations are
generally consistent (within uncertainties) in timing with appar-
ent sea-level variations in the well-dated and stratigraphically
coherent coral data from the Seychelles, and Bahamas3,22,23, but
with larger amplitudes. Northwestern Red Sea reef and coastal-
sequence architecture reconstructions offer both timing and
amplitude agreement (although age control needs rening)30,31
(Supplementary Note 1). The reef-architecture study in parti-
cular30 indicates an early-LIG sea-level rise with a post-128-ka
90°W 90°E 180°
<–10 >10–10 –2–4–6–8 0 108642
Multiple LIG highstands
LIG sea-level fall(s)
LIG sea-level oscillation(s)
LIG stillstand(s)
Multiple phases of LIG reef growth
Stratigraphic superposition
No stratigraphic superposition but reef
architecture/geomorphology consistent
with intra-LIG sea-level oscillation(s)
Hanish sill
Fig. 1 Global summary of stratigraphic evidence for Last Interglacial sea-level instability in coral-reef deposits and coastal-sediment sequences. Blue dot is
the location of Hanish Sill, the constraining point for the Red Sea sea-level record. Red squares with white centres are stratigraphically superimposed coral
reef or sedimentary archives for sea-level oscillations within the Last Interglacial (LIG). Solid red dots are locations where sea-level oscillations are inferred
but where there is no stratigraphic superposition. The underlying map is of the difference between maximum Last Interglacial (LIG) relative sea level (RSL)
values for glacio-isostatic adjustment (GIA) modelling results based on two contrasting ice models (ICE-1 and ICE-3) for the penultimate glaciation using
Earth model E1 (VM1-like set up). The ICE-1 model is a version of the ICE-5G ice history (LGM-like), whereas ICE-3 has both reduced total ice volume
relative to ICE-1, and a different ice-mass distribution (i.e., a smaller North American Ice Sheet complex and larger Eurasian Ice Sheet) that is consistent
with glaciological reconstructions of the penultimate glacial period4
NATURE COMMUNICATIONS | (2019) 10:5040 | | 3
Content courtesy of Springer Nature, terms of use apply. Rights reserved
110 115 120 125 130 135
Eirik drift δ18Osw (‰)
ΔSL (m)
–10 Ceiling
RSL (m)
RSL (m)
KL11 probabilistic analysis
Red sea RSL (m)
KL11 (central RS), KL23 (northern RS)
Antarctic ΔT (°C)
105 110 115 120 125
Age (ka BP)
130 135 140
<116 ka,
data density
too low in
just KL11
25 300
ODP976 SST (°C)
CO2 concentration (ppmv)
Fig. 2 Variability in Last Interglacial sea-level time-series. Yellow bar: time-interval of Heinrich Stadial 11 (HS11)19. Orange bar: approximate interval of
temporary sea-level drop in various records. Dashed line: end of main LIG highstand set to 118.5 ka (cross-bar indicates 95% condence limits of ±1.2 ka),
based on compilations in band the speleothem sea-level ceiling(c). aGrIS contributions to sea level from a model-based assessment of Greenland ice-
core data (blue)9, and changes in surface sea-water δ18O at Eirik Drift (black; this study) with uncertainties (2σ) determined from underpinning δ18O and
Mg/Ca measurement uncertainties and Mg/Ca calibration uncertainties. bNinety-ve per cent probability interval for coral sea-level markers above 0 m11
(brown), and LIG duration from a previous compilation (black)26.cRed Sea RSL stack (red, including KL23) with 1σerror bars. Smoothings are shown to
highlight general trends only, and represent simple polynomial regressions with 68% and 95% condence limits (orange shading and black dashes,
respectively). Purple line indicates the sea-level ceilingindicated by subaerial speleothem growth (Yucatan)24.dProbability maximum (PM, lines) and its
95% condence interval for Antarctic temperature changes (red)68, and proxy for eastern Atlantic water temperature (ODP976, grey)69. Blue crosses:
composite record of atmospheric CO
concentrations from Antarctic ice cores19.eIndividual records for Red Sea cores KL11 (blue, dots) and KL23 (red,
plusses), with 300-year moving Gaussian smoothings (as used in ref. 1). Also shown is a replication exercise to validate the single-sample earliest-LIG peak
in KL23 (grey, lled squares) with 1 standard error intervals (bars, σ/{N}, based on N=5, 5, 4, 4, and 5 replications, from youngest to oldest sample,
respectively). Separate blue cross indicates typical uncertainties (1σ) in individual KL11 datapoints prior to probabilistic analysis of the record. fProbabilistic
analysis of the KL11 Red Sea RSL record, taking into account the strict stratigraphic coherence of this record. Results are reported for the median (50th
percentile, dashed yellow), PM (modal value, black), the 95% probability interval of the PM (dark grey shading), and both the 68% and 95% probability
intervals for individual datapoints (intermediate and light grey shading, respectively)
4NATURE COMMUNICATIONS | (2019) 10:5040 | |
Content courtesy of Springer Nature, terms of use apply. Rights reserved
culmination at 510 m above present, followed by a millennial-
scale ~10 m sea-level drop to a lowstand centred on ~124 ka.
In more detail, the probabilistic Red Sea record suggests a
statistically robust dual substructure within the initial LIG sea-
level rise (Fig. 2f), which is replicated between Red Sea records
(Fig. 2e). It is not (yet) supported in wider global evidence
(Methods, Supplementary Note 1), but there are indications that
certain systems may have recorded it independently. For example,
southwestern Red Sea reef-architecture reveals two main reef
phases with a superimposed minor patch-reef phase1,32, reaching
total thicknesses up to 10 m. But more precise dating and support
from other locations are needed to be conclusive. In this context,
we calculate with a basic fringing-reef accretion model that the
rapid rises and short highstands inferred here (Fig. 2e, f) may
have left limited expressions in reef systems, except for rare ones
with exceptionally high accretion rates, or where rapid crustal
uplift offset some of the rapid sea-level rises (Supplementary
Note 5). Hence, we consider wider palaeoceanographic evidence
to evaluate the suggested sea-level history.
Palaeoceanographic support. AIS meltwater pulses implied by
sea-level rises R1 and R2 (Fig. 2f) should have left detectable
signals around Antarctica. The early-LIG AIS sea-level con-
tribution occurred immediately after Heinrich Stadial (HS) 11,
when overturning circulation had recovered from a collapsed
HS11 state (Figs. 24)28. This likely enhanced advection of rela-
tively warm northern-sourced deep water into the Circumpolar
Deep Water (CDW), which impinges on the AIS. At the same
time, there was a peak in Antarctic surface temperatures (Figs. 2d
and 4c) and Southern Ocean sea surface temperatures (ODP Site
1094 TEX
L, ODP Site 1089 planktic foraminiferal δ18O)
(Fig. 4ce), and Southern Ocean sea ice was reduced (Fig. 4b). We
infer that early-LIG AIS retreat resulted from both atmospheric
and (subsurface) oceanic warming, whichtogether with mini-
mal sea ice (important for shielding Antarctic ice shelves from
warm circumpolar waters, e.g., ref. 33)drove enhanced sub-
glacial melting rates and ice-shelf destabilisation, and thus strong
AIS sea-level contributions between 130 and 125 ka.
Wider palaeoceanographic evidence can be used to test
the concept that major AIS melt will provide freshwater to
the ocean surface, which density-straties the near-continental
Southern ocean, impeding Antarctic Bottom Water (AABW)
formation34,35, which in turn will lead to reduced AABW
ventilation/oxygenation and an increase in North Atlantic Deep
Water (NADW) proportion vs. AABW proportion in the Atlantic
Ocean28,36. Thus, we infer strong support for early-LIG AIS melt
from palaeoceanographic observations. For example, an anomaly
in authigenic uranium mass-accumulation rates (aU MAR) in
Southern Ocean ODP Site 1094 has been attributed to bottom-
water deoxygenation (AABW reduction/stagnation), due to
strong Antarctic meltwater releases and consequent water-
column stratication36 (Figs. 3c and 4g). Also, increased
bottom-water δ13C, due to expansion of high-δ13C NADW at
the expense of low-δ13C AABW, occurred at the end of HS11 in
both the abyssal North Atlantic (ODP Site 1063, core MD03-
2664) and South Atlantic (Sites 1089 and 1094) (Fig. 4i).
Moreover, ε
changes in Site 1063 (ref. 28) support the δ13C
interpretation (Fig. 4h). Given that intensication of relatively
warm NADW likely plays a key role in subglacial melting and
resultant AABW source-water freshening33,37, we infer a positive
feedback. In this feedback, meltwater-induced AABW reduction
warmed CDW through increased admixture of relatively warm
NADW, which then caused further subglacial melting and
AABW source-water freshening, driving additional AABW
decline. Finally, a distinct early-LIG minimum in the Site 1089
plankticbenthic foraminiferal δ18O gradient indicates a persis-
tent surface buoyancy anomaly, which agrees with strong AIS
meltwater input38 (Fig. 4cf). Surface buoyancy/stratication
increase would restrict airsea exchange and subsurface heat loss.
