Millennial-Scale Variability of a Major East Antarctic Outlet Glacier During the Last Glaciation Michelle Guitard University of South Florida, Mguitard@Mail.Usf.Edu

University of South Florida Scholar Commons

Graduate Theses and Dissertations Graduate School

10-19-2015 Millennial-scale Variability of a Major East Antarctic Outlet Glacier during the Last Glaciation Michelle Guitard University of South Florida, [email protected]

Follow this and additional works at: http://scholarcommons.usf.edu/etd Part of the Geology Commons

Scholar Commons Citation Guitard, Michelle, "Millennial-scale Variability of a Major East Antarctic Outlet Glacier during the Last Glaciation" (2015). Graduate Theses and Dissertations. http://scholarcommons.usf.edu/etd/5957

This Thesis is brought to you for free and open access by the Graduate School at Scholar Commons. It has been accepted for inclusion in Graduate Theses and Dissertations by an authorized administrator of Scholar Commons. For more information, please contact [email protected].

Millennial-scale Variability of a Major East Antarctic Outlet Glacier during the Last

Glaciation

Michelle E. Guitard

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Marine Science with a concentration in Geological Oceanography College of Marine Science University of South Florida

Major Professor: Amelia E. Shevenell, Ph.D. Eugene W. Domack, Ph.D. Brad E. Rosenheim, Ph.D.

Date of Approval: November 3, 2015

Keywords: Amery Ice Shelf, radiocarbon, cosmogenic beryllium-10, millennial-scale variability

ACKNOWLEDGMENTS

I want to say a huge thank you to those who helped and supported me through my

Masters degree. Thank you to my major advisor, Amelia Shevenell, who works so hard for her students. She inspires me to work hard and to always keep learning. Thank you to my committee members, Brad Rosenheim and Eugene Domack, for their insight and guidance on this project.

Thank you also to Amy Leventer at Colgate University, and Yusuke Yokoyama and his lab at the

Atmosphere and Ocean Research Institute, University of Tokyo for their mentorship in the laboratory and in the field. Thank you to the Australian researchers who have devoted their time to investigating the Prydz Bay region and for their scientific collaboration. Thank you to the crew and support staff aboard the Research Vessels Nathaniel B. Palmer and Aurora Australis for their assistance on cruises NBP01-01, AA-186, and KROCK. Thank you to the faculty, staff and students of the Paleo Lab at the College of Marine Science for making the lab experience a great one. Last, but certainly not least, thank you to my friends and family who have supported me through the triumphs and the tears. Your love and support mean the world to me.

TABLE OF CONTENTS

List of Tables ...... iii

List of Figures ...... iv

Abstract ...... v

Introduction ...... 1 Background ...... 1 Study Location ...... 2

Methods ...... 8 Physical Properties Analyses ...... 8 Development of Marine Composite Depth Scale ...... 8 Ramped Pyrolysis Preparation for Radiocarbon Analysis ...... 8 Radiocarbon Analysis, Correction and Calibration ...... 9 Beryllium-10 Preparation ...... 9

Results and Discussion ...... 11 Physical Properties ...... 11 Radiocarbon ...... 11 Beryllium-10 ...... 12 Regional Glacial and Oceanic Fluctuations: Timing and Variability ...... 14

Conclusions ...... 17

References ...... 18

Appendix A: Extended Data ...... 22 Extended Data Figure 1 ...... 22 Extended Data Figure 2 ...... 23 Extended Data Table 1 ...... 24 Extended Data Table 2 ...... 26 Extended Data Table 3 ...... 27 Extended Data Table 4 ...... 28

Appendix B: Supplementary Methods ...... 29 Radiocarbon Contamination Correction ...... 29 Calibration ...... 31 Beryllium-10 Concentration Corrections ...... 31

! i Beryllium-10 Budget Estimates ...... 32

Appendix C: Supplementary Discussion ...... 34 Diatom Assemblages ...... 34

Appendices References ...... 36

! ii

LIST OF TABLES

Extended Data Table 1: Percent diatom abundance (%abundance) from NBP01-01 JPC-34, -35, and -36 SMO units ...... 24

Extended Data Table 2: Prydz Channel 14C and cal BP ages, δ 13C, and %TOC ...... 26

Extended Data Table 3: NBP01-01 ramped pyrolysis 14C ages, RP temperature range, uncorrected and corrected Fm values ...... 27

Extended Data Table 4: Aliquot weight, uncorrected and corrected 10Be abundance with associated error, and flux for NBP01-01 cores ...... 28

! !

! iii

LIST OF FIGURES

Figure 1: Prydz Channel, East Antarctica, depicting multibeam swath bathymetry, core locations ...... 5

Figure 2: NBP01-01 core stratigraphy, magnetic susceptibility (MS), % diatom abundance, 10Be concentration ([10Be]) alongside Prydz Channel composite MS and radiocarbon (14C) ages ...... 6

Figure 3: Improved acid insoluble organic matter (AIOM) radiocarbon (14C) ages derived from the ramped pyrolysis (RP) technique ...... 7

Figure 4: Oceanic and climatic millennial-scale variability ...... 16

Extended Data Figure 1: Down-section NBP01-01, AA-186, and KROCK core lithology and normalized magnetic susceptibility data ...... 22

Extended Data Figure 2: Uncorrected Beryllium-10 concentration (10Be), Total Organic Carbon (TOC), and δ13C for NBP01-01 cores ...... 23

ABSTRACT

Ongoing retreat of Antarctica’s marine-based glaciers is associated with warm (~2° C) modified Circumpolar Deep Water intrusion onto the continental shelf, suggesting that Southern

Ocean temperatures may influence Antarctic ice sheet stability. Understanding past cryosphere response to environmental forcing is crucial to modeling future ice sheet behavior. Of particular interest is the response of the East Antarctic Ice Sheet (EAIS), which stands to contribute ~20 m to global sea level. However, marine sediment sequences recording timing and variability of

EAIS fluctuations through the last major climate shift, the Last Glacial Maximum (LGM), are either missing from the margin or have poor chronological control. Here we present three marine sediment cores that contain a record of pre-LGM fluctuations of the marine-based Lambert

Glacier-Amery Ice Shelf (LG-AIS) system into Prydz Channel, East Antarctica. Analyses of core lithology, physical properties, cosmogenic nuclide concentration and diatom assemblage demonstrate that Prydz Channel was characterized by alternating open-marine and sub-shelf deposition, implying repeated LG-AIS fluctuations through the LGM. Our radiocarbon chronology demonstrates that LG-AIS fluctuations occurred on millennial timescales. Our record corroborates regional marine and terrestrial records, which demonstrate millennial scale variability in Antarctic Circumpolar Current strength, ice-rafted debris deposition, sea ice extent,

Antarctic atmospheric temperature, and Southern Ocean sea surface temperature. This evidence suggests that the EAIS was sensitive to sub-orbital climate forcing in the past, and has implications for modeling future EAIS behavior.

! v

INTRODUCTION

Background

The response of Antarctica’s ice sheets to ongoing atmosphere and ocean warming is the largest uncertainty in future sea level rise projections. Whereas much has been learned about the impact of warming on the marine-based West Antarctic Ice Sheet (WAIS)1,2, recent observations of East Antarctic Ice Sheet (EAIS) outlet glacier retreat suggest that this ice sheet, with its 19.2 meters of marine-based sea level equivalent, may also respond sensitively to abrupt climate perturbations, including the warm ocean waters presently destabilizing the WAIS3. To reduce uncertainty surrounding future sea-level rise, predictive models require reliable ice-proximal observations of the orbital to millennial-scale EAIS response to past climate change.

Models and observations highlight the sensitivity of Antarctica’s ice sheets to abrupt oceanic perturbations (e.g. sea level rise and/or ocean warming) in regions of enhanced glacial flow2. Models that prescribe abrupt warmings and/or sea level rise during the last deglaciation reveal maximum ice flow and discharge rates via outlet glaciers that drain into the Weddell Sea,

Amundsen Sea, Ross Sea, and Prydz Bay and suggest that abrupt oceanic forcings propagate rapidly through associated glacial catchments4. Chronologically constrained observations of late

Quaternary (0-40 ka) EAIS variability are limited, but terrestrial records suggest episodes of thinning and retreat between 40 and 30 ka in catchments associated with Weddell Sea5 and Prydz

Bay6,7 marine-terminating outlet glaciers as well as during millennial-scale Antarctic Isotopic

Maxima (AIM) events, a proxy for atmospheric temperature8. Ice distal Southern Ocean marine sediment records reveal millennial-scale episodes of warming and IRD deposition, which suggest

! 1 either dynamic ice sheet responses to ocean perturbations and/or fluctuations in seasonal sea ice extent9.

Because the EAIS extended across the continental shelf in most locations during the last glaciation, preserved late Quaternary sediments are rare and dating these sequences is complex.

Thus, sub-orbital scale ice sheet variability is inferred from far-field marine records. Due to these limitations, existing studies of past outlet glacier behavior and forcings focus on understanding the timing and rates of retreat during the last deglaciation10,11 by radiocarbon (14C) dating the contact between the sub-glacial diamictites and overlying glacial marine sediments. This approach yields important paleoclimatic insights and is straightforward if sufficient marine calcium carbonate (CaCO3) is present, there is minimal sediment reworking, and a local reservoir age can be constrained through time. However, the ice-proximal Antarctic margin is a complex depositional environment; marine sediments are CaCO3-poor, reworking is common, and reservoir ages fluctuate. Bulk sedimentary acid insoluble organic matter (AIOM) 14C ages, used in the absence of CaCO3, often overestimate depositional age because they incorporate a matrix of autochthonous marine- and glacially-derived allochthonous organic carbon that encompasses a spectrum of ages12.

