Copyright Rebecca Lynn Minzoni 2015

Rebecca Lynn Minzoni

2015

ABSTRACT

The Antarctic Peninsula’s Response to Holocene Climate Variability:

Controls on Glacial Stability and Implications for Future Change

Rebecca Lynn Minzoni

The Antarctic Peninsula is one of the most rapidly changing regions in the cryosphere, with 87% of its glaciers receding and several ice shelves catastrophically collapsing since observations began in the 1960’s. These substantial, well-documented changes in the ice landscape have caused concern for the mass balance of the Antarctic Peninsula Ice Cap. To better understand the significance of these recent changes, I have assimilated a massive database of new and published marine sedimentary records spanning the Holocene Epoch (the last 11.5 kyrs). The database includes 9 coastal embayments with expanded sedimentary packages and well-dated cores. Each site represents an end-member in the wide range of Antarctic Peninsula oceanography, orography, meteorology, and glacial drainage basin characteristics. Multi-proxy analysis, including sedimentology, geochemistry, and micropaleontology, was conducted at each site to reconstruct glacial history at centennial-scale resolution, on par with ice-core data. The coastal sites were then compared in the context of published ice-core paleoclimate, paleoceanographic, and glaciological records.

The first of these sites, Herbert Sound, provides an unparalleled opportunity to compare the marine sediment record with a related ice-core in an Antarctic maritime setting. Herbert Sound is the southernmost embayment studied on the eastern side of the Antarctic Peninsula and represents an end-member with a cold, dry atmosphere and cold, saline ocean mass. The record from Herbert Sound indicates grounded ice receded quickly and early in the Holocene, followed by a floating ice phase

that collapsed 10  2.4 calendar kyrs before present (cal kyr BP, where present day is 1950 A.D.) and never re-advanced. The fjord remained open and productive during the prolonged warm intervals of the Mid Holocene, and began to experience greater glacial influence and sea ice cover during Late Holocene cooling, a period termed the Neoglacial.

The second site, Ferrero Bay of the Amundsen Sea, is the southernmost end-member on the western side of the Antarctic Peninsula and represents a truly polar coastal setting. Grounded ice receded from the deep basin much earlier than expected, ~10.7 cal kyr BP, likely due to advecting water masses that under-melted the extended ice sheet. Ferrero Bay was then covered by the extensive Cosgrove Ice Shelf, which remained stable during the Mid Holocene Hypsithermal despite incursion of warm water, and did not recede to its current position until 2.0 cal kyr BP, coincident with Neoglacial cooling. Thus, Ferrero Bay and the Cosgrove Ice Shelf were out of phase with the northern peninsula sites and apparently were not sensitive to Holocene climate variability but rather to impinging warm ocean currents, which are observed in Ferrero Bay today.

Comparison of 9 coastal sites, including Herbert Sound, Ferrero Bay, and several other embayments of various local settings, shows highly variable glacial response to Early Holocene warming, with difference in onset of  4.2 kyrs and difference in duration of  2.5 kyrs. The Mid Holocene more synchronous, with a difference in the onset of  1.7 kyrs and a difference in duration of  1.9 kyrs. The Late Holocene Neoglacial by contrast was highly synchronous with difference in onset and duration of  0.7 kyrs and  0.7 kyrs, respectively. Later historic events of shorter duration, such as the Little Ice Age, were more synchronous with difference in onset and duration of  0.3 kyrs and  0.2 kyrs, respectively. Regional glacial behavior became more synchronous during the Holocene as the glaciers and their drainage basins became progressively smaller, rendering them more sensitive to climate change and less influenced by local forcings such as ocean temperature, basin bathymetry, and precipitation effects. This increase in glacier sensitivity helps explain the current widespread glacial recession in response to the rapid regional warming of ~3.5 C over the last century.

Acknowledgments

First and foremost I am grateful to Dr. John Anderson for the numerous opportunities and challenges he has given me during this PhD. The research included in this dissertation was shaped by several enriching experiences that John facilitated, including my attendance to workshops around the globe to learn new skills, namely diatom micropaleontology. It takes a truly altruistic advisor to allow a student to pursue what might be considered tertiary interests to a project. John had the confidence in me to bring these various pursuits together into a holistic, large-scale analysis; fortunately, these external pursuits became the key proxies for our interpretations. I could not dream of a more exciting and fulfilling PhD experience. Thank you, John, for believing in me, for investing in me, and for growing me into a more confident and mature scientist.

I thank my committee members for investing countless hours and resources into my professional development. Thank you, Dr. André Droxler, for being a reliable sounding board for ideas, for allowing me to use your lab space with an open door policy, and for being an excellent teacher in the classroom and in the field. Thank you, Dr. Jeffrey Nittrouer, for constructive feedback and lively lab discussions. Thank you, Dr. Volker Rudolf, for serving as a mentor who truly has students’ best interest in mind.

I would like to especially thank Dr. Julia Wellner, who acted as a committee member in every sense of the word. Julia, you have been an excellent co- author, friend, and role model to me. I appreciate you incorporating me into your lab group at UH and facilitating collaboration and progressive thinking in our research. I also thank Dr. Amy Leventer (Colgate) who has been an exceptionally generous mentor to me in diatomology and in Antarctic field science, and a key supporter in my pursuit of micropaleontology. Thank you for connecting me with the welcoming community of the Polar Marine Diatom Workshops and teaching me skills that were essential to this research.

Thank you to my ever-supportive, high-spirited co-authors, Dr. Rodrigo Fernandez, Dr. Wojciech Majewski, Dr. Yusuke Yokoyama, and Dr. Martin Jakobsson.

This research builds upon three decades of hard work at sea in harsh conditions, as well as countless hours in the laboratory by dedicated scientists and technicians. I thank the crews of Oden Southern Ocean Cruise 09-10, Nathaniel B. Palmer cruises 05-02 and 07-03, Polar Duke 1990 and 1991, and Deep Freeze (USCG Glacier) 1982. Thank you to Dr. Charlotte Sjunneskog and Steve Petrushak for their assistance with sampling and data collection at the Antarctic Research Facility. Thank you to Dr. Tony Gary and TACSWorks for teaching me Fuzzy C-Means analysis and graciously allowing me to use their software for this project. Thank you to Dr. Caroline Masiello for allowing me to use her lab and advising me on geochemical methods.

This work was completed with funding from the National Science Foundation Award OPP99-09734 to John B. Anderson and Julia S. Wellner, and OPP0837925 to John B. Anderson. This research was also facilitated by the Douglas and Martha Lou Broussard Fellowship and travel grants from Rice University, the Scientific Committee on Antarctic Research, PAGES, and the International Symposium of Antarctic Earth Science through NSF. Martha Lou Broussard, you have created a legacy for students of the Earth Science Department that constantly inspires me; you are the model example of how alumni should give back to the community that shaped them. Thank you.

I am exceptionally grateful to Dr. Luis González (KU), the late Dr. Larry Martin (KU), and Dr. Kraig Derstler (UNO) for their openhanded, enthusiastic mentorship and for their belief in my early dreams of pursuing the noble, ever-evolving path of a research scientist.

I am especially grateful to Ruthie Halberstadt, Catherine Ross, Victoria Yuan, Kelsey Crocker, and Savannah Ezelle for working with me in the lab and for their unbridled enthusiasm, which served as motivation for me during PhD doldrums.

The support of my peers has been essential to my progress at Rice. Thank you Dr. Lauren Simkins, Lindsay Prothro, Yuribia Munoz, Brian Demet, Dr. Davin vi

Wallace, Travis Stolldorf, Karem Lopez, Dr. Lizette Leon-Rodriguez, Svetlana Burris, Dr. Xiaodong Gao, Krystle Hodge, Stacy Dwyer, Sarah Huff, Lily Seidman, Kaitlyn Moran, Tian Dong, Laura Carter, Jen Gabler, Dr. Lester King, Shibu Mathukutty, Ayca Agar Cetin, Dr. Brandon Harper, Clint Miller, Dr. Ananya Malik, Dr. Pranabendu Moitra, Dr. Matt Weller, Jacob Seigel, Dr. Ben Slotnik, and Pankaj Khanna. Many heart-felt thanks to the gracious ESCI administrators who were always available and encouraging to me, despite their busy schedules: Mary Ann Lebar, Soookie Sanchez, Sandra Flechsig, Roger Romero, Pat Jordan, and Lee Wilson. I also thank Dr. Malcolm Ross, Dr. Albert Bally, Dr. Steve Bergman, and Dr. James Eldrett for their mentorship and support at Shell and beyond.

Much like a healthy carbonate platform, my family has been the framework upon which my academic life was able to grow and mature. Thank you, Dr. Marcello Minzoni, my brilliant and ardent husband, who served in the trenches, picking up the slack at home when I needed to put extra hours into my research, and ultimately for understanding the joy and the challenge of completing a PhD. Thank you for being a loving, involved father to our beautiful boys and for consistently helping me see the big picture of what is most important in life.

I thank my supportive siblings, Nate, Louise, Matt, Cody, and Wyatt, for their interest in my research, and for always understanding when I missed events due to field excursions and writing deadlines. I also thank Dr. Nello Minzoni, Adriana Minzoni, Elise Bolton, Emily Copeland, and Sara Edwards for their loving support and encouragement.

I thank my father, Dr. Matthew Totten, for listening, advising, talking geology, and telling horrible dad jokes when necessary. Thank you for raising me to be iron-willed and for pushing me to think critically, to value healthy skepticism, to play my 4-string when I need a pick-me-up, and to see happiness as a learned skill rather than a good poker hand. Several times I pushed through what seemed impossible because I was inspired by your approach to adversity. vii

I thank my mother, Karen Parsons—the urban planner, the artist, the activist—for serving as emergency baby-sitter, as a caring listener, and as the inspiration for this research. Your chronicling of the raw aftermath of Hurricane Katrina—and your endless dedication to rebuilding the unique community of New Orleans—fuels my dedication to understand climate and sea level changes. The purpose you give to life gives me a sense of urgency to make a difference in our community and our society.

This dissertation is dedicated to my late mother, Dr. Iris Moreno Totten. (Yes, I have two smart and talented mothers.) Not only did her passion for teaching geology inspire me to become an earth scientist, but her relentless determination served as continual motivation during my PhD and my new role as a mother. Iris completed her PhD in geology and education with six children at home. The final year of her PhD she was diagnosed with breast cancer, and she didn’t slow down for one minute. She went on to become a professor alongside my father, who also had cancer. They fought the disease, and they beat it together. Her cancer came back, however; she fought it for the better part of a decade while raising kids, earning tenure, building new and exciting venues for teaching geology, and selflessly giving what energy she had left to her students, as well as to children and teachers at local schools. Iris, you may have been only physically present for the first half of my PhD marathon, but I know that you were with me every step of the way, cheering me on as you always have. I will always remember your wise words, your full, honest laugh, and your enthusiasm that won over so many minds to the pursuit of science. Thank you for showing me and so many others that a woman can be many things—a scientist, an educator, a professional, a nurturing mother, a devoted friend, and someone who can help change the world by investing in others and culturing an honest, imaginative vision for our future.

Contents

Acknowledgments ...... iv

Contents ...... viii

List of Figures ...... xi

List of Tables ...... xiii

Nomenclature ...... xiv

Introduction ...... 1 First opportunity to compare the marine sediment record with a related ice-core in coastal Antarctica1 ...... 4 2.1. Introduction ...... 6 2.2. Background ...... 11 2.2.1. Geologic setting ...... 11 2.2.2. Climate and oceanographic setting ...... 11 2.2.3. Glacial history ...... 12 2.2.4. Ice Shelf History of the NW Weddell Sea ...... 13 2.3. Methods ...... 14 2.3.1. Geophysical methods and coring ...... 14 2.3.2. Sedimentologic and stratigraphic methods ...... 15 2.3.3. Sedimentological, geochemical, and paleontological methods ...... 16 2.3.4. Radiocarbon dating ...... 18 2.4. Results ...... 20 2.4.1. Seismic and bathymetric surveys of Herbert Sound ...... 20 2.4.2. Radiocarbon results ...... 20 2.4.3. Sedimentological, geochemical, and paleontological results ...... 29 2.4.4. Unit 1 (11.4 to 9.9 mbsf) ...... 36 2.4.5. Unit 2 (9.9 to 9.7 mbsf) ...... 36 2.4.6. Unit 3 (9.7 to 8.4 mbsf ) ...... 37 2.4.7. Unit 4 (8.4 to 2.9 mbsf) ...... 37 2.4.8. Unit 5 (2.9 to 0 mbsf ) ...... 38

2.5. Discussion ...... 39 2.5.1. Diatom Assemblages (DA) ...... 39 2.5.2. The Last Glacial Maximum ...... 41 2.5.3. Permanent Ice canopy ...... 42 2.5.4. Early Holocene opening of the fjord ...... 43 2.5.5. Mid Holocene Hypsithermal ...... 48 2.5.6. Late Holocene Neoglacial ...... 50 2.6. Late Holocene ice shelf limits ...... 52 2.7. Conclusions ...... 53 Amundsen Sea ice shelf behavior out of phase with Antarctic Peninsula ice shelves during the Holocene ...... 55 3.1. Introduction ...... 56 3.2. Background ...... 60 3.2.1. Geologic setting ...... 60 3.2.2. Climatic and Oceanographic Setting ...... 60 3.3. Methods ...... 63 3.3.1. Geophysical methods and coring ...... 63 3.3.2. Radiocarbon analysis ...... 63 3.3.3. Sedimentological proxies ...... 64 3.3.4. Geochemical proxies...... 64 3.3.5. Micropaleontological proxies ...... 65 3.4. Results ...... 66 3.4.1. Geophysical results ...... 66 3.4.2. Radiocarbon results ...... 68 3.4.3. Kasten Core KC-15 ...... 69 3.4.4. Kasten Core KC-16 ...... 70 3.4.5. Kasten Core KC-17 ...... 70 3.5. Discussion ...... 74 3.5.1. Paleo-drainage of the West Antarctic Ice Sheet ...... 74 3.5.2. Unit 1: Glacial recession from Ferrero Bay ...... 76 3.5.3. Unit 2: Circumpolar Deepwater beneath the Cosgrove Ice Shelf ...... 76 x

3.5.4. Unit 3: Stable Cosgrove Ice Shelf during the Mid Holocene ...... 77 3.5.5. Unit 4: Late Holocene collapse of the Cosgrove Ice Shelf ...... 78 3.5.6. Glacial History of the ASE ...... 78 3.5.7. The Holocene history of small Antarctic ice shelves ...... 81 3.6. Conclusions ...... 84

Increasing glacial sensitivity to climate change in the Antarctic Peninsula ...... 85 4.1. Introduction ...... 86 4.2. Methods ...... 90 4.3. Discussion ...... 91 4.4. Conclusions ...... 95

References ...... 96

Appendix A ...... 118

List of Figures

Figure 2-1 Bathymetric and elevation map of the Antarctic Peninsula illustrating complex orography, oceanography, and climate...... 8

Figure 2-2 Study area map of James Ross Island and Herbert Sound...... 9

Figure 2-3 NBP0502 sub-bottom profile from southern Herbert Sound...... 15

Figure 2-4 Age model of Herbert Sound...... 25

Figure 2-5 Correlation of NBP-0502 Site 2 drill core and Kasten core with PD- 91 and DF-82 piston cores...... 26

Figure 2-6 Stratigraphic plot of key proxies used to interpret paleoenvironments in NBP0502 Site 2...... 27

Figure 2-7 Grain size and x-radiographs of Herbert Sound...... 31

Figure 2-8. PCA results from Herbert Sound diatom counts...... 32

Figure 2-9 Relative abundance of non-CRS diatom taxa ...... 33

Figure 2-10 Glacial reconstruction of Herbert Sound and Prince Gustav Channel (PGC)...... 46

Figure 2-11 Regional glacial and climatic history of the Antarctic Peninsula, James Ross Island, and Herbert Sound...... 47

Figure 3-1 Map of the Antarctic Peninsula (AP) ...... 58

Figure 3-2 Bathymetric map of the Amundsen Sea Embayment and Ferrero Bay ...... 59

Figure 3-3 Water column temperature and salinity measured during OSO- 0910 cruise in Ferrero Bay...... 62

Figure 3-4 OSO-0910 subbottom profile 98A from the axis of Ferrero Bay ..... 67

Figure 3-5 Multi-proxy analysis of the three OSO-0910 Kasten cores from Ferrero Bay...... 72

xii

Figure 3-6 Correlation of sedimentary facies in Kasten cores from Ferrero Bay...... 73

Figure 3-7 Glacial reconstruction of Ferrero Bay and outer Pine Island Bay ... 75

Figure 3-8 Ice shelf history and paleo-temperature proxies ...... 83

Figure 4-1 The Antarctic Peninsula and its dynamic climate and ocean interactions...... 88

Figure 4-2 Comparison of Antarctic Peninsula ice core temperature record and two ocean deep locations with the 9 coastal study sites...... 93

Figure 4-3 Difference in timing of onset of glacial response vs. difference in the duration of glacial response to Holocene climate events (± kyrs)...... 94

List of Tables

Table 2-1 Radiocarbon ages from Herbert Sound...... 24

Table 2-2 Summary of diatom assemblages (DA’s), their composition, and interpretations...... 35

Table 3-1 Radiocarbon ages from Ferrero Bay...... 69

Nomenclature

ACC Antarctic Circumpolar Current

AIO Acid Insoluble Organic fraction

AP Antarctic Peninsula

ASE Amundsen Sea Embayment cal kyr BP Thousand calendar years Before Present (where

present is 1950 A.D.)

CDW Circumpolar Deep Water

CRS Chaetoceros Resting Spores

EHCO Early Holocene Climatic Optimum

ESNO El Nino Southern Oscillation

JRI James Ross Island

LIA Little Ice Age

MAAT Mean annual atmospheric temperature

MHH Mid Holocene Hypsithermal

MCA Medieval Climate Anomaly

PIB Pine Island Bay

TN Total nitrogen

TOC Total organic carbon

SST Sea surface temperature

WSTW Weddell Sea Transitional Water

Chapter 1

Introduction

The Antarctic Peninsula (AP) is the most rapidly warming region in the Southern Hemisphere today. Its unique position as the northernmost extension of Antarctica creates a prominent barrier to ocean currents and wind patterns and creates a wide range of climatic, oceanographic, orographic, and glaciological environments. Thus, the AP is also perhaps the most dynamic region of the Southern Hemisphere and an ideal setting to study controls on glacial behavior through time. To do so, I have analyzed a dozen coastal embayments of the Antarctic Peninsula, reconstructing their Holocene glacial history from multi-proxy analysis of expanded sediment records. Seven study sites were published previously; where they were lacking, I have analyzed additional proxies (of sedimentology, geochemistry, and micropaleontology) on the sediment cores to bring all sites to a similar level of comparison. I have also added three new study locations to supplement important end-members of the AP’s wide range of climate, ocean, and glacial variability.

The second chapter of this dissertation details one such sediment record, from Herbert Sound, James Ross Island (Minzoni et al., 2015, Quaternary Science Reviews). This site is crucial to our understanding of climate effects on glacial stability because Herbert Sound presents the first-ever opportunity to compare the marine

2 sediment record with an ice-core from the same drainage basin, spanning the entire Holocene. Moreover, details in climate and glacial dynamics can be further deconvolved through the anomalously well-studied lacustrine and glacial drift records of northern James Ross Island. Herbert Sound records rapid early Holocene deglaciation and onset of a floating ice phase ~10  2.4 cal kyr BP (thousand calendar years Before Present, where present day is 1950 A.D.). Once the floating ice canopy collapsed, the fjord was open and productive for the remainder of the Holocene. The Late Holocene was characterized by colder, heavy sea-ice conditions and greater glacial influence in Herbert Sound; this is directly correlated with ice-core-derived temperature decrease and onshore glacial advances. Each major glacimarine change mimics ice-core temperature events, with the exception of a major ocean warming in the Mid Holocene that captures water mass changes.

The third chapter of this dissertation highlights the glacial history of the Cosgrove Ice Shelf in Ferrero Bay, the southern end-member in the regional comparison of AP glacial response to Holocene climate variability (in prep for PNAS). Ferrero Bay is a narrow embayment of the eastern Amundsen Sea and is a truly polar coastal site (-16C MAAT; mean annual atmospheric temperature). Due to its low latitude (73S), I hypothesized that Ferrero Bay would be insulated from Early Holocene warming, resulting in a delayed response. Remarkably, high-resolution chronology of three sediment cores indicate that grounded ice receded from Ferrero Bay around 10.7 cal kyr BP, prior to most northern AP fjords. Foraminiferal assemblages indicate that retreat may be linked to circulation beneath a large ice shelf and under-melting via relatively warm Circumpolar Deep Water. There are no indications of glacial response to climate events such as the Mid Holocene Hypsithermal and Neoglacial. In fact, the Cosgrove Ice Shelf collapsed during Late Holocene cooling, suggesting that it was out of phase with Antarctic Peninsula ice shelves and not sensitive to Holocene climate variability. Thus the Cosgrove glacial system was oceanographic-, not climatic- controlled during the Holocene.

Comparison of nine coastal embayments with well-dated, expanded Holocene sediment records—including Herbert Sound, Ferrero Bay and several new 3 datasets from previously studied sites—is presented in Chapter 4. This is the most extensive analysis to date on Antarctic Peninsula coastal environments and reveals vulnerability of the region to modern warming (in prep for Nature Geoscience). Sedimentary records are scrutinized closely with the James Ross Island ice-core, which provides a the best regional climate signal to date for the Antarctic Peninsula (Mulvaney et al., 2012). They are also analyzed alongside ocean deep records, the Palmer Deep and Bransfield Basin, which provide long-term regional oceanographic signals. Differences in the timing of glacial response to climate events, as well as differences in the duration of glacial response to climate events, decreased during the Holocene. The co-varying decrease of differences in timing and duration is a measure for increasing regional glacial synchronicity.

This finding illuminates increasing glacial sensitivity to climate change during the Holocene as glaciers became smaller and less susceptible to local forcings. The comparison also highlights dominant local controls on glacial stability during the Early and Mid Holocene warm intervals, including bathymetry and oceanography. Moreover, the progressive increase in regional glacial synchronicity indicates that the Antarctic Peninsula naturally became more sensitive to climate change, so that the recent rapid warming induced a widespread glacial recession. This causes concern for stability of not only the Antarctic Peninsula Ice Cap, but also the larger, marine-based West Antarctic Ice Sheet, which may contribute significant continental ice loss and, combined with thermal expansion of the ocean, poses a major threat to human sustainability. A key motivation for conducting this research is to provide data for linked Cryosphere-ocean- atmosphere models that may help predict future glacial and sea level response to modern warming. 4

Chapter 2 1

First opportunity to compare the 2

marine sediment record with a related 3 1 ice-core in coastal Antarctica 4

The sediment record offshore James Ross Island, northeast Antarctic 5 Peninsula presents an unparalleled opportunity to directly compare marine and terrestrial 6 climate records spanning the Holocene in maritime Antarctica. An 11 m drill core was 7 collected between Herbert Sound and Croft Bay as part of the SHALDRIL NBP-0502 8 initiative and produced the southernmost sediment record from the eastern side of the AP. 9 Thirty-eight radiocarbon ages are used to construct an age model of centennial-scale 10 resolution. Multi-proxy records, including magnetic susceptibility, pebble content, 11 particle size, total organic carbon, and diatom assemblages, were interrogated in the 12 context of nearby Holocene- age ice core, lake, and drift records from James Ross Island. 13

1 This chapter has been edited, reformatted, and reprinted from Quaternary Science 15 Reviews, Vol. 129, p. 239-259, Minzoni, R., Anderson, J.B., Fernandez, R., 16 Wellner J.S., 2015, Marine record of Holocene climate, ocean, and cryosphere 17 interactions: Herbert Sound, James Ross Island, Antarctica. Reproduced with 18 permission from Elsevier. 19 5

Differences in the timing and expression of Holocene events reflect marine controls on 20 tidewater glaciers, such as water mass configurations and sea ice. Glacial behavior 21 mimics ice-core paleotemperatures during the Holocene, with the exception of distinct 22 ocean warming events. Herbert Sound was fully occupied by grounded ice during the 23 Last Glacial Maximum, and experienced rapid lift-off, followed by a floating ice phase. 24 The canopy of floating ice receded by 10  2.4 cal kyr BP, presumably in response to 25 Early Holocene warming. Herbert Sound and Croft Bay fully deglaciated by 7.2 cal kyr 26 BP, when the Mid Holocene Hypsithermal commenced and the sound became open and 27 productive. An extreme peak in productivity ~6.1 cal kyr BP indicates an oceanic 28 warming event that is not reflected in atmospheric temperature or lacustrine sediment 29 records. Increase in sea ice cover and ice rafting mark the onset of the Neoglacial ~2.5 30 cal kyr BP, when pronounced atmospheric cooling is documented in the James Ross 31 Island ice-core. Our comparison facilitates more holistic understanding of atmosphere- 32 ocean-cryosphere interactions that may aid predictions of glacial response to future 33 warming and sea-level scenarios. 34 6

2.1. Introduction 35

The Antarctic Peninsula (AP; Fig. 2-1) is the most rapidly warming region 36 in the Southern Hemisphere, with a documented increase in atmospheric temperature of 37 3.7 ± 1.6 °C per century, ~6 times higher than the global mean (Houghton et al., 2001; 38 Vaughan et al., 2003; Turner et al., 2005; Zagorodnov et al., 2012; Vaughan et al., 2013). 39 The southward shift of isotherms is associated with recent catastrophic break-up of 40 several ice shelves, with the greatest losses in the Larsen Ice Shelf (LIS) in 1995 and 41 2002 (Figs. 2-1 and 2-2; Vaughan and Doake, 1996; Scambos et al., 2003; Rignot et al., 42 2004; Scambos et al., 2009; Vaughan et al., 2013). Today, ice shelves exist where mean 43 annual atmospheric temperatures average less than -9°C, and appear to become unstable 44 and disintegrate where temperatures rise above this threshold (Fig. 2-1; Mercer, 1978; 45 Morris and Vaughan, 2003). Less understood are oceanographic thresholds on ice shelf 46 stability, although upwelling of warmer water masses has been attributed to basal melting 47 and ice-shelf recession in the Ross Sea (Jacobs et al., 1979), Pine Island Bay (Jacobs et 48 al., 2011), and the western AP (Fig. 2-1; Cook and Vaughan, 2010; Pritchard et al., 49 2012). In drainage systems associated with disintegrated ice shelves, glaciers have 50 accelerated, causing concern for the mass balance of the AP Ice Cap (e.g., DeAngelis and 51 Skvarca, 2003; Scambos et al., 2004). In addition, 87% of AP glaciers (Cook et al., 2005) 52 are receding due to dynamic thinning of the AP Ice Cap under the influence of warming 53 oceans and climate (e.g., Rignot et al., 2008; Pritchard et al., 2009). 54