Analogous to explanations offered for high melt rates in some
regions of Antarctica today and for even higher melt rates in a
warmer future climate39, we therefore propose another positive
feedback for the LIG, in which melt-stratication led to
subsurface ocean warming, which then intensied ice-shelf
Finally, we note that the aU MAR variations in Southern
Ocean Site 1094 (ref. 36) also agree in more detail with our
inferred dual substructure in the AIS-related early-LIG highstand
(Fig. 3b, c). It is not yet possible to eliminate robustly the inferred
offsets (which fall within uncertainties) between the ODP 1094
ΔSL (m)
120 125
R2 R1
Red sea
KL11-based GMSL (m)
Southern ocean
aU MAR (μg cm–2 ky–1)
115 120 125
Age (ks BP)
Fig. 3 Identication of Greenland Ice Sheet and Antarctic Ice Sheet
contributions to Last Interglacial sea-level variations. aGlobal Mean Sea
Level (GMSL) approximation based on the probabilistically assessed KL11
PM (black line) and its 95% probability interval (grey). This record is
shown in terms of RSL in Fig. 2f, but here includes the glacio-isostatic
correction and its propagated uncertainty. Black triangles identify limits
between which sea-level rises R1, R2, and R3 were measured. Rates of rise
with 95% bounds: R1 =2.8 (1.23.7) m c1;R2=2.3 (0.93.5) m c1;R3=
0.6 (0.11.3) m c1.bBlue: GrIS sea-level contribution from the model-data
assimilation of ref. 9(shading represents the 95% probability interval).
Grey: GrIS contribution based on Eirik Drift δ18O
. Uncertainties as in
Fig. 2a. Orange: AIS contribution from subtraction of the blue GrIS
reconstruction from the record in a. Green: AIS contribution found by
subtracting the grey GrIS reconstruction from the record in a. Orange and
green AIS reconstructions are shown as medians (lines) and 95%
condence intervals (shading). Reconstructed AIS contributions cross
downward through a ne dashed when they fall below 10 m, which
indicates a rough maximum AIS growth limit in terms of sea-level lowering
(AIS growth is limited by Antarctic continental shelf edges). When the
green/orange curves fall below these limits, North American and/or
Eurasian ice-sheet growth is likely implied. The key result from the present
study lies in identication of GrIS and AIS sea-level contributions above
0m.cSouthern Ocean ODP (Ocean Drilling Program) Site 1094 authigenic
uranium mass accumulation rates, on its original, Antarctic Ice Core
Chronology (AICC2012) tuned, age model. Dashed lines indicate potential
offsets (within uncertainties) between the ODP 1094 AICC2012-based
chronology36 and our LIG chronology (see refs. 10,19 and this study)
NATURE COMMUNICATIONS | (2019) 10:5040 | | 5
Content courtesy of Springer Nature, terms of use apply. Rights reserved
AICC2012-based chronology36 and our LIG chronology (see
refs. 10,19 and this study) (Fig. 3b, c), but the offsets may also
(partly) arise from time-lags between meltwater input at the
surface and oxygenation decline at the sea oor. Given the
position of ODP Site 1094 (South Atlantic sector), the aU MAR
record may be to some extent site-specic, in which case it
suggests a likely meltwater source from the West Antarctic Ice
Sheet (WAIS). The lack of later aU MAR spikes for our further
inferred AIS contribution may then suggest either that most of
WAIS had been lost during the earliest LIG, or that it had at least
retreated far enough to stop contributions as is also indicated by
ice-sheet studies14,4043.
Site 1089
planktic-benthic foram.
δ18O (‰ VPDB)
115 120 125 130 135 140
Site 1094 TEX86
L (°C)
EDC ssNa (μg m–2 y–1)
Age (ka)
Benthic foraminifera δ13C (‰ VPDB)
Site 1063
atm. CO2 (ppm)
Site 1094
aU MAR (μg cm–2 ky–1)
retreat HS11
sea ice
δD (‰)
Site 1089
G. bulloides δ18O (‰ VPDB)
115 120 125 130 135 140
(3442 m)
(3442 m)
Site 1063 (4584 m)
benthic δ13C; Nd
Site 1063 (4584 m)
benthic δ13C; Nd
Site 1089
(4624 m)
Site 1094
(4624 m)
Site 1094
(4624 m)
6NATURE COMMUNICATIONS | (2019) 10:5040 | |
Content courtesy of Springer Nature, terms of use apply. Rights reserved
The summarised suite of palaeoceanographic observations offers
strong support to our reconstruction that early-LIG sea-level rise
above 0 m derived from the AIS, and that this meltwater input
occurred in several distinct pulses. Interruption of the rapid AIS
mass-loss rate during the main phase of ice-sheet/shelf reduction
may reect negative feedbacks of isostatic rebound and resultant
ice-shelf re-grounding that temporarily limited ice-mass loss (e.g.,
refs. 4449). The sea-level-lowering rates we nd in between the
LIG rapid-rise events range between multi-centennial means of
0.23 and 0.63 m c1(with peaks up to 1m c
1) (Fig. 2g,
Supplementary Fig. 10). These imply high rates of global net ice-
volume growth, but we note that LIG accumulation rates over the
AIS may have been ~30% higher than present50 (Supplementary
Note 6).
Our record (Fig. 3a) indicates a rst sea-level rise (R1) above
0 m at event-mean values of 2.8 (1.23.7) m c1, followed by R2
at 2.3 (0.93.5) m c1, and R3 at 0.6 (0.11.3) m c1, where the
ranges in brackets reect the 95% probability bounds. These
values lend credibility to similar rates inferred from ice modelling
that includes certain ice-shelf hydrofracturing and ice-cliff col-
lapse paramerisations51. These processes remain debated, but the
apparent reality of such extreme rates in pre-anthropogenic times
when climate forcing was slower, weaker, and more hemi-
spherically asynchronous than todayincreases the likelihood
that such poorly understood mechanisms may be activated under
anthropogenic global warming, to yield extreme sea-level rise.
In conclusion, we have reconstructed (Fig. 3) an initial sea-level
highstand (above 0 m) at ~129.5 to ~124.5 ka, which derived
almost exclusively from the AIS (in agreement with palaeocea-
nographic evidence), and which reached its highstand apex at
around 127 ka. We nd that the rise toward the apex occurred in
two distinct phases, which also agrees with a palaeoceanographic
record of AABW ventilation changes. Following the apex at
~127 ka, we reconstruct a sea-level drop to a relative lowstand
centred on 125124 ka, which in turn gave way to a minor rise
toward a small peak at or just above 0 m at ~123 ka. GrIS con-
tributions were differently distributed through time. These con-
tributions slowly ramped up from ~127 ka onward, reaching
maximum, sustained contributions to LIG sea level from ~124 ka
until the end of the LIG. Thus, we quantitatively reconstruct that
there was strong asynchrony in the AIS and GrIS contributions to
the LIG highstand, with an AIS-derived maximum that spanned
from ~129.5 to ~124.5 ka, a low centred on 125124 ka, and
variable, joint AIS +GrIS inuences from ~124 to ~119 ka.
We observe rapid rates of sea-level change within the LIG.
These may reect complex interactions through time between: (a)
enhanced accumulation during a regionally warmer-than-present
interglacial50; (b) persistent dynamic ice-loss due to long-term
heat accumulation (e.g., ref. 19); (c) negative glacio-isostatic
feedbacks to ice-mass loss (e.g., refs. 4449); and (d) positive
oceanic feedbacks to Antarctic meltwater releases (Discussion,
and refs. 35,52). Similar sequences may develop in future, given
that warmer CDW is encroaching onto Antarctic shelves, so that
future sea-level rise may become driven by increasingly rapid
mass-loss from the extant AIS ice sheet5356, in addition to the
well-observed GrIS contribution57,58.
Finally, we infer intra-LIG sea-level rises with event-mean rates
of rise of 2.8, 2.3, and 0.6 m c1. Such high pre-anthropogenic
values lend credibility to similar rates inferred from some ice-
modelling approaches51. The apparent reality of such extreme
pre-anthropogenic rates increases the likelihood of extreme sea-
level rise in future centuries.