Study Location

Here we present the first accurately-dated detailed late Quaternary (~40-0 ka) record of

EAIS outlet glacier variability. This record, from Prydz Channel, in Prydz Bay, East Antarctica, fills an important ice-proximal gap between terrestrial Antarctic and Southern Ocean records and enables direct investigation of late Quaternary EAIS variability. Within Prydz Channel, a 150 km wide and 700 m deep cross-shelf trough, a set of elongate streamlined ridges trend parallel to the channel axis. These ridges extend the length (~60 km) and width (40-150 km) of the central

! 2 trough, with a 0.5 to 2.5 km spacing from crest-to-crest (Fig. 1). Ridge relief is most pronounced at shallower depths (<500 m) near the shelf edge and muted landward in the deeper portion of

Prydz Channel. These ridges are interpreted as mega-scale glacial lineations (MSGLs). Large grounding zone wedges (GZWs) delineate Prydz Channel to the southwest, south, and southeast and are characterized by oblique fluting (Fig 1). Glacial landforms within Prydz Channel record the Pleistocene history of streaming ice and the retreat of the LG-AIS system. The MSGLs indicate that a fast, northwest flowing paleo ice stream occupied Prydz Channel The length and continuous nature of these MSGLs suggest rapid retreat of the LG-AIS system via ice-bed decoupling and iceberg calving. One clue to the timing of retreat comes from the Prydz Channel trough-mouth fan, where the transition from debrites to hemipelagic sediments dates to ~780 ka13.

The angle of Prydz Channel MSGLs relative to the angle of flutes Prydz Channel GZWs implies that MSGL formation predates the most recent deglaciation14.

Today, the marine-based Lambert Glacier-Amery Ice Shelf (LG-AIS) outlet glacier system, which drains 16% of the EAIS15, is situated at the head of Prydz Channel. The Amery

Ice Shelf, the largest embayed ice shelf in East Antarctica, is presently grounded 2500 m below sea level and experiences basal melt (~2 m ice/yr) due to the seasonal influx of modified

Circumpolar Deep Water (mCDW), which enters the ice shelf cavity on the eastern side of Prydz

Channel and exists on the western side as Ice Shelf Water16. Clockwise gyral circulation recognized in Prydz Channel and eastern Prydz Bay17,18 draws mCDW onto the shelf to compensate for dense shelf water outflow; this influx is particularly strong during the austral autumn and early winter. Regional mCDW presence is seasonally and inter-annually variable as it is associated with eddy flux and regional polynya dynamics19.

! 3 We investigate late Quaternary LG-AIS dynamics using three marine sediment cores collected along a southeast-northwest transect in inner Prydz Channel during United States

Antarctic Program cruise NBP01-01: JPC-34 (68°15.04’S, 72°43.82’E; 754 m), JPC-35

(68°17.13’S, 72°28.45’E; 854 m), and JPC-36 (68°03.94’S, 72°16.38’E; 752 m) (Fig. 1).

NBP01-01 JPC-34 and JPC-35 were collected from a ~10m hemipelagic to glacial marine sediment drape ~50 km seaward of a series of prominent grounding zone wedges, which delineate Prydz Channel; JPC-36 was collected 30 km farther north, towards the shelf edge. All cores exhibit a consistent sequence of up to four distinct siliceous mud and ooze (SMO) units interbedded with sandy muds and granulated sediments (Fig. 2; Extended Data Fig. 1). Initial 14C ages derived from bulk AIOM and foraminiferal calcite suggested this sequence contained a late

Quaternary record of LG-AIS dynamics20 (Fig 3). However, old to infinite 14C ages20 suggested either a low resolution sequence of limited paleoenvironmental utility or substantial contamination of the bulk AIOM 14C ages by pre-aged allochthonous organic carbon12.

! 4 a 72°E 73°E

mCDW

-500m 67° ISW

Prydz Channel

-500m

GC-22

GC-09 68° JPC-36 GC-24 GC-20 ISW JPC-34 JPC-35

Amery 69°S Ice Shelf

N mCDW 100 km

480 600 840 m

b 72°30'E 73°00'E

JPC-36 GC-09 GC-24 Ice flow GZW Depth (mbsf) GC-20 Prydz Channel

68°15'S m 480 JPC-34 600 JPC-35

GZW

N GZW

10 km 840

Figure 1. Prydz Channel, East Antarctica, depicting multibeam swath bathymetry, core locations. a b, a b S N

NBP01-01 NBP01-01 NBP01-01 Prydz Channel Chaetoceros F. curta and other T. antarctica E. antarctica Extinct JPC-34 JPC 35 JPC 34 JPC 36 mag. sus. stack Fragilariopsis spp. 10Be abundance 0 0 SMO-1 4 JPC-34 4 JPC-35 1 4 1 JPC-36 c 0.5 3

2 Depth (mcd) 3 2 SMO-2 1.5

3 2

3 3 2 2.5 4 0 5 10 15 20 25

Depth (mcd) 10 8 1 Be x 10 (atoms/g)

Depth (mbsf) 4 2 5 1 SMO-3 0 10 20

5 6 0 10 20 SMO-4

1 Magnetic susceptibility (10-5 cgs) 7 6 0 4 5 25 45 65 5 25 45 65 5 25 45 65 5 25 45 65 5 25 45 65 0 10 20 0 10 Normalized 10Be x 108 magnetic susceptibility (atoms g-1) 8 % Diatom abundance

Legend Foraminifer 14C siliceous mud and ooze (SMO) Bulk AIOM 14C sandy mud granulated texture RP AIOM 14C sandy clay granule-rich pebbly muddy sand

1 Core section

Figure 2. NBP01-01 core stratigraphy, magnetic susceptibility (MS), % diatom abundance, 10Be concentration ([10Be]) alongside Prydz Channel composite MS and radiocarbon ( 14C) ages. ab c

a Age (14C kyr) b Age (14C kyr) c Age (14C kyr)

0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50 0 NBP01-01 low RP AIOM NBP01-01 high RP AIOM 1,880 NBP01-01 bulk AIOM AA-186 mixed foraminifera 1 AA-186 and KROCK bulk AIOM

2 6,830

3 Depth (mcd)

10,490

9,410 6

Figure 3. Improved acid insoluble organic matter (AIOM) radiocarbon (14C) ages derived from the ramped pyrolysis (RP) technique. a, b c

METHODS

Physical Properties Analyses

To supplement lithologic analyses, we measured %Total Organic Carbon (TOC) and δ13C measurements. Measurements were made on a Carlo-Erba 2500 series II elemental analyzer

(EA) at the University of South Florida, College of Marine Science. Diatom counts were performed for all SMOs to determine past environmental conditions (ie. open-marine, glacial- marine, ice-proximal) in Prydz Channel (Extended Data Table 1, Supplementary Methods). We employed magnetic susceptibility as a proxy for sediment provenance20; NBP01-01 magnetic susceptibility was measured on a multi-sensor track at the Antarctic Marine Geology Research

Facility, Florida State University.

Development of Marine Composite Depth Scale

Individual Prydz Channel magnetic susceptibility (MS) records from NBP01-01, AA-186 and AA-KROCK were converted to CGS units, normalized, then stacked to create a composite depth scale. JPC-34 was selected as the reference core because it contains the most complete sequence record (Extended Data Fig. 1). Core depths were tied to magnetic high, lows, and clear transition points on the JPC-34 MS record, then converted from meters below sea floor (mbsf) to marine composite depth (mcd) using the computer program Analyseries.

Ramped Pyrolysis Preparation for Radiocarbon Analysis

To establish chronologic control in the Prydz Channel sequence, we generated 38 ramped pyrolysis (RP)-based radiocarbon (14C) ages derived from 12 acid insoluble organic matter

! 8 (AIOM) samples. We selected samples for dating from SMO units, from SMO-sandy mud transitions, and from units previously dated with bulk AIOM (Extended Data Table 2). Samples for RP analysis were dried, homogenized and treated with 2N HCl for 24 hours to remove any carbonate. Samples were centrifuged then decanted and rinsed with filtered water until a pH of

6.0-7.0 was achieved. Aliquots were dried at 50° C, then weighed. Samples underwent the RP method as described by Rosenheim et al.21 at the University of South Florida, College of Marine

Science. Because Prydz Channel samples are organically lean, we pyrolyzed ~200-570 mg of

14 sediment during each run to collect sufficient CO2 for C analysis. Four to five CO2 aliquots with a target concentration of 10-15 µmol C were collected; this concentration is optimal for limiting the incorporation of 14C-dead material in the youngest organic carbon fraction while maximizing Accelerator Mass Spectrometry (AMS) analytical precision22.

Radiocarbon Analysis, Correction and Calibration

Carbon dioxide aliquots were analyzed for 14C by AMS at Lawrence Livermore National

Laboratory. Fraction modern values were blank corrected offline; modern and dead 14C contributions from both the ramped pyrolysis technique23 and graphitization were included in the correction. Low temperature RP-based 14C ages were converted to calendar age using the

Marine13 calibration curve24 with a constant regional reservoir correction (1300±100 years;

Supplementary Methods).

Beryllium-10 Preparation

To identify LG-AIS grounding line and calving front fluctuations during the last glaciation, we extracted and measured cosmogenic beryllium-10 concentrations (hereafter denoted as [10Be]; half life: 1.387±0.012 Ma) in 23 sediment samples by AMS at the University of Tokyo. Beryllium-10 enters the Antarctic margin water column by direct atmospheric fallout, ! 9 upwelling of intermediate to deep waters, and glacial meltwater input; is scavenged from the water column by biogenic opal, clays, and Fe-Mn oxyhydroxides; and is deposited in glacial marine sediments25. We generated a meteoric 10Be record from 23 samples in Prydz Channel. We targeted SMOs and SMO to glacial-marine transitions to investigate how 10Be concentration fluctuated across facies previously interpreted as open-marine, glacial-marine, or ice-proximal to determine any correlation between 10Be concentration and lithology. Sediment samples were freeze-dried then homogenized using a ceramic mortar and pestle. Approximately 3000 mg of

9Be-carrier (100ppm) was added to 100 mg aliquots of dried sediment, which were then prepared according to Suganuma et al.26 Samples were analyzed for 10Be by AMS analysis at the

University of Tokyo, Micro Analysis Laboratory Tandem accelerator (MALT). 10Be counts were corrected for 10Be radioactive decay; pre-Holocene (>15 ka) intervals were also corrected for the

40% increase over modern production in 10Be production during the Pleistocene (Supplementary

Methods). Sample depths were converted to mcd for analysis alongside Prydz Channel lithology and 14C ages. To maximize the number of intervals analyzed, we did not run duplicates for any one sample; instead we ran multiple samples within various corresponding intervals from JPC34-

36 to investigate instrument error. We examine the effects of various depositional pathways on

10Be concentration in Prydz Channel sediments by analyzing down-section [10Be] in conjunction with estimates of 10Be input via LG-AIS melting, sample grain size, %TOC, δ13C, and diatom assemblage (Supplementary Methods).