To provide context for modern glacial recession in the AP, we use 55 glacimarine sediment records to reconstruct glacial response to climate events of the 56 Holocene (the last 11.5 kyrs). Due to their restricted drainage basins and relatively high 57 discharge, Antarctic tidewater glaciers are sensitive to both climatic and oceanographic 58 variability at centennial- to decadal time scales. Thus, bays and fjords contain good 59 sedimentary archives of Holocene climate and oceanographic change (Griffith and 60 Anderson, 1989). The basin including Herbert Sound and Croft Bay of James Ross 61 Island (JRI), to which we will collective refer to as Herbert Sound, is an ideal site to 62 7 investigate Holocene glacial history due to its unique environmental setting in the 63 Weddell Sea. 64

Due to a paucity of data, sediment records are often compared with 65 continental ice cores that are far afield and unrelated. Doing so creates an overly 66 simplistic perception of the cryosphere and its ocean-atmosphere interactions, while 67 marginalizing regional variability. The sedimentary record of Herbert Sound provides an 68 unparalleled opportunity to compare Holocene marine records not only with extensive 69 lacustrine and drift records from James Ross Island (JRI; Ingolfsson et al., 1992, 1998; 70 Bjork et al., 1996; Hjort et al., 1997; Strelin et al., 2006; Carrivick et al., 2011; Davies et 71 al., 2014), but also with a high-resolution Holocene ice core—the first time this has been 72 possible in an Antarctic maritime setting. We present core and seismic data from the 73 high-accumulation, well-preserved, and chronologically well-constrained sedimentary 74 record of Herbert Sound, and scrutinize this dataset with terrestrial climate records from 75 the same drainage basin (Figs. 2-1 and 2-2). SHALDRIL NBP-0502 Core Site 2 is 76 located just 27 km from the recently collected JRI ice-core drill site, which spans the 77 Holocene (Fig. 2-2; Mulvaney et al., 2012). This comparison allows delineation of 78 glacial response to well-documented climate and oceanographic events. 79 8

Figure 2-1 Bathymetric and elevation map of the Antarctic Peninsula illustrating complex orography, oceanography, and climate. Note ice shelf loss between the -5 and -9C isotherms. (Original figure with information compiled from Landsat; NASA; GeoMapApp; Reynolds, 1981; Hofmann et al., 1996; Smith et al., 1996; Morris and Vaughan, 2003; Martinson et al., 2008). 9

81 Figure 2-2 Study area map of James Ross Island and Herbert Sound. 82 a) Shaded relief map of NE Antarctic Peninsula with reconstructed LGM ice 83 sheet flow directions in grey arrows from Domack et al., 2005, and in black 84 arrows from Camerlenghi et al., 2001. Historic extent of the Prince Gustav Ice 85

Shelf (PGIS) in Prince Gustav Channel (PGC) and the former Larsen A Ice Shelf 86 (LIS-A) marked with dates A.D. b) SHALDRIL NBP-0502 multi-beam swath 87 bathymetry of Herbert Sound and Prince Gustav Channel, with Knudsen sub- 88 bottom profile location in the mid basin. Drainage basin area highlighted 89 yellow. c) Subtle linear features in the outer basin. d) Small basement high in 90 the mid basin that may have served as a pinning point during glacial retreat 91 (see Fig. 2-3). 92

2.2. Background 93

2.2.1. Geologic setting 94

Herbert Sound and Croft Bay comprise a large protected fjord that drains 95 the northern side of JRI, the largest island in the western Weddell Sea (Figs. 2-1 and 2-2). 96 Croft Bay is the southernmost basin of this fjord, and Herbert Sound connects Croft Bay 97 with Prince Gustav Channel (PGC) to the north and with Erebus and Terror Gulf to the 98 east (Fig. 2-2). We collectively refer to Croft Bay and the N-S axis of Herbert Sound as 99 the Herbert Sound. The fjord averages ~350 m water depth and is bounded by steep 100 walls and shallow sills. It is enclosed by Ulu Peninsula to the west, the Naze Peninsula to 101 the east, and Vega Island to the northeast. The drainage basin area of Herbert Sound is 102 ~400 km2, including both James Ross and Vega islands (Fig. 2-2). 103

JRI is a large and long-lived polygenetic shield volcano, formed in the 104 back-arc basin of the active volcanic arc in Bransfield Basin to the west. A large ice cap 105 ~100-400 m thick is centered at Mt. Haddington, which reaches over 1600 m elevation 106 (Fig. 2-2). Northern JRI is an anomalously well-mapped area of the AP, consisting of 107 Jurassic deep-marine mudstones and exposed dipping early Cretaceous to late Eocene 108 metasediments, overlain by the Neogene James Ross Island Volcanic Group, which 109 includes basalts, tuffs, volcaniclastic deltas and glacigenic diamictites (Bibby, 1966; 110 Nelson, 1975; Ineson et al., 1986; Smellie et al., 2006; Smellie et al., 2008; Hambrey et 111 al., 2008; Riley et al., 2011). Pleistocene and Holocene drift and modern glaciers, 112 including rock glaciers, cover the landscape (Rabassa, 1983; 1987; Rabassa et al., 1982; 113 Lundqvist et al., 1995; Davies et al., 2012a). 114

115

2.2.2. Climate and oceanographic setting 116

In contrast to the western AP, where relatively warm Circumpolar Deep 117 Water exerts a strong influence on ocean temperature, the eastern side of the AP is 118

12 influenced by colder, more saline Weddell Sea Transitional Water (Fig. 2-1; Hofmann et 119 al., 1996; Garcia et al., 2002). The NW region of the Weddell Sea is generally 120 characterized by extensive sea ice and abundant icebergs that pass through the area under 121 the influence of the Weddell Gyre (Venegas and Drinkwater, 2001). JRI is insulated 122 from the Westerlies by the AP Mountains, resulting in colder and drier conditions than 123 the north and west AP (Turner et al., 2002; King et al., 2003). Modern borehole 124 measurements from Mt. Haddington provide a mean annual temperature of -13.8°C, 125 about 9°C lower than the nearest meteorological measurements from Esperanza Station, 126 located at the northernmost tip of the AP (Fig. 2-2; Abram et al., 2013). The coast of JRI 127 is estimated to be more similar to Esperanza Station (~ -5.3°C), based on the altitudinal 128 lapse rate of the AP (Reynolds, 1981; NASA, 2004; Abram et al., 2013). Precipitation 129 averages 63 mm/yr (Abram et al., 2013). 130

131

2.2.3. Glacial history 132

During the Last Glacial Maximum (LGM), the AP Ice Sheet extended to 133 the continental shelf break on both sides of the AP, as evidenced by spectacular glacial 134 geomorphic features that extend across the shelf (Anderson et al., 2002; Evans et al., 135 2005; Heroy and Anderson, 2005; 2007; Wellner et al., 2006; O’Cofaigh et al, 2014; 136 Lavoie et al, 2014). Several large ice streams drained the northern AP, with ice 137 transporting erratics from Trinity Peninsula to JRI until 18 ka (Glasser et al., 2014). 138 Glacial drainage converged into an ice stream that flowed westward through PGC 139 (Camerlenghi et al., 2001) until 12 ka (Fig. 2-2; Glasser et al., 2014). Following the 140 LGM, recession of the ice sheet from the western side of the AP occurred in step-wise 141 fashion, with two significant steps coincident with meltwater pulses at 14 and 12 ka 142 (Heroy and Anderson, 2007; O’Cofaigh et al., 2014). Late stage recession on the 143 Weddell Sea side of the AP is poorly constrained, but appears to have had a similar 144 history to the western AP (O’Cofaigh et al., 2014). As the ice sheet receded across the 145 inner shelf and into inland passages, drainage progressively separated into numerous, 146

13 smaller outlets that appear to have experienced asynchronous recession (Heroy and 147 Anderson, 2007; Milliken et al.,2009; Michalchuck et al., 2009; Allen et al., 2010). 148

149

2.2.4. Ice Shelf History of the NW Weddell Sea 150

Ice shelves are much more extensive in the eastern AP than the western 151 AP, and these large expanses of floating ice are vulnerable to recent warming (Figs. 2-1 152 and 2-2). Since 1945, Larsen Ice Shelf-A (LIS-A) experienced steady westward 153 recession in the area between Robertson and James Ross islands, finally shrinking to a 154 small remnant, the Prince Gustav Ice Shelf (PGIS; Figs. 2-1 and 2-2; Reece, 1950; 155 Doake, 1982; Glasser et al., 2011). No ice shelf has been observed in the northern Prince 156 Gustav Channel (>64°S latitude), and early explorers identified the geographical 157 transition from ice shelf to fast sea ice by a major drop in elevation from >10 mabsl to ~3 158 mabsl (Reece, 1950). This small ice shelf included the Sjögren Glacier Ice Tongue, 159 which once merged with an ice tongue sourced from Röhss Bay that is open today (Fig. 160 2-2). LIS-A collapsed along with the final remnants of the PGIS in 1995 (Glasser et al., 161 2011). LIS-B lost >12,500 km2 area in 2002, the largest collapse in observed history 162 (e.g., Scambos et al., 2003). 163

Pudsey and Evans (2001) demonstrated that the PGIS was absent during 164 the mid to late Holocene (5-2 ka) and reformed during the late Holocene, based on ice 165 rafted debris (IRD) provenance from marine sediment cores. The LIS-A was also absent 166 during mid-Holocene time and re-formed in the Late Holocene, based on sediment cores 167 dated using geomagnetic palaeointensity (Brachfield et al., 2003). LIS-B was present 168 throughout the Holocene but progressively thinned, and 2002 was the first time it was 169 open since the last interglacial (Domack et al., 2005). There is currently a data gap on 170 Holocene floating ice stability north of the former PGIS. Herbert Sound complements 171 the ice shelf history of the NW Weddell Sea with data from a smaller, more restricted 172 glacial system of northern JRI. 173

2.3. Methods 174

2.3.1. Geophysical methods and coring 175

Multi-beam swath bathymetry data collected using a hull-mounted Simrad 176 EM-120 during the RV/IB Nathaniel B. Palmer -0502 SHALDRIL provided a 177 bathymetric map of the seafloor in Herbert Sound (Shipboard Scientific Party, 2005). An 178 ODEC Bathy 2000 (3.5 kHz chirp sonar) imaged the upper strata. Deeper strata were 179 imaged using a 15 in3 water-gun aboard the RV Polar Duke in 1991. Bathymetric and 180 seismic data were used to identify geomorphic features in Herbert Sound and to survey 181 ideal coring sites with thick, stratified sediment far from gravity flow deposits. A velocity 182 model (1500 m/s) was applied to seismic data and used to interpret stratigraphic 183 relationships. 184

Two cores collected on the NBP-0502 SHALDRIL cruise, a 11.5 m long 185 drill core and a 3.0 m long Kasten core, are the primary data set for detailed 186 environmental reconstructions. Five cores collected aboard RV Polar Duke in 1991 (PD- 187 91) and twelve cores collected aboard the USCGC Glacier in 1982 (Deep Freeze DF-82) 188 augmented our analysis of sediment facies and their distribution and complement the 189 chronology of the SHALDRIL cores. NBP0502 KC-2B had 100% recovery, while 190 NBP0502 drill core 2C had 73% recovery, with 0.5 to 1.0 m gaps due to the coring 191 process. At 10.5 m, the drill core encountered muddy diamicton at the base of the 192 acoustically layered seismic section (Fig. 2-3). 193

194

195

Figure 2-3 NBP0502 sub-bottom profile from southern Herbert Sound. 196 Un-interpreted (a) and interpreted (b) sections highlight two seismic facies: a 197 lower transparent till facies (purple) and a thick draping, laminated 198 glacimarine facies (blue). The profile crosses NBP0502 Site 2, marked in red. 199 Basement peaks are visible in bathymetric data (Fig. 2-2). 200

2.3.2. Sedimentologic and stratigraphic methods 201

Cores were split and processed onboard using standard procedures, 202 including Munsell soil color determination and shear strength measurements with a 203 Torvane. Core descriptions included grain size, grain shape, pebble lithology, texture, 204 and sedimentary structures. Smear slides were analyzed every ~0.5 m onboard for grain 205

16 shape, mineralogy, and microfossil assemblages. Refer to Shipboard Scientific Party 206 (2005) for additional information. 207

Core X-radiographs were taken of the archive core halves at the Antarctic 208 Research Facility in Tallahassee, Florida in order to characterize sedimentary structures 209 such as lamination and bioturbation, to determine gravel abundance, and to search for 210 macrofossils. Gravel abundance was determined by counting in-situ pebbles (diameter 211 >2 mm) over 5 cm intervals using X-radiographs. A Geotek Multi-Sensor Core Logger 212 (MSCL) measured magnetic susceptibility (MS) and Electric Resistivity (ER) every 2 213 cm, using a two-pronged probe (K. Burgdorf, unpublished B.A. thesis, Middlebury 214 College, Vermont). 215

216

2.3.3. Sedimentological, geochemical, and paleontological methods 217

Samples were taken every 10 cm for detailed sedimentological, 218 geochemical, and paleontological analyses at Rice University. Quantitative grain size 219 measurements were obtained by laser particle diffraction analysis using a Malvern 220 Mastersizer 2000 (McCave et al., 1986). Statistics were performed following Folk and 221 Ward (1957). 222

Total organic carbon (TOC), bulk carbon, nitrogen, and hydrogen 223 concentrations were determined for 126 samples using a Costek Elemental Analyzer, 224 following vaporous HCl acidification methods described by Komada et al. (2008). 225 Calibrations with proline standards yielded errors less than 0.01% for carbon and 226 nitrogen. 227

Nineteen samples were wet sieved at 63 µm for foraminiferal abundance. 228 Abundances are calculated by dividing the number of foraminifera tests by grams of 229 dried sediment (t/gds). Eighty-eight samples were analyzed for absolute diatom 230 abundance and assemblage analysis following the settling method of Scherer (1994). 231 Diatoms were counted in each prepared slide at 1000X magnification using an oil 232 immersion objective lens. Diatom assemblages of the AP are typically dominated by the 233

17 cosmopolitan Chaetoceros resting spores (CRS), which constitute 70 to 90% relative 234 diatom abundance (Leventer et al., 1996; Sjunneskog and Taylor, 2002). Two counts 235 were conducted; 88 slides were counted for high-resolution CRS abundance as indicators 236 of productivity, and 58 of those slides were counted for CRS-free assemblages in order to 237 extract meaningful ecological information from minor species that are more sensitive to 238 environmental conditions. Each prepared slide was counted with a minimum of 400 non- 239 CRS valves. Only species that contribute >2% relative abundance are presented for each 240 sample. Five slide transects were counted where non-CRS species abundances were low. 241 Absolute diatom abundances are calculated using the equation: 242

T = (NB/AF)/M, 243 where T= number of microfossils per unit mass, N = total number of microfossils 244 counted, B = area of bottom of beaker (mm2), A = area per field of view or transect 245 (mm2), F= number of fields of view or transects counted, M = mass of sample in grams 246 (Scherer, 1994). 247

The diatom assemblage data were examined statistically using Principal 248 Component Analysis (PCA) to describe principal stratigraphic zones in the post-glacial 249 sediment using Kovach Multivariate Statistical Package v. 3.22. Diatom assemblage data 250 were also analyzed using Fuzzy C-means (FCM) cluster analysis in the TACSWORKS 251 software suite (Gary et al., 2009). FCM is an especially valuable tool for 252 paleoenvironmental analysis because—unlike PCA, which uses a hard-clustered matrix 253 that poses species exclusivity constraints to the clusters—FCM relaxes the class 254 exclusivity constraint and allows species to have memberships in more than one cluster 255 (Zadeh, 1965; Bezdek, 1987; Bezdek and Pal, 1991), which is perhaps more 256 representative of the natural environment. FCM has been applied to Holocene-age 257 micropaleontological datasets where traditional PCA yields ambiguous results (cf. Erbs- 258 Hansen et al., 2011; 2012); for this reason we chose to apply FCM to glean more 259 environmental information from the Herbert Sound dataset. Relative species membership 260 is determined algorithmically by the Euclidean distance from cluster centers (or 261 “centroids”) in multi-variate space and has a value that ranges between 0 and 1, such that 262 for each sample all its cluster memberships will sum to one. Cluster hardness is manually 263

18 controlled by the “fuzzy exponent”, which ranges from 1.0 to 3.0, where 1.0 is equivalent 264 to a species-exclusive hard cluster. The fuzzy exponent value used in this data set is 1.4. 265 Before running the analysis, the data were filtered using a 95% confidence interval 266 criterion to exclude rare taxa. Navicula spp. were grouped due to ambiguity in species 267 identification. The fossil diatom assemblages were interpreted using modern 268 assemblages and environmental associations with substrate, sea ice conditions, water 269 temperature, water salinity, and turbidity. 270

2.3.4. Radiocarbon dating 271

Twenty-seven marine carbonate samples and eleven bulk sediment 272 samples were collected from six cores from Herbert Sound to derive high-resolution 273 chronostratigraphy (Table 2-1). Carbonate samples were sieved and cleaned with 274 distilled water, dried at 50°C, and analyzed at UC Irvine AMS (UCIAMS) facility. Four 275 carbonate samples from piston cores were sent to the Woods Hole National Ocean 276 Science AMS (NOSAMS) facility (Maher et al., 2004). A regional reservoir of 700 years 277 was applied to the 14C ages, entered into Calib 6.0 Calibration software, and calibrated 278 using the Marine09 curve, which has a built in global correction of 400 years, totaling a 279 1100 year correction (Stuvier and Reimer, 1998; Stuvier et al., 2005; Reimer et al., 2009). 280 The reservoir correction is conservative and is chosen by convention to compare Herbert 281 Sound with the two highest-resolution studies conducted in the AP, SHALDRIL sites 282 from Maxwell Bay (Milliken et al., 2009) and Firth of Tay (Michalchuk et al., 2009). 283

Bulk sediment was processed and analyzed by NOSAMS for AIO (acid 284 insoluble organic fraction) ages (McNichol et al., 1994). To calibrate AIO ages, some 285 authors subtract core top AIO ages from down-core ages (Domack, et al., 1999; Andrews 286 et al., 1999), whereas others calibrate down-core AIO ages with carbonate ages (Licht et 287 al., 1998). The surface age calibration requires an assumption that relative contributions 288 of old and autochthonous carbon are constant through time, even as the sedimentary 289 components vary, which is a tenuous assumption. Here AIO dates are calibrated 290 according to the procedure in Licht et al., 1998, with 6 coincident carbonate and bulk 291

AIO sample intervals, and surface age of -55, corresponding to the date when the core 292 was collected, 2005 AD, where present day is 1950 AD (Fig 2-4). 293

2.4. Results 294

2.4.1. Seismic and bathymetric surveys of Herbert Sound 295

Herbert Sound is characterized by a broad, gentle depression (~400 mbsl), 296 which shallows toward the outer fjord (~350-330 mbsl; Fig. 2-2). The mouth of the fjord 297 is characterized by a prominent grounding zone wedge, and an abrupt drop off into the 298 ~850 to 1250 m deep PGC. Seismic profiles show thick (up to ~30 m), continuous 299 deposits that drape the central fjord and pinch out along fjord walls and bedrock highs 300 (Fig. 2-3). The sediment package directly overlies steeply dipping strata, assumed to be 301 the same Cretaceous metasediments that are exposed in the walls of the fjord. A 302 transparent and chaotic seismic facies characterizes the 3-5 m thick basal till unit that 303 rests on acoustic basement that pinches out on highs (Fig. 2-3). Thinly laminated 304 glacimarine units rest above the till and exhibit draping, parallel reflections and subtle 305 onlap onto highs. This stratigraphic succession is similar to that observed in other AP 306 fjords (Griffith and Anderson, 1989; Simms et al., 2011) and indicates that pre-LGM 307 sediment was excavated during the LGM. 308 Multibeam swath bathymetry reveals subtle linear features that are 309 partially buried by thick glacimarine deposits. The linear features are parallel to the fjord 310 axis, extending from the mid basin to the mouth of the sound (Fig. 2-2). They average 311 heights of 60 m over a distance of ~1.0 to 1.6 km and ~0.6 km width at the apex. 312 Orientation shifts from N40E in the inner fjord to N15W in the mid fjord. A small 313 bedrock high marks the transition in flow direction, indicating a brief period of ice 314 stability at this location (Figs. 2-2 and 2-3). Johnson and colleagues (2011) refer to the 315 main axis of Herbert Sound as “Croft Trough” due to its elongate features documenting 316 concentrated ice flow from northern JRI during the LGM. 317

2.4.2. Radiocarbon results 318

Radiocarbon chronology is constrained by a total of 38 ages (Table 2-1; 319 Fig. 2-4). The shallowest carbonate-derived age was 1.505 ± 0.030 radiocarbon kyr BP, 320

21 calibrated in Calib 6.0 to 0.450 ± 0.140 cal kyr BP, obtained at 0.595 mbsf in the 321 SHALDRIL KC-2B. The ages are in stratigraphic order and provide an accumulation 322 rate of ~1.19 mm/yr (Fig. 2-4). Drill core NBP0502-2C yielded 13 carbonate-derived 323 radiocarbon ages, mostly from Yoldia eightsii and Laternula spp. shells, and included 324 three outliers (Fig. 2-4). These outliers were obtained from low-mass foraminifera 325 samples within the upper sections 4E-1 (lower glacimarine mud) and 5E-1 (diamicton). 326 Five articulated pelecypods, with periostrachum still intact, were extracted from the 327 center of the upper 50 cm of section 4E-1 in an effort to extract a reliable date and are 328 remarkably consistent, with a median radiocarbon age range of less than 10 years (Table 329 2-1; Figs. 2-4 and 2-5). Removal of foraminifera-derived ages results in a calibration 330 curve for NBP0502 Core 2C that is best fit by the line y = 0.00002x2 – 0.0507x + 76.756 331 with a linear regression of R2 = 0.9081 (Fig. 2-4). 332

In an effort to further constrain the age model, 13 bulk sediment samples 333 were dated using AIO AMS analysis (Table 2-1; Fig. 2-4). AIO dating has been applied 334 in regions where carbonate is sparse, such as the Ross Sea, Antarctica, with limited 335 success due to uncertainty in the reservoir correction (Hall et al., 2010) and reworking of 336 old carbon (Andrews et al., 1999; Licht and Andrews, 2002). AIO ages have proven to 337 be especially problematic in proximal glacimarine deposits where authigenic carbon is 338 low and the probability of reworking of old carbon is high. Distal diatomaceous 339 glacimarine sediments have proven more reliable for AIO analyses, particularly where 340 multiple carbonate ages are present in stratigraphic order, and sedimentation rates are 341 near-linear (e.g., Domack et al., 2001). The reason for obtaining AIO ages in Herbert 342 Sound is three-fold: to clarify discrepancies in the carbonate-derived age model, to 343 provide additional ages where carbonate is sparse, and to determine whether AIO dating 344 may be successfully applied to the Weddell Sea region. 345

The surface sample from NBP0502 KC-2B yielded an age of 5.210 ± 346 0.030 radiocarbon kyr BP (Table 2-1). This implies significant input of dead carbon into 347 the reservoir, but is not unexpected given high core-top ages (>6,000 yrs) from Prince 348 Gustav Channel, which drains southern JRI where dead carbon from Cretaceous strata are 349 calculated to contribute ~40% to the local reservoir (Pudsey and Evans, 2001; Evans et 350

22 al., 2005; Pudsey et al., 2006). Indeed, dark brown-tinted amorphous organic matter and 351 Cretaceous dinocysts similar to those described in Prince Gustav Channel were found in 352 Herbert Sound sediment. 353

KC-2B and drill core 2C were correlated using their age models in 354 addition to proxy data, such that core 2C was shifted down 1.5 mbsf (Figs. 2-4 and 2-5). 355 The surface age was assumed to be -55 yrs BP when the core was collected (by 356 convention 0 yr BP means 1950 AD), and 6 AIO ages were plotted vs. carbonate-derived 357 ages to provide a calibration curve of y = 1.0143x - 5210 with a linear regression of R2 358 =0.9532 (Fig. 2-4). For comparison, the foraminifera-derived outlier from 4E-1 was 359 plotted instead of the pelecypod age at the same interval and resulted in poor linear 360 regression (R2 = 0.5909; Fig. 2-4). In all cases, calibrated AIO ages strongly corroborate 361 the pelecypod ages (Table 2-1; Fig. 2-4). 362

One AIO age was taken from the top of the diamicton, recognizing that it 363 was likely to yield an erroneously old age due to reworked carbon in this facies. Even 364 though the top of the proximal glacimarine unit yielded a date of 31.200 ± 0.760 kyr BP, 365 which was calibrated to 25.568 cal kyr BP, it indicates that the underlying young 366 foraminifera-derived ages from this unit are indeed erroneous measurements due to low 367 mass (Table 2-1; Figs. 2-4 and 2-5). The foraminifera samples were taken at the top of 368 drill core sections and comprise the richest faunal diversity of the entire core, suggesting 369 the possibility of drilling contamination, as has been suggested for similar spurious 370 radiocarbon results from other SHALDRIL cores (Milliken et al., 2009; Michalchuck et 371 al., 2009). Thus, these outlier ages are not used in the age model and these intervals are 372 not used for paleoenvironmental reconstructions. 373

The base of the distal marine unit yielded an AIO radiocarbon age of 374 17.850 ± 0.140 cal kyr BP, which was calibrated to 12.406 cal kyr BP (Table 2-1; Fig. 2- 375 4). This represents the maximum age of deglaciation and onset of glacimarine 376 deposition. Dates extracted from legacy cores PD91 PC-06 and -08 and DF82 PC-192 377 and -190 complement the drill core chronology (Figs. 2-4 and 2-5). Unit 2 of NBP0502- 378 2C correlates with the high accumulation lower unit of core PD91- PC-08 (0.60 to 8.33 379 mbsf), which yielded a corrected carbonate-derived age of 8.610 ± 0.962 cal kyr BP at its 380

23 base, and 6.690 ± 0.444 cal kyr BP from a shell near its top (0.48 mbsf in PD91 PC-08; 381 Table 2-1; Fig. 2-4). Sedimentology of the units is similar, although PC-08 contains 382 abundant ash layers and has a significantly higher accumulation rate (Fig. 2-5). 383 Proximity to a local bathymetric high may explain the expanded section in PC-08, likely 384 resulting from sediment gravity flows. The carbonate age of 8.6 cal kyr BP taken from 385 the base of PD91- PC-08 is taken as the minimum age of deglaciation, while the AIO age 386 of 12.4 cal kyr BP is taken as the maximum age of deglaciation in Herbert Sound. This 387 results in a median age for glacial recession of 10 ± 2.4 cal kyr BP. 388