Red Sea relative sea level record. The Red Sea RSL record derives from con-
tiguous sampling of sediment cores and, thus, has tighter stratigraphic control than
samplings of reef systems, which consist of more complex three-dimensional fra-
meworks. Red Sea sediment cores consist of beige to dark brown hemipelagic mud
and silt, with high wind-blown dust contents in glacial/cold intervals and lower wind-
blown dust contents in interglacial intervals. This results in colour and sediment-
geochemistry variations that allow straightforward assessment of bioturbation. This
was found to be very limited in the cores used here, which agrees with extremely low
numbers of benthic microfossils (benthic numbers per gram are an order of mag-
nitude, or more, lower than planktonic numbers per gram59, reaching two orders of
magnitude lower in the LIG60), which in turn agree with extremely low Total Organic
Carbon contents (at or below detection limit)60. With limited bioturbation, the
stratigraphic coherence of the sediment record is well preserved.
The new KL23 δ18O analyses were performed on 30 specimens per sample of
the planktonic foraminifer Globigerinoides ruber (white) from the 320 to 350 µm
size fraction. Sample spacing and KL11-equivalent age model are indicated in the
data le. Prior to analysis, foraminiferal tests were crushed and cleaned by brief
ultrasonication in methanol. Measurements were performed at the Australian
National University using a Thermo Scientic DELTA V Isotope Ratio Mass
Spectrometer coupled with a KIEL IV Carbonate Device. Results are reported in
per mil deviations from Vienna PeeDee Belemnite using NBS-19 and NBS-18
carbonate standards. External reproducibility (1σ) was always better than 0.08.
Red Sea carbonate δ18O is calculated into RSL variations using a polynomial t
to the methods mathematical solution16,29 (see Supplement of ref. 17). The Red Sea
stack of records17 was dated in detail through the last glacial cycle based on the U/
Th dated Soreq Cave speleothem record10. Through the LIG, however, it was
constrained only by interpolation between tie-points at 135 and 110 ka. The age
model for the LIG-onset was later validated19, yet the LIG-end remained to be
better constrained. Here we make an important adjustment for the LIG-end, based
on radiometrically dated criteria described in the main text. This assignment is
based on a rst-order assessment of the entire Red Sea stack using a simple
polynomial and its 95% uncertainty envelope, and it is validated by the fact that in
the more precise probabilistic analysis of KL11 alone, the 95% probability zone for
individual datapoints (lightest grey) also crosses 0 m at 118.5 ka. We only use the
latter in validation, to avoid circularity in the age-model construction. This
reassigns the level originally dated (by interpolation) at 123 ka in the Red Sea
stack10, to 118.5 ka with 95% uncertainty bounds of ±1.2, where the uncertainties
relate to those of the original age model10 (Fig. 2, Supplementary Fig. 2). Initial age
uncertainties (at 95%) all derive from that study. Next, age interpolations using the
adjusted chronological control point are performed probabilistically using a
Monte-Carlo (MC)-style (n=2000) sequence of Hermite splines that impose
monotonic succession to avoid introduction of spurious age reversals
(Supplementary Fig. 2). Our new chronology for the Red Sea LIG record implies
low sediment accumulation rates without major uctuations within the LIG
(Supplementary Fig. 2). Finally, when performing the sea-level probabilistic
assessment for core KL11, we use the newly diagnosed age uncertainties from
Supplementary Fig. 2, which are wider (more conservative) through the interval
120110 ka than the originals (Supplementary Fig. 2).
The two separate high-resolution LIG sea-level records from the Red Sea
discussed here are an existing one from central Red Sea core KL11 (18°44.5N, 39°
20.6E)1, and a new one from northern Red Sea core KL23 (25°44.9N, 35°03.3E).
The new KL23 LIG record validates the KL11 record, but its early-LIG peak
Fig. 4 Timing of Antarctic Ice Sheet retreat relative to circum-Antarctic climate and ocean warming. LIG records of a. Antarctic ice core composite
atmospheric CO
(ref. 70), bEPICA Dome C sea-salt Na ux (on a logarithmic scale), which reects Southern Ocean sea-ice extent71,cVostok δD
(lilac)67,72,dSite 1089 planktic foraminiferal (G. bulloides)δ18O (red)38,eSite 1094 TEX
L-based sea surface temperatures (orange)36,fSite 1089
planktic minus benthic foraminiferal δ18O() plotted as 3-point running mean (red) and sample average including combined 1-sigma uncertainty (light
red shading)38,gSite 1094 authigenic uranium (aU) accumulation where higher values indicate bottom-water deoxygenation36,hSite 1063 ε
(dark blue,
measured by MC-ICP-MS; light blue, measured by TIMS)28, and ibottom-water δ13C records from Site 1063 (blue, 3-point running mean, based on benthic
foraminifera Cibicidoides wuellerstor,Melonis pompilioides, and Oridorsalis)28, MD032664 (yellow, 3-point running mean, C. wuellerstor)73, Site 1089 (red,
C. wuellerstor)36, and Site 1094 (orange, C. wuellerstor)36.hand iIndicate North Atlantic Deep Water (NADW) inuence as denoted. Map inset includes
marine core locations, plotted using Ocean Data View (
NATURE COMMUNICATIONS | (2019) 10:5040 | | 7
Content courtesy of Springer Nature, terms of use apply. Rights reserved
comprises only one sample/datapoint. The validity of this peak was conrmed with
a multiple replication exercise (Fig. 2e, grey).
Through its continuity, stratigraphic constraints, and consistently high signal-to-
noise ratio and sea-level variations are identied in the Red Sea record with limited
impacts from other factors10,1618,29. However, the Red Sea sea-level record still is only
a RSL record for the Hanish Sill, Bab-el-Mandab, and correction for glacio-isostatic
inuences is needed to obtain estimates of GMSL from this record (Supplementary
Note 4). Following these corrections, we estimate AIS sea-level contributions by
determining the difference between GMSL and two different estimates for the GrIS
contribution (see ref. 9and our Eirik Drift δ18O
approach), with full propagation of
the uncertainties involved (see below, and Supplementary Note 3).
The probabilistic analysis of the Red Sea core KL11 record (Fig. 2f) follows the
same approach as for the Red Sea RSL stack10,18, which gives similar results to an
independent Bayesian approach using the same dataset61. The method uses the full
probability distribution envelopes for both age and sea-level directions, as
characterised by the mean and standard deviation per sample point (see blue cross
in Fig. 2e for these 1σlimits in KL11), and performs 5000 MC-style resamplings of
the record. During this resampling, we here apply an additional criterion of strict
stratigraphic coherence within the contiguously sampled KL11 record (allowing no
age reversals during MC-resampling). The resultant suite of MC simulations is then
analysed at set time-steps to identify the probability maximum (modal value, with
95% probability window that depends on how well-dened the modal value is),
median, and the 16th, 84th, 2.5th, and 97.5th percentiles that demarcate the 68%
and 95% probability zones of the total MC-resampled distribution of individual-
sample points (Fig. 2f). Because of the stratigraphic coherence in the KL11 record
considered here, the modal value (and median) in each time-step probability
distribution through the MC simulations is tightly constrained, with the mode
(probability maximum) typically dened within 95% bounds of only ±2 to 2.5 m.
In the earlier studies for the Red Sea stack10,18, this was ±6 m, because a stack of
different records does not preserve strict stratigraphic coherence from one
datapoint to the next, so that relative age uncertainties between datapoints
remained much larger than in our new record.
Eirik Drift surface sea-water δ18O record (δ18O
). Our Eirik Drift surface sea-
water δ18Orecord(δ18O
) was determined for core MD03-2664 (57°26N, 48°36W,
3442 m) using the palaeotemperature equation of ref. 62, with a Vienna PeeDee
Belemnite to Standard Mean Ocean Water standards conversion of 0.27, using
δ18O (ref. 63) and Mg/Ca temperature data64 for the planktonic foraminiferal
species Neogloboquadrina pachyderma (sinistral; 150250 µm size fraction), on the
chronology of ref. 64. Previously published estimates for δ18O
covered only late
MIS 6 and early MIS 5e (26002850 cm core depth63), and are supplemented here
with new estimates for core depths ranging between 2350 and 2600 cm. Even today,
the location of MD03-2664 is dominated by currents carrying admixtures of 16O-
enriched Greenland melt water, with increased melt admixtures causing more
negative δ18O
values65,66. Specically, δ18O
at this site is highly sensitive to
changes in the net freshwater δ18O endmember65. Less GrIS meltwater discharge
and relative dominance of sea-ice meltwater yield a less negative net freshwater
endmember δ18O, whereas the opposite yields a very negative net freshwater
endmember δ18O (see ref. 65 and references there in). Regional freshwater end-
member changes span a range of ~10or more, so while marine endmember
changes are <0.565, sustained MD03-2664 δ18O
changes reect net freshwater
component changes, and therefore mainly GrIS melt. Using an endmember mixing
model, and fully propagating generous uncertainties, we nd that (all else being
constant) the observed 1.3δ18O
change in MD03-2664 corresponds to 4 ±
1.2 m GrIS-derived sea-level rise (Supplementary Note 3).