! 10

RESULTS AND DISCUSSION

Physical Properties

Prydz Channel jumbo piston cores (JPC-34, -35, and -36) exhibit a sequence of four distinct hemipelagic units interbedded with sandy muds, and granulated sediments that are characterized by down-section variations in physical properties, including magnetic susceptibility (MS), Total

Organic Carbon (%TOC), clay percentage (%clay) and diatom abundance (%abundance).

Granulated units exhibit relatively high MS and low %clay and %TOC, while transitional silty and sandy muds units demonstrate lower MS, and high %clay and %TOC. SMO units are identified as magnetic lows, and they exhibit mid-range %clay and %TOC values (Fig. 2;

Extended Data Table 2, Extended Data Fig. 2).

Radiocarbon

Our 14C results indicate that the upper 7m of Prydz Channel sediments were deposited between ~31 14C kyr and present (Fig. 3). At each sample depth, the spectrum of RP-based 14C ages indicates that the youngest age (range: 5.2 to 30.8 14C kyr) derives from the lowest temperature aliquot and the oldest age (range: 7.1 to >40.2 14C kyr) from higher temperature aliquots; paired high-temperature RP and bulk AIOM 14C ages are similar, within error (Fig. 3;

Extended Data Tables 2, 3). Therefore, low-temperature RP ages, derived from the most thermochemically-reactive organic carbon, accurately reflect the maximum depositional age of the sediment. Offsets between low- and higher temperature ages range between 1.2 to 14.0 14C kyr and vary down-section between different lithofacies. The largest and smallest offsets are associated with glacial-marine sediments (average: 12.0 14C kyr) and diatom-rich SMOs ! 11 (average: 7.7 14C kyr), respectively (Fig 3). This observation is consistent with the expectation that glacial-marine sediments contain more pre-aged allochthonous organic carbon than hemipelagic marine sediments (e.g. SMOs); low-temperature glacial-marine 14C variability likely reflects changes in the source and/or proportion of diagenetically stable organic carbon incorporated in the lowest temperature CO2 aliquot. Our results suggest that the RP-method improves the Prydz Channel chronology by up to 27 14C kyr when compared with bulk AIOM ages; a larger improvement from the other sediments dated with this method21,22,27.

RP-based 14C age-control enables us to establish LG-AIS grounding line retreat landward of the Prydz Channel sites by at least ~40 ka, The timing of initial grounding line retreat is consistent with terrestrial and lacustrine records from Dronning Maud Land28 to the Bunger

Hills29, each of which suggest a retracted EAIS margin and thinning between 40 and 30 ka, when

Southern Hemisphere insolation was at its lowest level of the last 100 ka and Marine Isotope

Stage (MIS) 3 sea levels were variable. In Prydz Channel, massive diamictites, indicative of grounded ice, are only observed landward of grounding zone wedges20, suggesting limited LG-

AIS re-advance during MIS2.

Beryllium-10

Radioactive decay-corrected [10Be] (Supplementary Methods) in Prydz Channel sediments are above detection limits in all facies and higher than Antarctic subglacial sediments

(~106)30,31, confirming that the sequence was deposited in a sub-ice shelf to open marine environment after retreat of the LG-AIS system at ~40 ka. Granulated diamictites have been interpreted as grounding-line proximal units deposited during melting of the ice shelf basal debris layer. Beryllium-10 concentrations in these low total organic carbon (TOC; 0.1-0.3%) units are the lowest of any Prydz Channel lithology (1.20±0.07x108-3.97±0.10x108; average:

! 12 2.23±0.08x108; Fig. 2; Extended Data Table 4; Extended Data Fig. 2), confirming our sub-ice shelf grounding-line proximal interpretation. Clay-rich (≤70%), higher TOC (0.3-0.6%) sandy muds with trace diatom abundances have an average [10Be] of 3.97±0.10x108 atoms g-1, suggesting hemipelagic settling of glacially-derived fines distal to the grounding line and beneath a sediment-starved ice shelf with limited sub-ice shelf oceanic influx.

Four distinct diatom-rich (13-33x106 v/g; 25-50% of total) SMO units were deposited in

Prydz Channel at 5.2, 18.2, 22.1, and 30.8 14C kyr±121 to 3016 14C yr (4.4, 20.5, 25.0, and 33.6 ka ±223 to 3118 yr; Extended Data Table 2; Supplementary Methods). The surface SMO (SMO-

1) diatom abundance and assemblage are typical of Holocene-age seasonally open-marine sediments ubiquitous on Antarctica’s continental shelves32. The average SMO-1 [10Be]

(10.8±0.2x108 atoms g-1) is similar to that of modern Antarctic open-marine surface sediments

(13.5-17.7x109 atoms g-1)30; SMO-1 10Be flux (18.5x109 atoms cm-2 kyr-1) is an order of magnitude greater than Holocene production (1.21±0.26x109 atoms cm-2 ky-1)33, suggesting that

10Be is most effectively scavenged from the Prydz Channel water column during intervals of enhanced biogenic export associated with seasonally open-marine conditions. Decay-corrected

[10Be] in SMOs 2-4 (7.89±0.15 to 23.84±0.35x108 atoms g-1) are similar to SMO-1 and production-corrected 10Be fluxes are greater than or equal to atmospheric production, suggesting that these units also reflect seasonally open-marine conditions. Miocene to Pliocene-age diatoms and pre-aged glacial-marine sediments dilute the [10Be] in SMO-2 to -4, an interpretation further supported by larger RP 14C age differences compared to SMO-1. High-resolution analysis of

NBP01-01 JPC-35 SMO-2 indicates a progressive increase in [10Be] up-section and a reduction of sea-ice associated diatoms and increase in open ocean species. Whereas the possibility exists that SMOs were deposited in a sub-ice shelf setting via lateral advection, high total diatom

! 13 percentages and elevated [10Be] support the interpretation that SMO units reflect seasonally open water conditions in Prydz Channel during the last glaciation.

Regional Glacial and Oceanic Fluctuations: Timing and Variability

Open water conditions in Prydz Channel and associated retreat of the LG-AIS grounding line and calving front on millennial timescales during the last glaciation cannot be explained by orbital variations, indicating the potential for remote forcing of this major EAIS outlet glacier system through changes in atmospheric-oceanic circulation. Within the limits of the respective age models, inner Prydz Channel is ice free when thinning and retreat are observed in the LG-

AIS catchment6,34, a response predicted by high-resolution ice sheet models perturbed by oceanic forcings (e.g. temperature and/or sea level)4. SMO deposition is coincident, within error, with millennial-scale warming in many Antarctic ice cores (AIM events)8 and Southern Ocean surface waters35, reduced Southern Ocean sea ice extent36, increased Weddell Sea IRD37, and variations in ACC transport38; due to uncertainties associated with 14C dating and glacial reservoir age variability, we cannot establish the absolute phasing of SMO deposition, Southern Hemisphere

AIM events, and Northern Hemisphere Dansgaard-Oeschger events (Fig 4).

Because ACC-derived mCDW presently impacts the LG-AIS, we speculate that the observed millennial-scale variability in LG-AIS extent during the last glaciation is related to variations in the volume and/or intrusion of warm nutrient-rich mCDW in Prydz Channel.

Increased mCDW influence results in frontal and basal ice shelf ablation, increased calving, and recession of the LG-AIS system, as well as the acceleration and thinning of associated catchment glaciers, consistent with our marine and local terrestrial6 records. Models and observations indicate that Antarctic outlet glaciers are sensitive to changes in mCDW2,4, which may result from fluctuations in the strength/position of the Southern Hemisphere Westerly Winds39, changes

! 14 in thermohaline circulation associated with the bipolar seesaw, and/or global sea levels40.

Because modern and paleoclimate observations and models reveal regional oceanographic changes associated with the intensification/poleward shift of the Southern Hemisphere

Westerlies, including: a southward shift in the ACC41, increased eddy heat flux and ventilation rates42, an increase in warm mCDW on Antarctica’s continental shelves43, and increased advection of warm maritime air onto the Antarctic continent44, we suggest that changes in the strength/position of the Southern Hemisphere Westerly Winds resulted in millennial-scale variations in Prydz Channel mCDW during the last glaciation.

Changes in global sea level are also implicated as a destabilizing factor for Antarctic Ice

Sheets, particularly during the last deglaciation. Far-field reconstructions indicate large millennial-scale sea level variations associated with Antarctic warmings45, suggesting EAIS involvement. The Prydz Channel record is the first ice-proximal record of EAIS variability during the last glaciation and suggests that the LG-AIS system may have initially retreated during the global sea level highstand associated with the AIM7-8 (A1) warming8. During the

LGM, Prydz Channel SMO-3 and -4 may, within chronological constraints, be associated with

AIM2 and AIM3 warmings and highstands. Interestingly, SMO-2 seems to be associated with the onset of Antarctic warming and deglaciation (21-20 ka), as reflected in the increased IRD input to the South Atlantic37 (AID8; 19-20 ka), and an abrupt short-lived sea level rise at ~19 ka45. Whereas the Prydz Channel record provides the first indication of major EAIS outlet glacier variability during the last glaciation and the onset of deglaciation, its resolution does not allow us to establish the phasing of warming, ice retreat, and sea level rise.

