2
East Antarctic Ice Sheet bed erosion indicates repeated large-scale retreat and
advance events.
3
Authors
1
Aitken, A.R.A.1, Roberts, J.L.2,3, Van Ommen, T2,3, Young, D.A.4, Golledge, N.R.5,6, Greenbaum, J.S.4,
Blankenship, D.D.4, Siegert, M.J.7
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
With changing climate, ice sheet retreat and advance is a major source of sea level change, but our
understanding of this is limited by high uncertainties. In particular, the contribution of the East Antarctic
Ice Sheet (EAIS) to past sea level change is not well defined. Several lines of evidence suggest possible
collapse of the Totten Glacier catchment of the EAIS during past warm periods, in particular, during the
Pliocene1-4, causing a multi-meter contribution to global sea level. However, the large-scale structure and
long-term evolution of the ice sheet have been insufficiently well known to constrain retreat-advance
extents in this region. Using aerogeophysical data5,6 we define the erosion of the ice sheet bed as an
indicator of past ice sheet activity. We show that deep erosion exposing basement rocks is restricted to
two regions: the head of the Totten Glacier, within 150 km of the present grounding line; and deep within
the Sabrina Subglacial Basin, 350-550 km from the present grounding line. These regions demarcate the
marginal zones of two distinct quasi-stable EAIS configurations corresponding to modern-scale and
retreated ice sheets. Despite experiencing both regimes, the 200 to 250 km broad transitional region
between the two is less eroded, suggesting a shorter-lived exposure to eroding conditions. Limited
transition-zone erosion is well explained by repeated retreat-advance events driven between the quasistable states by ocean-forced instabilities. Representative ice sheet models indicate that the global sealevel increase from retreat in this sector can be up to 0.9 m in the modern-scale configuration, and is in
excess of 2 m in the retreated configuration.
33
34
35
36
37
38
39
Satellite-based observations indicate that the margin of Totten Glacier may be experiencing greater ice loss
than anywhere else in East Antarctica7,8. This, coupled with the presence of the 1.5 km deep Aurora
Subglacial Basin upstream, means Totten Glacier and its drainage basin may be at risk of substantial ice loss
under ocean warming conditions. Vulnerability of this region to change may be driven by access of warm
modified circumpolar deep water to the ice-shelf cavity. At decadal timescales this access is controlled by a
combination of polynya activity and bathymetric constraints9,10. Recent studies indicate that potential access
pathways exist to access the cavity beneath the Totten Ice Shelf11.
40
41
42
Totten Glacier also possesses an inland ice stream that extends far into the continental interior. This ice
stream mostly overlies the Sabrina Subglacial Basin (SSB), which is a bowl-shaped depression bounded to the
east by the Terre Adelie Highlands, to the south by Highland C and to the west by Highland B (Figure 1). The
1 - School of Earth and Environment, The University of Western Australia, Perth, Western Australia, Australia
2 -Australian Antarctic Division, Kingston, Tasmania, Australia
3 -Antarctic Climate & Ecosystems Cooperative Research Centre, The University of Tasmania, Hobart,
Tasmania, Australia.
4- University of Texas Institute of Geophysics, The University of Texas at Austin, Austin, Texas, USA
5 - Antarctic Research Centre, Victoria University of Wellington, Wellington 6140, New Zealand.
6 - GNS Science, Avalon, Lower Hutt 5011, New Zealand
7 - The Grantham Institute; Imperial College London, London, United Kingdom
43
44
45
46
SSB is underlain by an extensive sedimentary basin (the SSSB) of moderate but quite variable thickness
(Figure 1c). This basin probably dates back to at least 40 Ma6, and therefore predates glaciation. Perhaps
because of this sedimentary basin, this region has widely distributed subglacial hydrology12 and little largescale topographic relief5,12,13, although the surface is quite rough at shorter wavelengths12.