Data availability
The new Red Sea KL23 δ18O and sea level data, Eirik Drift δ18O
data supporting the
ndings of this study, and source data for Figs. 2 and 3, are provided with the paper as a
Source Data le [] and via http://www. Further information is available from the corresponding author upon
reasonable request.
Received: 27 November 2018; Accepted: 7 October 2019;
1. Rohling, E. J. et al. High rates of sea-level rise during the last interglacial
period. Nat. Geosci. 1,3842, (2008).
2. Kopp, R. E., Simons, F. J., Mitrovica, J. X., Maloof, A. C. & Oppenheimer, M.
Probabilistic assessment of sea level during the last interglacial stage. Nature
462, 863867 (2009).
3. Dutton, A., Webster, J. M., Zwartz, D., Lambeck, K. & Wohlfarth, B. Tropical
tales of polar ice: evidence of Last Interglacial polar ice sheet retreat recorded by
fossil reefs of the granitic Seychelles islands. Quat. Sci. Rev. 107,182196 (2015).
4. Rohling, E. J. et al. Differences between the last two glacial maxima and
implications for ice-sheet, δ18O, and sea-level reconstructions. Quat. Sci. Rev.
176,128 (2017).
5. McKay, N. P., Overpeck, J. T. & Otto-Bliesner, B. L. The role of ocean thermal
expansion in Last Interglacial sea level rise. Geophys. Res. Lett. 38, L14605
6. Farinotti, D. et al. A consensus estimate for the ice thickness distribution of all
glaciers on Earth. Nat. Geosci. 12, 168173 (2019).
7. Cuffey, K. M. & Marshall, S. J. Substantial contribution to sea-level rise during
the last interglacial from the Greenland ice sheet. Nature 404, 591594 (2000).
8. Dahl-Jensen, D. et al. Eemian interglacial reconstructed from a Greenland
folded ice core. Nature 493, 489494 (2013).
9. Yau, A. M., Bender, M. L., Robinson, A. & Brook, E. J. Reconstructing the last
interglacial at Summit, Greenland: insights from GISP2. Proc. Natl. Acad. Sci.
USA 113, 97109715 (2016).
10. Grant, K. M. et al. Rapid coupling between ice volume and polar temperature
over the past 150,000 years. Nature 491, 744747 (2012).
11. Hibbert, F. D. et al. Coral indicators of past sea-level change: a global
repository of U-series dated benchmarks. Quat. Sci. Rev. 145,156 (2016).
12. Düsterhus, A., Tamisiea, M. E. & Jevrejeva, S. Estimating the sea level
highstand during the last interglacial: a probabilistic massive ensemble
approach. Geophys. J. Int. 2, 900920 (2016).
13. Kopp, R. E., Simons, F. J., Mitrovica, J. X., Maloof, A. C. & Oppenheimer, M.
A probabilistic assessment of sea level variations within the last interglacial
stage. Geophys. J. Int. 193, 711716 (2013).
14. Goelzer, H., Huybrechts, P., Marie-France, L. & Fichefet, T. Last Interglacial
climate and sea-level evolution from a coupled ice sheet-climate model. Clim
Past 12, 21952213 (2016).
15. Calov, R., Robinson, A., Perrette, M. & Ganopolski, A. Simulating the
Greenland ice sheet under present-day and palaeo constraints including a new
discharge parameterization. Cryosph 9, 179196 (2015).
16. Siddall, M. et al. Sea-level uctuations during the last glacial cycle. Nature 423,
853858 (2003).
17. Rohling, E. J. et al. Antarctic temperature and global sea level closely coupled
over the past ve glacial cycles. Nat. Geosci. 2, 500504 (2009).
18. Grant, K. M. et al. Sea-level variability over ve glacial cycles. Nat. Commun.
5, (2014).
19. Marino, G. et al. Bipolar seesaw control on last interglacial sea level. Nature
522, 197201 (2015).
20. Barlow, N. L. et al. Lack of evidence for a substantial sea-level uctuation
within the Last Interglacial. Nat. Geosci. 11, 627634 (2018).
21. Overpeck, J. T. et al. Paleoclimatic evidence for future ice-sheet instability and
rapid sea-level rise. Science 311, 17471750 (2006).
22. Vyverberg, K. et al. Episodic reef growth in the granitic Seychelles during the
Last Interglacial: implications for polar ice sheet dynamics. Mar. Geol. 399,
170187 (2018).
23. Thompson, W. G., Curran, H. A., Wilson, M. A. & White, B. Sea-level
oscillations during the last interglacial highstand recorded by Bahamas corals.
Nat. Geosci. 4, 684687 (2011).
24. Moseley, G. E., Smart, P. L., Richards, D. A. & Hoffmann, D. L. Speleothem
constraints on marine isotope stage (MIS) 5 relative sea levels, Yucatan
Peninsula, Mexico. J. Quat. Sci. 28, 293300 (2013).
25. Long, A. J. et al. Near-eld sea-level variability in northwest Europe and ice
sheet stability during the last interglacial. Quat. Sci. Rev. 126,2640 (2015).
26. Cutler, K. B. et al. Rapid sea-level fall and deep-ocean temperature change
since the last interglacial period. Earth Planet. Sci. Lett. 206, 253271 (2003).
27. Tzedakis, P. C. et al. Enhanced climate instability in the North Atlantic and
southern Europe during the Last Interglacial. Nat. Commun.9,
10.1038/s41467-018-06683-3 (2018).
28. Deaney, E. L., Barker, S. & van de Flierdt, T. Timing and nature of AMOC
recovery across Termination 2 and magnitude of deglacial CO
change. Nat.
Commun.8, (2017).
29. Siddall, M. et al. Understanding the Red Sea response to sea level. Earth
Planet. Sci. Lett. 225, 421434 (2004).
30. Plaziat, J.-C., Reyss, J.-L., Choukri, A. & Cazala, C. Diagenetic rejuvenation of
raised coral reefs and precision of dating. The contribution of the Red Sea
reefs to the question of reliability of the Uranium-series datings of middle to
late Pleistocene key reef-terraces of the world. Carnets Géol. Notebooks Geol. 4,
2008/04 (2008).
31. Orszag-Sperber, F., Plaziat, J. C., Baltzer, F. & Purser, B. H. Gypsum salina-
coral reef relationships during the Last Interglacial (Marine Isotopic Stage 5e)
on the Egyptian Red Sea coast: a Quaternary analogue for Neogene marginal
evaporites? Sediment. Geol. 140,6185 (2001).
32. Bruggemann, J. H. et al. Stratigraphy, palaeoenvironments and model for the
deposition of the Abdur Reef Limestone: context for an important
archaeological site from the last interglacial on the Red Sea coast of Eritrea.
Palaeogeogr. Palaeoclimatol. Palaeoecol. 203, 179206 (2004).
8NATURE COMMUNICATIONS | (2019) 10:5040 | |
Content courtesy of Springer Nature, terms of use apply. Rights reserved
33. Hellmer, H. H., Kauker, F., Timmermann, R., Determann, J. & Rae, J. Twenty-
rst-century warming of a large Antarctic ice-shelf cavity by a redirected
coastal current. Nature 485, 225228 (2012).
34. Fogwill, C. J., Phipps, S. J., Turney, C. S. M. & Golledge, N. R. Sensitivity of the
Southern Ocean to enhanced regional Antarctic ice sheet meltwater input.
Earths. Future 3, 317329 (2015).
35. Phipps, S. J., Fogwill, C. J. & Turney, C. S. M. Impacts of marine instability
across the East Antarctic Ice Sheet on Southern Ocean dynamics. Cryosphere
10, 23172328 (2016).
36. Hayes, C. T. et al. A stagnation event in the deep South Atlantic during the last
interglacial period. Science 346, 15141517 (2014).
37. Adkins, J. F. The role of deep ocean circulation in setting glacial climates.
Paleoceanography 28, 539561 (2013).
38. Ninnemann, U. S., Charles, C. D. & Hodell, D. A. In Mechanisms of Global
Climate Change at Millennial Time Scales, Geophysical Monograph Series (eds.
Clark, P. U., Webb, R. S. & Keigwin, L. D.) Vol. 112, 99112 (American
Geophysical Union, 1999).
39. Silvano, A. et al. Freshening by glacial meltwater enhances melting of ice
shelves and reduces formation of Antarctic Bottom Water. Sci. Adv. 4,
eaap9467, (2018).