! 15 Age (ka) 0 5 10 15 20 25 30 35 40

a -34 18

-36 O ( -38 ‰ S

NGRIP ice core M -40 (Greenland) O W -42 H2 H3 ) b 0 H1 H4 -44

-50

-100 MWP 19ka 16 sea level (m) c -150 Ice-volume equivalent

14 SS ODP 1233

(SE Pacific 41°S ) T (

12 °C ) d 10 MD07-3128 35 (SE Pacific 53°S ) 8 30 )

µm 25 CHC/northern ACC strength

SS ( 20 Lithic concentration

15 (x10 SA2 8

SA0 SA1 3 6 grains g SA3 e 4 TN057-13/1094 2 (SE Atlantic 53°S ) -1

0 ) 6 f

5 0.16 (grains cm

MD07-3133/3134 IBRD flux 4 0.12 3 (SW Atlantic 55-60°S ) 2 0.08 -2 1 0.04 yr -1

(relative to Holocene) 0 )

IBRD flux 500-yr average 0.00

g NBP01-01 JPC34-36 SMO-1 SMO-2 (Prydz Bay 68°S ) SMO-4 SMO-3 )

-1 3000 yr 2

2000 h

EDML ice core ( (Antarctica) 1000 flux +

0 nssCa

i -44 18 O (

EDML ice core -46 ‰ S M

-48 O

1 2 8 W 7 ) 0.5 j 3 4 5 6 -50 0 -0.5

Obliquity -1 -1.5

(°relative to modern) Holocene MIS 2 MIS 3 0 5 10 15 20 25 30 35 40 Age (ka) Figure 4. Oceanic and climatic millennial-scale variability. a bc de fg hi j

CONCLUSIONS

Output from high-resolution ice sheet models suggests that a combination of ocean heat flux and sea level drove millennial-scale LG-AIS variations during the last glaciation4. The

Prydz Channel sedimentary sequence provides the first marine evidence for millennial-scale variations of a major East Antarctic outlet glacier system during the last glaciation, with important implications for enhanced EAIS sensitivity to abrupt temperature and sea level changes anticipated with anthropogenic climate change. Recent increases in oceanic heat flux associated with rising atmospheric CO2 and the southward migration of the Southern Hemisphere

Westerlies46 are linked to recent Antarctic outlet glacier acceleration and ice sheet mass loss47.

On longer timescales, variations in the Southern Hemisphere Westerlies are linked to glacial- interglacial and millennial-scale variations in Southern Ocean circulation48 and temperature35, with important implications for carbon cycling49. Our Prydz Channel record provides a long-term perspective on EAIS sensitivity, suggesting that millennial-scale LG-AIS variations during the last glaciation may relate to changes in the supply of mCDW onto the Prydz Bay shelf driven by the strength/position of the westerlies. Our results suggest that chronologically constrained ice- proximal studies of outlet glacier systems will improve ice-sheet model parameterizations and advance understanding of climate forcings and feedbacks at the ocean-cryosphere interface.

Detailed paleotemperature and paleoventilation studies are required to confirm the hypothesized relationship between mCDW and SMO deposition and work is ongoing to improve East

Antarctic margin proxy-temperature calibrations.

! 17

REFERENCES

1 Rignot, E., Jacobs, S. S., Mouginot, J. & Scheuchl, B. Ice shelf melting around Antarctica. Sciencexpress, doi:10.1126/science.1235798 (2013). 2 Pritchard, H. D. et al. Antarctic ice-sheet loss driven by basal melting of ice shelves. Nature 484, 502-505, doi:10.1038/nature10968 (2012). 3 Vaughan, D. G. et al. Observations: Cryosphere. Climate Change 2013: The physical science basis. Contribution of the working group I to the Fifth Assessment Report of the Intergonvernmental Panel on Climate Change, 317-382 (2013). 4 Golledge, N., Fogwill, C., Mackintosh, A. & Buckley, K. M. Dynamics of the last glacial maximum Antarctic ice-sheet and its response to ocean forcing. Proc Natl Acad Sci 109, 16052-16056, doi:10.1073/pnas.1205385109/-/DCSupplemental (2012). 5 Anderson, J. B. & Andrews, J. T. Radiocarbon constraints on ice sheet advance and retreat in the Weddell Sea, Antarctica. Geology 27, 179-182 (1999). 6 White, D. A., Fink, D. & Gore, D. B. Cosmogenic nuclide evidence for enhanced sensitivity of an East Antarctic ice stream to change during the last deglaciation. Geology 39, 23-26, doi:10.1130/g31591.1 (2011). 7 Mackintosh, A. et al. Exposure ages from mountain dipsticks in Mac. Robertson Land, East Antarctica, indicate little change in ice-sheet thickness since the Last Glacial Maximum. Geology 35, 551, doi:10.1130/g23503a.1 (2007). 8 Members, E. C. et al. One-to-one coupling of glacial climate variability in Greenland and Antarctica. Nature 444, 195-198, doi:10.1038/nature05301 (2006). 9 Nielsen, S. H. & Hodell, D. A. Antarctic Ice-Rafted Detritus (IRD) in the South Atlantic: Indicators of Iceshelf Dynamics or Ocean Surface Conditions? A Keystone in a Changing World – Online Proceedings of the 10th ISAES, doi:10.3133/of2007-1047.srp020 (2007). 10 Mackintosh, A. et al. Retreat of the East Antarctic ice sheet during the last glacial termination. Nature Geoscience 4, 195-202, doi:10.1038/ngeo1061 (2011). 11 McKay, R. M. et al. Retreat history of the Ross Ice Sheet (Shelf) since the Last Glacial Maximum from deep-basin sediment cores around Ross Island. Palaeogeography, Palaeoclimatology, Palaeoecology 260, 245-261, doi:10.1016/j.palaeo.2007.08.015 (2008). 12 Andrews, J. T. et al. Problems and possible solutions concerning radiocarbon dating of surface marine sediments, Ross Sea, Antarctica. Quaternary Research 52, 206-216 (1999). 13 Passchier, S. et al. Pliocene–Pleistocene glaciomarine sedimentation in eastern Prydz Bay and development of the Prydz trough-mouth fan, ODP Sites 1166 and 1167, East Antarctica. Marine Geology 199, 279-305, doi:10.1016/s0025-3227(03)00160-9 (2003). 14 Guitard, M. E., Shevenell, A. E., Lavoie, C. & Domack, E. W. Mega-scale glacial lineations and grounding-zone wedges in Prydz Channel East Antarctica. Atlas of Submarine Glacial Landforms: Modern, Quaternary and Ancient, The Geological Society, London (accepted).

! 18 15 Allison, I. The mass budget of the Lambert Glacier drainage basin, Antarctica. Journal of Glaciology 22, 223-235 (1979). 16 Herraiz-Borreguero, L. et al. Circulation of modified circumpolar deep water and basal melt beneath the Amery Ice Shelf, East Antarctica. Journal of Geophysical Research: Oceans, n/a-n/a, doi:10.1002/2015jc010697 (2015). 17 Smith, N. R., Zhaoqian, D., Kerry, K. R. & Wright, S. Water masses and circulation in the region of Prydz Bay, Antarctica. Deep-Sea Research 31, 1121-1147 (1984). 18 Nunes Vaz, R. A. & Lennon, G. W. Physical Oceanography of the Pryz Bay region of Antarctic waters. Deep-Sea Research I 43, 603-641 (1996). 19 Oshima, K. et al. Antarctic Bottom Water production by intense sea-ice formation in the Cape Darnley polynya. Nature Geoscience 6, 235-240, doi:10.1038/ngeo1738 (2013). 20 Domack, E. et al. Late Quaternary sediment facies in Prydz Bay, East Antarctica and their relationship to glacial advance onto the continental shelf. Antarctic Science 10, 1445-1451, doi:10.1017/s0954102098000339 (1998). 21 Rosenheim, B. E. et al. Antarctic sediment chronology by programmed-temperature pyrolysis: Methodology and data treatment. Geochemistry, Geophysics, Geosystems 9, 1- 16, doi:10.1029/2007gc001816 (2008). 22 Subt, C., Fangman, K. A., Wellner, J. S. & Rosenheim, B. E. Sediment chronology in Antarctic deglacial sediments: Reconciling organic carbon 14C ages to carbonate 14C ages using Ramped PyrOx. The Holocene, doi:10.1177/0959683615608688 (2015). 23 Fernandez, A. et al. Blank corrections for ramped pyrolysis radiocarbon dating of sedimentary and soil organic carbon. Analytical chemistry 86, 12085-12092, doi:10.1021/ac502874j (2014). 24 Reimer, P. J. et al. IntCal and Marine13 radiocarbon age calibration curves 0-50,000 years cal BP. Radiocarbon 55, 1869-1887 (2013). 25 Aldahan, A. & Possnert, G. The 10Be marine record of the last 3.5 Ma. Nuclear Instruments and Methods in Physics Research B 172, 513-517 (2000). 26 Suganuma, Y. et al. 10Be evidence for delayed acquisition of remanent magnetization in marine sediments: Implication for a new age for the Matuyama–Brunhes boundary. Earth and Planetary Science Letters 296, 443-450, doi:10.1016/j.epsl.2010.05.031 (2010). 27 Rosenheim, B. E., Santoro, J. A., Gunter, M. & Domack, E. W. Improving Antarctic sediment 14C dating using ramped pyrolysis: an example from the Hugo Island Trough Radiocarbon 55, 115-126 (2013). 28 Hiller, A., Hermichen, W. D. & Wand, U. Radiocarbon-dated subfossil stomach oil deposits from petrel nesting sites: novel paleoenvironmental records from continental Antarctica. Radiocarbon 37, 171-180 (1995). 29 Gore, D. B. et al. Bunger Hills, East Antarctica: Ice free at the Last Glacial Maximum. Geology 29, 1103-1106 (2001). 30 Sjunneskog, C., Scherer, R., Aldahan, A. & Possnert, G. 10Be in glacial marine sediment of the Ross Sea, Antarctica, a potential tracer of depositional environment and sediment chronology. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 259, 576-583, doi:10.1016/j.nimb.2007.01.203 (2007). 31 Scherer, R. P. et al. Pleistocene Collapse of the West Antarctic Ice Sheet. Science 281, 82-85, doi:10.1126/science.281.5373.82 (1998).