47
48
49
50
51
Here we use geophysical data from the ICECAP program (extended data Figure 1) to define the erosion of
the SSB due to the activity of the East Antarctic Ice Sheet (EAIS). We use 2D gravity modelling along flight
lines, also including depth to magnetic basement estimates, to understand the thickness of sedimentary
rocks in the SSB region. Representative ice sheet models support the interpreted erosion regimes and
provide complementary estimates of global sea level change.
52
53
54
55
56
57
58
Figure 1: Interpretation of erosion in the Sabrina Subglacial Basin region: a) Base-ice elevations. Data from
ICECAP (bold) and Bedmap214 (muted). The Sabrina Subglacial Basin is the low-lying region bounded by
Highland B, Highland C and the Terre-Adelie Highlands. VF – Vanderford Glacier; CG – Cape Goodenough.
The sedimentary basin base (b) and sedimentary basin thickness (c) derived from gravity modelling. The SSSB
is the centrally located basin of moderate thickness. Much thicker sedimentary basins occur in the Aurora
Subglacial Basin and Vincennes Subglacial Basin (VSB). d) From these datasets we interpret regions of
59
60
differing erosion characteristics for the SSB. See text for further details of these regions. Inset shows the
study area location.
61
62
63
64
65
66
67
68
The gravity model results show the SSSB to extend southeast from the Totten Glacier for over 500 km. Its
base is as deep as -4000 m elevation (Figure 1b), but is typically much shallower. The basin base is tilted
towards the south or southeast, although several east-northeast oriented faults define smaller-scale
perturbations (Figure 2). Allowing for the subglacial topography (Figure 1a), the thickness of the SSSB is
defined (Figure 1c). The results suggest an initially broad and generally flat-based geometry, with a probable
pre-erosion thickness of approximately 3 km. Error analyses (Figure 2, extended data Figures 2, 4-7) indicate
the modelled SSSB thickness error is approximately ±500m. The ASB and VSB have larger uncertainties, up
to 2 km, and erosion patterns are not interpreted within these basins.
69
70
71
72
The thickness of the SSSB (Figure 1c) shows high variability within tectonic blocks, and also transgressions of
major tectonic structures (see extended data Figure 3, extended data Figure 8) and does not parallel the
tectonic structure of the region 6. Rather, the SSSB thickness defines distinct patterns that relate to the
erosion of the SSB by the East Antarctic Ice Sheet.
73
74
75
76
77
78
79
80
81
Glacial erosion occurs due to basal-sliding, which requires warm-based ice and a sufficiently high basal shear
stress to cause motion, and erosion-rate is primarily dependent on the basal velocity of the ice sheet 15.
Regions with high erosion potential are typically found near the margins of an ice sheet, and may be
selective or distributed. Under selective erosion, deep troughs occur in regions with high lateral
convergence of ice flux, as well as high basal-velocity . These convergent zones are often topographically
focused within pre-existing valley systems16. Thinner ice promotes stronger selective erosion because
highlands are often cold-based and are protected from erosion. The same highlands may be warm-based
under thicker ice cover. As a consequence, larger ice sheets are more likely to exhibit distributed erosion,
whereas smaller ice sheets are more likely to exhibit selective erosion.
82
83
84
85
86
For the SSB we interpret several regions of distinctive erosion (Figure 1d) that define two distinct quasistable EAIS configurations. We identify a modern-scale configuration, with a marginal zone near the presentday margin, and a retreated configuration, with a marginal zone located far inland. These configurations are
not time-specific, although global climate and sea level data suggest a retreated ice sheet was dominant in
the Oligocene to mid-Miocene, with a modern-scale ice sheet dominant since the mid-Miocene17.