40. Vaughan, D. G., Barnes, D. K. A., Fretwell, P. T. & Bingham, R. G. Potential
seaways across West Antarctica. Geochem., Geophys. Geosystems 12, Q10004
41. Holden, P. B. et al. Interhemispheric coupling, the West Antarctic Ice Sheet
and warm Antarctic interglacials. Clim 6, 431443 (2010).
42. Steig, E. J. et al. Inuence of West Antarctic Ice Sheet collapse on Antarctic
surface climate. Geophys. Res. Lett. 42, 48624868 (2015).
43. Holloway, M. D. et al. Antarctic last interglacial isotope peak in response to
sea ice retreat not ice-sheet collapse. Nat. Commun. 7, 12293 (2016).
44. Gomez, N., Mitrovica, J. X., Tamisiea, M. E. & Clark, P. U. A new projection
of sea level change in response to collapse of marine sectors of the Antarctic
Ice Sheet. Geophys. J. Int. 180, 623634 (2010).
45. Gomez, N., Pollard, D. & Mitrovica, J. X. A 3-D coupled ice sheetsea level
model applied to Antarctica through the last 40 ky. Earth Planet. Sci. Lett. 384,
8899 (2013).
46. Gomez, N., Pollard, D. & Holland, D. Sea-level feedback lowers projections of
future Antarctic Ice-Sheet mass loss. Nat. Commun. 6, 8798 (2015).
47. Konrad, H., Sasgen, I., Pollard, D. & Klemann, V. Potential of the solid-Earth
response for limiting long-term West Antarctic Ice Sheet retreat in a warming
climate. Earth Planet. Sci. Lett. 432, 254264 (2015).
48. Bradley, S. L., Hindmarsh, R. C. A., Whitehouse, P. L., Bentley, M. J. & King,
M. A. Low post-glacial rebound rates in the Weddell Sea due to Late Holocene
ice-sheet readvance. Earth Planet. Sci. Lett. 413,7989 (2015).
49. Kingslake, J. et al. Extensive retreat and re-advance of the West Antarctic Ice
Sheet during the Holocene. Nature 558, 430434 (2018).
50. Wolff, E. W. et al. Changes in environment over the last 800,000 years from
chemical analysis of the EPICA Dome C ice core. Quat. Sci. Rev. 29, 285295
51. DeConto, R. M. & Pollard, D. Contribution of Antarctica to past and future
sea-level rise. Nature 531, 591597 (2016).
52. Menviel, L., Timmermann, A., Timm, O. E. & Mouchet, A. Climate and
biogeochemical response to a rapid melting of the West Antarctic Ice Sheet
during interglacials and implications for future climate. Paleoceanography 25, (2010).
53. Rignot, E., Mouginot, J., Morlighem, M., Seroussi, H. & Scheuchl, B.
Widespread, rapid grounding line retreat of Pine Island, Thwaites, Smith, and
Kohler glaciers, West Antarctica, from 1992 to 2011. Geophys. Res. Lett. 41,
35023509 (2014).
54. Golledge, N. R. et al. The multi-millennial Antarctic commitment to future
sea-level rise. Nature 526, 421425 (2015).
55. Jenkins, A. et al. Decadal ocean forcing and Antarctic Ice Sheet response:
lessons from the Amundsen Sea. Oceanography 29, 106117 (2016).
56. The IMBIE team. Mass balance of the Antarctic Ice Sheet from 1992 to 2017.
Nature 558, 219222 (2018).
57. King, M. D. et al. Seasonal to decadal variability in ice discharge from the
Greenland Ice Sheet. Cryosphere 12, 38133825 (2018).
58. van den Broeke, M. R. et al. On the recent contribution of the Greenland ice
sheet to sea level change. Cryosphere 10, 19331946 (2016).
59. Rohling, E. J. et al. Magnitudes of sea-level lowstands of the past 500,000 years.
Nature 394, 162165 (1998).
60. Fenton, M. Late Quaternary History of Red Sea Outow. Ph.D. thesis,
Southampton University (1998).
61. Sambridge, M. Reconstructing time series and their uncertainty from
observations with universal noise. J. Geophys. Res. Solid Earth 121, 49905012
62. Shackleton, N. J. Attainment of isotopic equilibrium between ocean water and
the benthonic foraminifera genus Uvigerina: isotopic changes in the ocean
during the last glacial. Colloq. Int. Cent. Natl. Rech. Sci. 219, 203210 (1974).
63. Irvalı, N. et al. Rapid switches in subpolar North Atlantic hydrography and
climate during the Last Interglacial (MIS 5e). Paleoceanography 27,https:// (2012).
64. Irvalı, N. et al. Evidence for regional cooling, frontal advances, and East
Greenland Ice Sheet changes during the demise of the last interglacial. Quat.
Sci. Rev. 150, 184199 (2016).
65. Cox, K. A. et al. Interannual variability of Arctic sea ice export into the East
Greenland Current. J. Geophys. Res. Oceans 115, C12063 (2010).
66. Stanford, J. D., Rohling, E. J., Bacon, S. & Holiday, N. P. A review of the deep
and surface currents around Eirik Drift, south of Greenland: comparison of
the past with the present. Glob. Planet. Change 79, 244254 (2011).
67. Bazin, L. et al. An optimized multi-proxy, multi-site Antarctic ice and
gas orbital chronology (AICC2012): 120800 ka. Clim. Past 9, 17151731
68. Jouzel, J. et al. Orbital and millennial Antarctic climate variability over the
past 800,000 years. Science 317, 793796 (2007).
69. Martrat, B., Jimenez-Amat, P., Zahn, R. & Grimalt, J. O. Similarities and
dissimilarities between the last two deglaciations and interglaciations in the
North Atlantic region. Quat. Sci. Rev. 99, 122134 (2014).
70. Bereiter, B. et al. Revision of the EPICA Dome C CO
record from 800 to 600
kyr before present. Geophys. Res. Lett. 42, 542549 (2015).
71. Wolff, E. W. et al. Southern Ocean sea-ice extent, productivity and iron ux
over the past eight glacial cycles. Nature 440, 491496 (2006).
72. Petit, J. R. et al. Climate and atmospheric history of the past 420,000 years
from the Vostok ice core, Antarctica. Nature 399, 429436 (1999).
73. Galaasen, E. V. et al. Rapid reductions in North Atlantic Deep Water during
the peak of the Last Interglacial period. Science 343, 11291132 (2014).
This research contributes to Australian Research Council Laureate Fellowship
FL120100050 (to E.J.R.). UiB contribution (to E.V.G., N.I., K.K. and U.N.) supported by
RCN project THRESHOLDS (25496). G.M. acknowledges generous support from the
University of Vigo. All plotted new data will be made openly available via http://www.
Author contributions
E.J.R. and F.D.H. led the research. K.M.G., G.M., F.W. and J.Y. added wider doc-
umentation and context. H.S. contributed core curation, sampling, and processing
assistance. E.V.G., N.I., K.K., U.N. and Y.R. provided new oxygen isotope and microfossil
shell chemistry records for Eirik Drift. A.P.R. helped shape the initial concept and
focussed the presentation. All co-authors assisted in producing the manuscript.
Competing interests
The authors declare no competing interests.
Additional information
Supplementary information is available for this paper at
Correspondence and requests for materials should be addressed to E.J.R. or F.D.H.
Peer review information Nature Communications thanks Blake Dyer, Paul Blanchon
and the other, anonymous, reviewer(s) for their contribution to the peer review of this
work. Peer reviewer reports are available.
Reprints and permission information is available at
Publishers note Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional afliations.
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative
Commons license, and indicate if changes were made. The images or other third party
material in this article are included in the articles Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not included in the
articles Creative Commons license and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this license, visit
© The Author(s) 2019
NATURE COMMUNICATIONS | (2019) 10:5040 | | 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
... Neodymium isotope ratios are expressed as ε Nd values, the deviation of 143 Nd/ 144 Nd ratios from the Chondritic Uniform Reservoir value in parts per 10,000, and provide evidence on the locus of subglacial erosion 17 (see Methods). Authigenic Be isotope ratios are reported as 10 Be/ 9 Be ratios and are considered to be an indicator of meltwater inputs [23][24][25] (see Methods). In addition, to evaluate the extent of ice-sheet margin retreat in the WSB, we estimated changes in elevation at Talos Dome during the LIG based on the reported oxygen isotope (δ 18 O) records in Antarctic ice cores 22 . ...
... The production of 10 Be in the atmosphere is followed by its deposition onto the ocean and ice sheets, such that the melting of ice sheets and icebergs releases the accumulated 10 Be into the ocean, where it is scavenged by particles and accumulates in the authigenic fraction of marine sediments 23 . In contrast, 9 Be is the stable naturallyoccurring isotope found in bedrock and supplied via erosion and weathering. An increase in the 10 Be/ 9 Be ratio in marine sediments can therefore be interpreted as indicating a significant ice melting event 23 (see Methods). ...