! 19 32 Taylor, F. & Leventer, A. Late Quaternary palaeoenvironments in Prydz Bay, East Antarctica: interpretations from marine diatoms. Antarctic Science 15, 512-521, doi:10.1017/s0954102003001639 (2003). 33 Monaghan, M., Krishnaswami, S. & Turekian, K. The global-average production rate of 10Be. Earth and Planetary Science Letters 76, 279-287 (1985/86). 34 Hodgson, D. A. et al. Were the Larsemann Hills ice-free through the Last Glacial Maximum? Antarctic Science 13, 440-454, doi:10.1017/s0954102001000608 (2001). 35 Caniupán, M. et al. Millennial-scale sea surface temperature and Patagonian Ice Sheet changes off southernmost Chile (53°S) over the past ∼60 kyr. Paleoceanography 26, n/a- n/a, doi:10.1029/2010pa002049 (2011). 36 Collins, L. G., Pike, J., Allen, C. S. & Hodgson, D. A. High-resolution reconstruction of southwest Atlantic sea-ice and its role in the carbon cycle during marine isotope stages 3 and 2. Paleoceanography 27, n/a-n/a, doi:10.1029/2011pa002264 (2012). 37 Weber, M. E. et al. Millennial-scale variability in Antarctic ice-sheet discharge during the last deglaciation. Nature 510, 134-138, doi:10.1038/nature13397 (2014). 38 Lamy, F. et al. Glacial reduction and millennial-scale variations in Drake Passage throughflow. Proc Natl Acad Sci, doi:10.1073/pnas.1509203112 10.1594/ (2015). 39 Dinniman, M. S., Klinck, J. M. & Hofmann, E. E. Sensitivity of Circumpolar Deep Water Transport and Ice Shelf Basal Melt along the West Antarctic Peninsula to Changes in the Winds. Journal of Climate 25, 4799-4816, doi:10.1175/jcli-d-11-00307.1 (2012). 40 Siddall, M. et al. Sea-level fluctuations during the last glacial cycle. Nature 423, 853-858 (2003). 41 Gille, S. T. Warming of the Southern Ocean since the 1950s. Science 295, 1275-1277 (2002). 42 Meredith, M. P. & Hogg, A. M. Circumpolar response of Southern Ocean eddy activity to a change in the Southern Annular Mode. Geophysical Research Letters 33, doi:10.1029/2006gl026499 (2006). 43 Thoma, M., Jenkins, A., Holland, D. & Jacobs, S. Modelling Circumpolar Deep Water intrusions on the Amundsen Sea continental shelf, Antarctica. Geophysical Research Letters 35, doi:10.1029/2008gl034939 (2008). 44 Marshall, G. J., Orr, A., van Lipzig, N. P. M. & King, J. C. The impact of a changing Southern Hemisphere Annular Mode on Antarctic Peninsula summer temperatures. Journal of Climate 19, 5388-5404 (2006). 45 Siddall, M., Rohling, E. J., Thompson, W. G. & Waelbroeck, C. Marine isotope stage 3 sea level fluctuations: Data synthesis and new outlook. Reviews of Geophysics 46, doi:10.1029/2007rg000226 (2008). 46 Anderson, R. F. et al. Wind-driven upwelling in the Southern Ocean and the deglacial rise in atmospheric CO2. Science 323, 1443-1448, doi:10.1126/science.1167441 (2009). 47 Shepherd, A., Wingham, D. & Rignot, E. Warm ocean is eroding West Antarctic Ice Sheet. Geophysical Research Letters 31, 1-4, doi:10.1029/2004gl021106 (2004). 48 Muratli, J. M., Chase, Z., Mix, A. C. & McManus, J. Increased glacial-age ventilation of the Chilean margin by Antarctic Intermediate Water. Nature Geoscience 3, 23-26, doi:10.1038/ngeo715 (2009).

! 20 49 Chase, Z., Anderson, R. F. & Fleisher, M. Q. Evidence from authigenic uranium for increased productivity of the glacial subantarctic ocean. Paleoceanography 16, 468-478, doi:10.1029/2000pa000542 (2001).

! 21 APPENDIX A: EXTENDED DATA

W E

AA-186 NBP01-01 NBP01-01 NBP01-01 AA-186 AA-186 AA-KROCK GC 20 JPC 35 JPC 34 JPC 36 GC 22 GC 09 GC 24

0 !"

0 200 400 0 200

0 4 200 400 0 200 400 Depth (mbsf)

5 0 10 20 siliceous mud and ooze (SMO)

sandy mud granulated texture 6 0 10 20 sandy clay granule-rich pebbly muddy sand

diamicton Magnetic susceptibility (10-5 cgs) 7

0 10

Extended Data Figure 1. Down-section NBP01-01, AA-186, and KROCK core lithology and normalized magnetic susceptibility data. Siliceous mud and ooze (SMO) units are olive green to pale olive in color. Mud and silty units are olive grey in color with some yellow spotting. Granulated units are olive grey with some reddish-brown material present deeper in the stratigraphy. Low and high magnetic susceptibility corresponds to SMO and granulated units, respectively. Similar lithologic and magnetic variability represent a regional sedimentary sequence, which suggests repeated advance and retreat of the adjacent Lambert Glacier-Amery Ice Shelf (LG-AIS).

22 10 Extended DataFigure 2.Uncorrected Beryllium-10concentration( Be abundanceuncorrectedforagedecay andincreasedrelative

JPC 36 JPC 35 JPC 34 Depth (mbsf) Depth (mbsf) Depth (mbsf) 13 5 4 3 2 1 0 8 7 6 5 4 3 2 6 5 4 3 2 1 0 13 C values.SMOunitscontainthehighest C valuesco-varyandaregenerallyobserved inaSMO-proximalsandymudunitallcores. x10 5 8 10 atoms g Be 15 -1

25 10 Be production.Granulatedunitscontain thelowest 10 10 TOC 0.4 % 0.8 -24.0 ‰ 13 C 13 C for NBP01-01cores 10 Be abundanceand -26.0 13 C valuesin . 23 Extended Data Table 1 Percent diatom abundance (%abundance) from NBP01-01 JPC-34, -35, and -36 SMO units; average %abundance for each species included. Assemblage shifts down-section from an assemblage indicating open-marine/seasonal sea-ice conditions, to one containing robust or extinct taxa (4.9 Ma; 6.4–8.5 to 9.0 Ma), indicating ice-proximal or allochthonous deposition.

SMO-1 (%) SMO-2 (%) SMO-3 (%) SMO-4 (%) NBP01-01 JPC 34 Depth (cm) 1 8 19 Avg. 197 205 211 Avg. 427 430 432 Avg. 520 530 540 Avg. Chaetoceros - Hyalochaete 8.4 7.0 20.4 11.9 12.9 17.1 11.8 13.9 4.8 2.7 2.2 3.2 7.2 4.3 11.2 7.6 Chaetoceros - Phaeoceros 2.9 2.2 5.4 3.5 7.2 4.0 6.8 6.0 0.7 1.4 2.6 1.6 1.9 1.2 1.8 1.6 Fragilariopsis curta 47.7 46.2 14.0 36.0 19.1 16.8 19.8 18.6 0.9 2.2 1.8 1.6 2.8 0.2 0.7 1.2 Fragilariopsis kerguelensis 0.9 1.5 1.7 1.4 9.0 9.5 10.6 9.7 15.0 10.6 6.2 10.6 2.8 1.7 3.4 2.6 Eucampia antarctica var recta 0.9 2.6 3.9 2.5 8.0 12.6 12.1 10.9 18.2 18.4 15.4 17.3 44.0 73.4 35.2 50.9 Thalassiosira antarctica var antarctica 13.2 18.2 40.8 24.1 5.6 12.8 7.7 8.7 19.2 25.8 39.6 28.2 10.4 5.5 30.4 15.4 Dactyliosolen antarcticus 0.4 0.2 0.0 0.2 0.8 3.1 4.3 2.7 6.0 2.7 2.4 3.7 0.0 1.0 0.7 0.6 Fragilariopsis angulata 5.1 2.4 1.0 2.8 2.4 0.5 0.5 1.1 0.7 0.0 0.0 0.2 0.0 0.2 0.0 0.1 Fragilariopsis obliquecostata 2.0 1.2 1.5 1.6 1.2 0.9 1.4 1.2 2.8 0.0 0.0 0.9 0.2 0.7 0.5 0.5 Fragilariopsis separada 0.7 1.0 0.7 0.8 2.6 0.2 1.0 1.3 0.2 0.2 0.0 0.1 0.0 0.0 0.0 0.0 Fragilariopsis sublineata 4.4 3.1 0.5 2.7 5.4 0.7 6.0 4.0 0.5 0.7 0.4 0.5 1.4 0.2 0.2 0.6 Fragilariopsis van huerckii 2.2 0.5 0.2 1.0 1.4 1.7 1.0 1.4 0.2 0.0 0.0 0.1 0.0 0.0 0.0 0.0 Porosira glacialis 0.9 3.1 2.9 2.3 0.8 1.2 1.4 1.1 1.2 1.9 2.2 1.8 1.2 0.5 3.0 1.6 Rhizosolena spp. 0.2 1.0 0.5 0.6 1.0 0.0 1.0 0.7 2.3 0.5 2.0 1.6 0.9 0.0 0.2 0.4 Stellarima microtrias 1.3 1.0 0.5 0.9 3.4 2.1 0.0 1.8 2.1 2.7 2.0 2.3 2.1 0.1 0.0 0.7 Thalassiosira gracilis 0.4 1.2 1.2 0.9 2.6 1.4 1.7 1.9 0.9 0.0 0.6 0.5 0.7 0.0 0.2 0.3 Thalassiosira lentiginosa 0.9 1.7 0.5 1.0 7.8 5.7 4.8 6.1 3.0 3.1 1.8 2.6 1.9 1.7 0.9 1.5 Thalassiosira torokina 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 12.7 21.3 16.4 16.8 16.7 4.8 7.8 9.8 Thalassiosira tumida 0.7 0.0 0.0 0.2 2.6 2.6 0.5 1.9 0.0 0.0 0.0 0.0 0.2 0.0 0.7 0.3 Thalassiothrix antarctica 0.2 0.2 0.2 0.2 0.4 1.4 1.7 1.2 2.1 1.0 1.0 1.4 0.9 0.5 0.5 0.6 Total Extinct spp. 0.0 0.0 0.0 0.0 0.6 0.2 0.7 0.5 15.5 28.6 20.5 21.5 21.3 5.3 0.2 8.9

SMO-2 (%)