87
88
89
90
91
92
Cumulative glacial erosion is generally low in elevated coastal regions, including Law Dome, the Knox Coast
and the Cape Goodenough region. Erosion is also low on the highlands surrounding the SSB and ASB. Ice
sheet models suggest that the coastal highlands are either covered with slow-moving ice, or are ice free
(Figure 3; Extended data figure 9). Highlands A and C, the upper part of Highland B, Ridge B and Dome C
remain ice covered in all models, and it is likely that these highlands were also covered by the early ice sheet
5
93
94
95
96
97
98
99
100
101
102
103
104
Figure 2: Gravity model along flight line R06Ea (location, Figure 1b). a) Observed and calculated gravity
data, including model components from ice and topography, the deep crust and Moho, and the sedimentary
basins. b) The model, also showing prior depth to magnetic basement estimates, the tensioned spline fit,
and cross-ties with other models. The gravity disturbance was modelled at true observation elevations as
indicated by the uppermost grey line c) SSSB thickness and current surface ice velocity (vertically flipped).
Average error in ice sheet velocity on this line is 10.5 m/yr18. A|B, B|C and C|HC (HC-Highland C) indicate
the locations of the boundaries demarcated in Figures 1 and 3. Note the thin to absent basin downstream of
A|B, the sloped base from A|B to B|C, and the anti-correlation between basin thickness and ice-sheet
velocity. Dashed lines indicate the inferred prior 3 km thickness of the SSSB, which has been almost entirely
removed in Region A, and eroded in line with present-day ice velocity in Region B. See extended data figures
4 through 7 for additional representative models throughout the SSB region.
105
106
107
108
Region A surrounds the coastal glaciers, and is typified by a broad region with moderate elevation, cut by a
number of smaller channels that are aligned with modern ice sheet flow (Figure 1a). The SSSB thickness is nil
for much of Region A and rarely exceeds 1 km (Figure 1c). The A|B boundary is located near a topographic
ridge that extends WSW from Cape Goodenough. The ridge is at ca. -200 m elevation, but is cut by channels
109
110
as deep as -600 m elevation (Figure 1a). The ice-sheet bed within region A is ocean-sloping, with an average
measured gradient along flight lines of +2 to +4 m/km.
111
112
113
114
Region B is defined by an almost linear inland-thickening trend to the base of the SSSB, with an upstream
limit defined where this thickening trend ceases to be evident. Over ca. 200-250 km, the thickness of the
SSSB increases in co-variance with the reduction in ice sheet velocity (Figure 2 and extended data Figures 47).
115
116
117
118
119
120
121
Non-selective and inland-reducing erosion within regions A and B points to the activity of an ice sheet with a
modern-scale configuration15. Under current climate conditions our ice sheet model indicates high basal
velocities within Region A (extended data Figure 9a). With an air temperature, dTa=+4°C and an ocean
temperature, dTo=+1°C, the model indicates a retreat of the ice sheet margin to the A|B boundary, with high
basal velocities mostly within Region B (Figure 3a). These results indicate a history of repeated smaller-scale
retreat and advance cycles within Region A. These are likely orbitally forced2, with a typical length-scale of
less than 200 km, and do not invoke collapse.
122
123
124
125
126
127
Region B comprises two subregions. Subregion B2 preserves only the modern-scale ice sheet erosion
signature, and the SSSB is typically over 2 km thick. In Subregion B1, the SSSB is typically 1-2 km thick and, in
addition to the modern-scale ice sheet signature, a muted dendritic pattern is evident (Fig 1). The ice sheet
bed in Subregion B1 is beneath sea level and typically inland-sloping, with an average gradient of -1 to -3
m/km. Therefore, once the topographic ridge at the A|B boundary is reached, retreat into the inland SSB is
subject to ocean-driven instabilities.
128
129
130
131
132
133
Ocean-driven instabilities include the marine ice sheet instability (MISI), moderated by ice-shelf
buttressing19, and ice-cliff failure augmented by hydrofracturing20,21. From a modern-scale configuration,
highly unstable ice sheet reconstructions for the Pliocene, especially those including ice-cliff failure20,21, can
breach the A|B boundary. Under similar climate conditions, more stable reconstructions22,23 typically fail to
breach this boundary. This suggests that a significant degree of ice-sheet instability is necessary to cause
retreat into the SSB interior under Pliocene conditions.