... Our data provide a unique opportunity to compare a well-resolved iceproximal record of changes in the East Antarctic Ice Sheet to global Article sea-level records for this interglacial period (Fig. 4) [1][2][3][4][5][6][7][8][9] . Sea-level variability during the LIG remains uncertain, but three broad scenarios have been proposed: (1) a high-stand during the early LIG 8,9 ; (2) a high-stand for the initial several thousand years followed by a secondary rise around 120 ka 1,6 ; and (3) bimodal sea-level peaks during the early and late LIG 2-4 ( Fig. 4a-c). ...
Full-text available
The Last Interglacial (LIG: 130,000–115,000 years ago) was a period of warmer global mean temperatures and higher and more variable sea levels than the Holocene (11,700–0 years ago). Therefore, a better understanding of Antarctic ice-sheet dynamics during this interval would provide valuable insights for projecting sea-level change in future warming scenarios. Here we present a high-resolution record constraining ice-sheet changes in the Wilkes Subglacial Basin (WSB) of East Antarctica during the LIG, based on analysis of sediment provenance and an ice melt proxy in a marine sediment core retrieved from the Wilkes Land margin. Our sedimentary records, together with existing ice-core records, reveal dynamic fluctuations of the ice sheet in the WSB, with thinning, melting, and potentially retreat leading to ice loss during both early and late stages of the LIG. We suggest that such changes along the East Antarctic Ice Sheet margin may have contributed to fluctuating global sea levels during the LIG.
... However, recent reconstructions based on both observations and ice sheet modeling suggest that MIS 5e GMSL might have been lower than this range (Clark et al., 2020;Dyer et al., 2021;Polyak et al., 2018). These recent results imply that polar ice sheets may have been less susceptible to melting than previously thought, or, as we propose here, that Antarctica and Greenland were out-of-phase with one another during this time interval, a pattern also suggested by others (i.e., Dyer et al., 2021;Rohling et al., 2019). Here we present a set of well-dated sea-level records that provide more accurate estimates for the magnitude and timing of MIS 5e GMSL, which can be used to test and calibrate ice-sheet models and ultimately to improve projections of the sea-level rise by 2100 and beyond. ...
... Assuming the remaining signal purely reflects AIS melt, this scenario would imply that AIS volume reached a minimum around 128-127 ka, then ice regrew until ~121 ka and finally AIS melting recommenced and lasted until ~118 ka before ice growth towards stage MIS-5d. These simulations are consistent with results from Rohling et al. (2019), but there is no agreement in the contribution magnitude. ...
Accurate characterization of Last Interglacial (MIS 5e; ∼129–116 ka) sea level is important for understanding ice sheet sensitivity to climate change, with implications for predicting future sea-level rise. Here we present a record of MIS 5e sea level based on high-precision U-series ages of 23 corals with precise elevation measurements from reefs around Crooked Island, Long Cay, Long Island, and Eleuthera, The Bahamas. Rigorous screening criteria identified the most pristine samples, and nearly all samples show a narrow δ234Uinitial range between 143.8 and 151.3‰. We infer global mean sea level (GMSL) from these local observations by correcting them for glacial isostatic adjustment (GIA) and long-term subsidence. For GIA, we consider a range of ice histories and Earth viscosity structures. We identify, via Bayesian inference, the range of isostatic and GMSL histories that are consistent with MIS 5e observations across The Bahamas. When applying an open-system correction to our ages, we find that MIS 5e GMSL likely peaked higher than 1 m, but very unlikely exceeded 2.7 m. Our posterior GMSL is lower than previous estimates, but consistent with recent results of modeling and observations. Additionally, sea level observations at other locations (Seychelles, Western Australia, Yucatan) are only slightly above/within the 95% range of predicted local sea level, i.e., GIA plus GMSL, for our open-system/closed-system results. Our relatively constant MIS 5e GMSL indicates that Greenland and Antarctica melted beyond their present extents and, given the insolation forcing, that their contributions to GMSL were likely out-of-phase. These results indicate that the ice sheets may be very sensitive to regional temperature, which has important implications for their combined impact on global sea levels at a time when greenhouse gases increases are causing simultaneous warming at both poles.
... Higher accuracy in defining the distribution in space and time of RSL indicators is of prime importance to evaluate the possibility of meterscale sea-level fluctuations within the LIG highstand, that would imply periods of melting and regrowth of ice caps (e.g. Blanchon et al., 2009;Kopp et al., 2013;Barlow et al., 2018;Polyak et al., 2018;Rohling et al., 2019). ...
... The thinness of the fossil reef (about 2 m) and the presence of corals of different ages around the same elevation suggests a prograding reef growing below a stable sea level. These observations are inconsistent with the significant meter-scale sealevel fluctuations proposed during the LIG, which have been estimated to range from 4 m to 15 m according to Kopp et al. (2013) and Rohling et al. (2019). While it is possible to consider occasional drops in sea level within the thickness of the reef (as indicated by Fig. 5. Paleo RSL calculation from the fossil reef used as a sea-level proxy in Lembetabe (following the methodology of Rovere et al., 2016a). ...
Full-text available
The study of geological sea-level proxies formed during previous interglacials is a common approach to assess how global sea level will evolve under warmer climate conditions. Over the last decades, technical advancements in both survey and geochronology have allowed improving our knowledge of past sea-level highstands. This is of prime importance to refine our understanding of processes contributing to sea-level changes, and ultimately to improve both local and global sea-level projections. Last Interglacial sea-level proxies in the Western Indian Ocean (and more specifically the island nation of Madagascar), have been less investigated than in other intertropical oceans over the last decades. As a result, paleo sea-level data in this region are less abundant and less precise than elsewhere. Here, we report the results of two field campaigns aimed at studying the site of Lembetabe, southwest Madagascar, where a fossil reef was first described by the researcher Ren e Battistini more than 50 years ago. We estimate paleo relative sea level history in space and time from 15 new U-series ages from a fossil reef platform mapped with differential GNSS and drone photogrammetry. Our data suggest that, between 129 ka and 115 ka, paleo relative sea level at this location was about 3.4 ± 1.4 m above modern. Once corrected for glacial isostatic adjustment, we find that paleo global mean sea level did not exceed 3 m above modern. Only slight crustal subsidence would reconcile the peak Last Interglacial sea level measured at Lembetabe with the 5 e10 m range reported in the literature.
... importance of the assessment of Antarctic tipping points (Rohling et al., 2014(Rohling et al., , 2019. One of these warmer periods is the mid-Pliocene Warm Period ( 3.3-3.0 ...
Full-text available
Tipping elements, including the Antarctic Ice Sheet (AIS), are Earth system components that can reach critical thresholds due to anthropogenic emissions. Increasing our understanding of past warm climates can help to elucidate the future contribution of the AIS to emissions. The mid-Pliocene warm period (mPWP, 3.3–3.0 million years ago) serves as an ideal benchmark experiment. During this period, CO2 levels were similar to present-day (350–450 ppmv), but global mean temperatures were 2.5–4.0 degrees higher. Sea-level reconstructions from that time indicate a rise of 10–20 meters compared to the present, highlighting the potential crossing of tipping points in Antarctica. In order to achieve a sea-level contribution far beyond 10 m not only the West Antarctic Ice Sheet (WAIS) needs to largely decrease, but a significant response in the East Antarctic Ice Sheet (EAIS) is also required. A key question in reconstructions and simulations is therefore which of the AIS basins retreated during the mPWP. In this study, we investigate how the AIS responds to climatic and bedrock conditions during the mPWP. To this end we use the Pliocene Model Intercomparison Project, Phase 2 (PlioMIP2) general circulation model ensemble to force a higher-order ice-sheet model. Our simulations reveal that the West Antarctic Ice Sheet experiences collapse with a 0.5 K oceanic warming, the Wilkes basin shows retreat at 3 K oceanic warming, although higher precipitation rates could mitigate such a retreat. Totten glacier shows slight signs of retreats only under high oceanic warming conditions (greater than 4 K oceanic anomaly). We also examine other sources of uncertainty related to initial topography and ice dynamics. we find that the climatologies yield a higher uncertainty than the dynamical configuration, if parameters are constrained with PD observations and that starting from Pliocene reconstructions lead to smaller ice-sheet configurations due to hysteresis behaviour of marine bedrocks. Ultimately, our study concludes that cliff instability is not a prerequisite for the retreat of Wilkes basin. Instead, a significant rise in oceanic temperatures can initiate such a retreat. Our research contributes to a better understanding of Antarctic tipping points and the likelihood of crossing them under future emission scenarios.