389

Table 2-1 Radiocarbon ages from Herbert Sound. 390

391

Figure 2-4 Age model of Herbert Sound. 392 a) Carbonate-derived calibrated radiocarbon ages from six cores in Herbert 393 Sound. Note three outlier foraminifera-derived ages (x). The basal age from 394 PC-08 (denoted by closed square) provides a minimum age of glacial 395 recession. b) Calibration of AIO ages was conducted by plotting uncalibrated 396 AIO ages vs. calibrated carbonate ages from cores 2B and 2C. Removing 397 foraminifera-derived ages results in good linear regression. c) Combined age 398 model of NBP-0502 Site 2. Characteristic hockey-stick shape of radiocarbon 399 curve in the basal units is due to reworked carbon in the proximal setting. 400 Basal AIO ages are taken as maximum ages of glacial retreat. Lithologic legend 401 included in Figure 2-5. 402

403

404 Figure 2-5 Correlation of NBP-0502 Site 2 drill core and Kasten core with PD-91 and DF-82 piston cores. 405 Correlations based on radiocarbon ages, magnetic susceptibility curves, and lithology. 406

407

Figure 2-6 Stratigraphic plot of key proxies used to interpret paleoenvironments in NBP0502 Site 2. 408

Proxies include: magnetic susceptibility (MS), maximum shear, pebble counts from x-radiographs, foraminifera 409 abundace per gram of dried sediment (gds), total organic carbon (TOC), total diatom abundance and Chaetoceros 410 Resting Spore (CRS) abundance in millions of valves per gram of dried sediment (million v/gds), and relative diatom 411 assemblage membership from Fuzzy C-means. Explanation of diatom assemblages (DA’s) given in Table 2-2. Gaps in 412 the core are shaded light gray. 413

2.4.3. Sedimentological, geochemical, and paleontological results 414

Fluctuations in sedimentologic, geochemical, and paleontological 415 parameters of the SHALDRIL Site 2 cores are used to subdivide the core into integrated 416 litho-, chemo-, and bio- stratigraphic units (Fig. 2-6). Four main units are recognized, 417 with subunits where appropriate. Magnetic susceptibility is highly variable and generally 418 lower in the basal diamicton than in the overlying glacimarine silts. Shear strength 419 increases down-core as a result of de-watering and compaction. Particle grain size 420 generally decreases from the basal units to core top, while carbon and nitrogen generally 421 increase (Figs. 2-6 and 2-7). Silt dominates the overall grain size distribution within the 422 glacimarine interval. However, total grain size analysis reveals distinct populations that 423 reflect variations in marine influence through time (Fig. 2-7). 424

Foraminiferal abundance is low in Herbert Sound (<40 tests/gds; Fig. 2-6). 425 Foraminifera have highest abundance and diversity in the glacimarine units, with peaks 426 where calcareous percentages are highest (up to 150 tests/gds; Fig. 2-6). The upper 10-30 427 cm of drill core 2C sections 4E-1 and 5E-1 are calcareous foraminifer-rich, but were 428 likely affected by contamination during the drilling process (Figs. 2-4, 2-5, and 2-6). 429 Assemblages are similar to those found in Vega Drift of northern Prince Gustav Channel 430 (Szymcek et al., 2007) and Firth of Tay (Majewski and Anderson, 2009), but are entirely 431 benthic. Arenaceous taxa include Milammina arenacea, which dominates and often 432 forms monospecific assemblages, and minor species such as Paratrochammina lepida, 433 Portatrochammina bipolaris, Spiroplectammina spp., and other unidentifiable forms that 434 may be fragments of Reophax spp. Calcareous assemblages are dominated by robust 435 Globocassidulina biora tests with evidence of partial dissolution, which, along with 436 abundance of arenaceous species, is a strong indicator of preservation bias. Other less 437 abundant calcareous species include Fursenkoina fusiformis, Stainforthia concava, 438 Procerolagena gracillis and Uvigerina spp. Due to preservation bias, the foraminifera 439 were not analyzed for assemblage information. 440

Diatoms are well preserved and abundant in Herbert Sound. Absolute 441 diatom concentrations are broadly correlated with lithology, as they are absent to sparse 442

30 in the diamicton unit, and relatively abundant in glacimarine sediments (Fig. 2-6). 443 Significant variation in diatom concentrations and assemblages exist within the 444 glacimarine sediments, however (Figs. 2-6, 2-8, and 2-9). Silty mud dominates the upper 445 9 m of core with minor lithologic change, yet diatoms increase mid core and decrease in 446 the upper core with significant shifts in non-CRS species associations (Fig. 2-6). Low 447 valve concentrations of ~0.7 x 106 valves per gram of dried sediment (v/gds) characterize 448 the basal diamicton and increase vertically in the overlying thin clay unit to 115 x 106 449 v/gds. The overlying glacial marine units contain high and variable diatom 450 concentrations, ranging from 155 to 3087 x 106 v/gds, with an average concentration of 451 665 x 106 v/gds. Despite fluctuating total diatom concentration, CRS compose the 452 majority of the assemblages, with average values of ~85% in the glacimarine units (Fig. 453 2-6). The diatom concentrations in NBP0502 Site 2 are consistent with Holocene marine 454 sediment from other fjords of the AP region (e.g. Taylor et al., 2001; Sjunneskog and 455 Taylor, 2002; Allen et al., 2010; Christ et al., 2014). 456

The diatom species abundance dataset was interrogated using PCA, both 457 to identify major axes in multi-variate space that explain the variability in the data set, 458 and to evaluate the relative importance of each species on the extracted axes (variable 459 loadings). The first three axes have eigenvalues of 50.6%, 12.4%, and 10.0%, 460 respectively, accounting for 73.1% of the total variability in the data set (Fig. 2-8). The 461 first and second axes have the highest eigenvalues and account for the most significant 462 assemblage transitions. The relative loadings of these axes are used to help delineate 463 trends in the diatoms and identify units (Fig. 2-8). 464

Figure 2-7 Grain size and x-radiographs of Herbert Sound. Particle size frequency distribution for all the samples analyzed in each unit of NBP-0502 Site 2 cores. Distributions are shown in log-linear graphs from the Malvern Mastersizer 2000 software. Volume is reported as percent of the total sample. Representative x-radiographs from each unit are displayed to the right to illustrate sedimentary structures and pebble abundance.

Figure 2-8. PCA results from Herbert Sound diatom counts. a) Ordination bi-plot of variable loadings for first two axes of the Principal Components Analysis, with significant species variables (>1σ) labeled. b) Stratigraphic plot of component scores for Axes 1 and 2.

Figure 2-9 Relative abundance of non-CRS diatom taxa Taxa with significant variable loadings on Axis 1 and Axis 2 from the PCA are plotted.

Diatom counts were next analyzed using Fuzzy C-Means analysis with a fuzzy exponent of 1.4. Four diatom assemblages (DAs) were defined by fuzzy cluster memberships, and ranked by most abundant species (Figs. 2-6, 2-8, and 2-9). Diatom assemblages are summarized in Table 2-2, and relative membership down-core is plotted in Figure 6. Thalassiosira antarctica T2 (warm variety) composed the greatest percentage of 3 DA’s, while Navicula spp. dominated the fourth DA. DA’s are ranked by % abundance and accessory species composition. They include: (DA1) the Thalassiosira antarctica T2 dominant DA (31.7% membership of T. antarctica T2, followed by 16.2% T. antarctica T1); (DA2) the Thalassiosira antarctica T2 and T1 DA (20.2% T2 followed by 13.7% T1 with accessory Navicula spp. and F. curta); (DA3) the T. antarctica and F. cylindrus mixed DA (13.6% T2, 11.2% F. cylindrus, 9.9% T1, and 8.7% F. curta with accessory F. vanheurkii and E. antarctica); and (DA4) the Navicula spp. DA (15.34% Navicula spp., 11.0% F. curta, followed by 9.3% T1 and 8.1% T2). Stratigraphic Fuzzy cluster membership plots were used in conjunction with PCA loadings, sedimentology, geochemistry, and geotechnical properties of SHALDRIL NBP0502 Site 2 to define four major stratigraphic units and their subunits (Figs. 2-6, 2-7, 2-8, and 2-9).

Table 2-2 Summary of diatom assemblages (DA’s), their composition, and interpretations.

2.4.4. Unit 1 (11.4 to 9.9 mbsf)

The basal unit of the NBP0502 drill core, Unit 1 (11.4 to 9.9 mbsf) is a very dark greenish grey (gley 1; 3/N) diamicton that is characterized by high pebble concentration (Fig. 2-6). Grain size analysis of the matrix reveals higher levels of sorting than is typical of till and proximal glacimarine sediments (Anderson, 1999), which is due to the abundance of a sorted fine sand component that is likely derived from Cretaceous metasediments in the area (Fig. 2-7). The base of Unit 1 is characterized by ~60% sand, 30% silt, and 10% clay (Fig. 2-6). Relative percentage of sand decreases significantly up-section to ~9%, while silt and clay increase to 68% and 23%, respectively. Unit 1 contains several large pebbles (greater than 10cm; Fig. 2-6). Pebble content increases up- section, while pebble size decreases. Pebble lithology is typical of the JRI volcanic group. Maximum shear strength is 0.18 kg/cm2, which is typical of diamicton that has proximal glacimarine origin or is “soft” till (Anderson, 1999). Due to the paucity of diatoms in Unit 1, samples did not meet the criteria for 95% confidence and were not included in the statistical analysis of diatom assemblage information. The contact between Unit 1 and Unit 2 is marked by abrupt decrease in grain size and the sorted fine sand component, decrease in pebbles, increase in TOC, and increase in faunal abundance (Fig. 2-6).

2.4.5. Unit 2 (9.9 to 9.7 mbsf)

Unit 2 consists of a thin, clay-rich dark grey mud with laminations (Figs. 2-6 and 2-7). Pebbles are sparse and small. Unit 2 is poorly sorted, with the highest percentage of clay in the core (Figs. 2-6 and 2-7). Total diatom abundance is low and increases at the top of the unit (Fig. 2-6).

2.4.6. Unit 3 (9.7 to 8.4 mbsf )

Unit 3 consists of dark greenish grey (1 gley; 3/N) silty, sandy mud with black mottles throughout (9.7 to 8.4 mbsf; Fig. 2-6). Sediment accumulation rate in Unit 3 is low (>0.097 mm/yr; Fig. 2-4). Silt dominates the unit (~60%), while sand (~20- 30%) has noteworthy abundance (Figs. 2-6 and 2-7). Pebbles increase up-section, coinciding with increasing sand and silt concentration and decreasing clay concentration (Fig. 2-6). TOC increases substantially in Unit 3 to 0.87%. Total diatom concentrations mirror this trend, increasing with CRS. Unit 3 is characterized by DA3 at its base, and transitions to DA1 at its top (Table 2-2; Fig. 2-6). Unit 3 has statistically significant concentrations of Eucampia antarctica var. antarctica. The contact between Unit 3 and Unit 4 is defined by decrease in diatom abundance, decrease in MSCL, increase in silt concentration to ~70%, and loss of calcareous foraminifera (Fig. 2-6).

2.4.7. Unit 4 (8.4 to 2.9 mbsf)

Unit 4 is a dark greenish grey (1; 3/10Y) silty mud with faint black mottling and green laminations (Fig. 2-6). Particle size is dominated by ~80 % silt, with minor variations in clay (10-20 %), and occasional sandy lenses (up to 25%; Fig. 2-6). Grain size data show moderate sorting and decrease in fine sand and increase in fine silt relative to Unit 3 (Fig. 2-7). Pebble concentration is low throughout Unit 4, with the exception of one peak at 4.7 mbsf (Figs. 2-6 and 2-7). Foraminifera are mostly arenaceous, and of low abundance. Unit 4 is dominated by DA1 and DA2 (Table 2-2; Fig. 2-6). The top of Unit 4 shifts much greater membership of DA4. Significant changes in geochemistry and faunal abundances and assemblages are used to subdivide Unit 4 into three subunits. Unit 4a (8.4 to 6.4 mbsf) is characterized by a gradual decrease in TOC with punctuated intervals of higher TOC that correlate with diatom abundance peaks (Fig. 2-6 and 2-8). E. antarctica var. antarctica decreases up-section and is near absent throughout the overlying core sediment. Unit 4b (6.4 to 5.5 mbsf) is characterized by a sharp increase to the highest TOC values throughout the core (1.93 % at 5.9 mbsf; Fig. 2-6). This productivity peak is accompanied by high membership of

DA1 (Fig. 2-6). Unit 4c (5.5 to 2.9 mbsf) is characterized by TOC similar to the base of Unit 4b (Fig. 2-6) and a decrease in DA1 and increase in DA2 and DA4 (Fig. 2-6).

2.4.8. Unit 5 (2.9 to 0 mbsf )

Unit 5 is a soft dark olive grey (5Y 3/1) silty mud with black (5Y; 2.5/1) and brown (10YR; 5/6) subhorizontal mottles, scattered small pebbles, and abundant pelecypods (Fig. 2-6). Silt dominates the size distribution, ranging from 57 to 76 %, and sand content is higher than Unit 4 (14-28%; Figs. 2-6 and 2-7). Unit 5 is characterized by a bimodal size distribution, fining upwards from a coarse to a fine silt mode (Fig. 2-7). Sand increases up-section to 1.5 mbsf, then gradually decreases; this is inversely correlated with TOC (Fig. 2-6). Pebble concentration is high. Peaks in pebble concentration correlate with peaks in sand concentration at 1.5 to 1.6 mbsf. Unit 5 is dominated by DA4. Unit 5 is subdivided into 4 units. Unit 5a (2.9 to 1.5 mbsf) is characterized by decreasing TOC up-section and increasing in total diatom abundance, accompanied by a transition from DA1 to DA4 (Fig. 2-6). Unit 5b (1.5 to 0.5 mbsf) is characterized by increasing TOC and decreasing diatoms, sand, and pebbles up-section (Fig. 2-6). Unit 5c (0.5 to 0.25 mbsf) is characterized by a negative peak in TOC and total diatoms accompanied by slight increase in DA1 (Fig. 2-6). Unit 5d (0.25 to 0 mbsf) is characterized by high but variable TOC and dominant DA4 membership (Fig. 2-6). TOC and diatom abundance decrease and sand increases at the seafloor.

2.5. Discussion

2.5.1. Diatom Assemblages (DA)

Modern ecological affinities are used to interpret environmental significance of DAs defined by Fuzzy C-means analysis. Thalassiosira antarctica is the most abundant species in the core (Fig. 2-9), and has two morphotypes: T1, which is associated with cold water masses, often found as sea-ice is forming during late summer and autumn, and T2, which is the warm variety, common in warmer water masses and in open ocean-type environments with a well-mixed water column (Taylor et al., 2001; Taylor and Sjunneskog, 2002; Domack et al., 2003; Crosta et al., 2005). The Thalassiosira antarctica T2 dominant DA1 has the highest abundance of T. antarctica, with more T2 variety than the other DA’s. It is dominated by Thalassiosira spp., with 32.7% T2, 16.2% T1, 6.1% Thalassiosira spp., and an accessory of 5.8% Fragilariopsis curta. F. curta has strong sea-ice affinity and has long been used as a proxy for reconstructing past sea-ice conditions in Antarctic marine sediment (e.g. Zielinski and Gersonde, 1997; Armand, et al., 2005; Gersonde et al., 2005; Buffen et al., 2006; Pike et al., 2009). The low abundance of F. curta indicates limited seasonal sea-ice duration. The Thalassiosira antarctica T2 dominant DA1 also has positive correlation with T. antarctica vegetative cells, T. tumida, and Odontella weissflogii. The Thalassiosira antarctica T2 dominant DA1 species memberships suggest persistent open water conditions, with a mixed water column, and warm conditions associated with high surface productivity, reflected in the greater diatom abundances.

The Thalassiosira antarctica T2 and T1 DA2 is also dominated by T. antarctica, with more equal membership of T2 and T1 (20.2% T2 followed by 13.7% T1), and accessory Navicula spp. and F. curta (9.2% and 7.7%, respectively). Navicula glaceii makes up a large component of Navicula spp. in the samples, and, much like F. curta, is associated with sea-ice conditions. Thus, DA2 is interpreted to represent a second assemblage associated with open water conditions, but cooler water masses, less surface mixing, and longer sea-ice duration than DA1.

The T. antarctica and F. cylindrus mixed DA3 also has high percentage of T. antarctica, but it is more evenly mixed with Fragilariopsis spp. (13.6% T2, 11.2% F. cylindrus, 9.9% T1, and 8.7% F. curta). F. cylindrus is found in modern sea-ice as epibenthos, and in stratified water near sea-ice, as well as in surface water following sea- ice melt (Pike et al., 2008). Although it is commonly used as a sea-ice indicator, it has also been found in cold water without sea-ice and in glacial meltwater (Kang and Fryxell, 1993). F. cylindrus frustrules are less robust than F. curta and can be subject to valve dissolution; thus, its presence indicates good preservation. F. curta is also found in modern epibenthic sea-ice communities, as well as in ice-edge water, and in melted sea- ice and associated surface water (Garrison et al., 1987; Garrison, 1991; Scott et al., 1994; Leventer and Dunbar, 1996; Armand et al., 2005). F. curta may also seed in the water column from fast ice during spring sea ice recession, and it is found in high abundance where sea-ice persists in the summer (Leventer and Dunbar, 1996; Taylor et al., 1997; Cunningham and Leventer, 1998). The T. antarctica and F. cylindrus mixed DA3 also consists of accessory species F. vanheurkii, Porosira glacialis, and Eucampia antarctica resting spores, which all have been shown to have sea-ice affinity (Garrison and Buck, 1985; Meddlin and Priddle, 1990; Burckle et al., 1984; Zielinski and Gersonde, 1997; Armand et al., 2005; Pike et al., 2009). F. vanhuerkii may also be associated retreating ice shelf conditions as it was observed in surface sediment following the Larsen B break up (Domack et al., 2005; Leventer, personal comm.). E. antarctica, too, is a diatom of multifaceted affinity, as two varieties may be related to endemism rather than ecology; E. antarctica var. antarctica has subpolar distribution, whereas E. antarctica var. recta is found only in Antarctic waters, typically at the ice edge (Zeilinski and Gersonde, 1997). E. antarctica resting spores are very robust, and when found in high abundance, it may be used as a proxy for current winnowing and reworking (Truesdale and Kellogg, 1979; Taylor et al., 1997; Cunningham and Leventer, 1998). In some cases, E. antarctica has been associated with iron-fertilization, and associated with melting icebergs (Burckle et al., 1984; Armand et al., 2008). In Herbert Sound, its elevated abundance may indicate nutrient enrichment and iron-fertilization from meltwater into the fjord, since most specimens are of the polar variety. E. antarctica can also indicate current winnowing when it is found in high abundance, and that is a possibility given the low total diatom

41 abundance associated with this DA. Thus, T. antarctica and F. cylindrus mixed DA3 is interpreted to represent a condition of moderate sea-ice duration with enhanced glacial meltwater production and/or potential sorting by bottom currents resulting in elevated E. antarctica resting spore abundance.

The Navicula spp. DA4 (15.34% Navicula spp., 11.0% F. curta, followed by 9.3% T1 and 8.1% T2) has the highest percentages of sea-ice related species, and is indicative of prolonged sea-ice duration, cold-water masses, and a stratified water column with stable surface water. It is the only DA with greater percentage of T1 than T2, indicating cold, open water masses and enhanced sea-ice formation in the late summer and autumn (Taylor et al., 2001; Taylor and Sjunneskog, 2002; Maddison et al., 2005; Buffen et al., 2006). The Navicula spp. DA4 is associated with moderate total diatom abundance and low %CRS, indicating lower primary productivity and/or dilution by terrigenous sedimentation due to increased glacial influence.

2.5.2. The Last Glacial Maximum

Ice-core paleotemperatures indicate that extreme cold conditions (~6.1  1.0 C colder than present) associated with the LGM persisted until ~12 cal kyr BP, when temperatures rapidly increased nearly 7.5 C (Fig. 2-11; Mulvaney et al., 2012). Cosmogenic 3He ages from the western coast of JRI indicate that the island was completely ice covered during the LGM (Johnson et al., 2009), and erratics were transported from Trinity Peninsula to JRI (Glasser et al., 2014). Subglacial geomorphic features, including mega-scale glacial lineations, megaflutes and grounding zone wedges, occur on the outer continental shelf of the NW Weddell Sea, providing evidence of significant APIS expansion during the LGM (Bentley and Anderson, 1998; Pudsey and Evans, 2001; Anderson et al., 2002; Heroy and Anderson, 2005; Lavoie et al., 2014). Ice drainage was focused into ice streams and was grounded at depths >900 m within the PGC, with a drainage divide separating ice draining the Trinity Peninsula from ice

42 draining from JRI (Fig. 2-2; Camerlenghi et al., 2001). A major ice stream channeled flow from JRI to the north through Croft Bay (Johnson et al., 2011; Glasser et al., 2014).

Linear glacial bedforms provide compelling evidence for ice having grounded throughout Herbert Sound during the LGM (Fig. 2-2 and 2-10; Section 2.4.1; Camerlenghi et al., 2001; Johnson et al., 2011). Poor sorting and the presence of large angular pebbles imply that the Unit 1 diamicton was deposited sub-glacially or in proximity to the grounding line (Figs. 2-6 and 2-7). Low but present TOC (<0.5%) and diatoms support proximal glacimarine origin. The sand component is sorted (~100 microns, Fig. 2-7) and quartz-rich, which implies source control from JRI Mesozoic sedimentary deposits (Riley et al., 2011).

The fjord notably lacks moraines, suggesting that recession of the grounding line was rapid and continuous in Herbert Sound (Fig. 2-2 and 2-10). A calibrated AIO age of 25.6 cal kyr BP is known to be erroneously old but provides a maximum age of deposition that suggests the diamicton was deposited during or soon following the LGM (Table 2-1; Fig. 2-4 and 2-10). Herbert Sound would have deglaciated following the outer Weddell Sea shelf between 20.4 (Smith et al., 2010) and 18.3 cal kyr BP (Heroy and Anderson, 2005). Moreover, 10Be ages from the northwestern tip of JRI (Cape Lachman) provide a minimum age of 12.9 ± 1.2 ka for for decoupling of the Prince Gustav ice stream and deglaciation of low lying coastal areas (Nyvlt, 2014). Post-LGM retreat models suggest that eustatic sea- level rise and its interaction with seafloor bathymetry were the main controls on ice decoupling from outer to inner shelf areas (Heroy and Anderson, 2005; Johnson et al., 2011; O’Cofaigh et al, 2014).

2.5.3. Permanent Ice canopy

Much like recessional patterns in the PGC (Pudsey and Evans, 2001) and the Larsen-A and -B areas (Brachfield et al., 2003; Domack et al., 2005), Herbert Sound experienced a permanent floating ice phase following lift-off of grounded ice (Figs. 2-2,

2-10 and 2-11). Unit 2 is a thin, stratified, silty clay layer with sparse diatoms, low TOC, and virtually no ice-rafted debris (Fig. 2-6), which indicates deposition beneath a canopy of permanent floating ice. It is unclear from the sediment record whether the ice canopy was nourished by glaciers or formed by thick perennial sea ice and surface accumulation (for simplicity it is depicted as an ice tongue in Fig. 2-10). Poor sorting indicates minimal marine current influence (Fig. 2-7). This facies is consistent with sub-ice shelf deposits beneath the former Larsen ice shelves (Pudsey and Evans, 2001; Pudsey et al., 2001; Evans et al., 2005; Domack et al., 2005; Buffen et al., 2007) and other regions of Antarctica (Anderson et al., 1991). Diatom frustrules are characteristically small and transport easily in the water column, so they can easily be advected under ice shelves (Domack et al., 2001; Pudsey et al., 2001; Domack et al., 2005). Low abundance, low diversity, and notable abundance of E. antarctica supports the interpretation of a floating ice environment (Figs. 2-6, 2-8, and 2-9; Cunningham and Leventer, 1998; Cunningham et al., 1999). Presence of the Miocene diatom Denticulopsis spp in Unit 2 indicates reworking and terrigenous flux, as was noted in core top glacimarine muds from beneath the former LIS-A (Pudsey et al., 2001).

2.5.4. Early Holocene opening of the fjord

Average atmospheric temperatures recorded in the JRI ice core during the early Holocene were ~1.3 ± 0.3 °C warmer than present (Mulvaney et al., 2012; Fig. 2- 11), suggesting that extreme warming during the Early Holocene Climatic Optimum, combined with rising sea level, was likely responsible for loss of floating ice and opening of Herbert Sound. Deglaciation, ice canopy disintegration, and onset of open marine conditions in Herbert Sound began ~10  2.4 cal kyr BP (See Section 2.4.2; Table 2-1; Fig. 2-4). This is consistent with lake sediment cores in the Beak Island Group (~30 km north of JRI; Fig. 2-2), which record the onset of marine sedimentation and deglaciation of northern Prince Gustav Channel ~10.6  0.05 cal kyr BP (Sterken, 2012). Furthermore, glacial lift-off took place in the southern PGC ~12 -11 cal kyr BP (Evans et al., 2005), in Greenpeace Trough ~10.7 cal kyr BP (Brachfield et al., 2003), and in the

44 area beneath the former Larsen B Ice Shelf ~10.6 cal kyr BP (Domack et al., 2005; Figs. 2-2 and 2-11). Due to the bathymetric connectivity of the >900 m deep PGC with the relatively shallow fjord (~350 m; Fig. 2-2), we hypothesize that Herbert Sound deglaciated after PGC. Lack of a developed sediment fan at the mouth of the fjord suggests that Herbert Sound deglaciated soon after PGC (Figs. 2-2 and 2-10).

Cosmogenic nuclide exposure (10Be) ages from northern JRI range from 10.8 ± 1.0 cal kyr BP to 6.6 ± 0.7 cal kyr BP (Johnson et al., 2011). Johnson and colleagues argue that JRI became ice-free ~ 8.0 cal kyr BP, when Cape Lachman (120 m above present day sea level) of the eastern Ulu Peninsula deglaciated (2011; Fig. 2-2). They further argue that the southern Naze (Fig. 2-2) deglaciated ~6.6 and ~6.8 cal kyr BP, respectively. The combined results from offshore and onshore studies indicate that deglaciation occurred in step-wise fashion from northern JRI with small, independent ice domes persisting until Mid Holocene time. Lakes of the Naze provide further evidence that the northern coast of JRI deglaciated by 7.4 cal kyr BP (Fig. 2-11; Hjort et al., 1997). Coastal outcrops indicate glacial re-advance to Cape Lachman 7.3 to 7.0 cal kyr BP (Figs. 2-2 and 2-11; Strelin et al., 2006; recalibrated by Johnson et al., 2011).