134
135
136
137
138
139
140
Following retreat into the SSB interior, a new ice sheet margin is established in front of Highlands B and C, in
a more advanced location than a previous interpretation of the early ice sheet5. Region C (Figure 1) is
characterised by a well-developed dendritic erosion pattern, with sufficient erosion to expose basement.
This region is consistent with the occupation of a fluvial valley network by a smaller-scale ice sheet 5.
Highland C and the Terre Adelie Highlands are significant topographic highs extending well above sea level.
Average measured gradient on their frontal slopes is ~+2.5 m/km. Consequently, these highlands are
persistent barriers to further ice sheet retreat, even under highly unstable conditions15,21.
141
142
143
144
Highland B, however, is pierced by three deep fjords providing access to the Aurora Subglacial Basin (ASB)5.
This additional vulnerability limits the residence-time of an ice margin in front of Highland B and permits ice
sheet retreat into the ASB5,15,21. The B|C boundary links to the central fjord, suggesting that this fjord is the
most significant for facilitating further collapse5.
145
146
147
148
The erosion characteristics of regions B1 and C point to the activity of a retreated ice sheet. With dTa=+8°C
and dTo=+2°C (Figure 3b) the ice sheet margin retreats to Subregion B1, with high basal velocities mostly
within Subregion B1 and Region C. With dTa=+12°C and dTo=+5°C further retreat to the B|C boundary is
indicated, with only Region C of the SSB being ice covered (Figure 3c). This retreat scenario is similar in
149
150
151
152
153
extent to a previous interpretation of the early ice sheet, prior to advance into the SSB5, although the
locations of the ice margins differ in location. With dTa=+15°C and dTo=+5°C full retreat into the ASB is
indicated (extended data Figure 9b), with ice coverage only on inland highlands. This retreat scenario is
somewhat more extensive than a previous interpretation of the early ice sheet, prior to advance into the
ASB5.
154
155
156
157
158
159
160
161
162
Figure 3: Ice sheet models with differing climate forcing. Ice sheet models showing ice sheet surface and
bed elevations and basal velocity contours. Models were run with air (dTa) and ocean (dTo) temperatures
above today’s. Sea level contributions including ice mass loss and isostatic rebound are estimated for all
Antarctica (dVA) and for the SSB/ASB sector (dVl), indicated by the green area (inset). With dTa=+4°C and
dTo=+1°C (a) the ice sheet margin is located near the A|B boundary and high basal velocities are focused in
region B. With dTa=+8°C and dTo=+2°C (b) the ice sheet margin is located in region B and high basal velocities
are focused in regions B1 and C. With dTa=+12°C and dTo=+5°C (c) the ice sheet margin is located near the
163
164
B|C boundary and high basal velocities are focused in region C and the ASB. Additional models are shown in
extended data figure 9.
165
166
167
168
169
Overall, SSB erosion points to large periods of time with a modern-scale ice sheet, with similar structure to
todays, and also large periods of time with a retreated ice sheet. Despite experiencing both regimes,
Subregion B1 is less eroded than both regions A and C, suggesting that less time has been spent transitional
between the two states. Thus, although Highland B is a significant barrier, retreat of the EAIS into the ASB is
necessary to explain our observed erosion, and is a characteristic of a fully retreated ice sheet.
170
171
172
173
174
There is no direct evidence in the erosion record for Pliocene ice-sheet retreat events at Totten Glacier. Icerafted detritus from ODP site 1165 at Prydz Bay1 includes a detrital signature that matches closely both the
detrital record in Totten-proximal sediment cores24 and the predicted basement geology of Region A6,25. This
suggests basement rocks were being eroded in Region A at 7 Ma and at 3.5 Ma1. Therefore, the modernscale ice sheet has been a regular feature of the EAIS at Totten Glacier since well before 7 Ma.