... The timing of peak GMSL within the LIG is controversial with different studies arguing for an early peak, a late peak, and even multiple peaks due to orbitally driven asynchronicity in ice loss from the Antarctic and Greenland Ice Sheets (Horton et al., 2018). Ice-sheet modeling and ice core-based reconstructions are beginning to reach consensus on a 0.5-3.5 m LIG sea-level contribution from the Greenland Ice Sheet (e.g., Plach et al., 2019), while an additional $1 m is expected due to thermal expansion and melting of mountain glaciers (Rohling et al., 2019). The Antarctic contribution to LIG GMSL remains unclear since large uncertainties in peak LIG GMSL combined with a lack of direct evidence for contemporaneous mass loss permit values anywhere between 0 and 6 m GMSLE. ...
... These characteristics are therefore useful for understanding future sealevel rise (6,14,15). However, in general, there remains low agreement in LIG GMSL related to its magnitude [e.g., 6.6 to 9.4 m based on a global data compilation (16) to 1.2 to 5.3 m based on data from the Bahamas (17)], timing [e.g., sea-level peak early in the LIG in the Seychelles (18) versus late in Western Australia (19)], structure [e.g., unimodal (20), dual-peaked (16), or multi-peaked (21)], and melt source [GIS or AIS (4)]. Currently, constraints for the Antarctic contribution to LIG GMSL have been inferred from relative sealevel (RSL) curves from single regions [e.g., the Seychelles (6)] or probabilistic assessments of globally distributed datasets (16). ...
Full-text available
Polar temperatures during the Last Interglacial [LIG; ~129 to 116 thousand years (ka)] were warmer than today, making this time period an important testing ground to better understand how ice sheets respond to warming. However, it remains debated how much and when the Antarctic and Greenland ice sheets changed during this period. Here, we present a combination of new and existing absolutely dated LIG sea-level observations from Britain, France, and Denmark. Because of glacial isostatic adjustment (GIA), the LIG Greenland ice melt contribution to sea-level change in this region is small, which allows us to constrain Antarctic ice change. We find that the Antarctic contribution to LIG global mean sea level peaked early in the interglacial (before 126 ka), with a maximum contribution of 5.7 m (50th percentile, 3.6 to 8.7 m central 68% probability) before declining. Our results support an asynchronous melt history over the LIG, with an early Antarctic contribution followed by later Greenland Ice Sheet mass loss.
... The amount, timing and sources of polar meltwater contributions to Last Interglacial (Marine Isotope Stage 5e MIS5e, 130 to 115 ka BP) sea-levels are poorly constrained Barlow et al., 2018;Rohling et al., 2019). Best estimates place the total (eustatic and steric) sea-level rise (SLR) during MIS5e to be 6e9 m of sea-level equivalent above modern sea-level (Kopp et al., 2009(Kopp et al., , 2013. ...
Full-text available
A unique index-record of Last Interglacial (Marine Isotope Stage 5e MIS5e) relative sea level (RSL) and wave climate history in Southeast Australia is presented from Robbins Island, in western Bass Strait. This is applied to interpret the wider MIS5e coastal evidence around Bass Strait. At Robbins Island, the combination of low wave and wind energy, a tide-modified regime and a sand supply resulted in the shoreline progradation throughout MIS5e. This preserved a time-series of paleo-sea level across a 7 km wide strandplain (Remarkable Banks). After a highstand, MIS5e RSL attained a stillstand of þ5.75 ± 0.5 m above modern mean sea level during 126 to~119 ka BP. The MIS5e RSL interpretation is underpinned by modern analogues and hydrodynamic modelling of waves, tides and currents. A high resolution LiDAR Digital Elevation Model (DEM) supported by morpho-sedimentary studies, ground-penetration radar (GPR) surveys and a geochronology based upon Optically Stimulated Luminescence (OSL) methods were used to define the proxy RSL record. The observed RSL history was compared to modelled RSL history that accounted for the theoretical fall in RSL (regression) throughout MIS5e, due to the Glacio-Isostatic Adjustment (GIA) forcing. Three stages of RSL change occurred during MIS 5e: (i) RSL fall during phase 1 from~129 to 126 ka BP, and during phase 3 between~118 and 114 ka BP.; and, (ii) a multi-millennial stillstand during the intervening phase 2 from 126 to~119 ka BP. The stillstand departure from GIA theory, points unam-biguously to persistent polar meltwater contributions to sea level of~2 m from 126 to 119 ka BP, where the component of RSL fall due to GIA was balanced by the RSL rise from meltwater. The potential contributions of paleo wave climate (direction) and boundary current histories were reconstructed from across all Bass Strait sites to determine an RSL budget. In addition, the paleo wave climate history allowed the triangulation of directional ocean wave synoptic sources and identified a 5 poleward shift in the Subtropical Ridge during MIS5e.
Full-text available
The Last Interglacial (LIG, ∼ 130-116 ka) was one of the warmest interglacials of the past 800 000 years and an important test bed for future climate conditions warmer than today. LIG temperature reconstructions from marine records and paleoclimate models show that middle and high northern latitudes were considerably warmer (by about 2 to 5 • C) compared to today. In central Europe, the LIG has been widely studied using pollen and more recently chironomids preserved in lake sediments. While these bio-archives document temperature changes across the LIG, they are commonly poorly constrained chronologically. Speleothems and fluid inclusions contained therein offer superior age control and provide information on past climate, including qualitative and partly also quantitative records of temperature and precipitation. Here, we present a precisely dated fluid-inclusion record based on seven speleothems from two caves in the southeastern Alps (Obir and Katerloch) and use a δ 2 H/T transfer function to reconstruct regional LIG temperatures. We report a temperature change across the glacial-interglacial transition of 5.2 ± 3.1 • C and peak temperatures at ∼ 127 ka of 2.4±2.8 • C above today's mean (1973-2002). The fluid-inclusion δ 2 H record of these speleothems exhibits millennial-scale events during the LIG that are not well expressed in the δ 18 O calcite. The early LIG in the southeastern Alps was marked by an important climate instability followed by progressively more stable conditions. Our record suggests that the southeastern Alps predominantly received Atlantic-derived moisture during the early and middle LIG, while more Mediterranean moisture reached the study site at the end of the LIG, buffering the speleothem δ 18 O calcite signal. The return towards colder conditions is marked by an increase in δ 13 C starting at ∼ 118 ka, indicating a decline in the vegetation and soil activity.
Full-text available
Knowledge of the ice thickness distribution of the world’s glaciers is a fundamental prerequisite for a range of studies. Projections of future glacier change, estimates of the available freshwater resources or assessments of potential sea-level rise all need glacier ice thickness to be accurately constrained. Previous estimates of global glacier volumes are mostly based on scaling relations between glacier area and volume, and only one study provides global-scale information on the ice thickness distribution of individual glaciers. Here we use an ensemble of up to five models to provide a consensus estimate for the ice thickness distribution of all the about 215,000 glaciers outside the Greenland and Antarctic ice sheets. The models use principles of ice flow dynamics to invert for ice thickness from surface characteristics. We find a total volume of 158 ± 41 × 10 ³ km ³ , which is equivalent to 0.32 ± 0.08 m of sea-level change when the fraction of ice located below present-day sea level (roughly 15%) is subtracted. Our results indicate that High Mountain Asia hosts about 27% less glacier ice than previously suggested, and imply that the timing by which the region is expected to lose half of its present-day glacier area has to be moved forward by about one decade. © 2019, The Author(s), under exclusive licence to Springer Nature Limited.
Full-text available
Rapid changes in thickness and velocity have been observed at many marine-terminating glaciers in Greenland, impacting the volume of ice they export, or discharge, from the ice sheet. While annual estimates of ice-sheet-wide discharge have been previously derived, higher-resolution records are required to fully constrain the temporal response of these glaciers to various climatic and mechanical drivers that vary in sub-annual scales. Here we sample outlet glaciers wider than 1km (N=230) to derive the first continuous, ice-sheet-wide record of total ice sheet discharge for the 2000–2016 period, resolving a seasonal variability of 6%. The amplitude of seasonality varies spatially across the ice sheet from 5% in the southeastern region to 9% in the northwest region. We analyze seasonal to annual variability in the discharge time series with respect to both modeled meltwater runoff, obtained from RACMO2.3p2, and glacier front position changes over the same period. We find that year-to-year changes in total ice sheet discharge are related to annual front changes (r²=0.59, p=10-4) and that the annual magnitude of discharge is closely related to cumulative front position changes (r²=0.79), which show a net retreat of >400km, or an average retreat of >2km, at each surveyed glacier. Neither maximum seasonal runoff or annual runoff totals are correlated to annual discharge, which suggests that larger annual quantities of runoff do not relate to increased annual discharge. Discharge and runoff, however, follow similar patterns of seasonal variability with near-coincident periods of acceleration and seasonal maxima. These results suggest that changes in glacier front position drive secular trends in discharge, whereas the impact of runoff is likely limited to the summer months when observed seasonal variations are substantially controlled by the timing of meltwater input.