Unsorted gravel and sand (IRD) and low TOC (Fig. 2-6) indicate that NBP-0502 Site 2 remained relatively ice-proximal with significant glacial influence until 7.2 cal kyr BP. Peaks in IRD content indicate iceberg calving that may have resulted from grounding line retreat events. Increasing diatom abundance and diversity signify productivity in the absence of a permanent floating ice canopy (Figs. 2-6 and 2-10). Mixed diatom assemblages (DA1, 2, and 3) suggest increasing seasonality of sea-ice formation in open fjord conditions (Fig. 2-6; Table 2-2). Abundance of E. antarctica, as well as minor taxa including Thalossionema and Trichotoxon spp., likely reflect nutrient- rich meltwater influx during deglaciation (Figs. 2-6 and 2-9; Table 2-2). Sea surface temperatures may have decreased as a result of enhanced meltwater input, creating a freshwater lens and stratification that could support the community of DA3 (Table 2-2; Fig. 2-6). The shift to greater membership of DA2 up-section suggests transition to open marine conditions by 7.2 cal kyr BP. At this time, re-advance at nearby Cape Lachman was documented by Strelin et al. (2006), suggesting local advances associated with

45 increase in precipitation or with dynamic stretching and thinning of ice domes or ice aprons in northern JRI, perhaps explaining the final phases of ice berg rafting documented in the marine record (Fig. 2-6).

Figure 2-10 Glacial reconstruction of Herbert Sound and Prince Gustav Channel (PGC). Ice flow direction denoted by red arrows. James Ross Island (JRI) Ice Core shown at summit of Mt. Haddington. Marine sediment cores from Herbert Sound shown and labeled in red.

Figure 2-11 Regional glacial and climatic history of the Antarctic Peninsula, James Ross Island, and Herbert Sound. Proxy datasets from ice-core, moraines, and lacustrine records are compared with the multi-proxy sediment analysis of Herbert-Croft Fjoyd. Prince Gustav Ice Shelf (PGIS) , Larsen A Ice Shelf (LIS-A), and Larsen B Ice Shelf (LIS-B) records are superimposed on the JRI ice-core record in gray (Pudsey and Evans, 2001; Brachfield et al., 2003; Domack et al., 2005).

2.5.5. Mid Holocene Hypsithermal

The high abundance of open water diatoms (DA1 and DA2; Table 2-2; Fig. 2-6) and high TOC indicate primary productivity in an open marine setting, while fine-skewed, silt-dominated grain size denotes pelagic sedimentation with minimal glacial influence (Figs. 2-6 and 2-7) beginning 7.2 cal kyr BP, associated with the Mid Holocene Hypsithermal (MHH) event (Fig. 2-10). Sorting within the silt fraction further reflects marine influence (Fig. 2-7). An absence of calcareous microfossils during this interval may be the result of enhanced dissolution either in the water column or in pore fluids due to high organic production. In the northern PGC, the Mid Holocene is also characterized by exclusively arenaceous foraminiferal assemblages that may indicate presence of corrosive Hyper Saline Shelf Water (Szymcek et al., 2007). Foraminifera from the Firth of Tay, ~100 km north of JRI, are also dominated by similar arenaceous assemblages during the Mid Holocene, indicating peak primary productivity from 7.7 to 6.0 cal kyr BP (Fig. 2-2; Majewski and Anderson, 2009).

Productivity decreased and seasonality increased in the upper Unit 4a, evidenced by a TOC decrease and a DA2 membership increase (Figs. 2-6 and 2-9; Table 2-2). A substantial peak in both diatom abundance and TOC in Unit 4b followed, associated with DA1, indicating more persistent warm, ice-free conditions. The extreme peak in productivity, corresponding to maxima of TOC and DA1 membership ~6.1 cal kyr BP, represents an event within the MHH when Herbert Sound was exceptionally warm. Continued dominance of DA1 in Unit 4, with variable membership of DA2, indicates that warm and open-marine conditions persisted in Herbert Sound (Figs. 2-6 and 2-9; Table 2-2). A decrease in warm, open-water diatoms (DA1) and corresponding increase in cold, open-water diatoms and sea-ice diatoms (DA4) indicates colder sea- surface temperatures with somewhat greater seasonal sea-ice duration ~5.7 cal kyr BP (4.00 mbsf, Unit 4c), while the core site remained in a distal marine setting under minimal glacial influence until ~2.5 cal. yr BP (Fig. 2-10). A significant gap in the Herbert Sound SHALDRIL core from 5.3 to 2.7 cal kyr BP is an unfortunate limitation

49 of the dataset, but nearby cores do not record significant glacial advance during that interval (Figs. 2-5 and 2-6).

Discrepancies in local climate records result from the responses of terrestrial vs. marine systems to differential forcings. While Herbert Sound records persistent warm open ocean conditions during the Mid Holocene, lake records from Ulu Peninsula suggest that JRI experienced cold and arid conditions from 7.5 to 4.8 cal kyr BP, followed by warm and humid conditions until 3.0 cal kyr BP (Fig. 2-11; Björck et al., 1996; Hjort et al., 1997; Ingolfsson et al., 2003). Moraines date re-advance of Glacier IJR-45 into Brandy Bay ~5.4 and 4.7 cal kyr BP (Hjort et al., 1997; recalibrated by Johnson et al., 2011). Davies and others documented significant radiocarbon contamination resulting in erroneously old ages from their onshore samples, however, and conclude that IJR-45 instead advanced in the Late Holocene (2012b; 2014). Proxy temperature data from the JRI ice-core are better dated, higher-resolution, and indicate stable warmth from ~8.6 to 2.5 kyr BP that was similar to present day conditions (0.2 ± 0.2°C anomaly; Fig. 2-11; Mulvaney et al., 2012). Southern Hemisphere summer insolation was at a maximum ~5.9 to 2.6 cal kyr BP (Berger, 1978), which may have resulted in greater ocean warming due to albedo effects and high thermal capacity of water. Furthermore, an extreme peak in Herbert Sound’s productivity and warm, open water conditions took place ~6.1 cal kyr BP, but no significant variation in the ice core temperature anomaly is observed. This suggests that the unusually warm conditions in Herbert Sound were not the result of climate amelioration, but rather changing oceanographic conditions that led to increased sea surface temperatures and phytoplankton blooms. It is possible that the 6.1 cal kyr BP event may be related to significant warm events in a TEX-86 derived sea surface temperature record from the Bellinghausen Sea of the western AP (Fig. 2-1; Palmer Deep) at ~6.5 and ~5.5 cal kyr BP of ~+5 and ~+3°C, respectively (Shevenell et al., 2011). Moreover, oxygen isotopes from diatoms in the same core suggest frontal melting, implying that ocean warming was an important forcing during this time (Pike et al., 2013). Characteristic benthic foraminifera observed in shelf areas with impinging warm deep water were not observed in Herbert Sound (although preservation bias cannot be ruled out), however, and Herbert Sound is far removed from warm Circumpolar Deep Water today.

2.5.6. Late Holocene Neoglacial

Unit 5 is characterized by decreasing TOC and diatom abundance relative to Unit 4, indicative of suppressed productivity (Fig. 2-6). This is accompanied by an increase in IRD, indicative of greater glacial influence. A coarser overall grain size relative to units 3 and 4 and absence of distinct silt modes is consistent with more persistent sea ice cover and diminished marine current influence (Fig. 2-7). DA4 dominates Unit 5, with an overwhelming abundance of sea-ice affinity diatoms and cold- water diatoms, signifying prolonged sea-ice duration and stratification of surface water (Figs. 2-6 and 2-9; Table 2-2). Unit 5, therefore, marks the onset of cooling and enhanced glacial conditions ~2.5 cal kyr BP (Fig. 2-10). These more severe sea-ice conditions, suppressed productivity, and increased ice rafting are the manifestation of the Neoglacial event in Herbert Sound.

Marine cooling is coincident with atmospheric cooling that began 2.5 cal kyr BP, as documented in the JRI ice core, marking the onset of the Neoglacial (Fig. 2- 11; Mulvaney et al., 2012). This is consistent with JRI lake records that indicate cooling and aridity ~3.0 cal kyr BP (Fig. 2-11; Björck et al., 1996; Hjort et al., 1997). The Neoglacial has been recognized in the AP region not as a period of far-reaching glacial advances, like those observed in the northern hemisphere and Patagonia, but rather as one of relatively cool and dry conditions with suppressed biogenic productivity and increased ice-rafting (Leventer et al., 1996, 2002; Domack et al., 2001; Yoon et al., 2003; Heroy et al., 2008; Milliken et al., 2009; Michalchuk et al., 2009; Barnard et al., 2014). While cooling promoted sea-ice formation and ice rafting from the ice margin, the bathymetry of Herbert Sound, combined with insufficient precipitation on JRI, prevented tidewater glaciers from advancing and grounding in the ~400 m-deep fjord.

Although tidewater glacier advance is not observed in Herbert Sound, land-terminating glaciers of JRI notably advanced during the Late Holocene (Carrivick et al., 2012; Davies et al., 2014). The mass balance of land-terminating glaciers is directly influenced by regional climate perturbations, as they lack interaction with the sea; therefore, JRI glacier advance can be directly related to Neoglacial cooling (Carrivick et

51 al., 2012; Davies et al., 2012b, 2014). Modeling experiments demonstrate that advance of Glacier IJR45 was driven by mean annual temperature decrease despite precipitation starvation (Davies et al., 2014).

The TOC and diatom abundance record in Herbert Sound generally mimics the JRI atmospheric temperature anomaly (Fig. 2-11; Mulvaney et al., 2012). Atmospheric warm peaks at 1.350 and 2.300 cal kyr BP coincide with productivity peaks (from TOC and diatoms) in Herbert Sound, suggesting that the marine and atmospheric systems were coupled. A long-term increase of TOC and decrease of IRD recorded in Unit 5b signal warming from 1.250 to 0.380 cal kyr BP (Fig. 2-6). An increase of TOC and of DA1 membership from 0.900 to 0.380 cal kyr BP indicate warming in Herbert Sound that coincides with the Medieval Climate Anomaly (Fig. 2-6 and 2-11). The ice core record, however, continues on a cooling trend until 0.600 cal kyr BP, followed by rapid warming similar to present rates (Fig. 2-11; Mulvaney et al., 2012). The marine warming observed in Herbert Sound appears to be a precursor to atmospheric warming, potentially as a result of albedo effects. Thus, the Medieval Climate Anomaly was not prominent in the AP region.

High TOC with increased membership of warm, open DA1membership in 5c and 5d likely result from the prolongued warming from 600 cal yr BP to present in the ice core (Fig. 2-11; Mulvaney et al., 2012). High variability may reflect seasonal variation in sea ice and ice rafting. Short-lived increase in IRD and decrease in TOC in Unit 5c indicate decreased productivity and increased glacial influence that may be associated with the Little Ice Age ~380 cal yr BP (Fig. 2-6). These variations were identified post-hoc, however, and are within the background variability during the Holocene. Moreover, the JRI ice core lacks evidence for Little Ice Age cooling (Mulvaney et al., 2012).

Unit 5d is characterized by variable TOC, and subtle increase in membership of DA1, which indicates short-lived warm and open-water conditions ~32 cal yr BP (0.1 mbsf; Figs. 2-6 and 2-9; Table 2-2). The core top is characterized by an increase in sea-ice related DA4 and decrease in TOC, suggesting relatively more prolonged sea-ice conditions and decreased productivity at present, although additional

52 dating is needed to confirm that seafloor sediment is preserved. Today, persistent sea-ice and cool sea surface temperatures (-1.0 to -0.5 °C, some of the lowest documented sea surface temperatures in the AP; Locarnini, 2010) characterize the fjord. Glaciers that drain into Herbert Sound are receding at rates similar to or faster than other glaciers of the northern AP (Davies et al., 2012a), and summer snow melt has increased over the last century (Abram et al., 2013; Barrand et al., 2013). Meltwater contribution lowers sea surface temperatures, creates a freshwater lense, and promotes sea-ice growth; this phenomenon has been presented as a negative feedback to recent ocean warming and ice shelf melting in Antarctica (Bintanja et al., 2013). Ongoing research is aimed at dating surface sediments with Pb-210 and Cs-137 profiles to better constrain recent chronology (and surface sediment preservation) and to characterize the sedimentary manifestation of the recent rapid warming and associated glacial recession in Herbert Sound and other Antarctic Peninsula fjords.

2.6. Late Holocene ice shelf limits

The Mid to Late Holocene was characterized by dynamic ice shelf behavior in the NW Weddell Sea, and Herbert Sound helps fill a gap in knowledge of northern ice shelf limits. Cosmogenic exposure ages from nunataks of Sjörgren and Boydell glaciers (Fig. 2-2), which nourished the PGIS from the Trinity Peninsula, experienced progressive lowering between ~7.0 and 3.5 cal kyr BP (Balco et al., 2013). The former PGIS disintegrated during the Mid to Late Holocene ~5 cal kyr BP (Fig. 2- 11; Pudsey and Evans, 2001). The Larsen-A embayment also lost its ice shelf during the Mid to Late Holocene ~3.8 kyr BP (Domack et al., 2001; Brachfield et al., 2003), presumably due to prolonged Mid Holocene warmth. The Prince Gustav and Larsen-A ice shelves did not grow back to their historic extents until ~2.0 and 1.8 cal kyr BP, respectively (Fig. 2-11; Pudsey and Evans, 2001; Brachfield et al., 2003). Lag time between atmospheric cooling and ice shelf formation confirms that ice shelf growth is more gradual than ice shelf loss, due to sensitivity to bottom melting, tidal forcings, sea ice formation and duration, and accumulation patterns. By contrast, the floating ice canopy in Herbert Sound (for which it is unclear whether it was glacially derived or

53 permanent pack-ice) disintegrated by 10  2.4 cal kyr BP (See Section 2.4.2; Table 2-1; Fig. 2-4), and it never reformed due to lack of glacial advance into the fjord.

Because Herbert Sound has been free of a floating ice canopy throughout the Holocene, it is unlikely that the PGIS extended significantly northward of its historic extent, since the mouth of Herbert Sound would reduce confinement and stability of an ice shelf in PGC. While prolonged warmth during the Mid Holocene likely resulted in loss of the Prince Gustav and Larsen-A ice shelves (Pudsey and Evans, 2001; Brachfield et al., 2003; Mulvaney et al., 2012), the former Larsen-B Ice Shelf was present throughout the Holocene, but progressively thinning, so that the modern collapse was the first time the area opened since the last interglacial (Domack et al., 2005). Moreover, marine geophysics have demonstrated long-term grounding zone stability in the Larsen-B area, confirming that its collapse was driven by atmospheric warming (Rebesco et al., 2014). Therefore, the threshold of ice shelf stability in the eastern AP—and perhaps the position of the -9°C isotherm (Morris and Vaughan, 2003)—has occupied a narrow geographical range between 64.0 and 64.5°S during the Holocene. Recent warming may be shifting that zone southward, rendering AP ice shelf systems more vulnerable to collapse that would lead to acceleration and deflaction of glaciers that drain into these ice shelves (e.g., Vaughan and Doake, 1996; Scambos et al., 2004; Rignot et al., 2004).

2.7. Conclusions

The high-resolution marine sediment record from Herbert Sound, together with extensive terrestrial climate records from JRI, provides unparalleled detail in climatic, glacial, and marine variability during the Holocene in an Antarctic maritime setting. Deglaciation of Herbert Sound occurred by 10  2.4 cal kyr BP. Early Holocene deglaciation implicates sea-level rise (Bard et al., 1996) and extreme atmospheric- temperature rise (Mulvaney et al., 2012) following the LGM as causes for the rapid decoupling and subsequent loss of permanent floating ice in Herbert Sound. The fjord fully deglaciated by 7.2 cal kyr BP, when warm and open marine conditions of the Mid Holocene Hypsithermal initiated. Herbert Sound remained open and productive for the

54 remainder of the Holocene, with no evidence of major glacial re-advance. A key productivity event occurred in Herbert Sound ~6.1 cal kyr BP, but ice core data show little change throughout the Mid Holocene (Mulvaney et al., 2012), suggesting that it was a marine warming event. Increase in sea-ice cover and ice rafting mark the onset of the Neoglacial ~2.5 cal kyr BP in Herbert Sound, coincident with atmospheric temperature decrease ~2.5 cal kyr BP on JRI (Mulvaney et al., 2012), cooling and aridity on Ulu Peninsula ~3.0 cal kyr BP (Bjork et al., 1996), and advance of land-terminating JRI glaciers (Carrivick et al., 2011; Davies et al., 2014). Historic events including the Medieval Climate Anomaly and Little Ice Age remain unsupported by the atmospheric temperature record, although increase in productivity ~900 cal yr BP coincides with the MCA and greater abundance of sea ice may coincide with the LIA ~380 cal yr BP in Herbert Sound.

Chapter 3

Amundsen Sea ice shelf behavior out of phase with Antarctic Peninsula ice shelves during the Holocene

Ferrero Bay, located in eastern Pine Island Bay of the Amundsen Sea Embayment, is the largest and southernmost fjord yet studied in Antarctica. High- resolution multi-beam swath bathymetry data, 3.5 kHz subbottom profiler data, and three Kasten cores were collected in Ferrero Bay during the Oden Southern Ocean 2009-2010 cruise. Core KC-15 from the inner bay yielded two carbonate ages that record ice sheet recession from this sector of Pine Island Bay by 10.7 cal kyr BP. Seven additional acid insoluble organic fraction radiocarbon ages provide a combined linear age model with an R2 of 0.98. Holocene accumulation rates in Ferrero Bay (0.11 mm/yr) were up to one order of magnitude lower than temperate and subpolar fjords of the northern Antarctic Peninsula. Variations in magnetic susceptibility, grain size, total organic carbon and nitrogen, diatom abundance, and foraminiferal assemblage and abundance are used to interpret glacial history and paleoceanographic conditions. Planktic foraminiferal assemblages indicate that grounding line retreat may have coincided with an episode of

56 increased circulation beneath a large, Amundsen Sea-wide ice shelf, which may have experienced undermelting in the presence of relatively warmer water. Following initial deglaciation, the Cosgrove Ice Shelf covered Ferrero Bay, and productivity was virtually absent during the Mid Holocene while benthic foraminiferal assemblages indicate incursion of warm Circumpolar Deep Water. The ice shelf persisted until 2.0 cal kyr BP, when TOC and diatom abundance increased as the bay opened and coastal areas deglaciated. Abundant diatoms demonstrate open marine conditions and seasonal sea ice during the recent open water phase, while high foraminiferal diversity indicates active benthos. The Cosgrove Ice Shelf was out of phase with Antarctic Peninsula ice shelves and with ice-core proxy temperatures, implying that it did not respond to Holocene climate events, but rather to the influence of oceanography and internal glacial dynamics.

3.1. Introduction

Several Antarctic ice shelves collapsed over the last three decades and are documented through satellite and radar data. These events have been linked to poleward shift of isotherms due to atmospheric warming (Fig. 3-1; Morris and Vaughan, 2003; Cook et al., 2010). The largest losses are taking place in the rapidly warming Antarctic Peninsula (AP). Its poster child is the Larsen B Ice Shelf which collapsed in March of 2002, causing outlet glaciers to accelerate, increase ice mass losses, and contribute to sea level rise (Morris and Vaughan, 2003; Scambos et al., 2003; De Angelis and Skvarca, 2003; Scambos et al., 2004; Cook et al., 2010). Today, the remaining AP ice shelves exist in mean annual atmospheric temperatures (MAAT) below -9°C; above that threshold, ice shelves collect meltwater in their crevasses and collapse (Fig. 3-1; Hughes, 1978; Doake and Vaughan, 1991; Morris and Vaughan, 2003). However, recent hydrographic studies demonstrate that warmer water masses also destabilize ice shelves. The strongest case is in the Amundsen Sea Embayment (ASE), where recent autonomous underwater vehicle missions and conductivity, temperature, and depth (CTD) measurements have causally linked melting of the underside of Pine Island Glacier and the Getz Ice Shelf to Circumpolar Deep Water (CDW) at rates >40m/yr (Figs. 3-1 and 3-

2; Rignot and Jacobs, 2002; Jenkins et al., 2010, 2012; Jacobs et al., 2011, 2012, 2013). CDW is also observed within Ferrero Bay of the eastern ASE, where the receded Cosgrove Ice Shelf drains today. The Cosgrove Ice Shelf, which averages -16°C MAAT, is well below the -9°C threshold at which Antarctic Peninsula ice shelves have collapsed in recent years (Fig. 3-1; Morris and Vaughan, 2003). Ferrero Bay, therefore, presents an important site in which climatic and oceanographic controls on ice shelf stability can be tested. Here we investigate the Holocene history of the Cosgrove Ice Shelf through multi-proxy analysis of sediment cores from Ferrero Bay and discuss the behavior of the Cosgrove Ice Shelf during the Holocene within the context of regional terrestrial and marine data of the ASE. The Cosgrove Ice Shelf is then compared with recently collapsed AP ice shelves. While small ice shelves that are nourished by outlet glaciers and ice streams are receding in unison today, the question that remains is how they behaved during natural climate oscillations of the Holocene, and whether CDW affected their stability. Understanding ice shelf response to climate and ocean forcings— and their relative roles as causal mechanisms for recession—will improve our ability to predict future stability of the marine-based West Antarctic Ice Sheet (WAIS). These results will help steer models of glacial response and sea level contribution under future ocean and climate change scenarios.

Figure 3-1 Map of the Antarctic Peninsula (AP) and the Amundsen Sea Embayment (ASE), illustrating the climatic limit of ice shelf stability (Morris and Vaughan, 2003). Ice shelf collapse (in pink) has followed the southward migration of the -9C MAAT isotherm (Vaughan, 2003).

(Original figure with information compiled from: Landsat; NASA; GeoMapApp; Hofmann et al., 1996; Smith et al., 1999; Martinson et al., 2008; Jacobs et al., 2013.)

Figure 3-2 Bathymetric map of the Amundsen Sea Embayment and Ferrero Bay modified from Jakobsson et al., 2012 (a). Multi-beam swath bathymetry collected during the OSO-0910 cruise in Ferrero Bay, with core locations marked with red stars and furrow orientations with black arrows (b). Oblique view of bathymetry, highlighting the glacially sculpted sea-scape, rife with bedrock drumlins and scours (c).

3.2. Background

3.2.1. Geologic setting

Ferrero Bay is a narrow embayment located within eastern Pine Island Bay (PIB) of the Amundsen Sea Embayment (ASE; Figs. 3-1 and 3-2). The bathymetry of Ferrero Bay is structurally controlled; it is part of a series of fault-bounded rift basins related to Antarctica’s break-up with New Zealand in the Late Cretaceous (e.g. Dalziel and Elliot, 1982; Storey, 1991; Gohl, 2012). The northern perimeter of Ferrero Bay is characterized by great depths (up to 1300 mbsl, some of the greatest depths of PIB), due to an east-west oriented fault on the northern side of the basin that is associated with the diffuse Bellinghausen Plate Boundary (Fig. 3-2; Gohl, 2012; Cochran et al., 2015). This is bounded by the E-W elongate bedrock high, which forms King Peninsula and separates the Cosgrove Ice Shelf from the Abbot Ice Shelf to the north (Fig. 3-2). The southern side of Ferrero Bay averages ~700 mbsl and is bounded by Canisteo Peninsula. King Peninsula and the inland Hudson Mountains (Fig. 3-2) are composed of olivine basalts and tuffs. By contrast, the Canisteo Peninsula to the south is dominated by basement complex consisting of granites, diorites, and gneisses, likely of Triassic and Cenozoic age (Wade and LaPrade, 1969).

3.2.2. Climatic and Oceanographic Setting

The Cosgrove Ice Shelf is located within the Eights Coast, which is characterized by a polar climate, with -16 C MAAT (Fig. 3-1; Morris and Vaughan, 2003). Precipitation on the Eights Coast is relatively high, however, due to low-pressure systems that travel ashore during the austral winter (Vaughan et al., 2003). The ASE is one of the most remote and under-studied regions of Antarctica, due in large part to persistent sea-ice cover (e.g., Jones, 1982; Jacobs et al., 2012). Sea ice has decreased significantly during recent decades (Parkinson and Cavalieri, 2012), however, and during the austral summer of 2009 the IB Oden

61 encountered open seas, resulting in successful coring and geophysical efforts (OSO-0910 Cruise Report). In fact, this is the first time the sedimentary record of Ferrero Bay has been examined. One previous cruise to outer Ferrero Bay collected Piston cores (Kellogg and Kellogg, 1987a,b), but the cores were never studied and were eventually compromised. Mean summer sea surface temperature in Ferrero Bay was -0.9 C when the cores were collected (Locarnini et al., 2010). Water mass temperatures within PIB typically range between -1.5 and 0 C. The exception to this is dense, relatively warm (by up to 3.5 C) Circumpolar Deep Water (CDW) that impinges deep into the PIB along troughs that are up to 1000 m deep (Figs. 3-1 and 3-2; Walker, et al., 2007; Jacobs et al., 2011, 2013). Ferrero Bay is part of a long, narrow, landward sloping trough that connects to the shelf edge in the eastern ASE (Figs. 3-1 and 3-2). Water column profiles from CTD and XBT data collected during the 2009 Oden cruise clearly show impinging CDW, with sharp temperature and salinity increase from 275 to 500 m water depth to ~+1.2 C and ~34.6 ppm, respectively (Figs. 3-3 and 3-4; OSO-0910 Cruise Report). Productivity in the Amundsen Sea is among the highest in the Southern Ocean, with long-lived phytoplankton blooms related to the Pine Island and Amundsen polynyas (Arrigo et al., 2008; Smith and Comiso, 2008; Thuroczy et al., 2012). In contrast to the well-studied Ross Sea, little is known about modern distributions of phytoplankton communities in the Amundsen Sea. Recent studies of summer phytoplankton compositions and hydrography have linked blooms to melt events of Pine Island Glacier (Arrigo et al., 2003; Fragoso et al., 2011; Thuroczy et al., 2012; Gerringa et al., 2012). Modified CDW has also been argued to bring iron to the shelf as it travels along deep troughs into PIB, re-suspending sediment that may facilitate blooms (Dinniman et al., 2003). Diatoms make up a major fraction of these blooms and are present in open-ocean and ice-marginal settings (Leventer and Dunbar, 1996; Fragoso et al., 2011).

Figure 3-3 Water column temperature and salinity measured during OSO-0910 cruise in Ferrero Bay. Note temperature and salinity increase below 275 m water depth, indicating the presence of Upper Circumpolar Deep Water. (Y-axis is pressure-corrected depth below sea level.)