175
176
177
178
179
180
Nonetheless, a significant retreat here may be necessary to generate a sufficient EAIS contribution to
Pliocene sea-level highstands of up to 20m above present26. Our representative ice sheet models provide
guidance as to the contribution of this sector to global sea level, although it is important to note that
different ice sheet models will resolve these pinning points with differing amounts of sea level contribution
27428
. In addition, transient effects, including the lag-time between ice removal and isostatic response, may
generate short-term sea level fluctuations that we do not consider here27.
181
182
183
184
With the bi-stable EAIS configurations we have defined, any Pliocene retreat event was likely restricted to
within Region A (< 150 km retreat from present) or involved over 350 km of retreat. A stable long-term ice
sheet margin in the intervening zone is less likely, excepting transient pinning at Highland B. Numerical
models suggest this transient period in front of Highland B may persist for several thousand years21.
185
186
187
188
The largest retreat under the modern-scale ice sheet configuration (Figure 3a) is associated with a total
Antarctic contribution to global sea level of 8.39 m, of which the SSB/ASB sector provides 0.89 m. Allowing
for 7.3 m of sea level rise obtained from complete collapse of the Greenland ice sheet29, this is insufficient to
explain a 20m highstand.
189
190
191
192
The retreated ice sheet models (Figures 3b and 3c) are associated with a total Antarctic contribution to
global sea level of 16.5 m and 21.5 m respectively, of which the SSB/ASB sector provides 2.18 m and 2.89 m
respectively. Full retreat into the ASB (extended data Figure 9b) is associated with a total Antarctic
contribution of 29.1 m, with 4.29 m being sourced from the SSB/ASB sector.
193
194
195
196
197
198
199
Therefore, the influence of Totten Glacier instability on global sea level is clearly significant, but for any
particular warm period, it is also highly uncertain and subject to progressive instability. Our results suggest
that the first discriminant is the development of sufficient retreat to breach the A|B boundary ridge. This
causes an instability-driven transition from the modern-scale configuration to the retreated configuration.
Under ongoing retreat, the breaching of Highland B causes further collapse into the ASB. Each of these
changes in state is associated with a significant increase in both the absolute and proportional contribution
of this sector to global sea level.
200
METHODS * To be redone*
201
DATA
202
Collection, processing, imaging, availability
203
GRAVITY MODELLING
204
UNCERTAINTIES ASSOCIATED WITH GRAVITY MODELLING
205
ICE SHEET MODELLING
206
CODE AVAILABILITY
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
The ice sheet surface and base were derived from laser and radar sounding as described in earlier ICECAP
publications5, interpolated by kriging. Gravity modelling used the Free-Air gravity anomaly corrected for
latitude, instrument drift, elevation, aircraft kinematic accelerations and the Eotvös Effect. Modelling along
flight lines used an iterative combination of forward modelling and inversion using a commercial twodimensional code based upon the principles of Talwani30. This code can successfully model 1D and 2D
structures, but not complex 3D structures. Our approach is valid for the ice sheet surface and the SSSB base,
but some topographic features are three-dimensional in form and cause localised errors. Only relative
gravity differences can be modeled, and results were levelled to a common baseline. Cross-line differences
were reduced by applying a constant value to a tensioned spline fit through the basin base on each line so as
to minimise the misfit between that line and all cross-ties. Median cross line difference is 200 m, and the
arithmetic mean is 300 m, with the highest values associated with rugged topography, certain flight lines,
and the deepest basin regions. The central Sabrina Subglacial Basin provides consistently low cross-line
differences (Extended Data Fig. 2). We imposed the following density structure: Ice sheet density 920 kg m-3,
crystalline basement density 2670 kg m-3, lower crust density 2800 kg m-3, mantle density 3200 kg m-3.