Full-text available
Considerable ambiguity remains over the extent and nature of millennial/centennial-scale climate instability during the Last Interglacial (LIG). Here we analyse marine and terrestrial proxies from a deep-sea sediment sequence on the Portuguese Margin and combine results with an intensively dated Italian speleothem record and climate-model experiments. The strongest expression of climate variability occurred during the transitions into and out of the LIG. Our records also document a series of multi-centennial intra-interglacial arid events in southern Europe, coherent with cold water-mass expansions in the North Atlantic. The spatial and temporal fingerprints of these changes indicate a reorganization of ocean surface circulation, consistent with low-intensity disruptions of the Atlantic meridional overturning circulation (AMOC). The amplitude of this LIG variability is greater than that observed in Holocene records. Episodic Greenland ice melt and runoff as a result of excess warmth may have contributed to AMOC weakening and increased climate instability throughout the LIG.
Full-text available
During the Last Interglacial, global mean sea level reached approximately 6 to 9 m above the present level. This period of high sea level may have been punctuated by a fall of more than 4 m, but a cause for such a widespread sea-level fall has been elusive. Reconstructions of global mean sea level account for solid Earth processes and so the rapid growth and decay of ice sheets is the most obvious explanation for the sea-level fluctuation. Here, we synthesize published geomorphological and stratigraphic indicators from the Last Interglacial, and find no evidence for ice-sheet regrowth within the warm interglacial climate. We also identify uncertainties in the interpretation of local relative sea-level data that underpin the reconstructions of global mean sea level. Given this uncertainty, and taking into account our inability to identify any plausible processes that would cause global sea level to fall by 4 m during warm climate conditions, we question the occurrence of a rapid sea-level fluctuation within the Last Interglacial. We therefore recommend caution in interpreting the high rates of global mean sea-level rise in excess of 3 to 7 m per 1,000 years that have been proposed for the period following the Last Interglacial sea-level lowstand.
Full-text available
The Antarctic Ice Sheet is an important indicator of climate change and driver of sea-level rise. Here we combine satellite observations of its changing volume, flow and gravitational attraction with modelling of its surface mass balance to show that it lost 2,720 ± 1,390 billion tonnes of ice between 1992 and 2017, which corresponds to an increase in mean sea level of 7.6 ± 3.9 millimetres (errors are one standard deviation). Over this period, ocean-driven melting has caused rates of ice loss from West Antarctica to increase from 53 ± 29 billion to 159 ± 26 billion tonnes per year; ice-shelf collapse has increased the rate of ice loss from the Antarctic Peninsula from 7 ± 13 billion to 33 ± 16 billion tonnes per year. We find large variations in and among model estimates of surface mass balance and glacial isostatic adjustment for East Antarctica, with its average rate of mass gain over the period 1992–2017 (5 ± 46 billion tonnes per year) being the least certain.
Full-text available
To predict the future contributions of the Antarctic ice sheets to sea-level rise, numerical models use reconstructions of past ice-sheet retreat after the Last Glacial Maximum to tune model parameters 1 . Reconstructions of the West Antarctic Ice Sheet have assumed that it retreated progressively throughout the Holocene epoch (the past 11,500 years or so)2-4. Here we show, however, that over this period the grounding line of the West Antarctic Ice Sheet (which marks the point at which it is no longer in contact with the ground and becomes a floating ice shelf) retreated several hundred kilometres inland of today's grounding line, before isostatic rebound caused it to re-advance to its present position. Our evidence includes, first, radiocarbon dating of sediment cores recovered from beneath the ice streams of the Ross Sea sector, indicating widespread Holocene marine exposure; and second, ice-penetrating radar observations of englacial structure in the Weddell Sea sector, indicating ice-shelf grounding. We explore the implications of these findings with an ice-sheet model. Modelled re-advance of the grounding line in the Holocene requires ice-shelf grounding caused by isostatic rebound. Our findings overturn the assumption of progressive retreat of the grounding line during the Holocene in West Antarctica, and corroborate previous suggestions of ice-sheet re-advance 5 . Rebound-driven stabilizing processes were apparently able to halt and reverse climate-initiated ice loss. Whether these processes can reverse present-day ice loss 6 on millennial timescales will depend on bedrock topography and mantle viscosity-parameters that are difficult to measure and to incorporate into ice-sheet models.
Full-text available
Strong heat loss and brine release during sea ice formation in coastal polynyas act to cool and salinify waters on the Antarctic continental shelf. Polynya activity thus both limits the ocean heat flux to the Antarctic Ice Sheet and promotes formation of Dense Shelf Water (DSW), the precursor to Antarctic Bottom Water. However, despite the presence of strong polynyas, DSW is not formed on the Sabrina Coast in East Antarctica and in the Amundsen Sea in West Antarctica. Using a simple ocean model driven by observed forcing, we show that freshwater input from basal melt of ice shelves partially offsets the salt flux by sea ice formation in polynyas found in both regions, preventing full-depth convection and formation of DSW. In the absence of deep convection, warm water that reaches the continental shelf in the bottom layer does not lose much heat to the atmosphere and is thus available to drive the rapid basal melt observed at the Totten Ice Shelf on the Sabrina Coast and at the Dotson and Getz ice shelves in the Amundsen Sea. Our results suggest that increased glacial meltwater input in a warming climate will both reduce Antarctic Bottom Water formation and trigger increased mass loss from the Antarctic Ice Sheet, with consequences for the global overturning circulation and sea level rise.
p>A palaeoceanographic study is carried out on cores from the central Red Sea and western Gulf of Aden. Time stratigraphic frameworks are determined using oxygen isotope ratios, from the test of planktonic foraminifera Globigerinoides ruber , and combined with AMS <sup>14</sup>C dates. Down-core variations in planktonic and benthic foraminifera, organic carbon content, and δ<sup>18</sup>O are used to deliver a comprehensive history of changes in the Red Sea basin during the late Quaternary. The short time span of the Gulf of Aden core 1006 and the presence of a redepositional event in core 1005 rendered both cores unsuitable for the study of glacial Red Sea outflow. Instead the study is focussed on changes in Red Sea circulation and deep water formation relying on evidence provided in the Red Sea cores. The Red Sea is a marginal basin of the NW Indian Ocean. Today, water exchange with the open ocean only takes place across the shallow Hanish Sill at the Strait of Bab el Mandab. River inflow along with precipitation into the basin is negligible with respect to the high evaporation rate of 200 cm yr<sup>-1</sup>. Thus, the basin is extremely sensitive to global climate change and sea level variation. Circulation in the basin is anti-estuarine, with a surface water inflow compensated by a subsurface outflow. Surface flow alters seasonally according to the monsoon. During the summer SW monsoon, northwesterly winds over the entire basin drive a south flowing surface current. At this time inflow into the basin continues as a shallow, subsurface current. During the winter NE monsoon, winds are northwesterly, north of 20<sup>o</sup>N, driving a weak southward surface water flow, and southeasterly to the south, driving a strong northward surface water flow. The result is a zone of surface water convergence at around 25<sup>o</sup>N which migrates south as intensity of the SW monsoon increases. Deep water is a 1:1 mixture of Red Sea surface water with Gulf of Suez outflow and is formed mainly in winter in the north of the basin, as well as the Gulf of Suez. The waters of the present day Red Sea are oligotrophic, supporting a population of tropical-subtropical, spinose foraminifera dominated by Globigerinoides ruber (plate 1) in the south and Globigerinoides sacculifer (plate 2) in the north. The species distribution is controlled by their individual dietary requirements and the seasonal availability of food (mainly zooplankton), in turn controlled by the position of the current convergence zone and monsoonal intensity.</p
Studies of past glacial cycles yield critical information about climate and sea-level (ice-volume) variability, including the sensitivity of climate to radiative change, and impacts of crustal rebound on sea-level reconstructions for past interglacials. Here we identify significant differences between the last and penultimate glacial maxima (LGM and PGM) in terms of global volume and distribution of land ice, despite similar temperatures and radiative forcing. Our analysis challenges conventional views of relationships between global ice volume, sea level, seawater oxygen isotope values, and deep-sea temperature, and supports the potential presence of large floating Arctic ice shelves during the PGM. The existence of different glacial ‘modes’ calls for focussed research on the complex processes behind ice-age development. We present a glacioisostatic assessment to demonstrate how a different PGM ice-sheet configuration might affect sea-level estimates for the last interglacial. Results suggest that this may alter existing last interglacial sea-level estimates, which often use an LGM-like ice configuration, by several metres (likely upward).