3.3. Methods

3.3.1. Geophysical methods and coring

High-resolution multibeam swath bathymetry was collected using a Kongsberg 12 kHz EM122 1x1 multibeam echo sounder onboard the Ice Breaker Oden (Fig. 3-2). Bathymetry data were integrated and spatially analyzed in ArcMap GIS 10.1 software. The upper sediment was imaged to study sediment distribution and determine optimal coring sites using subbottom profiler 120 3°x3° chirp sonar, and analyzed in Knudsen software (Fig. 3-4). The chirp sonar system was operated continuously between 2.5 and 7 kHz at 35 ms pulses. Three cores were collected in Ferrero Bay during the IB Oden OSO-0910 cruise (Fig. 3-2). KC-15 (73.3603S, 101.8362W) was taken in the inner most part of the fjord at 1274 mbsl, and recovered 1.3 m of sediment. KC-16 (73.454S, - 102.0792W; 706 mbsl) was taken on a structural high ~700 mbsl and recovered only 0.4 m of sediment. KC-17 (73.4197S, 102.8267W; 855 mbsl) was taken in the outer bay and recovered 1.4 m of sediment. Detailed core descriptions include Munsell soil color determination, preliminary grain size and shape, pebble lithology, sedimentary structures, wet sieving for carbonate material, Torvane strain measurements, and magnetic susceptibility. When KC-15 and KC-17 were archived, the core catcher material was appended to the bottom of the cores.

3.3.2. Radiocarbon analysis

Two marine carbonate samples were collected from KC-15 for radiocarbon analysis (Table 1). Samples were wet sieved for calcareous shells and fragments, dried at 50°C, and sent to the UC Irvine Accelerator Mass Spectrometry (UCIAMS) facility for analysis. A reservoir of 900 years was applied to the 14C ages, entered into Calib 6.0 Calibration software, and calibrated using the Marine09 curve,

64 which has a built in global correction of 400 years (Stuvier and Reimer, 1998; Stuvier et al., 2005; Reimer et al., 2009). The carbonate ages were reported by Kirshner et al. (2012). To build a higher-resolution chronology, 7 bulk sediment samples were selected from KC-15 and analyzed by University of Tokyo AMS laboratory (YAUT) for acid insoluble organic fraction (AIO) 14C dating (refer to YAUT for procedures).

3.3.3. Sedimentological proxies

X-radiographs were taken of the archive core halves at the Antarctic Research Facility in Tallahassee, Fl. to indentify sedimentary structures, to determine pebble abundance, and to search for additional carbonate macrofossils. Pebble abundance was determined by counting in-situ pebbles (diameter >2 mm) over 5 cm intervals on X-radiographs. Each core was sampled onboard at 10 cm intervals for grain size, geochemistry, and microfaunal abundance and assemblage prior to archiving. Grain size samples were analyzed at Rice University using a Malvern Mastersizer 2000 Laser Particle Size Analyzer (McCave et al., 1986) and grain size statistics were performed following Folk and Ward (1957).

3.3.4. Geochemical proxies

Total organic carbon (TOC), nitrogen, and hydrogen were measured using the Rice Costech Elemental Analyzer after removing carbonate following the vaporous HCl decarbonation method (Komada, 2008). Samples were dried, crushed, and weighed into silver capsules following standard procedures. Twenty microliters of DI water were added to each sample, which was then placed in a dessicator adjacent to open beakers of 20-30% HCl for 17 hours, then dried overnight prior to analysis. Calibrations with proline standards yielded errors less than 0.01% for C and N.

3.3.5. Micropaleontological proxies

Foraminiferal samples were wet-sieved at 63 microns. All specimens were picked and arranged by taxa on micropaleontological slides. Taxonomic classification is that used in Majewski (2013). The collection is hosted at the Institute of Paleobiology, Warsaw, Poland. Total foraminiferal abundances are calculated by dividing the number of all foraminiferal tests by grams of dried sediment (t/gds). Faunal diversities are expressed by the Shannon diversity index,

H = −∑ ni/n ln (ni/n), where ni is the number of individuals of species i. Additionally, samples were analyzed for absolute diatom abundances and assemblages, and prepared following the settling method (Scherer, 1994). Diatoms were counted in each prepared slide at 400X magnification using a transmitted light microscope. Each prepared slide is counted with a minimum of 400 valves. Diatom abundance was too low for most of the core sections to conduct a robust statistical analysis on assemblages; instead, qualitative observations of assemblages were noted, and pennate vs centric diatoms were counted. Absolute diatom abundances were calculated using the equation: T = (NB/AF)/M, where T= number of microfossils per unit mass, N = total number of microfossils counted, B = area of bottom of beaker (mm2), A = area per field of view or transect (mm2), F= number of fields of view or transects counted, M = mass of sample (g; Scherer, 1994). The combined information from geochemical, sedimentological, and paleontological data was used to interpret depositional environment. Diatoms and foraminifera were analyzed in the context of modern affinities to marine environments.

3.4. Results

3.4.1. Geophysical results

PIB has a broad, fore-deepened shelf (Fig. 3-2) that is characterized by basement-floored inner shelf and a sediment-floored outer shelf, which is composed of several stratigraphic packages that onlap basement and thicken seaward (e.g. Wellner et al., 2001; Lowe and Anderson, 2002; Gohl et al., 2013). Ferrero Bay is entirely crystalline basement-floored (Fig. 3-2). Subbottom profiler data show very thin, draping sediment accumulation in small pockets within the glacially sculpted basement topography (Fig. 3-4). The deep, inner basin is characterized by linear glacial features that orient E-W. Bedrock drumlins are prominent in the inner basin and taper to the west. Outer Ferrero Bay is characterized by linear furrows and cross-cutting lineations that reflect convergence of ice flowing from the bay with ice flowing out of PIB (Fig. 3-2). The southern side of the swath area includes prominent bedrock highs (up to 600 relief) that include networks of linear scours and mini-basins at the crest of the highs.

Figure 3-4 a) OSO-0910 subbottom profile 98A from the axis of Ferrero Bay 2 showing the small basin with <2 m sediment where KC-15 was collected. b) 3 Bathymetric map with cruise tracklines, CTD and XBT sites, and Line 98A 4 highlighted in white. 5

3.4.2. Radiocarbon results 6

Shell fragments from the base of KC-15 yielded an age for deglaciation of 7 10695  35 radiocarbon years BP (where present day is 1950 A.D.), which was calibrated 8 to 10736  219 cal yrs BP (calendar yrs BP; Table 1; Kirshner et al., 2012). Another age 9 was extracted from planktonic foraminifera in lower units (110 cmbsf), yielding an age of 10 10000  120 radiocarbon years BP, which was calibrated to 9855  337 cal yrs BP 11 (Kirshner et al., 2012). The carbonate-derived ages at the base of the core provide a more 12 robust chronology than is typical for Antarctic marine sediments, where terrigenous flux 13 in proximal facies often creates an exponential increase in radiocarbon age, a 14 phenomenon known as the “hockey stick effect” that results from reworking of old 15 carbon (e.g., Andrews et al., 1999). Here, there is no significant increase in radiocarbon 16 ages, and both AIO-fraction and carbonate material from the proximal glacimarine facies 17 are consistent. 18

The two carbonate-derived radiocarbon ages and seven AIO-fraction ages 19 from KC-15 provide a linear age model best fit by the line y = 0.0111x + 6.588 (where y 20 is depth and x is age) with remarkable linear regression of R² = 0.98329 (Table 1; Fig. 3- 21 5). The AIO-fraction ages were calibrated using the carbonate-derived age and an AIO- 22 fraction age from the same depth (110 cmbsf), and a surface age assumption of -60 cal 23 yrs BP (2010 A.D. when the core was collected; following Licht et al., 1998). 24

When the AIO-fraction age of 2124 ± 57 cal yr BP at the top of KC-15 is 25 subtracted from the age at 110 cmbsf, the result matches well with the carbonate-derived 26 calibrated age and corroborates a total marine reservoir correction of 1300 years on the 27 Marine09 curve (Stuvier and Reimer, 2009; after Kirshner et al., 2012). The resulting 28 age model from subtraction of the AIO surface age is described by the line y = 0.0109x + 29 5.626 with R² = 0.98212, and is virtually identical to the age model derived by carbonate 30 calibration. 31

Table 3-1 Radiocarbon ages from Ferrero Bay. 32

3.4.3. Kasten Core KC-15 33

Four units of different sedimentological, geochemical, and paleontological 34 properties were identified in KC-15. 35 The base of KC-15, Unit 1 (137-127 cmbsf), is characterized by a dark 36 brown, clayey sand with abundant pebbles. Magnetic susceptibility is high in Unit 1 37 (Fig. 3-5). The core bottomed out in this unit likely due to its relatively stiff, cohesive 38 properties. Based on the subbottom profile of this area, the 137 cm core (including 39 catcher material) appears to capture the entire post-glacial sediment record (Fig. 3-4). 40 The contact of Unit 1 and Unit 2 is characterized by a transition to light 41 grayish-brown sandy clay (Fig. 3-5). Magnetic susceptibility is significantly lower in 42 Unit 2 (127 to 104 cmbsf), but peaks with increasing sand content at 110 cmbsf. TOC 43 and TN are near zero in this unit (0 to 0.2 %). Diatoms are virtually absent in Unit 2, 44 while foraminifera are abundant. The planktonic species Neogloboquadrina pachyderma 45 dominate the foraminiferal assemblage of Unit 2, and calcareous foraminifera comprise 46 near 100% of the assemblage. 47 Unit 3 (104 to 30 cmbsf) has a sharp base with color change to greenish- 48 bluish gray and significant decrease in grain size (Fig. 3-5). Unit 3 is a silty glacimarine 49 mud with low magnetic susceptibility, sparse pebbles, and gradually increasing TN and 50 TOC. Diatoms are present but remain in low abundance throughout Unit 3. Foraminifera 51

70 are sparse in Unit 3 but peak in abundance and diversity at 50 and 70 cmbsf. At these 52 intervals, benthic foraminifera dominate, and the CDW-affinity species Buluminea 53 aculeata is present (Fig. 3-5). 54 Productivity indicators TOC, TN, and diatom abundance increase 55 significantly in Unit 4 (30 cmbsf to core top), which is sedimentologically very similar to 56 Unit 3 (Fig. 3-5). TOC increases above 0.5% to 0.75% at the top of the core. Both 57 diatoms and foraminifera abundances increase steadily to their highest values in the core- 58 top sediment. Pennate to centric diatom ratios increase up-section. The upper 2 cm of 59 core sediment has a noticeable color change to orange-brown, resembling iron oxidation. 60

3.4.4. Kasten Core KC-16 62

Collected from a bedrock high, KC-16 is significantly shorter in length 63 than KC-15 and KC-17, which were acquired within small depocenters of Ferrero Bay 64 (Figs. 3-2 and 3-4). KC-16 is much coarser overall and consists of pebble-rich and sand- 65 rich mud. Magnetic susceptibility is an order of magnitude higher in KC-16 than in KC- 66 15 and KC-17, clearly due to its high pebble concentration (up to 20% volume; Fig. 3-5). 67 Increasing TOC and diatom abundance in this short core, in addition to its striking 68 similarity in color to the top of KC-15, indicate that it may correlate temporally with Unit 69 4 of KC-15 and KC-17 (Fig. 3-6). The high pebble concentration, however, suggests that 70 KC-16 captures a different facies that may be related to its location on a bathymetric high 71 (Fig. 3-2). Live benthic organisms in the upper 2 cm indicate that the seafloor surface 72 was preserved. 73

3.4.5. Kasten Core KC-17 75

KC-17 is characterized by a similar facies progression as in KC-15, with 76 some differences in microfaunal assemblage (Fig. 3-5 and 6). Unit 1 (147 to 130 cmbsf) 77 is a dark greenish pebble-rich clayey sand with high magnetic susceptibility (Fig. 3-5). 78

Unit 1 is characterized by the highest sand content measured in Ferrero Bay (~40%) and 79 virtually absent TOC and TN. Diatoms identified in Unit 1 include reworked 80 Denticulopsis spp., as has been described by Kellogg and Kellogg in PIB (1987a). 81 Foraminifera are present in the top of this unit are dominated by planktic species 82 Neogloboquadrina pachyderma. 83 A decrease in magnetic susceptibility and low foraminifera abundance 84 characterize KC-17 Unit 2, a greenish sandy clay that contains layers of light greenish 85 silty clay (130 to 97 cmbsf; Fig. 3-5). Magnetic susceptibility peaks at ~100 cmbsf. 86 Foraminiferal assemblages include Neogloboquadrina pachyderma, similar to Unit 2 of 87 KC-15, but also includes the common agglutinated species Milliamina arenacea. Other 88 parallels to KC-15 Unit 2 include virtually absent diatoms, TOC, and TN, while sand 89 concentration is high (~35%; Fig. 3-5). 90 KC-17 Unit 3 (97 to 30 cmbsf) is a greenish silty clay with few pebbles 91 (Fig. 3-5). Grain size decreases to a silt dominated population from Unit 2 to Unit 3, 92 much like the facies transition in KC-15. Magnetic susceptibility decreases and stabilizes 93 to low values ~70 cmbsf. TOC, TN, and diatoms are low and steadily increase up- 94 section. Foraminifera are sparse, and Buluminea aculeata is not present in this core. 95 Milliamina arenacea is present at the base of Unit 3, however (Fig. 3-5). 96 Unit 4 (30 cmbsf to core top) of KC-17 is a brownish green silty clay, 97 characterized by fine grain size and increasing diatom, TOC, and TN percentages (Fig. 3- 98 5). This is identical in grain size, geochemistry, and diatom abundance to KC-15 Unit 4. 99 Foraminifera abundance and diversity increase from 40 cmbsf to the core top and is 100 dominated by Milliamina arenacea. Neogloboquadrina pachyderma is present only in the 101 core top. Pennate to centric diatom ratios increase up-section. All productivity proxies 102 are highest in the core top (Fig. 3-5). 103

Figure 3-5 Multi-proxy analysis of the three OSO-0910 Kasten cores from Ferrero Bay. Unit distinctions are based on changes in magnetic susceptibility, sedimentology, TOC and TN, and diatom and foraminiferal abundance and assemblages.

104

105

106

Figure 3-6 Correlation of sedimentary facies in Kasten cores from Ferrero Bay. 107 Differences in sediment character within facies suggests some lateral 108 variability, likely related to proximity to the grounding line and water depth. 109

110

3.5. Discussion 111

3.5.1. Paleo-drainage of the West Antarctic Ice Sheet 112

The bathymetric survey of Ferrero Bay reveals a paleo-trough with 113 bedrock drumlins, scours, and deep furrows into basement and reveals that Ferrero Bay 114 channeled grounded ice into Pine Island Bay during the LGM, perhaps contributing to a 115 small ice stream that is believed to have occupied a trough that crosses the continental 116 shelf north of the Abbot Ice Shelf (Figs. 3-2 and 3-7; Gohl, 2012; Jakobsson et al., 2012). 117 Paleo-ice flow was directed from east to west, with significant curvature in the deep, 118 inner basin. Linear features exist up to 1300 mbsl and signify grounding of the WAIS to 119 the deep seafloor (Gohl, 2012). Overlapping furrows in the outer bay potentially record 120 convergence of ice streaming from Ferrero Trough with ice streaming from PIB (Fig. 3- 121 2). Potential meltwater storage basins are observed in the middle basin, where a major 122 (>500m relief) basement high has a connected network of linear scours that may have 123 been generated by subglacial meltwater drainage when ice was grounded in the bay (Fig. 124 3-2). The basement highs likely served as important pinning points, where the ice could 125 ground and remain relatively stable during deglaciation and sea level rise. 126

Figure 3-7 Glacial reconstruction of Ferrero Bay and outer Pine Island Bay showing a) Ferrero Bay completely filled with grounded ice until 10.7 cal kyr BP, following deglaciation of the outer ASE after the Last Glacial Maximum. b) The grounding line receded rapidly to near its current position, while the Cosgrove Ice Shelf persisted during the entire mid Holocene warm period, despite advection of CDW. c) The Cosgrove Ice Shelf receded to its current position by 2.0 cal kyr BP, likely forced by persistent under-melting by CDW. Core locations are projected onto the profile; KC-16 was collected at a local high of ~700 mbsl, which is to the south of this profile (Fig. 3-2). Profile was built from OSO-0910 multibeam swath bathymetry, from Bedmap 2 (Fretwell et al., 2013), and from Ice Bridge data (Cochran et al., 2015). Dashed lines are inferred subglacial topography, due to 127 limited radar coverage and resolution. Ice thicknesses are best estimates based on inferred bathymetry.

3.5.2. Unit 1: Glacial recession from Ferrero Bay

Despite minor lateral variation, Unit 1 can be correlated from the proximal core (KC-15) to the distal core (KC-17) with remarkable precision (Figs. 3-5 and 3-6). The magnetic susceptibility is high at the base of both cores and is followed by a low peak, which is tightly correlated along with coarse grain size and color change. KC-15 and KC-17 bottomed out in relatively stiff sandy clay indicating the basal Unit 1 is a proximal glacimarine facies and records deglaciation of the fjord. The calibrated radiocarbon age of 10736  219 cal yrs BP (Table 1; Fig. 3-5; Kirshner et al., 2012) provides a minimum age for deglaciation of proximal Ferrero Bay. Ferrero Bay deglaciated almost 2 kyrs after outer PIB, but both systems maintained a floating ice shelf for some time following grounding line retreat and lift off (Kirshner et al., 2012). The difference in timing suggests that Ferrero Bay was stable for a longer period of time. Greater thickness of the expanded WAIS in proximal coastal settings likely allowed ice to remain grounded in the trough of Ferrero Bay until 10.7 cal kyr BP, after it had already receded from the outer shelf (Fig. 3-7). Furthermore, Ferrero Bay is characterized by a low Holocene accumulation rate of 0.011 mm/yr, an order of magnitude less than northern AP fjords (e.g. Milliken et al., 2009; Michalchuck et al., 2009; Allen et al., 2010; Minzoni et al., 2015). This may be due to latitudinal decrease in erosion rate associated with increasing polar conditions and the freezing of the glaciers to their beds (Fernandez et al., in press, GSA Bull.), but basin-wide volume calculations are needed to test this idea.

3.5.3. Unit 2: Circumpolar Deepwater beneath the Cosgrove Ice Shelf

Observed in both KC-15 and KC-17, Unit 2 is a proximal glacimarine facies with coarse, poorly sorted grain size and low productivity (Fig. 3-5). Unit 2 represents a second phase of grounding line recession in Ferrero Bay from ~8.8 cal kyr BP, resulting in a more distal sub-ice shelf setting at the core sites. Presence of planktonic foraminifera in Unit 2 (Fig. 3-5) may indicate ocean circulation into the innermost bay beneath an ice shelf (Fig. 3-7). Abundant, well- 76

77 preserved planktonic foraminifera were attributed to landward advection beneath the Amery Ice Shelf (Hemer, 2007), and Majewski et al. (2012) demonstrated that an abundance of planktonic foraminifera in fjord settings can result from stronger inflow of open ocean water when turbidity is low. Here we rule out the second case due to low productivity during this interval. Abundance of planktonic foraminifera Neogloboquadrina pachyderma in the sub-ice shelf facies associated with deglaciation event from 12.3 to 10.6 cal kyr BP was argued by Kirshner and others (2012) and Majewski and others (2013) as the evidence of offshore currents advecting under an extensive Amundsen Sea ice shelf. The temperature and salinity of the water is unknown, but it may have been up to 3-4° C warmer than sub-ice shelf water based on modern measurements of CDW in the ASE (Jacobs et al., 1996; 2011; 2013). Moreover, the abundance of Neogloboquadrina pachyderma in both Ferrero Bay and across the outer shelf of PIB, as described by Kirshner et al. (2012), supports the idea that this was a basin-wide oceanographic event.

3.5.4. Unit 3: Stable Cosgrove Ice Shelf during the Mid Holocene

Unit 3 represents a significant shift to distal sub-ice shelf setting during the Mid Holocene (Figs. 3-5 and 3-7). Unit 3, observed in both KC-15 and KC-17, is composed of silty mud and lacks pebbles and sand that might indicate ice rafting or proximity to the grounding line. Low TOC, TN, and diatom abundance suggest long- term stability of the Cosgrove Ice Shelf from 8.8 to 2.0 cal kyr BP. Progressive increase in these productivity indicators indicate gradual recession of the ice margin. Sparse but present diatoms were likely transported under the ice shelf by advection, as they are light and easily suspended (e.g., Anderson et al., 1991; Evans et al., 2002; Domack et al., 2005; Hemer et al., 2007). While Neogloboquadrina pachyderma is not present in Unit 3, the benthic foraminifera Buluminea aculeata is present at 70 and 50 cmbsf in KC-15 (Fig. 3-5). It may suggest advection of warm deep water onto the shelf during the Mid Holocene, as this foraminifer is strongly associated with presence of CDW (Ishman and Domack 1994; Majewski et al., in review). Continued warm deep incursion would have been channeled

78 underneath the Cosgrove Ice Shelf during deposition of Unit 3 (Fig. 3-7), much like what is observed below the Pine Island Glacier’s ice shelf today (Jenkins et al., 2010, 2012). Fine grain size, low productivity and lack of biogenic particles indicate sub-ice shelf conditions persisted throughout the Early and Mid Holocene.

3.5.5. Unit 4: Late Holocene collapse of the Cosgrove Ice Shelf

Unit 4 marks recession of the Cosgrove Ice Shelf and the opening of Ferrero Bay to its present configuration by 2.0 cal kyr BP (Fig. 3-7). Substantial increase in TOC, TN, and diatom abundance show increased productivity in absence of the Cosgrove Ice Shelf (Fig. 3-5). Moreover, the pennate to centric diatom ratios increase to the core top, indicating sea ice was common after the bay opened. This is consistent with historic observations of heavy seasonal sea-ice cover (e.g., Parkinson and Cavalieri, 2012; Jacobs et al., 2012). The uppermost 2 cm in KC-15 and KC-17 are high in TOC and hosted benthic organisms and modern biological activity when they were collected. KC-15 also shows evidence of iron oxidation and oxygenated bottom water that promotes benthic fauna. This indicates that the seafloor surface is well preserved in these cores.

3.5.6. Glacial History of the ASE

PIB is one of the most heavily glaciated regions of West Antarctica today, but also one of the least stable due to its marine-based configuration and exposure to de- stabilizing oceanographic processes (Fig. 3-1). These combined vulnerabilities led Hughes et al. (1981) to refer to Pine Island Glacier as “the weak underbelly of the West Antarctic Ice Sheet”. Recent velocity measurements indicate that the main trunks of ice drainage in the region, the ice streams Pine Island and Thwaites glaciers, are accelerating at unsustainable rates, with total mass loss of 81 ± 17 km3/ year (e.g., Rignot et al., 2004). Over the last few decades, massive calving events were observed via satellite imagery (e.g., Kellogg and Kellogg, 1987b; Rignot, 2002; Bigg et al., 2013). Current rates of ice

79 loss will cause Pine Island Glacier to completely lose contact with its base within the next 100 years (Wingham et al., 2009) and contribute to destabilization of the WAIS. If all the ice currently draining into the Amundsen and Bellinghausen seas discharged, sea level would rise 2 to 5 m (Vaughan, 2008; Bamber et al., 2009; Joughin and Alley, 2011; Larter et al., 2014). Concerns about stability of the WAIS have driven extensive work to understand the long-term behavior of Pine Island Glacier. Recent geophysical and coring efforts focused on improving the chronology of recessional events through coring and high-resolution bathymetry of the shelf. Detailed geomorphology demonstrates that the WAIS was grounded near the continental shelf edge during the Last Glacial Maximum (LGM, ~20 ka; Fig. 3-2; Lowe and Anderson, 2002). Deglaciation was episodic across the ASE shelf to mid PIB, with grounding zone wedges marking pauses in recession (Fig. 3-2; Lowe and Anderson, 2002; Graham et al., 2010; Kirshner, et al., 2012; Jakobsson et al., 2012; Larter et al., 2014). Deglaciation was largely bathymetrically controlled, with the smooth sedimentary seafloor facilitating synchronous retreat during phases from 16.4 and 12.3 cal kyr BP that coincide with meltwater pulses during the Late Pleistocene and Early Holocene (Bard et al., 1996; Lowe and Anderson, 2002; Jakobsson et al., 2011, 2012; Kirshner et al., 2012; Larter et al., 2014). This was followed by sub-ice shelf conditions from 12.3 to 10.6 cal kyr BP, in which an expanded Amundsen Sea Ice Shelf is hypothesized to have covered mid PIB, while Ferrero Bay was still glaciated (Kirshner et al., 2012). Relict channels on the deglaciated inner shelf provide compelling evidence for subglacial meltwater drainage under the expanded WAIS during the LGM (Lowe and Anderson, 2003; Graham et al., 2010; Witus et al., 2014). Subglacial meltwater storage and drainage networks may have destabilized the marine ice sheet via reduction in basal friction and outbursts from beneath Pine Island Glacier, similar to what has been observed at modern grounding lines of East Antarctica (Stearns et al., 2008). A widespread terrigenous silt facies was deposited between 7.8 and 7.0 cal kyr BP and suggests a Mid Holocene phase of rapid ice recession related to subglacial meltwater outburst (Lowe and Anderson, 2003; Kirshner et al., 2012; Witus et al., 2014). Bedrock incisions in Ferrero Bay may provide evidence for subglacial meltwater in the mid basin

80 when the WAIS was grounded in Ferrero Bay (Fig. 3-2), but better spatial bathymetric coverage is needed to test this hypothesis. Well-sorted silt plumites associated with meltwater outbursts described by Lowe and Anderson (2003) and Witus et al. (2014) were not observed in Ferrero Bay, but that does not preclude the Cosgrove glacial system as a meltwater contributor during the Early Holocene when ice was grounded in the Ferrero Bay trough. While recessional history for Pine Island Bay is well constrained on the outer shelf, there is considerable data gap in the coastal environments of the ASE. Conflicting data leave ambiguity regarding glacial recession from the inner PIB, ranging from sometime before 11.7 (Hillenbrand et al., 2013) to 1.3 cal kyr BP (Kirshner et al., 2012). Ferrero Bay contains a complete Holocene sedimentary record with no evidence of a hiatus, and provides an age of 10.7 cal kyr BP for glacial recession of the eastern ASE coast. Cosmogenic exposure ages of the Hudson Mountains (northeast of Pine Island Glacier; Fig. 3-2) record ice surface elevation decrease in the Early Holocene (Bentley et al., 2011; Johnson et al., 2012). A second grounding line recession ~8.8 cal kyr BP in Ferrero Bay correlates with 10Be exposure ages of the Hudson Mountains that record lowering to near modern levels by 8.0 cal kyr BP (Johnson et al., 2014). Deglaciation of Ferrero Bay also coincides with recession of the Antarctic Peninsula Ice Sheet in the Early Holocene (Heroy and Anderson, 2007), and specifically from AP fjords from 14 cal kyr BP (Milliken et al., 2009) to 9 cal kyr BP (Allen et al., 2010). Moreover, the George VI basin was free of grounded ice by 12 ka (Figs. 3-1 and 3-8; Smith et al., 2007), and the Larsen A and B embayments experienced grounding line retreat ~10.5 ka (Figs. 3-1 and 3-8; Brachfield et al., 2003; Domack et al., 2005). Each of these embayments experienced a subsequent ice shelf phase, much like Ferrero Bay. The Cosgrove Ice Shelf differed from Antarctic Peninsula ice shelf histories in their Mid to Late Holocene behavior, however.