Sedimentary rock density was varied between a lower limit of 2200 kg m-3 and an upper limit of 2500 kgm-3
to create the ranges of sediment thickness shown in Fig. 2c and extended figures 4b-7b. The former was
based on a reasonable maximum porosity of 25% given the thickness of ice and likely age of the rocks; the
latter was defined by the limit where basin thickness departs considerably from magnetic depth to basement
estimates6. This uncertainty is the error shown in Figure 2. It varies linearly and predictably with density
contrast, and has no influence on the pattern of basin thickness, only the magnitude. The Moho and a 13 km
thick lower crust were defined by a flexural model that accounts for ice and topographic loads. An elastic
thickness of 25 km was used, although due to the long-wavelength of the loads involved little variation in
Moho structure is observed between 10 km and 50 km elastic thickness.
230
REFERENCES
231
232
233
234
235
236
237
238
239
240
241
242
243
1
2
3
4
5
6
Williams, T. et al. Evidence for iceberg armadas from East Antarctica in the Southern Ocean during
the late Miocene and early Pliocene. Earth and Planetary Science Letters 290, 351-361 (2010).
Patterson, M. O. et al. Orbital forcing of the East Antarctic ice sheet during the Pliocene and Early
Pleistocene. Nature Geosci 7, 841-847, doi:10.1038/ngeo2273 (2014).
Cook, C. P. et al. Dynamic behaviour of the East Antarctic ice sheet during Pliocene warmth. Nature
Geoscience 6, 765-769 (2013).
Pollard, D., DeConto, R. M. & Alley, R. B. Potential Antarctic Ice Sheet retreat driven by
hydrofracturing and ice cliff failure. Earth and Planetary Science Letters 412, 112-121,
doi:10.1016/j.epsl.2014.12.035 (2015).
Young, D. A. et al. A dynamic early East Antarctic Ice Sheet suggested by ice-covered fjord
landscapes. Nature 474, 72-75 (2011).
Aitken, A. R. A. et al. The subglacial geology of Wilkes Land, East Antarctica. Geophysical Research
Letters, 2014GL059405, doi:10.1002/2014GL059405 (2014).
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Pritchard, H. D., Arthern, R. J., Vaughan, D. G. & Edwards, L. A. Extensive dynamic thinning on the
margins of the Greenland and Antarctic ice sheets. Nature 461, 971-975 (2009).
Rignot, E., Jacobs, S., Mouginot, J. & Scheuchl, B. Ice Shelf Melting Around Antarctica. Science 341,
266-270, doi:10.1126/science.1235798 (2013).
Gwyther, D. E., Galton-Fenzi, B. K., Hunter, J. R. & Roberts, J. L. Simulated melt rates for the Totten
and Dalton ice shelves. Ocean Science 10, 267-279, doi:10.5194/os-10-267-2014 (2014).
Khazendar, A. et al. Observed thinning of Totten Glacier is linked to coastal polynya variability. Nat
Commun 4, doi:10.1038/ncomms3857 (2013).
Greenbaum, J. S. et al. Ocean access to a cavity beneath Totten Glacier in East Antarctica. Nature
Geosci 8, 294-298, doi:10.1038/ngeo2388 (2015).
Wright, A. P. et al. Evidence of a hydrological connection between the ice divide and ice sheet
margin in the Aurora Subglacial Basin, East Antarctica. Journal of Geophysical Research F: Earth
Surface 117 (2012).
Roberts, J. L. et al. Refined broad-scale sub-glacial morphology of Aurora Subglacial Basin, East
Antarctica derived by an ice-dynamics-based interpolation scheme. Cryosphere 5, 551-560 (2011).
Fretwell, P. et al. Bedmap2: Improved ice bed, surface and thickness datasets for Antarctica.
Cryosphere 7, 375-393 (2013).
Jamieson, S. S. R., Sugden, D. E. & Hulton, N. R. J. The evolution of the subglacial landscape of
Antarctica. Earth and Planetary Science Letters 293, 1-27,
doi:http://dx.doi.org/10.1016/j.epsl.2010.02.012 (2010).
Rose, K. C. et al. Early East Antarctic Ice Sheet growth recorded in the landscape of the Gamburtsev
Subglacial Mountains. Earth and Planetary Science Letters (2013).