NBP01-01 JPC 35 Depth (cm) 410 418 435 487 492 Avg. Chaetoceros (Hyalocheate) 5.3 7.1 15.6 6.4 7.8 8.44 Chaetoceros (spores) 10.3 9.1 6.1 2.2 28.6 11.26 Eucampia antarctica var recta 15.6 5.3 2.6 5.6 20.1 9.84 Fragilariopsis curta 12.3 32 38 62.8 7.8 30.58 Fragilariopsis kerguelensis 5.9 4.4 3.7 3.2 0.7 3.58 Thalassiosira antarctica var antarctica 3.3 4.4 2 1 16.9 5.52 Dactyliosolen antarcticus 2.6 3.3 1.5 2.9 3.9 2.84 Fragilariopsis angulata 2.9 4 2.6 0 0 1.9 Fragilariopsis obliquecostata 0.9 3.1 1.3 1.5 1.1 1.58 Fragilariopsis separada 3.3 2.9 2.4 0.2 0 1.76 Fragilariopsis sublineata 3.1 5.5 7.6 3.2 0.7 4.02 Porosira glacialis 1.1 0.9 0.4 0 2.3 0.94 Stellarima microtrias 3.3 3.5 2.4 1 0.2 2.08 Thalassiosira gracilis 2.4 3.1 3.7 14.7 1.4 5.06 Thalassiosira lentiginosa 9.9 2.9 3 1.7 3.2 4.14 Thalassiosira oestrupii 2.9 0.7 0 0.2 0.2 0.8 Total Extinct spp. 5.5 1.3 0.2 0.2 0.2 1.48

! 24 SMO-1 (%) SMO-2 (%) SMO-3 (%) NBP01-01 JPC 36 Depth (cm) 0 4 8 Avg. 160 168 175 Avg. 246 350 355 Avg. Chaetoceros (Hyalocheate) 7.6 7.2 8.7 7.8 1.8 2.6 0.7 1.7 0.0 0.0 0.4 0.1 Chaetoceros (spores) 0.2 5.6 0.9 2.2 0.0 3.3 0.7 1.3 0.0 0.0 0.0 0.0 Eucampia antarctica var recta 0.6 0.2 0.9 0.6 27.3 33.6 64.6 41.8 84.4 92.0 71.0 82.5 Fragilariopsis curta 53.5 58.4 53.3 55.1 0.9 4.9 0.8 2.2 0.0 0.0 0.9 0.3 Fragilariopsis kerguelensis 0.4 0.7 0.4 0.5 16.9 12.5 6.9 12.1 0.9 0.8 4.2 2.0 Thalassiosira antarctica var antarctica 14.5 8.1 2.7 8.4 13.8 9.2 5.4 9.5 1.9 0.3 11.9 4.7 Actinocyclus actinochilius 0.2 0.4 0.4 0.3 3.4 2.4 1.1 2.3 0.0 0.0 0.0 0.0 Dactyliosolen antarcticus 0.4 0.4 0.2 0.3 3.8 10.1 1.8 5.2 0.9 0.3 0.7 0.6 Fragilariopsis angulata 4.8 5.1 5.8 5.2 0.0 0.2 0.0 0.1 0.0 0.0 0.0 0.0 Fragilariopsis obliquecostata 2.4 2.2 1.1 1.9 0.7 0.7 0.4 0.6 0.0 0.2 0.0 0.1 Fragilariopsis sublineata 3.2 4.0 3.8 3.7 0.5 1.9 0.4 0.9 0.0 0.0 0.2 0.1 Fragilariopsis van huerckii 0.4 0.2 1.1 0.6 2.3 0.0 0.0 0.8 0.0 0.0 0.0 0.0 Porosira glacialis 2.4 0.7 1.8 1.6 0.7 0.0 0.0 0.2 0.0 0.0 0.0 0.0 Rhizosolena spp. 0.9 0.0 0.0 0.3 0.7 1.4 2.2 1.4 0.5 0.0 0.2 0.2 Stellarima microtrias 0.4 0.7 0.9 0.7 5.6 1.4 0.4 2.5 0.5 0.5 0.7 0.6 Thalassiosira gracilis 1.3 0.7 2.4 1.5 0.5 0.7 0.0 0.4 0.0 0.0 0.0 0.0 Thalassiosira lentiginosa 1.1 0.2 0.4 0.6 14.9 8.5 6.1 9.8 0.5 1.3 0.4 0.7 Thalassiosira torokina 0.0 0.0 0.0 0.0 0.0 0.2 2.2 0.8 9.5 3.9 8.6 7.3 Other 17.5 14.6 17.9 16.7 33.1 27.5 14.6 25.1 11.9 6.2 10.8 9.6 Total Extinct spp. 0.0 0.0 0.0 0.0 0.2 0.5 4.5 1.7 10.7 4.1 9.6 8.1

! 25 Extended Data Table 2. Prydz Channel 14C and cal BP ages, 13C, and %TOC. *Samples were background corrected in house according to individual lab AMS procedure. **Sample had, post graphitization, anomalously high residual pressure and may have fractionated during graphitization. It is reported for completeness but without a formal CAMS number.

Lab No. Core mbsf mcd δ 13C %TOC 14C age cal BP Carbon Source

AA34660* AA-186 GC09 1.72-1.74 0.4-0.41 - - 14120±120 15330±260 MF

R21977/2* AA-186 GC20 1.1-1.2 0.97-1.24 - - 33590±440 36170±540 MF

AA34661* AA-186 GC20 3.22-3.24 3.71 - - 37500±1400 40660±1300 MF

R21977/3* AA-186 GC22 0.27-0.32 0.76-1.3 - - 20230±120 22770±190 G. crassa

R21977/4* AA-186 GC22 0.27-0.32 0.76-1.3 - - 20770±170 23420±270 N. pachyderma

AA34662* AA-186 GC22 0.42-0.44 1.48-1.49 - - 33630±490 36230±630 MF

AA34663* AA-186 GC22 0.52-0.54 1.54-1.56 - - 43900±3200 46070±2530 MF

R21751/7* AA-186 GC22 1.0-1.01 2.1 - - 22170±180 25190±300 AIOM (bulk)

R21308/18* AA-KROCK 24 0.05-0.06 1.6-1.7 - - 2506±66 1160±120 AIOM (bulk)

R21308/19* AA-KROCK 24 1.0-1.01 14.6-14.8 - - 8541±85 8100±140 AIOM (bulk)

R21308/17* AA-KROCK 24 1.38-1.39 0.2 - - 12680±110 13250±140 AIOM (bulk)

Unknown* NBP01-01 JPC34 4.2-4.3 4.2-4.3 - - 32000±360 34610±340 AIOM (bulk)

Unknown* NBP01-01 JPC34 5.2-5.3 5.2-5.3 - - 37500±390 40840±430 AIOM (bulk)

Unknown* NBP01-01 JPC36 3.5-3.6 3.5-3.6 - - 46400±1100 48470±1090 AIOM (bulk)

CAMS 168131 NBP01-01 JPC34 1.5-1.55 1.5-1.55 -23.42 0.27 20560±90 23200±190 AIOM (RP-low)

CAMS 168017 NBP01-01 JPC34 1.95-2.0 1.95-2.0 -23.41 0.18 17390±70 19400±160 AIOM (RP-low)

CAMS 162050 NBP01-01 JPC34 2.25-2.3 2.25-2.3 - 0.31 24350±1060 27140±1020 AIOM (RP-low)

CAMS 168134 NBP01-01 JPC34 3.3-3.35 3.3-3.35 -23.46 0.30 19360±80 21870±190 AIOM (RP-low)

CAMS 162031 NBP01-01 JPC34 4.35-4.4 4.35-4.4 - 0.20 24510±1030 27330±1000 AIOM (RP-low)

CAMS 168136 NBP01-01 JPC34 5.0-5.05 5.0-5.05 - 0.20 18540±90 20780±180 AIOM (RP-low)

CAMS** NBP01-01 JPC34 5.3-5.35 5.3-5.35 -24.412 0.18 30800±3020 33630±3120 AIOM (RP-low)

CAMS 162040 NBP01-01 JPC35 3.95-4.0 1.19-1.22 - 0.26 23780±1050 26720±980 AIOM (RP-low)

CAMS 168129 NBP01-01 JPC35 4.4-4.45 1.5-1.53 -24.08 0.14 19080±120 21540±230 AIOM (RP-low)

CAMS 162161 NBP01-01 JPC36 0.08-0.15 0.09-0.17 -23.34 0.79 5190±120 4410±220 AIOM (RP-low)

CAMS 168138 NBP01-01 JPC36 1.75-1.8 2.02-2.1 -22.85 0.23 21330±120 24100±190 AIOM (RP-low)

CAMS 168019 NBP01-01 JPC36 3.4-3.45 4.18-4.22 -23.64 0.17 19700±90 22260±160 AIOM (RP-low)

! 26 14 Extended Data Table 3. NBP01-01 ramped pyrolysis C ages, RP temperature range, uncorrected and corrected Fm values. **Sample had, post graphitization, anomalously high residual pressure and may have fractionated during graphitization. It is reported for completeness but without a formal CAMS number.