3.5.7. The Holocene history of small Antarctic ice shelves

The Cosgrove Ice Shelf remained stable in outer Ferrero Bay until 2.0 cal kyr BP, when it receded to its current configuration (Figs. 3-6, 3-7, and 3-8). Late Holocene collapse of the Cosgrove Ice Shelf is consistent with cosmogenic exposure ages from an island seaward of the Canisteo Peninsula on the southern boundary of Ferrero Bay, where boulders near sea level record onshore deglaciation by 2.2 ± 0.2 ka (Fig. 3-2; Johnson et al., 2008). This indicates that the Cosgrove Ice Shelf completely filled Ferrero Bay, and ice remained stable at sites seaward of Canisteo Peninsula, likely providing further buttress to an advanced Cosgrove Ice Shelf, until the Late Holocene. Meanwhile, the Prince Gustav Ice Shelf of the northern Antarctic Peninsula destabilized and collapsed by 5.0 cal kyr BP (Figs. 3-1 and 3-8; Pudsey et al., 2001) and the Larsen A Ice Shelf collapsed by 3.8 cal kyr BP (Figs. 3-1 and 3-8; Brachfield et al., 2003). Both reformed by 2.0 and 1.8 cal kyr BP, respectively, during climatic cooling observed in the James Ross Island and EPICA Dome C ice-cores (Figs. 3-1 and 3-8; Pudsey et al., 2001; Monnin et al., 2001; Brachfield et al., 2003; Mulvaney et al., 2012). By contrast, the Larsen B Ice Shelf was present but progressively thinning throughout the Holocene and only recently collapsed in 2002 (Figs. 3-1 and 3-8; Domack et al., 2005). The George VI Ice Shelf collapsed much earlier in the Early Holocene ~9 to 7.8 cal kyr BP, following a period of rapid climate warming and impinging warm CDW onto the inner shelf (Figs. 3-1 and 3-8; Smith et al., 2007; Bentley et al., 2009; Allen et al., 2010). Thus, while AP ice shelf collapse was not strictly synchronous, ice shelves generally receded during the warm intervals of the Early or Mid Holocene, and reformed during the Late Holocene Neoglacial. Furthermore, the largest Arctic ice shelf, the Ward-Hunt Ice Shelf, was absent throughout the Early and Mid Holocene warm periods, then formed 4.0 cal kyr BP and collapsed 1.4 cal kyr BP (Fig. 3-8; Antoniades et al., 2011). The Ward-Hunt Ice Shelf re-formed again 800 years ago during the Little Ice Age (Antoniades et al., 2011), broadly coinciding with a small advance of the Müller Ice Shelf in the Antarctic Peninsula (Fig. 3-1; Shevenell et al., 1996; Taylor et al., 2001).

Ferrero Bay provides critical polar representation in the global history of small ice shelves that were nourished by outlet glaciers and ice streams during the Holocene, and records the behavior of the Cosgrove Ice shelf. The Cosgrove Ice Shelf retreated when most AP ice shelves were either stable or re-forming during Late Holocene climatic cooling (Fig. 3-8). The out-of-phase behavior of the Cosgrove Ice Shelf indicates that it was either insensitive to climate change, or perhaps climatic shifts were less pronounced in the more polar climate of the Amundsen Sea coast. In either case, ocean circulation, especially warm CDW, appears to have been the dominant control on glacial stability of the Cosgrove glacial system throughout the Holocene. Likewise, the George VI Ice Shelf was largely out of phase with northern AP ice shelves due to influence of CDW (Fig. 3-1, 3-2; Smith et al., 2007). The asynchronous global pattern of ice shelf retreat and re-growth during the Holocene contrasts with the synchronous nature of historic ice shelf loss (Fig. 3-8), in which 28,000 km2 of AP ice shelf area has been irreparably lost to the sea since the 1980’s (Cook et al., 2010; Hodgson et al., 2006, 2011).

Figure 3-8 Ice shelf history and paleo-temperature proxies from EPICA Dome C and James Ross Island ice cores (Monnin et al., 2001; Mulvaney et al., 2012). The Cosgrove Ice Shelf adds a polar endmember to the dynamic history of ice shelves in the Antarctic Peninsula (Pudsey and Evans, 2001; Brachfield et al., 2003; Domack et al., 2005; Smith et al., 2007) and the Arctic (Ward-Hunt Ice Shelf; Antoniades, 2011). The Cosgrove Ice Shelf retreated when most Antarctic Peninsula ice shelves were either stable or re- forming. Figure modified from Hodgson, 2011, PNAS.

3.6. Conclusions

To date, Ferrero Bay provides the southern-most and highest-resolution Holocene record of an outlet glacier-fed ice shelf in Antarctica (Fig. 3-1). Following lift- off and grounding line retreat 10.7 cal kyr BP, the Cosgrove glacial system sustained an ice shelf that filled Ferrero Bay throughout the Early and Mid Holocene warm periods, when many Antarctic Peninsula ice shelves collapsed (Fig. 3-8; Pudsey and Evans, 2001; Brachfield et al., 2003; Smith et al., 2007). Presence of foraminifera with an affinity for environmental conditions associated with Circumpolar Deep Water (Kirshner et al., 2012; Majewski et al., 2013; Majewski et al., in review), which is present in Ferrero Bay today, demonstrates that persistent upwelling of relatively warm water during the Holocene caused under-melting of the Cosgrove Ice Shelf and forced its eventual collapse. The Cosgrove Ice Shelf did not recede to its current position until 2.0 cal kyr BP, when climate was cooling. This coincides with deglaciation of the southern perimeter of Ferrero Bay from a near-sea level cosmogenic exposure age of 2.2 ± 0.2 ka (Fig. 3-2; Johnson et al., 2008). Thus, the Cosgrove Ice Shelf was out of phase with other Antarctic ice shelves and did not respond to climate variability, but rather to oceanographic forcing during the Holocene. Amundsen Sea and Western AP ice shelves will likely experience further recession under the influence of CDW, impacting velocity and stability of outlet glaciers in the ASE, which comprise the third-largest drainage system of the West Antarctic Ice Sheet today.

Chapter 4 1

Increasing glacial sensitivity to 2

climate change in the Antarctic 3

Peninsula 4

Recent warming and accelerated melting in the Antarctic Peninsula has 5 incited widespread concern for the mass balance of its ice cap (e.g., Abram et al., 2013). 6 Warming at 3.6C per century, about 5 times higher than the global mean (Houghton et 7 al., 2001; Zagorodnov et al., 2012), the Antarctic Peninsula is one of the most rapidly 8 warming regions in the world. Dynamic thinning of the ice cap and recession of 87% of 9 the glaciers in the peninsula (Cook et al., 2005) are hypothesized to be the direct result of 10 atmospheric warming and upwelling of relatively warm Circumpolar Deep Water that 11 flows through deep troughs that incise the western peninsula shelf (Pritchard et al., 2012; 12 Cook et al., 2010). To better understand the extent and implications of glacial response 13 to modern warming, we provide geologic context to recent trends through the analysis of 14 glacimarine sediments that record glacial behavior during the past 14,000 years. Here we 15 provide results from the most extensive comparative analysis to date of high-resolution 16 marine sediment records from coastal embayments of the Antarctic Peninsula. This 17 novel comparison reveals that tidewater glaciers responded asynchronously to climate 18

86 events during the Early and Mid Holocene, but synchronously to Late Holocene events as 19 these glaciers and their drainage areas decreased in size. An increase in synchronicity of 20 glacial behavior throughout the Holocene implies that the retreating glacial systems are 21 exhibiting vulnerability to recent rapid atmospheric warming. 22

4.1. Introduction 24

The Antarctic Peninsula (AP) is among the Earth’s most dynamic climate 25 systems, with a prominent mountain chain and a large ice cap acting as major influences 26 on regional climate (Fig. 4-1). The 1.5 km-elevation AP mountains form a barrier to both 27 atmospheric and oceanographic processes (Fig. 4-1). The western AP is exposed to 28 strong, moisture-laden westerly winds and cyclonic low-pressure systems that, combined 29 with warmer water of the Bellinghausen Sea, create a relatively temperate maritime 30 climate (Turner et al., 2002). Upwelling of warm Circumpolar Deep Water onto the shelf 31 (Fig. 4-1) is controlled by changing strength of westerly winds and sea-ice cover related 32 to low- and high- latitude teleconnections that are sensitive to major climate phenomena, 33 such as the El Nino Southern Oscillation and the Southern Oscillation Mode (White and 34 Peterson, 1996; Stammerjohn and Smith, 1996). By contrast, the eastern Antarctic 35 Peninsula is in the precipitation shadow of the AP mountains and experiences colder 36 temperatures as a result of low-level barrier winds (Turner et al., 2002). Precipitation on 37 the western AP coast can be up to 8 times higher than the eastern AP (Turner et al., 2002; 38 Morris and Vaughan, 2003). Atmospheric temperatures are typically 7°C higher on the 39 western side than the eastern side of the AP at the same latitude (Morris and Vaughan, 40 2003). The latitudinal temperature gradient is significant along the AP, with temperate 41 conditions on islands of the northwestern AP and subpolar and polar climates in the 42 southern AP and continental Antarctica(Morris and Vaughan, 2003). Southward 43 migration of the -5°C isotherm has been linked to catastrophic ice shelf collapse in recent 44 decades (Morris and Vaughan, 2003), with 28,000 km2 of ice shelf area lost since the 45 1960’s (Fig. 4-1; Cook et al., 2010), suggesting that atmospheric temperature imposes a 46

87 key control on floating ice stability. Accelerated flow of outlet glaciers has been shown 47 to follow loss of the protective buttressing and dampening effects provided by ice shelves 48 (e.g., DeAngelis and Skvarca, 2003; Scambos et al., 2004). Glacier size and drainage 49 basin area, elevation, hypsometry show considerably variability (Davies et al, 2012a), 50 which influences temperature and precipitation effects on glacial mass balance, and thus 51 on their response to environmental forcings. A poorly understood element of the AP 52 climate is sea ice, which forms by freezing of ocean surface water and yields significant 53 control on water column stratification, temperature, and salinity. Generally, sea ice is 54 more extensive and forms permanent pack ice in the colder, more saline Weddell Sea, but 55 has variable distribution in the Bellinghausen Sea (Fetterer et al., 2002). 56

Figure 4-1 The Antarctic Peninsula and its dynamic climate and ocean 58 interactions. 59 Study sites in black boxes. JRI ice core is located within HCF box. Note areas of 60 recent ice shelf loss are located between -5 and -9 °C isotherms (Morris and 61 Vaughan, 2003). Study areas in boxes. (Original figure with information 62 compiled from Landsat, NASA, and GeoMapApp; Hofmann et al., 1996; Smith et 63 al., 1999; Martinson et al., 2008; Jacobs et al., 2013.) 64

Given the considerable regional variability in modern Antarctic Peninsula 65 climate and oceanography, as well as the considerable variability in glacial drainage 66 basin characteristics, glaciers should respond uniquely to climate events. The modern 67 warming, however, is expressed as synchronous, widespread recession of AP glaciers 68 (Cook et al., 2005). To understand the significance of the modern recession, we 69 investigate glacial records from a variety of coastal embayments to determine timing and 70 magnitude of tidewater glacier response to past climate events. Tidewater glaciers, which 71 terminate in the sea, are ideal for investigating climate histories because they are sensitive 72 to both climate and ocean perturbations on decadal timescales (Griffith and Anderson, 73 1989). Restricted bays and fjords function as traps for sediment produced by tidewater 74 glaciers and by organisms in the water column; as such, they provide high-resolution 75 (sub-decadal) records of glacial and oceanographic variations (Griffith and Anderson, 76 1989). To date, only one early Holocene to recent-age ice core exists in the AP region 77 (Mulvaney et al., 2012). These high-resolution marine sedimentary archives offer an 78 essential supplement to ice core-derived temperature records by providing a holistic 79 regional archive of past environmental changes reflecting glacial response to climate 80 change. Moreover, expanded sedimentary records and modern dating techniques allow 81 remarkable precision in chronology, with error in ages of less than 200 years for the 82 oldest sediments (Appendix A). 83

Over three decades of seismic acquisition, bathymetric mapping, and 84 extensive multi-proxy core analysis of several coastal embayments have culminated in a 85 massive database with which we have reconstructed regional glacial and oceanographic 86 history of the peninsula. Here, we present records from nine coastal embayments, 87 extending from 62°S to 73°S, from both sides of the AP, and whose modern glaciers are 88 under the influence of variable oceanography, orography, sea-ice cover, atmospheric 89 temperature, drainage basin size, hypsometry, and seafloor bathymetry (Fig. 4-1). We 90 compare these sites with two deep ocean sites that provide regional oceanographic 91 context for the Holocene (Domack et al., 2001; Heroy et al., 2008; Barnard et al., 2014) 92 and with a recently collected ice core from James Ross Island that provides the first 93 record spanning the Late Pleistocene and Holocene from the AP (Mulvaney et al., 2012). 94

4.2. Methods 95

Cores and marine geophysical data were collected during several cruises 96 to the Antarctic Peninsula and Pine Island Bay, including: USCG Glacier DF-82, RV 97 Polar Duke-91, RV NB Palmer-0201, -0502, -0602, -0703, and Icebreaker Oden Southern 98 Ocean-0910. Multibeam swath bathymetry and Knudsen 3.5 kHz subbottom profiler data 99 were used to target expanded sedimentary sections in fjords and bays for coring. Cores 100 were processed onboard and stored at the Antarctic Research Facility at Florida State 101 University. 102

Paleoclimate records were obtained by conducting centennial scale- 103 resolution, multi-proxy analyses on well-dated cores, including: grain size, magnetic 104 susceptibility, diatom and foraminiferal assemblages, organic carbon and nitrogen 105 content, and stable isotopes (Appendix A). Our study integrates and builds upon the 106 published record by re-sampling cores for additional proxies and adding new study areas 107 that meet the criteria of having expanded Holocene stratigraphy and well-constrained 108 chronology. Multi-proxy analysis facilitates consistent comparison across sites, while 109 supplementary sites of varied oceanographic and climatic settings provide a means to test 110 relative consequence of local forcings on tidewater glacial environments through time. 111 Study areas include: Maxwell Bay (Milliken et al., 2009; Majewski et al., 2012), 112 Beascochea Bay (Hardin, 2011), Barilari Bay (Christ et al., 2014), Lallemand Fjord 113 (Shevenell et al., 1996), Neny Fjord (Allen et al., 2009), and Ferrero Bay, Firth of Tay 114 (Michalchuk et al., 2009), and Herbert Sound (Minzoni et al., in review). These records 115 were scrutinized closely with the Late Pleistocene-Holocene- age ice core from James 116 Ross Island (Mulvaney et al., 2012) and with continental shelf records from Palmer Deep 117 (Domack et al., 2001; Taylor and Sjunneskog, 2002; Shevenell et al., 2011) and 118 Bransfield Basin (Heroy et al., 2008; Barnard et al., 2014). 119

120

121

4.3. Discussion 122

Our study demonstrates that regional glacial response to climate, 123 particularly to major warm events, was highly asynchronous during the Early Holocene 124 (Figs. 4-2 and 4-3). For example, the Early Holocene Climatic Optimum (EHCO) was 125 characterized by rapid warming to temperatures +1.3  0.3C warmer than present from 126 ~13.3 to ~11.5 thousand calendar years before present (cal kyrs BP, where present day is 127 1950 AD; Mulvaney et al., 2012). Yet, initial recession of tidewater glaciers was highly 128 asynchronous, ranging from ~14.1 in Maxwell Bay (Milliken et al., 2009; Majewski et 129 al., 2012) to ~5.7 cal kyrs BP in Beascochea Bay (Fig. 4-2; Hardin, 2011). Hence, the 130 difference in onset of tidewater glacial response to warming climate and sea level rise 131 following the Last Glacial Maximum was ± 4.2 kyrs ((first site to retreat – last site to 132 retreat)/2; Fig. 4-3; Appendix A). Likewise, difference in duration of glacial response 133 was large, ± 2.5 kyrs ((longest duration of response – shortest duration of response)/2; 134 Fig. 4-3; Appendix A). Seafloor bathymetry and orography were likely the dominant 135 controls during initial deglaciation of the rugged inner continental shelf and inland 136 waters. The first site to experience glacial recession, Maxwell Bay, is particularly 137 temperate today and in a location directly influenced by Westerlies and frequent storms 138 (Milliken et al., 2009). Maxwell Bay’s glaciers may not have responded directly to 139 atmospheric warming, since the EHCO post-dates their recession; rather, open and deep 140 bathymetry (Milliken et al., 2009) likely made Maxwell Bay susceptible to warm water 141 intrusion from the south and to sea level rise (Figs. 4-1 and 4-2). This might explain the 142 similar timing of deglaciation of Maxwell Bay and the Palmer Deep, as these sites are 143 both subject to the influence of northward flowing ACC-CDW (Figs. 4-1 and 4-2). The 144 last site to deglaciate was Beascochea Bay, which is characterized by shallow, stepped 145 basins and connected by bathymetric highs that served to stabilize the glacier grounding 146 line, resulting in prolonged, stepped retreat until the mid Holocene ~5.7 cal kyr BP 147 (Hardin, 2011). Interestingly, Beascochea Bay is one of the exceptional locations where 148 glaciers are advancing today (Cook et al., 2005), perhaps owing to its shallow bathymetry 149 (Hardin, 2011), large drainage basin, and high precipitation (Turner et al., 2002). 150

A minor cooling, while only recognized in 4 of the 9 sites, has similar 151 synchronicity of glacial response (difference in onset of ± 1.0 kyrs and difference in 152 duration of ± 1.1 kyrs; Fig. 4-2; Appendix A). Since this event was not recognized 153 everywhere, it is considered less significant and highlights the variability in AP coastal 154 climate in the Mid Holocene. 155

The Mid Holocene Hypsithermal (MHH) was characterized by 156 atmospheric temperature similar to present day (0.2 ± 0.2°C) from ~8.2 to ~2.5 cal kyr 157 BP (Mulvaney et al., 2012). By this time, most tidewater glaciers in the study area 158 receded to their inner basins. Minimal glacial activity, high marine productivity, and 159 warm, open water conditions signal glacial and marine response to the MHH. Difference 160 in timing of response to the MHH was ± 1.7 kyrs—less than half that of initial glacial 161 recession (Fig. 4-3). Difference in duration of response was also much smaller, ± 1.9 162 kyrs (Fig. 4-3). While causes for differences in timing and duration of response are still 163 emerging, key forcings likely include meteorology and oceanography. Glacial systems 164 with more warmth and moisture delivery via westerly winds, such as northerly Maxwell 165 Bay (Milliken et al., 2009) and Firth of Tay (Majewski et al., 2009), responded sooner 166 than those insulated by the AP Mountains and by cooler water masses (Figs. 4-1 and 4-2). 167

Atmospheric cooling began ~3 cal kyr BP (Mulvaney et al., 2012), an 168 event referred to as the Late Holocene Neoglacial (Fig. 4-2). Response of tidewater 169 glaciers to this event was an order of magnitude more synchronous than initial glacier 170 recession, with difference in onset of ± 675 yrs and difference in duration of ± 675 yrs 171 (Fig. 4-3). Glacier response to the Neoglacial is recognized by advance, by increased ice- 172 rafting and sea-ice cover, and accompanied by decreased productivity (e.g. Michalchuk et 173 al., 2009). The Neoglacial is expressed in 8 of the 9 coastal study sites, in both ocean 174 deep sites (Domack et al., 2001; Heroy et al., 2008; Barnard et al., 2014), and in 175 lacustrine records ~3 to 2.5 cal kyr BP (Ingolfsson et al., 2003; Fig. 4-2). Since the 176 Neoglacial is documented at almost all sites within a narrow time interval that coincides 177 with onshore cooling, we interpret the main control on glacial expression to be regional 178 climate deterioration, rather than internal glacial instability or local influences that might 179 precipitate advance. 180

Figure 4-2 Comparison of Antarctic Peninsula ice core temperature record and two ocean deep locations with the 9 coastal study sites. Each column represents an individual site, and each box denotes the duration of glacial marine response to climate events (Shevenell et al., 1996; Domack et al., 2001; Taylor and Sjunneskog, 2002; Milliken et al., 2009; Michalchuk et al., 2009; Allen et al., 2010; Hardin, 2011; Majewski et al., 2012; Majewski et al., 2013; Christ et al., 2014; Minzoni et al., in review). LIA—Little Ice Age; MCA—Medieval Climate Anomaly; NGL—Neoglacial; MHH—Mid Holocene Hypsithermal; MWP—Meltwater Pulse. 181

Figure 4-3 Difference in timing of onset of glacial response vs. difference in the duration of glacial response to Holocene climate events (± kyrs). These values are calculated for each event and labeled with the number of coastal study sites that recorded a response to the event, as shown in Fig. 4-2. Each consecutive climate event decreased significantly in differences in timing and duration of glacial response. Thus, climate events are naturally in chronological order due to this phenomenon.

Historic climate events including the Medieval Climate Anomaly (MCA, ~AD 950-1250) and the Little Ice Age (LIA, ~AD 1150-1550), were much more synchronous, with difference in onset of ± 400 and ± 250 yrs, respectively, and difference in duration of ± 825 and ± 180 yrs, respectively (Fig. 4-3; Appendix A). The MCA was only recognized in 3 of the 9 sites (Fig. 4-3; Appendix A), however, perhaps because the magnitude of this event was minor in southern high latitudes (e.g., Neukom et al., 2014). The LIA, by contrast, was recognized in 7 of the 9 sites (Fig. 4-3; Appendix A).

In contrast to the peninsular coastal embayments, Ferrero Bay, the southernmost site located in Pine Island Bay of the Amundsen Sea (Fig. 4-1), did not respond to Holocene climate events (see Chapter 3; Fig. 4-2). We hypothesize that the extreme polar setting of Ellsworth Land was insulated from climate oscillations, allowing

Ferrero Bay to maintain a stable ice shelf until ~2.0 cal kyrs BP (Fig. 4-2). Both initial glacial recession (~10.7 cal kyr BP) and Late Holocene ice shelf loss are likely related to changing hydrographic conditions in Ferrero Bay, where upwelling of relatively warm deep water is observed today (OSO-0910 Cruise Report). This finding suggests that Holocene variations in atmospheric conditions were not significant enough to affect the mass balance or flow dynamics of polar Antarctic glaciers. Thus, polar tidewater systems were not as sensitive to Holocene climate variability as sub-polar and temperate Antarctic glaciers.

4.4. Conclusions

Decrease of differences in glacial response to climate events during the Holocene and increase in regional glacial synchronicity (Fig. 4-3) suggest that tidewater glaciers became increasingly sensitive to climate events as the AP Ice Sheet receded from its expanded Last Glacial Maximum configuration to inner- shelf embayments. We conclude that once glaciers receded into their inner drainage basins, they became less susceptible to maritime effects and other local forcings, and their smaller size rendered them more susceptible to large-scale climate forcings. This study suggests that the progressive increase in glacial synchronicity from the Early Holocene to the present may have resulted in greater sensitivity of the receded AP Ice Cap to the rapid recent warming of ~3.5 C over the last century (Houghton et al., 2001; Zagorodnov et al., 2012). This helps explain the widespread retreat of 87% of all glaciers in the AP and loss of over 28,000 km2 of ice shelf area over the last few decades (Cook et al., 2005, 2010).

References

Abram, N. J., Mulvaney, R., Wolff, E. W., Triest, J., Kipfstuhl, S., Trusel, L. D., Vimeux, F., Fleet, L., and Arrowsmith, C., 2013. Acceleration of snow melt in an Antarctic Peninsula ice core during the twentieth century. Nature Geoscience 6, 404-411.

Allen, C.S., Oakes-Fretwell, L., Anderson, J.B., and Hodgson, D.A., 2010. A record of Holocene glacial and oceanographic variability in Neny Fjord, Antarctic Peninsula, The Holocene 20,. 551-564.

Anderson, J. B., 1999. Antarctic Marine Geology. Cambridge University Press, Cambridge 289 pp.

Anderson, J.B., Jakobsson, M., and Party, O.S., 2010. Oden Southern Ocean 0910 OSO0910 Cruise Report. University of Stockholm, Stockholm.

Anderson, J.B., Kennedy, D.S., Smith, M.J., and Domack, E.W., 1991. Sedimentary facies associated with Antarctica's floating ice masses: in Anderson, J.B. and Ashley, G.M. (eds.), Paleoclimatic Interpretation of Glacial Marine Deposits, Special Publication 261, Geological Society of America, 1-25.

Anderson, J.B., Shipp, S.S., Lowe, A.S., Wellner, J.S., and Mosola, A.B., 2002. The Antarctic Ice Sheet during the LGM and its subsequent retreat history: a review Quaternary Science Reviews 21, 49–70.

Andrews, J.T., Domack, E.W., Cunningham, W.L., Leventer, A., Licht, K.J., Jull, A.J.T., DeMaster, D.J., and Jennings, A.E., 1999. Problems and possible solutions concerning radiocarbon dating of surface marine sediments, Ross Sea, Antarctica Quaternary Research 52, 206-216.

Antoniades, D., Francus, P., Pienitz, R., St-Onge, G., and Vincent, W.F., 2011. Holocene dynamics of the Arctic’s largest ice shelf. Proc Natl Acad Sci USA 108, 18899– 18904.

Armand, L.K., Crosta, X., Romero, O., and Pichon, J.J., 2005. The biogeography of major diatom taxa in Southern Ocean sediments: 1. Sea ice related species. Palaeoceanography, Palaeoclimatology, Palaeoecology 233, 93–126.

Armand L.K., Cornet-Barthau V., Mosseri J. and u guiner B., 2008. Late summer diatom biomass and community structure on and aroundthe naturally iron- fertilized Kerguelen Plateau in the Southern Ocean. Deep-Sea Research II, 55, 677-692.

Arrigo, K.R., Worthen, D.L., and Robinson, D.H., 2003. A coupled ocean-ecosystem model of the Ross Sea: 2. Iron regulation of phytoplankton taxonomic variability and primary production. J. Geophys. Res. 108.

Arrigo, K.R., van Dijken, G.L., and Bushinsky, S., 2008. Primary production in the Southern Ocean, 1997–2006. J. Geophys. Res. 113.

Balco, G., Schaefer, J.M., and the LARISSA group., 2013. Exposure-age record of Holocene ice sheet and ice shelf change in the northeast Antarctic Peninsula Quaternary Science Reviews 59, 101-111.