Flower, B. P. & Kennett, J. P. The middle Miocene climatic transition: East Antarctic ice sheet
development, deep ocean circulation and global carbon cycling. Palaeogeography,
Palaeoclimatology, Palaeoecology 108, 537-555, doi:10.1016/0031-0182(94)90251-8 (1994).
Rignot, E., Mouginot, J. & Scheuchl, B. Ice flow of the antarctic ice sheet. Science 333, 1427-1430
(2011).
Gudmundsson, G. H. Ice-shelf buttressing and the stability of marine ice sheets. The Cryosphere 7,
647-655, doi:10.5194/tc-7-647-2013 (2013).
Gasson, E., DeConto, R. & Pollard, D. Antarctic bedrock topography uncertainty and ice sheet
stability. Geophysical Research Letters 42, 2015GL064322, doi:10.1002/2015GL064322 (2015).
Pollard, D., DeConto, R. M. & Alley, R. B. Potential Antarctic Ice Sheet retreat driven by
hydrofracturing and ice cliff failure. Earth and Planetary Science Letters 412, 112-121,
doi:http://dx.doi.org/10.1016/j.epsl.2014.12.035 (2015).
Pollard, D. & DeConto, R. M. Modelling West Antarctic Ice Sheet growth and collapse through the
past five million years. Nature 458, 329-332 (2009).
de Boer, B. et al. Simulating the Antarctic ice sheet in the late-Pliocene warm period: PLISMIP-ANT,
an ice-sheet model intercomparison project. The Cryosphere 9, 881-903, doi:10.5194/tc-9-881-2015
(2015).
Pierce, E. L. et al. Characterizing the sediment provenance of East Antarctica's weak underbelly: The
Aurora and Wilkes sub-glacial basins. Paleoceanography 26 (2011).
Aitken, A. R. A. et al. The Australo-Antarctic Columbia to Gondwana transition. Gondwana Research
(2015).
Miller, K. G. et al. High tide of the warm Pliocene: Implications of global sea level for Antarctic
deglaciation. Geology 40, 407-410, doi:10.1130/g32869.1 (2012).
Golledge, N. R. et al. The multi-millennial Antarctic commitment to future sea-level rise. Nature 526,
421-425, doi:10.1038/nature15706 (2015).
De Boer, B. et al. Simulating the Antarctic ice sheet in the late-Pliocene warm period: PLISMIP-ANT,
an ice-sheet model intercomparison project. Cryosphere 9, 881-903, doi:10.5194/tc-9-881-2015
(2015).
Alley, R. B. et al. History of the Greenland Ice Sheet: paleoclimatic insights. Quaternary Science
Reviews 29, 1728-1756, doi:http://dx.doi.org/10.1016/j.quascirev.2010.02.007 (2010).
Talwani, M. Computation with the help of a digital computer of magnetic anomalies caused by
bodies of arbitrary shape. Geophysics 30, 797-817 (1965).
294
295
296
297
29
298
ACKNOWLEDGEMENTS
299
300
301
302
303
304
305
ICECAP data collection was supported by National Science Foundation grant PLR-0733025, National
Aeronautics and Space Administration grants NNX09AR52G, NNG10HPO6C and NNX11AD33G (Operation Ice
Bridge and the American Recovery and Reinvestment Act), Australian Antarctic Division projects 3013 and
4077, NERC grant NE/D003733/1, the Jackson School of Geosciences and the Australian Governm.ent's
Cooperative Research Centres Programme through the Antarctic Climate & Ecosystems Cooperative
Research Centre (ACE CRC). Stewart Jamieson is thanked for his comments and for supplying model images
for comparison. This is UTIG contribution #.
306
AUTHOR CONTRIBUTIONS
307
308
309
310
ARAA undertook data processing, analysis and modelling and wrote the paper. JLR, TVO, MJS,DAY, DDB, JSG
coordinated the field program and/or undertook data collection, quality control and processing. N.R.G.