14 Lab No. Core mbsf mcd Temperature CO2 evolved Uncorrected Fm Corrected Fm C age range (° C) (µmol) CAMS 168131 NBP01-01 JPC34 1.5-1.55 1.5-1.55 ambient-325 14.35 0.0815±0.0004 0.0773±0.0009 20562±94 CAMS 168132 NBP01-01 JPC34 1.5-1.55 1.5-1.55 371-409 16.11 0.0216±0.0003 0.0169±0.0008 32766±379 CAMS 168133 NBP01-01 JPC34 1.5-1.55 1.5-1.55 512-844 18.94 0.0315±0.0003 0.0274±0.0007 28892±219 CAMS 168017 NBP01-01 JPC34 1.95-2.0 1.95-2.0 ambient-356 15.41 0.1183±0.0005 0.1148±0.0010 17388±67 CAMS 168018 NBP01-01 JPC34 1.95-2.0 1.95-2.0 652-953 14.51 0.0469±0.0004 0.0423±0.0009 25407±167 CAMS 162050 NBP01-01 JPC34 2.25-2.3 2.25-2.3 ambient-324 11.12 0.0601±0.0005 0.0482±0.0064 24354±1062 CAMS 162051 NBP01-01 JPC34 2.25-2.3 2.25-2.3 324-386 16.94 0.0195±0.0003 0.0101±0.0040 36923±3181 CAMS 162052 NBP01-01 JPC34 2.25-2.3 2.25-2.3 386-439 26.34 0.0110±0.0006 0.0019±0.0038 >37447 CAMS 162159 NBP01-01 JPC34 2.25-2.3 2.25-2.3 439-606 24.31 0.0222±0.0003 0.0139±0.0035 34323±1990 CAMS 162162 NBP01-01 JPC34 2.25-2.3 2.25-2.3 606-947 17.42 0.0249±0.0004 0.0172±0.0032 32622±1480 CAMS 168134 NBP01-01 JPC34 3.3-3.35 3.3-3.35 ambient-362 15.84 0.0936±0.0005 0.0898±0.0009 19358±81 CAMS 168135 NBP01-01 JPC34 3.3-3.35 3.3-3.35 564-853 9.98 0.0307±0.0009 0.0246±0.0014 29761±459 CAMS 162031 NBP01-01 JPC34 4.35-4.4 4.35-4.4 ambient-336 11.69 0.0588±0.0005 0.0473±0.0061 24509±1028 CAMS 162032 NBP01-01 JPC34 4.35-4.4 4.35-4.4 336-410 18.39 0.0218±0.0004 0.0140±0.0032 34302±1844 CAMS 162033 NBP01-01 JPC34 4.35-4.4 4.35-4.4 410-462 19.92 0.0203±0.0005 0.0149±0.0021 33781±1152 CAMS 162038 NBP01-01 JPC34 4.35-4.4 4.35-4.4 462-606 21.43 0.0283±0.0004 0.0214±0.0028 30891±1071 CAMS 162039 NBP01-01 JPC34 4.35-4.4 4.35-4.4 606-946 23.53 0.0228±0.0005 0.0157±0.0029 33395±1502 CAMS 168136 NBP01-01 JPC34 5.0-5.05 5.0-5.05 ambient-345 12.68 0.1036±0.0005 0.0995±0.0011 18538±87 CAMS 168137 NBP01-01 JPC34 5.0-5.05 5.0-5.05 725-938 7.86 0.0265±0.0005 0.0192±0.0014 31760±593 CAMS** NBP01-01 JPC34 5.3-5.35 5.3-5.35 ambient-346 10.78 0.0355±0.0045 0.0216±0.0081 30802±3016 CAMS 162046 NBP01-01 JPC34 5.3-5.35 5.3-5.35 346-420 16.45 0.0200±0.0003 0.0098±0.0044 37143±3613 CAMS 162047 NBP01-01 JPC34 5.3-5.35 5.3-5.35 420-484 16.55 0.0174±0.0004 0.0071±0.0044 39721±4999 CAMS 162048 NBP01-01 JPC34 5.3-5.35 5.3-5.35 484-663 19.15 0.0274±0.0004 0.0185±0.0038 32048±1666 CAMS 162049 NBP01-01 JPC34 5.3-5.35 5.3-5.35 663-950 18.22 0.0163±0.0003 0.0067±0.0040 40212±4847 CAMS 162040 NBP01-01 JPC35 3.95-4.0 1.19-1.22 ambient-307 12.29 0.0641±0.0005 0.0518±0.0068 23778±1047 CAMS 162041 NBP01-01 JPC35 3.95-4.0 1.19-1.22 307-389 23.21 0.0183±0.0003 0.0082±0.0043 >38600 CAMS 162042 NBP01-01 JPC35 3.95-4.0 1.19-1.22 389-443 29.36 0.0065±0.0001 0.0000±0.0029 >40269 CAMS 162043 NBP01-01 JPC35 3.95-4.0 1.19-1.22 443-592 26.41 0.0159±0.0003 0.0082±0.0031 38600±3015 CAMS 162044 NBP01-01 JPC35 3.95-4.0 1.19-1.22 592-947 25.86 0.0130±0.0002 0.0030±0.0042 >35897 CAMS 168129 NBP01-01 JPC35 4.4-4.45 1.5-1.53 ambient-375 10.69 0.0977±0.0008 0.0930±0.0014 19081±117 CAMS 168130 NBP01-01 JPC35 4.4-4.45 1.5-1.53 727-954 14.79 0.0505±0.0003 0.0460±0.0008 24728±147 CAMS 162161 NBP01-01 JPC36 0.08-0.15 0.09-0.17 ambient-391 31.03 0.5213±0.0018 0.5239±0.0079 5194±121 CAMS 162053 NBP01-01 JPC36 0.08-0.15 0.09-0.17 443-534 30.55 0.4870±0.0019 0.4889±0.0074 5749±122 CAMS 162054 NBP01-01 JPC36 0.08-0.15 0.09-0.17 534-835 25.89 0.4138±0.0015 0.4148±0.0075 7070±146 CAMS 168138 NBP01-01 JPC36 1.75-1.8 2.02-2.1 ambient-329 12.84 0.0748±0.0005 0.0702±0.0010 21334±117 CAMS 168139 NBP01-01 JPC36 1.75-1.8 2.02-2.1 519-794 16.06 0.0170±0.0002 0.0124±0.0008 35297±507 CAMS 168019 NBP01-01 JPC36 3.4-3.45 4.18-4.22 ambient-340 14.09 0.0902±0.0005 0.0861±0.0010 19696±90 CAMS 168020 NBP01-01 JPC36 3.4-3.45 4.18-4.22 536-928 17.18 0.0236±0.0002 0.0191±0.0008 1784±323 !

! 27 Extended Data Table 4. Aliquot weight, uncorrected and corrected 10Be concentration with associated error, and flux for NBP01-01 cores. *Sieved duplicate (<150 µm). ! Core mbsf mcd Aliquot wt. Uncorrected 10Be±1σ Corrected 10Be±1σ 10Be flux Sediment Type (mg) (x108 atoms g-1) (x108 atoms g-1) (atoms cm-2 kyr-1) JPC 34 0.06 0.06 100.23 11.29±0.24 11.32±0.24 1.94E+10 Diatom mud JPC 34 0.35* 0.35 102.03 7.69±0.16 7.71±0.16 1.48E+10 Sandy mud JPC 34 0.35 0.35 101.24 7.01±0.13 7.04±0.13 1.35E+10 Sandy mud JPC 34 0.65* 0.65 101.29 1.81±0.05 1.81±0.05 4.57E+09 Granule-rich muddy sand JPC 34 0.65 0.65 101.2 1.55±0.06 1.56±0.06 3.93E+09 Granule-rich muddy sand JPC 34 1.33 1.33 102.59 11.86±0.19 11.95±0.19 1.41E+10 Sandy mud JPC 34 1.97 1.97 99.17 19.74±0.33 19.95±0.34 1.12E+11 Diatom mud JPC 34 2.18 2.18 102.35 11.41±0.20 11.53±0.20 6.50E+10 Diatom mud JPC 34 2.72 2.72 101.26 2.20±0.05 2.23±0.05 1.32E+10 Sandy mud JPC 34 3.37* 3.37 100.08 2.09±0.05 2.11±0.05 1.46E+10 Sandy mud JPC 34 3.37 3.37 100.4 1.73±0.06 1.75±0.06 1.21E+10 Sandy mud JPC 34 4.22 4.22 105.98 15.47±0.31 15.66±0.31 1.65E+10 Diatom mud JPC 34 4.55 4.55 104.39 3.91±0.10 3.97±0.10 4.80E+09 Granule-rich muddy sand JPC 34 5.33 5.33 102.71 7.76±0.15 7.89±0.15 8.89E+09 Diatom mud JPC 34 5.71 5.71 98.26 3.36±0.09 3.42±0.09 4.46E+09 Granulated JPC 35 3.77* 0.82 103.55 17.73±0.28 17.83±0.29 3.35E+10 Sandy mud JPC 35 3.77 0.82 103.11 17.29±0.28 17.39±0.28 3.26E+10 Sandy mud JPC 35 4.13 1.22 99.21 23.66±0.35 23.84±0.35 1.34E+11 Diatom mud JPC 35 4.5 1.63 99.26 15.36±0.21 15.51±0.21 8.74E+10 Diatom mud JPC 35 4.89 2.09 102.46 11.28±0.20 11.40±0.20 6.42E+10 Diatom mud JPC 35 5.37 2.36 98.86 1.75±0.08 1.77±0.08 1.06E+10 Granulated JPC 36 0.14 0.15 99.27 10.26±0.17 10.29±0.17 1.76E+10 Sandy/Diatom mud JPC 36 0.45 0.36 100.37 1.20±0.07 1.20±0.07 2.36E+09 Granulated JPC 36 1.8 2.11 98.96 7.39±0.16 7.47±0.16 4.27E+10 Sandy mud JPC 36 2.33 3.69 102.67 11.46±0.19 11.59±0.20 8.51E+10 Silty clay JPC 36 2.92 3.84 100.68 1.86±0.09 1.88±0.09 1.46E+10 Granulated JPC 36 3.37 4.17 107.34 12.25±0.19 12.40±0.20 8.72E+10 Sandy clay ! !

! 28

APPENDIX B: SUPPLEMENTARY METHODS

Radiocarbon Contamination Correction

For all NBP01-01 Ramped pyrolysis (RP) samples, we applied a blank correction in house according to a study done by Fernandez et al.3, who observed that the modern blank contribution is dependent on the duration of the ramped pyrOx run, whereas the dead contribution is constant

1 for all samples. The modern contribution likely arises from CO2 introduction through valves in the collection apparatus. Dead contribution is most likely introduced through combustion of

14 Modern a dead contribution from CO collection apparatus is residual C-dead material within the apparatus. nd 2

2.8±0.6 and 1.4±0.8 µg C, respectively, for samples prepared at College of Marine Science; 8.8±4.4 and 4.1±5.4 µg C, respectively, for samples prepared at Tulane University. Modern 14C is also introduced to our samples through graphitization

(2.0±1.0 µg) of the material during AMS analysis. We account for all 3 components in our offline calculation. The measured fraction modern (subscript M) is written as a mixture of three components, the actual sample material (subscript A), the dead blank (subscript d), the modern blank (subscript mod), and the graphitization blank (subscript g):

!! = !!!! + !!!"!!"# + !!!! + !!!! (1)

Where:

! !!! !! = 1 (2)

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ! ! 29 We know the mass (m) of the sample and the masses of the three blank contamination components, therefore the variable f is written as follows:

!!! !! = ! (3) !!!!!!"#!!!!!!!