Bárcena M.A., Fabrés J., Isla E., Flores J.A., Sierro F.J., Canals M., and Palanques A., 2006. Holocene neoglacial events in the Bransfield Strait (Antarctica). Palaeoceanographic and palaeoclimatic significance. Scientia Marina 70, 607- 619.

Bard, E., Hamelin, B., Arnold, M., Montaggioni, L., Cabioch, G., Faure, G., and Rougerie, F., 1996. Deglacial sea-level record from Tahiti corals and the timing of global meltwater discharge. Nature 382, 241-244.

Barnard, A., Wellner, J., and Anderson, J., 2014. Late Holocene climate change recorded in proxy records from a Bransfield Basin sediment core, Antarctic Peninsula. Polar Research 33.

Barrand, N.E., Vaughan, D.G., Steiner, N., Tedesco, M., Kuipers Munneke, P., van den Broeke, M.R., and Hosking, J.S., 2013. Trends in Antarctic Peninsula surface melting conditions from observations and regional climate modeling. Journal of Geophysical Research: Earth Surface 118(1), 315-330.

Bentley, M. J., and Anderson, J. B., 1998. Glacial and marine geological evidence for the ice sheet configuration in the Weddell Sea Antarctic Peninsula region during the Last Glacial Maximum, Antarctic Science 10, 309-325.

Bentley, M. J., Hodgson, D. A., Smith, J. A., Cofaigh, C. Ó., Domack, E. W., Larter, R. D., Roberts, S. J. Brachfeld, S., Leventer, A., Hjort, C., Hillenbrand, C-D., and Evans, J., 2009. Mechanisms of Holocene palaeoenvironmental change in the Antarctic Peninsula region. The Holocene 19(1), 51-69.

Bentley, M. J. and 76 others, The RAISED Consortium, 2014. A community-based geological reconstruction of Antarctic Ice Sheet deglaciation since the Last Glacial Maximum, Quaternary Science Reviews 100, 1-9.

Berger, A., 1978. Long-term variations of daily insolation and Quaternary climatic changes. Journal of the Atmospheric Sciences 35(12), 2362-2367.

Bezdek, J. C., 1987. Pattern Recognition with Fuzzy Objective Function Algorithms. New York Plenum Press, New York. 256 pp.

Bezdek, J.C., and Pal, S.K., 1991. Fuzzy Models For Pattern Recognition. Methods That Search for Structures in Data. New York: IEEE Press.

Bibby, J.S., 1966. The Stratigraphy of Part of North-East Graham Land and the James Ross Island Group. British Antarctic Survey Scientific Report 53, 37 pp.

Bintanja, R., Van Oldenborgh, G. J., Drijfhout, S. S., Wouters, B., and Katsman, C. A., 2013. Important role for ocean warming and increased ice-shelf melt in Antarctic sea-ice expansion. Nature Geoscience 6(5), 376-379.

Björck, S., Håkansson, H., Olsson, S., Ellis-Evans, C., Humlum, O., and Lirio, J. M., 1996. Late Holocene palaeoclimatic records from lake sediments on James Ross Island, Antarctica. Paleogeography, Paleoclimatology, Paleoecology 113, 195– 220.

Brachfield, S., Domack, E., Kissel, C., Laj, C., Leventer, A., Ishman, S., Gilbert, R., Camerlenghi, A., and Eglinton, L.B., 2003. Holocene History of the Larsen-A Ice Shelf constrained by geomagnetic paleointensity dating. Geology 31, 749-752.

Bromwich, D.H., Nicolas, J.P., Monaghan, A.J., Lazzara, M.A., Keller, L.M., Weidner, G.A., and Wilson, A.B., 2012. Central west Antarctica among the most rapidly warming regions on Earth. Nat. Geosci. 6, 139e145.

Buffen, A., Leventer A., Rubin, A. and Hutchins, T., 2007. Diatom assemblages in surface sediments of the northwestern Weddell Sea, Antarctic Peninsula. Marine Micropaleontology 62, 7–30.

Burckle, L. H., 1984. Ecology and paleoecology of the marine diatom Eucampia antarctica (Castr.) Mangin. Marine Micropaleontology 9(1), 77-86.

Camerlenghi, A., Domack, E., Rebesco, M., Gilbert, R., Ishman, S., Leventer, A., Brachfield, S., and Drake, A., 2001. Glacial morphology and post-glacial contourites in northern Prince Gustav Channel (NW Weddell Sea, Antarctica). Marine Geophysical Researches 22, 417–443.

Carrivick, J.L., Davies, B. J.,Glasser, N.F., Nyvlt, D., and Hambrey, M. J., 2012. Late- Holocene changes in character and behavior of land-terminating glaciers on James Ross Island, Antarctica, Journal of Glaciology 58 (212), 1176-1190.

Christ, A.J., Talaia-Murray, M., Elking, N., Domack, E.W., Leventer, A., Lavoie, C., Brachfeld, S., Yoo, K., Gilbert, R., Jeong, S., Petrushak, S., Wellner, J., and the LARISSA group, 2014. Late Holocene glacial advance and ice shelf growth in Barilari Bay, Graham Land, west Antarctic Peninsula, Geological Society of America Bulletin 127.1-2, 297-315.

Cochran, J. R., Tinto, K. J., and Bell, R. E., 2015. Abbot Ice Shelf, structure of the Amundsen Sea continental margin and the southern boundary of the Bellingshausen Plate seaward of West Antarctica, Geochem. Geophys. Geosyst. 16, 1421–1438

Cook, A. J., and Vaughan, D. G., 2010. Overview of areal changes of the ice shelves on the Antarctic Peninsula over the past 50 years, The Cryosphere 4, 77-98.

Cook, A., Fox, A. J., Vaughan, D. G., and Ferrigno, J. G., 2005. Retreating glacier-fronts on the Antarctic Peninsula over the last 50 years, Science 22, 541–544.

Crosta, X., Romero, O.E., Armand, L.K., and Pichon, J.-J., 2005. The biogeography of major diatom taxa in Southern Ocean surface sediments: 2. Open-ocean related species. Palaeogeography, Palaeoclimatology, Palaeoecology 223, 66–92.

Cunningham, W. L., and Leventer, A., 1998. Diatom assemblages in surface sediments of the Ross Sea: relationship to present oceanographic conditions. Antarctic Science 10(2), 134-146.

Cunningham, W.L., Leventer, A., Andrews, J.T., Jennings, A.E., and Licht, K.J., 1999. Late Pleistocene-Holocene marine conditions in the Ross Sea, Antarctica: evidence from the diatom record, The Holocene 9(2), 129-139.

Dalziel, I. W., and Elliot, D. H., 1982. West Antarctica: problem child of Gondwanaland. Tectonics 1(1), 3-19.

Davies, B.J., Carrivick, J.L., Glasser, N.F., Hambrey, M.J., and Smellie, J.L., 2012a. Variable glacier response to atmospheric warming, northern Antarctic Peninsula, 1988-2009. The Cryosphere 6, 1031-1048.

Davies, B.J., Hambrey, M.J., Smellie, J.L., Carrivick, J.L., and Glasser, N.F., 2012b. Antarctic Peninsula Ice Sheet evolution during the Cenozoic Era. Quaternary Science Reviews 31, 30-66.

100

Davies, B.J., Golledge, N.R., Glasser, N.F., Carrivick, J.L., Ligtenberg, S.R.M., Barrand, N.E., van den Broeke, M.R., Hambrey, M.J., and Smellie, J.L., 2014. Modelled glacier response to centennial temperature and precipitation trends on the Antarctic Peninsula. Nature Climate Change 4, 993–998.

De Angelis, H., and Skvarca, P., 2003. Glacier surge after ice shelf collapse. Science 299, 1560–1562.

Dinniman, M.S., Klinck, J.M., and Smith Jr., W.O., 2003. Cross-shelf exchange in a model of the Ross Sea circulation and biogeochemistry. Deep Sea Res. II 50, 3103–3120.

Doake, C.S., 1982. State of Balance of the ice sheet in the Antarctic Peninsula, Annals of Glaciology 3.

Doake, C. S. M., and Vaughan, D. G., 1991. Rapid disintegration of the Wordie Ice Shelf in response to atmospheric warming, Nature 350, 328–330.

Domack, E.W., and Ishman, S.E., 1993. Oceanographic and physiographic controls on modern sedimentation within Antarctic fjords: Geological Society of America Bulletin 105 (9), 1175–1189.

Domack, E.W., Ishman, S.E., Stein, AB., McClennen, C.E., and Jull, A.J.T., 1995. Late Holocene advance of the M ller Ice Shelf, Antarctic Peninsula: sedimentological, geochemical and palaeontological evidence. Antarctic Science 7 (2), 159-170.

Domack, E., Leventer, A., Dunbar, R., Taylor, F., Brachfeld, S., and Sjunneskog, C., 2001. Chronology of the Palmer Deep site, Antarctic Peninsula: a Holocene palaeoenvironmental reference for the circum-Antarctic. Holocene 11, 1-9.

Domack, E.W., Burnett, and Leventer, A., 2003. Environmental Setting of the Antarctic Peninsula, In: Antarctic Peninsula Climate Variability (E.W. Domack, Ed.), Antarctic Research Series, American Geophysical Union 79, 1-13.

Domack, E., Duran, D., Leventer, A., Ishman, S., Doane, S., McCallum, S., Amblas, D., Ring, J., Gilbert, R., and Prentice, M., 2005. Stability of the Larsen B ice shelf on the Antarctic Peninsula during the Holocene epoch. Nature 436, 681-685.

Erbs-Hansen, D.R., Knudsen, K.L., Gary, A.C., Jansen E., Gyllencreutz, G., Scao, V., and Lambeck, K., 2011. Late Younger Dryas and early Holocene palaeoenvironments in Skagerrak, eastern North Atlantic: A multiproxy study. Boreas 40, 660–680.

101

Erbs-Hansen D.R., Knudsen, K.L., Gary, A.C., Gyllencreutz, G., and Jansen, E., 2012. Holocene climatic development in Skagerrak, eastern North Atlantic: Foraminiferal and stable isotopic evidence, The Holocene 22 (3), 301-312.

Evans, J., and Pudsey, C.J., 2002. Sedimentation associated with Antarctic Peninsula ice shelves: implications for palaeoenvironmental reconstructions of glacimarine sediments. Journal of the Geological Society 159, 233-237.

Evans, J., Pudsey, C.J., OCofaigh, C., Morris, P., and Domack, E., 2005. Late Quaternary history, flow dynamics and sedimentation along the eastern margin of the Antarctic Peninsula Ice Sheet, Quaternary Science Reviews 24, 741-774.

Fetterer, F., Knowles, K., Meier, W., and Savoie, M., 2002. Sea Ice Index. National Snow and Ice Data Center. National Snow and Ice Data Center, Boulder, CO, digital media. Updated 2007.

Folk, R.L., and Ward, W.C., 1957. Brazos River bar [Texas]; a study in the significance of grain size parameters. Journal of Sedimentary Research 27 (1), 3-26.

Fragoso, G. M., and Smith, W. O., 2012. Influence of hydrography on phytoplankton distribution in the Amundsen and Ross Seas, Antarctica. Journal of Marine Systems 89(1), 19-29.

Fretwell, P., Pritchard, H.D., Vaughan, D.G., and 57 others, 2013. Bedmap2: improved ice bed, surface and thickness datasets for Antarctica. Cryosphere 7, 375-393.

Garcıa, M. A., Castro, C. G., Rıos, A. F., Doval, M. D., Rosón, G., Gomis, D., and López, O., 2002. Water masses and distribution of physico-chemical properties in the Western Bransfield Strait and Gerlache Strait during Austral summer 1995/96. Deep Sea Research Part II: Topical Studies in Oceanography 49(4), 585- 602.

Garrison, D. L., and Buck, K. R., 1985. Sea-ice algal communities in the Weddell Sea: species composition in ice and plankton assemblages. Marine biology of polar regions and effects of stress on marine organisms. John Wiley and Sons, Ltd., Chichester, United Kingdom 103-122.

Garrison, D. L., Buck, K. R., and Fryxell, G. A., 1987. Algal assemblages in Antarctic pack ice and in ice-edge plankton. Journal of Phycology 23(4), 564-572.

Garrison, D.L., 1991. Antarctic sea ice biota. American Zoologist 31, 17–33.

102

Gary, A. C., Wakefield, M. I., Johnson, G. W. and Ekart, D. D., 2009. Application of fuzzy c-means clustering to paleoenvironmental analysis: example from the Jurassic, central North Sea, UK. Society for Sedimentary Geology, Special Publication 93, 9–20.

Gerringa, L. J., Alderkamp, A. C., Laan, P., Thuroczy, C. E., De Baar, H. J., Mills, M. M., van Dijken, G.L., van Haren, H., and Arrigo, K. R., 2012. Iron from melting glaciers fuels the phytoplankton blooms in Amundsen Sea (Southern Ocean): Iron biogeochemistry. Deep Sea Research Part II: Topical Studies in Oceanography 71, 16-31.

Gersonde, R., Crosta, X., Abelmann, A., and Armand, L., 2005. Sea-surface temperature and sea ice distribution of the Southern Ocean at the EPILOG Last Glacial Maximum—a circum-Antarctic view based on siliceous microfossil records. Quaternary Science Reviews 24, 869-896.

Glasser, N.F., Scambos, T.A., Bohlander, J., Truffer, M., Pettit, E., and Davies, B.J., 2011. From ice-shelf tributary to tidewater glacier: continued rapid recession, acceleration, and thinning of Rohss Glacier following the 1995 collapse of the Prince Gustav Ice Shelf, Antarctic Peninsula, Journal of Glaciology 57 (203), 397-406.

Glasser, N. F., Davies, B. J., Carrivick, J. L., Rod s, A., Hambrey, M. J., Smellie, J. L., and Domack, E., 2014. Ice-stream initiation, duration and thinning on James Ross Island, northern Antarctic Peninsula Quaternary Science Reviews 86, 78–88.

Gohl, K., 2012. Basement control on past ice sheet dynamics in the Amundsen Sea Embayment, West Antarctica. Palaeogeography, Palaeoclimatology, Palaeoecology 335, 35-41.

Gohl, K., Uenzelmann-Neben, G., Larter, R. D., Hillenbrand, C. D., Hochmuth, K., Kalberg, T., Weigelt, E., Davy, B., Kuhn, G., and Nitsche, F. O., 2013. Seismic stratigraphic record of the Amundsen Sea Embayment shelf from pre-glacial to recent times: evidence for a dynamic West Antarctic ice sheet. Marine Geology 344, 115-131.

Graham, A.G.C., Larter, R.D., Gohl, K., Dowdeswell, J.A., Hillenbrand, C.D., Smith, J.A., Evans, J., Kuhn, G., and Deen, T., 2010. Flow and retreat of the Late Quaternary Pine Island-Thwaites palaeo-ice stream, West Antarctica. Journal of Geophysical Research-Earth Surface 115.

Griffith, T.W. and Anderson, J.B., 1989. Climatic control on sedimentation in bays and fjords of the northern Antarctic Peninsula: Marine Geology 85, 181-204.

103

Hall, B. L., Henderson, G. M., Baroni, C., and Kellogg, T. B., 2010. Constant Holocene Southern-Ocean 14 C reservoir ages and ice-shelf flow rates. Earth and Planetary Science Letters 296(1), 115-123.

Hambrey, M. J., Smellie, J. L., Nelson, A. E. and Johnson, J. S., 2008. Late Cenozoic glacier–volcano interaction on James Ross Island and adjacent areas, Antarctic Peninsula region. Geological Society of America Bulletin 120, 709–731.

Hardin, L., 2011, University of Houston Master’s thesis.

Hemer, M., Post, A.L., O’Brien, P.E., Craven, M., Truswell, E.M., Roberts, D., and Harris, P.T., 2007. Sedimentological signatures of the sub-Amery Ice Shelf circulation. Antarctic Science 19, 497-506.

Heroy, D. C., and Anderson, J. B., 2005. Ice-sheet extent of the Antarctic Peninsula region during the Last Glacial Maximum (LGM)—Insights from glacial geomorphology. Geological Society of America Bulletin 117, 1497-1512.

Heroy, D. C., Sjunneskog, C., and Anderson, J. B., 2008. Holocene climate change in the Bransfield Basin, Antarctic Peninsula: evidence from sediment and diatom analysis. Antarctic Science 20, 69-87.

Heroy, D.C., and Anderson, J.B., 2007. Radiocarbon constraints on Antarctic Peninsula ice sheet retreat following the Last Glacial Maximum (LGM). Quaternary Science Reviews 26 (25-28), 3286-3297.

Hillenbrand, C.D., Smith, J.A., Kuhn, G., Esper, O., Gersonde, R., Larter, R.D., Maher, B., Moreton, S.G., Shimmield, T.M., and Korte, M., 2010. Age assignment of a diatomaceous ooze deposited in the western Amundsen Sea Embayment after the Last Glacial Maximum. Journal of Quaternary Science 25, 280-295.

Hjort, C., Ingolfsson, O., Moller, P., and Lirio, J. M., 1997. Holocene glacial history and sealevel changes on James Ross Island, Antarctic Peninsula. Journal of Quaternary Science 12, 259-273.

Hodgson, D. A., Bentley, M. J., Roberts, S. J., Smith, J. A., Sugden, D. E., and Domack, E. W., 2006. Examining Holocene stability of Antarctic Peninsula ice shelves. Eos, Transactions American Geophysical Union 87(31), 305-308.

Hodgson, D. A., 2011. First synchronous retreat of ice shelves marks a new phase of polar deglaciation. Proceedings of the National Academy of Sciences 108(47), 18859-18860.

104

Hofmann E.E., Klinck J.M., Lascara C.M. and Smith D.A., 1996. Water mass distribution and circulation west of the Antarctic Peninsula and including Bransfield Strait. In: Ross R.M., Hofmann E.E. and Quetin L.B. (eds), Foundations for Ecological Research West of the Antarctic Peninsula. Antarctic Research Series 70, 81–104.

Houghton, J. T., Ding, Y., Griggs, D. J., Noguer, M., van den Linden, P. J., Dai, X., Maskell, K., and Johnson, C. A., 2001. Climate Change 2001: The Scientific Basis. Cambridge University Press, Cambridge, 881 pp.

Hughes, T.J., 1981. The weak underbelly of the West Antarctic ice sheet. Journal of Glaciology 27(97), 518-525.

Ineson, J. R., Crame, J. A. and Thomson, M. R. A., 1986. Lithostratigraphy of the Cretaceous Strata of West James Ross Island, Antarctica. Cretaceous Research 7, 141–159.

Ingolfsson, O., Hjort, C., Björck, S. and Smith, L. R. I., 1992. Late Pleistocene and Holocene glacial history of James Ross Island, Antarctic Peninsula. Boreas 21, 209–222.

Ingólfsson, O., Hjort, C., Berkman, P. A., Björck, S., Colhoun, E., Goodwin, I. D., Hall, B., Hirakawa, K., Melles, M., Moller, P., and Prentice, M. L., 1998. Antarctic glacial history since the Last Glacial Maximum: an overview of the record on land. Antarctic Science 10, 326-344.

Ingolfsson, O., Hjort, C., and Humlum, O., 2003. Glacial and climate history of the Antarctic Peninsula since the Last Glacial Maximum. Arctic, Antarctic, and Alpine Research 35, 175–186.

Ishman, S. E. and Szymcek, P., 2003. Foraminiferal distributions in the former Larsen-A ice shelf and Prince Gustav Channel region, eastern Antarctic Peninsula margin: a baseline for Holocene paleoenvironmental change. Antarctic Research Series 79, 239-260.

Jacobs, S. S., Gordon, A. L., and Ardai, J. L. Jr., 1979. Circulation and melting beneath the Ross Ice Shelf, Science 203, 439-445.

Jacobs, S.S., and Comiso, J.C., 1997. Climate variability in the Amundsen and Bellingshausen Seas. J. Clim. 10, 697–709.

Jacobs, S.S., Hellmer, H.H., and Jenkins, A., 1996. Antarctic ice sheet melting in the Southeast Pacific. Geophysical Research Letters 23(9), 957-960.

105

Jacobs, S.S., Jenkins, A., Giulivi, C.F., and Dutrieux, P., 2011. Stronger ocean circulation and increased melting under Pine Island Glacier ice shelf. Nature Geoscience 4 (8), 519-523.

Jacobs, S., Jenkins, A., Hellmer, H., Giulivi, C., Nitsche, F., Huber, B., and Guerrero, R., 2012. The Amundsen Sea and the Antarctic Ice Sheet. Oceanography 25, 154– 163.

Jacobs, S., Giulivi, C., Dutrieux, P., Rignot, E., Nitsche, F., and Mouginot, J., 2013. Getz Ice Shelf melting response to changes in ocean forcing, J. Geophys. Res. Oceans 118, 4152–4168.

Jakobsson, M., Anderson, J.B., Nitsche, F.O., Dowdeswell, J.A., Gyllencreutz, R., Kirschner, N., Mohammad, R., O'Regan, M., Alley, R., Anandakrishnan, S., Eriksson, B., Kirshner, A.E., Fernandez, R., Stolldorf, T., Minzoni, R., and Majewski, W., 2011. Geological record of ice shelf break-up and grounding line retreat, Pine Island Bay, West Antarctica, Geology 39(7), 691-694.

Jenkins, A., Dutrieux, P., Jacobs, S., McPhail, S., Perrett, J., Webb, A., and White, D. 2012. Autonomous underwater vehicle exploration of the ocean cavity beneath an Antarctic ice shelf. Oceanography 25(3), 202–203.

Johnson, J.S., Bentley, M.J., and Gohl, K., 2008. First exposure ages from the Amundsen Sea embayment, west Antarctica: the late quaternary context for recent thinning of Pine Island, Smith, and Pope Glaciers. Geology 36(3), 223-226.

Johnson, J.S., Smellie, J.L., Nelson, A.E., and Stuart, F.M., 2009. History of the Antarctic Peninsula Ice Sheet since the Early Pliocene e evidence from cosmogenic dating of Pliocene lavas on James Ross Island, Antarctica. Global and Planetary Change 69, 205-213.

Johnson, J.S., Bentley, M.J., Roberts, S.J., Binnie, S.A., and Freeman, S.P., 2011. Holocene deglacial history of the northeast Antarctic Peninsula -A review and new chronological constraints, Quaternary Science Reviews 30, 3791-3802.

Johnson, J.S., Bentley, M.J., Smith, J.A., Finkel, R.C., Rood, D.H., Gohl, K., Balco, G., Larter, R.D. and Schaefer, J.M., 2014. Rapid thinning of Pine Island Glacier in the early Holocene, Science 343(6174), 999-1001.

Joughin, I., and Alley, R. B., 2011. Stability of the West Antarctic ice sheet in a warming world. Nature Geoscience 4(8), 506-513.

106

Kang, S.-H., and Fryxell, G.A., 1993. Phytoplankton in the Weddell Sea, Antarctica. Composition, abundance and distribution in the water-column assemblages of the marginal ice-edge zone during austral summer. Marine Biology 116, 335–348.

Kellogg, D.E. and Kellogg, T.B., 1987a. Microfossil distributions in modern Amundsen Sea sediments. Mar. Micropaleontol. 12: 203-222.

Kellogg, T. B., and Kellogg, D. E., 1987b. Recent glacial history and rapid ice stream retreat in the Amundsen sea. J. Geophys. Res. 92, 8859–8864.

King, J. C., Turner, J., Marshall, G. J., Connelly, W. M.and Lachlan-Cope, T. A., 2003. Antarctic Peninsula climate variability and its causes as revealed by analysis of instrumental records. In: Domack, E. W., Leventer, A., Burnett, A., Bindschadler, R., Convey, P. and Kirby, M. (eds) Antarctic Peninsula Climate Variability: Historical and Palaeoenvironmental Perspectives. Antarctic Research Series. American Geophysical Union, Washington, DC 79, 17–30.

Kirshner, A.E., Anderson, J.B., Jakobsson, M., O’Regan, M., Majewski, M., Nitsche, F.O., 2012. Post-LGM deglaciation in Pine Island Bay, West Antarctica, Quaternary Science Reviews 38(30), 11-26.

Komada, T., Anderson, M.R., and Dorfmeier, C.L., 2008. Carbonate removal from coastal sediments for the determination of organic carbon and its isotopic signatures, δ13C and Δ14C: comparison of fumigation and direct acidification by hydrochloric acid. Limnology and Oceanography: Methods 6(6), 254-262.

Larter, R. D., Anderson, J. B., Graham, A. G., Gohl, K., Hillenbrand, C. D., Jakobsson, M., Johnson, J.S., Kuhn, G., Nitsche, F.O., Smith, J.A., Witus, A.E., Bentley, M.J., Dowdeswell, J.A., Ehrmann, W., Klages, J.P., Lindow, J., Ó Cofaigh, C., and Spiegel, C., 2014. Reconstruction of changes in the Amundsen Sea and Bellingshausen sea sector of the West Antarctic ice sheet since the last glacial maximum. Quaternary Science Reviews 100, 55-86.

Lavoie, C., Domack, E. W., Pettit, E. C., Scambos, T. A., Larter, R. D., Schenke, H. -W., Yoo, K.C., Gutt, J., Wellner, J., Canals, M., Anderson, J., and Amblas, D., 2014. Paleo-ice flow directions of the Northern Antarctic Peninsula Ice Sheet based upon a new synthesis of seabed imagery. The Cryosphere Discussions 8(5), 5321- 5360.

Leventer, A., and Dunbar, R.B., 1996. Factors influencing the distribution of diatoms and other algae in the Ross Sea. Journal of Geophysical Research 101, 18489–18500

Leventer, A., Domack, E.W., Ishman, S.E., Brachfeld, S., McClennen, C.E., Manley, P., 1996. Productivity cycles of 200–300 years in the Antarctic Peninsula region:

107

understanding linkages among the sun, atmosphere, oceans, sea ice, and biota. Geological Society of America Bulletin 108, 1626–1644.

Leventer, A., Domack, E., Barkoukis, A., McAndrews, B., and Murray, J., 2002. Lamination from the Palmer Deep: A diatom-based interpretation, Paleoceanography 17 (3), 8002.

Licht, K. J., Cunningham, W. L., Andrews, J. T., Domack, E. W., and Jennings, A. E., 1998. Establishing chronologies from acid-insoluble organic 14C dates on antarctic (Ross Sea) and arctic (North Atlantic) marine sediments. Polar Research 17(2), 203-216.

Licht, K. J., and Andrews, J. T., 2002. The 14 C record of Late Pleistocene ice advance and retreat in the central Ross Sea, Antarctica. Arctic, Antarctic, and Alpine Research 324-333.

Locarnini, R. A., Mishonov, A. V., Antonov, J. I., Boyer, T. P., Garcia, H. E., Baranova, O. K., Zweng, M. M., and Johnson, D. R., 2010. World Ocean Atlas 2009, Volume 1: Temperature. S. Levitus, Ed. NOAA Atlas NESDIS 68, U.S. Government Printing Office, Washington, D.C., 184 pp.