Conducted the ice sheet modelling. All Authors contributed significantly to interpretation, discussion and
manuscript preparation.
311
30
312
Extended Data
313
314
315
Extended Data Figure 1: Geophysical Data in the study region. Free-air gravity data (a) and magnetic
intensity (b); also showing interpreted basement faults6.
316
317
318
319
Extended Data Figure 2: Model intersection misfits for the gravity models. The dots show the tensioned
spline fit after levelling. Levelling applied a constant value to each line to achieve optimal least-squares fit to
the cross-ties.
320
321
322
323
Extended Data Figure 3: SSSB base morphology. This map also shows basement faults (thick dashed), and
the locations for line R06Ea and additional representative model profiles along Y07b (Extended Data Figure
4), R03Ea (Extended Data Figure 5), R08Eb (Extended Data Figure 6) and GL0092a (Extended Data Figure 7).
324
325
326
327
328
329
Extended Data Figure 4: Gravity model along flight line Y07b (location, Extended Data Figure 3). a)
Observed and calculated gravity data, including model components from ice and topography, the deep crust
and Moho, and the sedimentary basins. b) The model, also showing depth to magnetic basement estimates,
the tensioned spline fit, and cross-ties with other models. c) SSSB thickness and current surface ice velocity
(flipped). Average error in ice sheet velocity on this line is 8.4 m/yr18.
330
331
332
333
334
335
Extended Data Figure 5: Gravity model along flight line R03Ea (location, Extended Data Figure 3). Observed
and calculated gravity data, including model components from ice and topography, the deep crust and
Moho, and the sedimentary basins. b) The model, also showing depth to magnetic basement estimates, the
tensioned spline fit, and cross-ties with other models. c) SSSB thickness and current surface ice velocity
(flipped). Average error in ice sheet velocity on this line is 9.7 m/yr18. TAH - Terre-Adelie Highlands.
336
337
338
339
340
341
Extended Data Figure 6: Gravity model along flight line R08Eb (location, Extended Data Figure 3).
Observed and calculated gravity data, including model components from ice and topography, the deep crust
and Moho, and the sedimentary basins. b) The model, also showing depth to magnetic basement estimates,
the tensioned spline fit, and cross-ties with other models. c) SSSB thickness and current surface ice velocity
(flipped). Average error in ice sheet velocity on this line is 8.6 m/yr18. HC - Highland C.
342
343
344
345
346
347
Extended Data Figure 7: Gravity model along flight line GL0092a (location, Extended Data Figure 3).
Observed and calculated gravity data, including model components from ice and topography, the deep crust
and Moho, and the sedimentary basins. b) The model, also showing depth to magnetic basement estimates,
the tensioned spline fit, and cross-ties with other models. c) SSSB thickness and current surface ice velocity
(flipped). Average error in ice sheet velocity on this line is 10.6 m/yr18. ASB - Aurora Subglacial Basin.
348
349
350
351
352
353
354
Extended Data Figure 8: Gravity models including basement density variations. a) line R03Ea, b) line
R06Ea, c) line R08Eb. The models involve an attempt to generate the flattest basin topography with blockscale density contrasts. Long-dashed dashed lines indicate major faults. The dotted line indicates the basin
structure with a homogenous basement density of 2670 kg/m3. Blocks are changed within 25 kg/m3 of this
value. Differences are substantial in places but the overall pattern is preserved.
355
356
357
358
359
360
361
Extended Data Figure 9: Representative ice sheet models for retreat states. Each image represents the ice
sheet extent and thickness after a long-term 20 ky run under consistent climate forcing. Each model has
different air/ocean temperatures. a) with air/ocean temperatures of today (0/0°C), b) with air/ocean
temperatures of 4/1°C above today's, c) with air/ocean temperatures of 8/2°C above today's, d) with
air/ocean temperatures of 12/5°C above today's, e) with air/ocean temperatures of 15/5°C above today's.
Stated values indicate the total Antarctic contribution and the component from western Wilkes Land.