!!!"# !!"# = ! (4) !!!!!!"#!!!!!!!

!!! !! = ! (5) !!!!!!!"#!!!!!!!

!!! !! = ! (6) !!!!!!!"#!!!!!!!

Where:

! !!!!!"#!!!!!! !!! !! = !! + !!"# + !! + !! = = 1 (7) !!!!!"#!!!!!!

Fractions modern above are written as deltas to convey the similarities of a multi-component isotopic mixing model. To solve equation 1 for the actual value of the sample, we rearrange:

!!!!!!!!!!"#!!"#!!!!! !! = (8) !!

We then simplify equation 8 by writing all fraction terms as masses (substituting equations 3-6 in for all f variables in equation 8), and cancelling terms that are multiplied by the fraction modern of the dead blank (δd = 0):

!! !!"# !! !!"# !! !! = !! + !! + !! + !! − !!"# − !! (9) !! !! !! !! !!

! 30

From equation 8, we derive an expression for the propagated analytical uncertainty as well as the uncertainties of the masses of blank contamination from:

! ! ! ! !" !! = !!! !!! (10) !!!

Where y=f(x1, x2,…,xi).

Substituting equation 9 into equation 10, we have:

! ! ! ! ! ! ! !!!!!"#!!! ! !! ! !!!!!"# ! !!!!! !!!!!!!!!!!!!!!!!!! = !! + !! + !!! + !!!"# + !!! + ! ! ! !! !! !! !!

! ! !!"#!!"#!!!!!!!!!!!!!!!!!!!!"# !!!! ! (11) !!

Calibration

Calibration to calendar years was performed using Calib7.1 Radiocarbon Calibration program and the Marine13 calibration curve4. We applied a 1300±100 yr reservoir correction, based on the average 14C age of pre-bomb benthic marine carbonates from the Southern Ocean5. We selected this reservoir age, which is within the error of a locally derived reservoir correction of

1280±200 yr6, because it enables comparison of our record to a larger number of circum-

Antarctic marine-based records.

Beryllium-10 Concentration Corrections

We correct all 10Be concentrations ([10Be]) for radioactive decay using the updated 10Be half-life 1.387±0.0012 Ma8. 10Be flux values based on age-corrected [10Be] were calculated from

! 31 measured sediment density and sedimentation rates established from average SMO 14C ages. Pre-

Holocene (<15 ka) flux values were corrected further to account for increased 10Be production9, which we assume to be ~40% excess over modern 10Be production (1.21±0.26x106 atoms cm-2 yr-1). Although reconstructions of 10Be production from ice cores and marine sediments demonstrate a 20-60% excess over modern values through the LGM and late MIS-3 (40-18 ka)9,10, we correct for 40% increase over modern production (1.69±0.36 x106 atoms cm-2 yr-1) because this increase corresponds to the timing of SMO-2 to -4 deposition. Holocene 10Be production is close to modern10, but assumptions concerning Holocene 10Be production cannot be made because timing of maximum radionuclide flux still unresolved11.

Beryllium-10 Budget Estimates

To estimate maximum 10Be input from LG-AIS basal melt. We assume uniform LG-AIS

10Be concentration, based on the average 10Be concentration in the Law Dome ice core record12.

Exposure ages from the Amery Oasis indicate ~250 m lowering since the LGM ( ~25 ka)13. LG-

AIS lowering since the LGM due to basal melt14 contributed 2.34x108 atoms/cm2 to Prydz

Channel. The length of the sediment sequence deposited since 25 ka is ~430 cm. Assuming

100% of 10Be was captured in the sediment, we calculate a sediment accumulation of 5.44x105 atoms/cm3 (3.11x105 atoms/g (density 1.75 g/cm3)). Our estimate accounts for an average 20% increase in 10Be production over the present during the LGM9.

To investigate if grain size affects the accuracy of 10Be concentration measurements, we sieved aliquots from the coarsest intervals. Following Sjunneskog et al15, aliquots were dry sieved using a 150 µm mesh, then weighed. In all sieved aliquots, the mass of the >150 µm fraction was <5% of the total aliquot mass. Results demonstrate that, while 10Be concentration between sieved and non-sieved fractions is significantly different at the 95% confidence interval,

! 32 the difference between values is still within 0.67x108 atoms g-1 when values range from 1.2x108 atoms g-1 to 23.8x108 atoms g-1. The effects of grain size inhomogeneity should be considered in future studies, but we observe little effect on accuracy of Prydz Channel 10Be measurements.

! 33

APPENDIX C: SUPPLEMENTARY DISCUSSION

Diatom Assemblages

The SMO-1 diatom assemblage in NBP01-01 JPC-34 and JPC-36 is dominated by Fragilariopsis curta, but also includes other Fragilariopsis spp., Chaetoceros spp. (Hyalochaete) and

Thalassiosira antarctica, suggesting sea ice associated productivity within a polar coastal marine environment, typical of Holocene sediments in Prydz Bay1. At all sites, we observe a decreased

%abundance of F. curta and T. antarctica and an increased %abundance of Eucampia antarctica in SMO-2. However, E. antarctica, a heavily silicified, robust taxa indicative of strong bottom water currents that winnow smaller frustules during sedimentation1, dominates the assemblage in

JPC-36, which is located farthest from Amery Ice Shelf. We observe minimal variability between SMO-1 and SMO-2 %abundance in JPC-34 and -35. The presence of Pliocene and late

Miocene (4.0 Ma and ~7 to 9.0 Ma, respectively) diatom taxa in SMO-2 suggests influx of glacially-derived sediment to Prydz Channel, likely from erosion of the Pagodroma Group and/or

Sørsdal Formation, present in the LG-AIS catchment. In SMO-3, %abundance of extinct taxa increases in both JPC-34 and -36. Again, E. antarctica dominates the JPC-36 assemblage. In

JPC-34, F. curta and Chaetoceros %abundance both decrease, but T. antarctica dominates the assemblage, suggesting reduced sea ice influence on the Prydz Channel paleoenvironment.

Although the environmental preferences of T. antarctica remain largely unknown, there is some evidence from the Ross Sea to suggest that this species is associated with platelet ice formation proximal to the calving front of the Ross Ice Shelf2. SMO-4 is only observed in JPC-34 and is

! 34 characterized by E. antarctica, which is the dominant species. Chaetoceros, T. antarctica, and extinct taxa demonstrate similar %abundance (Fig. 2; Supplementary Table 1).

! 35

APPENDICES REFERENCES

1 Taylor, F. & Leventer, A. Late Quaternary palaeoenvironments in Prydz Bay, East Antarctica: interpretations from marine diatoms. Antarctic Science 15, 512-521, doi:10.1017/s0954102003001639 (2003). 2 Cunningham, W. L., Leventer, A., Andrews, J. T., Jennings, A. E. & Licht, K. J. Late Pleistocene-Holocene marine conditions in the Ross Sea, Antarctica: evidence from the diatom record. The Holocene 9, 129-139 (1999). 3 Fernandez, A. et al. Blank corrections for ramped pyrolysis radiocarbon dating of sedimentary and soil organic carbon. Analytical Chemistry 86, 12085-12092, doi:10.1021/ac502874j (2014). 4 Reimer, P. J. et al. IntCal and Marine13 radiocarbon age calibration curves 0-50,000 years cal BP. Radiocarbon 55, 1869-1887 (2013). 5 Berkman, P. A. & Forman, S. L. Pre-bomb radiocarbon and the reservoir correction for calcareous marine species in the Southern Ocean. Geophysical Research Letters 23, 363- 366, doi:10.1029/96gl00151 (1996). 6 Leventer, A. et al. Marine sediment record from the East Antarctic margin reveals dynamics of ice sheet recession. GSA Today 16, 4, doi:10.1130/gsat01612a.1 (2006). 7 Blaauw, M. & Christen, A. J. Flexible Paleoclimate Age-Depth Models Using an Autoregressive Gamma Process. Bayesian Analysis 6, 457-474, doi:10.1214/11-ba618 (2011). 8 Korschinek, G. et al. A new value for the half-life of 10Be by Heavy-Ion Elastic Recoil Detection and liquid scintillation counting. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 268, 187-191, doi:10.1016/j.nimb.2009.09.020 (2010). 9 Frank, M. et al. A 200 kyr record of cosmogenic radionuclide production rate and geomagnetic field intensity from 10Be in globally stacked deep-sea sediments. Earth and Planetary Science Letters 149, 121-129 (1997). 10 Christl, M., Lippold, J., Steinhilber, F., Bernsdorff, F. & Mangini, A. Reconstruction of global 10Be production over the past 250ka from highly accumulating Atlantic drift sediments. Quaternary Science Reviews 29, 2663-2672, doi:10.1016/j.quascirev.2010.06.017 (2010). 11 Muscheler, R., Beer, J., Kubik, P. W. & Synal, H. A. Geomagnetic field intensity during the last 60,000 years based on 10Be and 36Cl from the Summit ice cores and 14C. Quaternary Science Reviews 24, 1849-1860, doi:10.1016/j.quascirev.2005.01.012 (2005). 12 Smith, A. M. et al. 7Be and 10Be concentrations in recent firn and ice at Law Dome, Antarctica. Nuclear Instruments and Methods in Physics Research B 172, 847-855 (2000). 13 White, D. A., Fink, D. & Gore, D. B. Cosmogenic nuclide evidence for enhanced sensitivity of an East Antarctic ice stream to change during the last deglaciation. Geology 39, 23-26, doi:10.1130/g31591.1 (2011).

! 36 14 Herraiz-Borreguero, L. et al. Circulation of modified circumpolar deep water and basal melt beneath the Amery Ice Shelf, East Antarctica. Journal of Geophysical Research: Oceans, n/a-n/a, doi:10.1002/2015jc010697 (2015). 15 Sjunneskog, C., Scherer, R., Aldahan, A. & Possnert, G. 10Be in glacial marine sediment of the Ross Sea, Antarctica, a potential tracer of depositional environment and sediment chronology. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 259, 576-583, doi:10.1016/j.nimb.2007.01.203 (2007). 16 Scherer, R. P. et al. Pleistocene Collapse of the West Antarctic Ice Sheet. Science 281, 82-85, doi:10.1126/science.281.5373.82 (1998).

! !

! 37