Lowe, A.L., and Anderson, J.B., 2002. Reconstruction of the West Antarctic Ice Sheet in Pine Island Bay during the last glacial maximum and its subsequent retreat history. Quaternary Science Reviews 21 (16-17), 1879-1897.

Lowe, A.L., and Anderson, J.B., 2003. Evidence for abundant subglacial meltwater beneath the paleo-ice sheet in Pine Island Bay, Antarctica. Journal of Glaciology 49 (164), 125-138.

Lundqvist, J., Lillieskold, M., and Ostmark, K., 1995. Glacial and periglacial deposits of the Tumbledown Cliffs area, James Ross Island, West Antarctica. Geomorphology 11, 205–214.

Maddison, E. J., Pike, J., Leventer, A., and Domack, E. W., 2005. Deglacial seasonal and sub-seasonal diatom record from Palmer Deep, Antarctica. Journal of Quaternary Science 20(5), 435-446.

Maher, J., Wellner, J.S., and Anderson, J.B., 2004. Seismic and sediment data analysis from Herbert Sound, Northwestern Weddell Sea, indicate a long-lived ice canopy, American Geophysical Union Fall Meeting abstract #PP51E-1361.

108

Majewski, W., Wellner, J.S., Szczuciński, W., and Anderson J.B., 2012. Holocene oceanographic and glacial changes recorded in Maxwell Bay, West Antarctica. Marine Geology 326–328, 67–79.

Majewski, W., 2013. Benthic foraminifera from Pine Island and Ferrero bays, Amundsen sea. Polish Polar Research, 34(2), 169-200.

Mankoff, K. D., Jacobs, S. S., Tulaczyk, S. M., and Stammerjohn, S. E., 2012. The role of Pine Island Glacier ice shelf basal channels in deep-water upwelling, polynyas and ocean circulation in Pine Island Bay, Antarctica. Annals of Glaciology 53(60), 123-128.

Martinson, D. G., Stammerjohn, S. E., Iannuzzi, R. A., Smith, R. C., and Vernet, M., 2008. Western Antarctic Peninsula physical oceanography and spatio-temporal variability. Deep Sea Research Part II: Topical Studies in Oceanography 55, 1964-1987.

McCave, I.N., Bryant, R.J., Cook, H.F., and Coughanowr, C.A., 1986. Evaluation of a laserdiffraction- size analyzer for use with natural sediments. Journal of Sedimentary Petrology 56 (4), 561-564.

McNichol, A.P., Osborne, E.A., Gagnon, A.R., Fry, B., and Jones, G.A., 1994. Tic, Toc, Dic, Doc, Pic, Poc - Unique Aspects in the Preparation of Oceanographic Samples for C-14 Ams. Nuclear Instruments and Methods in Physics Research Section B- Beam Interactions with Materials and Atoms 92, 162-165.

Medlin, L.K., and Priddle, J., 1990. Polar marine diatoms. British Antarctic Survey, Cambridge, 214 pp.

Mercer, J.H., 1978. West Antarctic ice sheet and CO2 greenhouse effect: A threat of disaster: Nature 271, 321–325.

Michalchuk, B.R., Anderson, J.B., Wellner, J.S., Manley, P.L., Majewski, W., and Bohaty, S., 2009. Holocene climate and glacial history of the northeastern Antarctic Peninsula: the marine sedimentary record from a long SHALDRIL core, Quaternary Science Reviews 28, 3049–3065.

Milliken, K.T., Anderson, J.B., Wellner, J.S., Bohaty, S.M., and Manley, P.L., 2009. High resolution climate record from Maxwell Bay, South Shetland Islands, Antarctic. Geological Society of America Bulletin 121, 1711–1725.

109

Minzoni, R.T., Anderson, J.B., Fernandez, R., and Wellner J.S., 2015. Marine record of Holocene climate, ocean, and cryosphere interactions: Herbert Sound, James Ross Island, Antarctica, Quaternary Science Reviews 129, 239-259.

Monnin, E., Indermühle, A., Dällenbach, A., Flückiger, J., Stauffer, B., Stocker, T. F., Raynaud, D., and Barnola, J. M., 2001. Atmospheric CO2 concentrations over the last glacial termination. Science 291, 112–114.

Morris, E. M., and Vaughan, D. G., 2003. Spatial and temporal variation of surface temperature on the Antarctic Peninsula and the limit of viability of ice shelves. Antarctic Research Series 79, 61-68.

Mulvaney, R., Abram, N. J., Hindmarsh, R. C. A., Arrowsmith, C., Fleet, L., Triest, J., and Foord, S., 2012. Recent antarctic peninsula warming relative to Holocene climate and ice-shelf history. Nature 489(7414), 141-144.

NASA, 2004. Air temperature conditions and climatic geomorphological characteristics of James Ross Island, Antarctic Peninsula Entry ID: CDA_AR_GEO_J.ROSS_CLIMATOLOGY. Data Centre, Centro de Datos Antarcticos, Argentina: Antarctic Master Directory, http://gcmd.nasa.gov

Nelson, P. H. H., 1975. The James Ross Island Volcanic Group of north-east Graham Land. British Antarctic Survey Scientific Reports 54, 1–62.

Nelson, A.E., Smellie, J.l., Hambrey, M.J., Williams, M., Vautravers, M., Salzmann, U., McArthur, J.M., and Regelous, M., 2009. Neogene glacigenic debris flows on James Ross Island, northern Antarctic Peninsula, and their implications for regional climate history. Quaternary Science Reviews 28, 3138-3160.

Neukom R., Gergis J., Karoly D., Wanner, H., Curran, M., Elbert, J., González-Rouco , F., and Frank, D., 2014. Inter-hemispheric temperature variability over the past millennium. Nature Climate Change doi: 10.1038/NCLIMATE2174.

Nitsche, F.O., Jacobs, S.S., Larter, R.D., Gohl, K., 2007. Bathymetry of the Amundsen Sea continental shelf: implications for geology, oceanography, and glaciology. Geochemistry Geophysics Geosystems 8.

Nitsche, F.O., Gohl, K., Larter, R.D., Hillenbrand, C.-D., Kuhn, G., Smith, J.A., Jacobs, S., Anderson, J.B., and Jakobsson, M., 2013. Paleo ice flow and subglacial meltwater dynamics in Pine Island Bay,West Antarctica. Cryosphere 7, 249-262.

Nývlt, D., Braucher, R., Engel, Z., Mlčoch, B., and ASTER Team, 2014. Timing of the Northern Prince Gustav Ice Stream retreat and the deglaciation of northern James

110

Ross Island, Antarctic Peninsula during the last glacial–interglacial transition. Quaternary Research 82(2), 441-449.

Ó Cofaigh, C., Davies, B.J., Livingstone, S.J., Smith, J.A., Johnson, J.S., Hocking, E.P., Hodgson, D.A., Anderson, J.B., Bentley, M.J., Canals, M., Domack, E., Dowdeswell, J.A., Evans, J., Glasser, N.F., Hillenbrand, C.-D., Larter, R.D., Roberts, S.J., and Simms, A.R., 2014. Reconstruction of ice-sheet changes in the Antarctic Peninsula since the Last Glacial Maximum, Quaternary Science Reviews 100, 87-110.

Parkinson, C.L., and Cavalieri, D.J., 2012. Antarctic sea ice variability and trends, 1979– 2010. The Cryosphere 6, 871–880.

Payne, A.J., Vieli, A., Shepherd, A.P., Wingham, D.J., and Rignot, E., 2004, Recent dramatic thinning of largest West Antarctic Ice Stream triggered by oceans: Geophysical Research Letters 31, L23401.

Pike J., Crosta X., Maddison E.J., Stickley C.E., Denis D., Barbara L. and Renssen H., 2009. Observations on the relationship between the Antarctic coastal diatoms Thalassiosira antarctica Comber and Porosira glacialis (Grunow) Jørgensen and sea ice concentrations during the late Quaternary. Marine Micropaleontology 73, 14-25.

Pike, J., Allen, C.S., Leventer, A., Stickley, C.E. and Pudsey, C.J., 2008. Comparison of contemporary and fossil diatom assemblages from the western Antarctic Peninsula shelf, Marine Micropaleontology 67, 274–87.

Pike, J., Swann, G. E., Leng, M. J., and Snelling, A. M., 2013. Glacial discharge along the west Antarctic Peninsula during the Holocene. Nature Geoscience 6(3), 199- 202.

Pritchard, H. D., Arthern, R. J., Vaughan, D. G., and Edwards, L. A., 2009. Extensive dynamic thinning on the margins of the Greenland and Antarctic ice sheets. Nature 461(7266), 971-975.

Pritchard, H.D., S.R.M. Ligtenberg, H.A. Fricker, D.G. Vaughan, M.R. van den Broeke, and L. Padman, 2012. Antarctic ice-sheet loss driven by basal melting of ice shelves. Nature 484(7395), 502-505.

Pudsey, C.J., and Evans, J., 2001. First survey of Antarctic sub-ice shelf sediments reveals mid-Holocene ice shelf retreat. Geology 29, 787-790.

111

Pudsey, C.J., Evans, J., Domack, E.W., Morris, P, and del Valle, R.A., 2001. Bathymetry and acoustic facies beneath the former Larsen-A and Prince Gustav ice shelves, north-west Weddell Sea. Antarctic Science 13, 312–322.

Pudsey, C. J., Murray, J. W., Appleby, P., and Evans, J., 2006. Ice shelf history from petrographic and foraminiferal evidence, Northeast Antarctic Peninsula, Quaternary Science Reviews 25, 2357-2379.

Rabassa, J., Skvarca, P., Bertani, L. and Mazzoni, E., 1982. Glacier inventory of James Ross and Vega Islands, Antarctic Peninsula. Annals of Glaciology 3, 260–264.

Rabassa, J., 1983. Stratigraphy of the glacigenic deposits in Northern James Ross Island, Antarctic Peninsula. INQUA Symposium on the Genesis and Lithology of Quaternary Deposits 329–339.

Rabassa, J., 1987. Drumlins and Drumlinoid forms in northern James Ross Island, Antarctic Peninsula. In: Menzies, J. and Rose, J. (eds) Drumlin Symposium. Balkema, Rotterdam 267–288.

Rebesco, M., Domack, E., Zgur, F., Lavoie, C., Leventer, A., Brachfeld, S., Willmott, V., Halverson, G., Truffer, M., Scambos, T., Smith, J., and Pettit, E., 2014. Boundary condition of grounding lines prior to collapse, Larsen-B Ice Shelf, Antarctica. Science 345(6202), 1354-1358.

Reece, A., 1950. The ice of Crown Prince Gustav Channel, Graham Land, Antarctica, Journal of Glaciology 1(8), 404-409.

Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Ramsey, C.B., Buck, C.E., Burr, G.S., Edwards, R.L., Friedrich, M., Grootes, P.M., Guilderson, T.P., Hajdas, I., Heaton, T.J., Hogg, A.G., Hughen, K.A., Kaiser, K.F., Kromer, B., McCormac, F.G., Manning, S.W., Reimer, R.W., Richards, D.A., Southon, J.R., Talamo, S., Turney, C.S.M., van der Plicht, J., and Weyhenmeye, C.E., 2009. Intcal09 and Marine09 radiocarbon age calibration curves, 0-50,000 years Cal BP. Radiocarbon 51, 1111-1150.

Reynolds, J., 1981. The distribution of mean annual temperatures in the Antarctic Peninsula: British Antarctic Survey Bulletin 54, 123–133.

Rignot, E.J., 1998. Fast recession of a West Antarctic glacier. Science 281 (5376), 549- 551.

Rignot, E., Vaughan, D.G., Schmeltz, M., Dupont, T., and MacAyeal, D., 2002. Acceleration of Pine Island and Thwaites Glaciers, West Antarctica. Annals of

112

Glaciology 34, 189-194.

Rignot, E., Casassa, G., Gogineni, P., Krabill, W., Rivera, A., and Thomas., R., 2004. Accelerated ice discharge from the Antarctic Peninsula following the collapse of Larsen B ice shelf, Geophysical Research Letters 31.

Rignot, E., 2006. Changes in ice dynamics and mass balance of the Antarctic ice sheet. Philosophical Transactions of the Royal Society A-Mathematical Physical and Engineering Sciences 364 (1844), 1637-1655.

Rignot, E., Bamber, J.L., Van Den Broeke, M.R., Davis, C., Li, Y.H., Van De Berg, W.J., and Van Meijgaard, E., 2008. Recent Antarctic ice mass loss from radar interferometry and regional climate modelling. Nature Geoscience 1, 106-110.

Riley, T. R., Flowerdew, M. J. and Haselwimmer, C. E., 2011. Geological Map of Eastern Graham Land, Antarctic Peninsula (1:6 25 000 Scale). BAS GEOMAP 2 Series, Sheet 1. British Antarctic Survey, Cambridge.

Scambos, T., Hulbe, C., and Fahnestock, M., 2003. Climate-induced ice shelf disintegration in the Antarctic Peninsula. Antarctic Research Series 79, 79-92.

Scambos, T. A., Bohlander, J.A., Shuman, C.A., and Skvarca, P., 2004. Glacier acceleration and thinning after ice shelf collapse in the Larsen B embayment, Antarctica, Geophysical Research Letters 31, 18.

Scambos, T., H.A. Fricker, C.-C. Liu, J. Bohlander, J. Fastook, A. Sargent, R. Massom, and A.-M. Wu, 2009. Ice shelf disintegration by plate bending and hydro-fracture: Satellite observations and model results of the 2008 Wilkins ice shelf break-ups. Earth and Planetary Science Letters 280(1-4), 51-60.

Scherer R.P., 1994. A new method for the determination of absolute abundance of diatoms and other silt-sized sedimentary particles. Journal of Paleolimnology 12(2), 171–179.

Scott P., McMinn A., and Hosie G., 1994. Physical parameters influencing diatom community structure in eastern Antarctic sea ice. Polar Biology 14, 507–517.

Shevenell, A.E., Domack, E.W., and Kernan, G.M., 1996. Record of Holocene palaeoclimate change along the Antarctic Peninsula: Evidence from glacial marine sediments, Lallemand Fjord. In Banks, M.R. and Brown, M.J. (Eds): Climatic Succession and Glacial History of the Southern Hemisphere ocer the Last Five Million Years. Pap. Proc. R. Soc. Tasm. 130(2): 55-64.

113

Shevenell, A. E., and Kennett, J. P., 2002. Antarctic Holocene climate change: A benthic foraminiferal stable isotope record from Palmer Deep, Paleoceanography 17.

Shevenell, A. E., Ingalls, A. E., Domack, E. W., and Kelly, C., 2011. Holocene Southern Ocean surface temperature variability west of the Antarctic Peninsula. Nature 470 (7333), 250-254.

Shipboard Scientific Party, 2005. SHALDRIL 2005 Cruise Report: http://SHALDRIL.rice.edu/PDFs/total.pdf (April 2009).

Simms, A.R., Milliken, K.T., Anderson, J.B., and Wellner, J.S., 2011. The marine record of deglaciation of the South Shetland Islands, Antarctica since the Last Glacial Maximum, Quaternary Science Reviews 30, 1583-1601.

Sjunneskog, C., and Taylor, F., 2002. Postglacial marine diatom record of the Palmer Deep, Antarctic Peninsula (ODP Leg 178, Site 1098) 1: total diatom abundance. Paleoceanography 17, 8003.

Smellie, J. L., McArthur, J. M., McIntosh, W. C. and Esser, R., 2006. Late Neogene interglacial events in the James Ross Island region, northern Antarctic Peninsula, dated by Ar/Ar and Sr-isotope stratigraphy. Palaeogeography, Palaeoclimatology, Palaeoecology 242, 169–187.

Smellie, J. L., Johnson, J. S., McIntosh, W. C., Esser, R., Gudmundsson, M. T., Hambrey, M. J. and van Wyk de Vries, B., 2008. Six million years of glacial history recorded in volcanic lithofacies of the James Ross Island Volcanic Group, Antarctic Peninsula. Palaeogeography, Palaeoclimatology, Palaeoecology 260, 122–148.

Smith, A.M., Jordan, T.A., Ferraccioli, F., and Bingham, R.G., 2013. Influence of subglacial conditions on ice stream dynamics: seismic and potential field data from Pine Island Glacier, West Antarctica. J. Geophys. Res. 118, 1471-1482.

Smith, D. A., Hofmann, E. E., Klinck, J. M., and Lascara, C. M., 1999. Hydrography and circulation of the west Antarctic Peninsula continental shelf. Deep-Sea Research Part I Oceanographic Research Papers 46, 925-949.

Smith, D.A., Hofmann, E.E., Klinck, J.M., and Lascara, C.M., 1999. Hydrography and circulation of the west Antarctic Peninsula continental shelf. Deep-Sea Research Part I Oceanographic Research Papers 46, 925-949.

Smith, J.A., Hillenbrand, C.-D., Pudsey, C.J., Allen, C.S., and Graham, A.G.C., 2010. The presence of polynyas in the Weddell Sea during the Last Glacial period with

114

implications for the reconstruction of sea-ice limits and ice sheet history. Earth and Planetary Science Letters 296, 287-298.

Smith JA, Bentley, M. J., Hodgson, D. A., Roberts, S. J., Leng, M. J., Lloyd, J. M., Barrett, M.S., Bryant, C., and Sugden, D. E., 2007. Oceanic and atmospheric forcing of early Holocene ice shelf retreat, George VI Ice Shelf, Antarctica Peninsula. Quaternary Science Reviews 26, 500–516.

Smith, J.A., Hillenbrand, C.D., Kuhn, G., Larter, R.D., Graham, A.G.C., Ehrmann, W., Moreton, S.G., and Forwick, M., 2011. Deglacial history of the West Antarctic Ice Sheet in the western Amundsen Sea Embayment. Quaternary Science Reviews 30, 488-505.

Smith Jr.,W.O., and Comiso, J.C., 2008. Influence of sea ice on primary production in the Southern Ocean: a satellite perspective. J. Geophys. Res. 113.

Stammerjohn, S., and Smith, R. C., 1996. Spatial and temporal variability of western Antarctic Peninsula sea ice coverage. In: Foundations for ecological research west of the Antarctic Peninsula, American Geophysical Union, 81-104.

Stearns, L.A., Smith, B.E., and Hamilton, G.S., 2008. Increased flow speed on a large East Antarctic outlet glacier caused by subglacial floods. Nature Geoscience 1 (12), 827-831.

Steig, E.J., Ding, Q., Battisti, D.S., and Jenkins, A., 2012. Tropical forcing of Circumpolar deep water inflow and outlet glacier thinning in the Amundsen sea embayment, West Antarctica. Ann. Glaciol. 53 (60), 19-28.

Sterken, M., Roberts, S.J., Hodgson, D.A., Vyverman, W., Balbo, A.L., Sabbe, K., Moreton, S.G., and Verleyen, E., 2012. Holocene glacial and climate history of Prince Gustav Channel, northeastern Antarctic Peninsula, Quaternary Science Reviews 31, 93-111.

Storey, B. C., R. J. Pankhurst, I. L. Millar, I. W. D. Dalziel, and A. M. Grunow, 1991. A new look at the geology of Thurston Island, in Geological Evolution of Antarctica, edited by M. R. A. Thomson, J. A. Crame, and J. W. Thomson, Cambridge Univ. Press, Cambridge, U. K, pp. 399–403.

Strelin, J. A. and Sone, T., 1998. Rock glaciers on James Ross Island, Antarctica. Permafrost – Seventh International Conference (Proceedings) 55, 1027–1033.

Strelin, J. A., Sone, T., Mori, J., Torielli, C. A. and Nakamura, E., 2006. New data related to Holocene landform development and climatic change from James Ross Island,

115

Antarctic Peninsula. In: Futterer, D. K., Damaske, D., Kleinschmidt, G., Miller, H. and Tessensohn, F. (eds) Antarctica: Contributions to Global Earth Sciences. Proceedings of the IX International Symposium of Antarctic Earth Sciences, Potsdam, 2003. Springer, New York, 455–460.

Stuiver, M., Reimer, P., Bard, B., Beck, J.W., Burr, G.S., Hughen, K.A., Kromer, B., McCormac, G., Van Der Plicht, J., and Spurk, M., 1998. INTCAL.98 radiocarbon age calibration, 24,000–0 cal. BP, Radiocarbon 1041–1083.

Stuiver, M., Reimer, P.J., and Reimer, R.W., 2005. CALIB 6.0: www program and documentation, http://calib.qub.ac.uk/calib (January 2013).

Szymcek, P., Ishman, S.E., Domack, E.W., and Leventer, A., 2007. Holocene Oceanographic and Climatic Variability of the Vega Drift Revealed Through Foraminiferal Interpretation – Online Proceedings of the 10th ISAEC, edited by A.K. Cooper and C.R. Raymond et., USGS Open-File Report 2007-1047, Short Research Paper 010, 4. doi:10.3133/of2007-1047.srp010.

Taylor, F., Mcminn, A., and Franklin, D., 1997. Distribution of diatoms in surface sediments of Prydz Bay, Antarctica. Marine Micropaleontology 32, 209-229

Taylor, F., Whitehead, J. and Domack, E., 2001. Holocene paleoclimate change in the Antarctic Peninsula: evidence from the diatom, sedimentary and geochemical record, Marine Micropaleontology 41, 25–43.

Taylor, F., and Sjunneskog, C., 2002. Postglacial marine diatom record of the Palmer Deep, Antarctic Peninsula (ODP leg 178, Site 1098). Paleoceanography 17, 1026.

Thomas, R., Rignot, E., Casassa, G., Kanagaratnam, R., Acuña, C., Akins, T., Brecher, H., Frederick, E., Gogineni, P., Krabill, W., Manizade, S., Ramamoorthy, H., Rivera, A., Russell, R., Sonntag, J., Swift, R., Yungel, J., and Zwally, J., 2004. Accelerated sea-level rise from west Antarctica. Science 306.

Thuróczy, C. E., Alderkamp, A. C., Laan, P., Gerringa, L. J., Mills, M. M., Van Dijken, G. L., DeBaar, H.J.W., and Arrigo, K. R., 2012. Key role of organic complexation of iron in sustaining phytoplankton blooms in the Pine Island and Amundsen Polynyas (Southern Ocean). Deep Sea Research Part II: Topical Studies in Oceanography 71, 49-60.

Truesdale, R. S., and Kellogg, T. B., 1979. Ross Sea Diatoms: Modern assemblage distributions and their relationship to ecologic, oceanographic and sedimentary conditions. Marine Micropaleontology 4, 13-31.

116

Turner, J., Lachlan-Cope, T.A., Marshall, G.J., Morris, E.M., Mulvaney, R., and Winter, W., 2002. Spatial variability of Antarctic Peninsula net surface mass balance, Journal of Geophysical Research 107 (D13), 4173.

Turner, J., Colwell, S.R., Marshall, G.J., Lachlan-Cope, T.A., Carleton, A.M., Jones, P.D., Lagun, V., Reid, P.A., Iagovkina, S., 2005. Antarctic climate change during the last 50 years. International Journal of Climatology, 2005 25, 279-294.

Vaughan, D. G., and Doake, C. S. M., 1996. Recent atmospheric warming and retreat of ice shelves on the Antarctic Peninsula, Nature 379, 328–331.

Vaughan, D. G., Marshall, G. J., Connolley, W. M., Parkinson, C., Mulvaney, R., Hodgson, D. A., King, J. C., Pudsey, C. J., and Turner, J., 2003. Recent rapid regional climate warming on the Antarctic Peninsula. Climatic Change 60, 243- 274.

Vaughan, D.G., J.C. Comiso, I. Allison, J. Carrasco, G. Kaser, R. Kwok, P. Mote, T. Murray, F. Paul, J. Ren, E. Rignot, O. Solomina, K. Steffen and T. Zhang, 2013. Observations: Cryosphere. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

Venegas, S.A., Drinkwater, M.R., 2001. Sea ice, atmosphere and upper ocean variability in the Weddell Sea, Antarctica. Journal of Geophysical Research 106 (C8), 16747-16766.

Wade, F. A., and LaPrade, K., 1969. Geology of King Peninsula, Canisteo Peninsula, and Hudson Mountains areas, Ellsworth Land, Antarctica. Antarctic Journal of the United States 4(4), 92.

Walker, D. P., Brandon, M. A., Jenkins, A., Allen, J. T., Dowdeswell, J. A., and Evans, J., 2007. Oceanic heat transport onto the Amundsen Sea shelf through a submarine glacial trough, Geophys. Res. Lett. 34, L02602.

Wellner, J. S., Lowe, A. L., Shipp, S. S., and Anderson, J. B., 2001. Distribution of glacial geomorphic features on the Antarctic continental shelf and correlation with substrate: Implications for ice behavior, J. Glaciol. 47, 397–411.

Wellner, J.S., Heroy, D.C., and Anderson, J.B., 2006. The death mask of the Antarctic ice sheet: Comparison of glacial geomorphic features across the continental shelf, Geomorphology 75, 157-171.

117

White, W.B., and Peterson, R.G., 1996. An Antarctic circumpolar wave in surface pressure, wind, temperature and sea-ice extent. Nature 380, 699–702.

Wingham, D.J., Shepherd, A., Muir, A., and Marshall, G.J., 2006. Mass balance of the Antarctic ice sheet. Philosophical Transactions of the Royal Society A- Mathematical Physical and Engineering Sciences 364 (1844), 1627-1635.

Wingham, D.J., Wallis, D.W., and Shepherd, A., 2009. Spatial and temporal evolution of Pine Island Glacier thinning, 1995-2006. Geophysical Research Letters 36.

Witus, A. E., Branecky, C. M., Anderson, J. B., Szczuciński, W., Schroeder, D. M., Blankenship, D. D., and Jakobsson, M., 2014. Meltwater intensive glacial retreat in polar environments and investigation of associated sediments: example from Pine Island Bay, West Antarctica. Quaternary Science Reviews 85, 99-118.

Yoon, H.I., Park, B.K., Kim, Y., Kang, Cheon Yun, and Kang, S.H., 2003. Origins and paleoceanographic significance of the layered diatom ooze interval from the Bransfield Straight in the northern Antarctic Peninsula around 2500 years BP. Antarctic Research Series 79, 225–238.

Zadeh, L.A., 1965. Fuzzy Sets, Information and Control 8, 338-353.

Zielinski, U., and Gersonde, R., 1997. Diatom distribution in Southern Ocean surface sediments (Atlantic sector): Implications for paleoenvironmental reconstructions Palaeogeography, Palaeoclimatology, Palaeoecology 129, 213–250.

Zagorodnov, V., Nagornov, O., Scambos, T. A., Muto, A., Mosley-Thompson, E., Pettit, E. C., and Tyuflin, S., 2012. Borehole temperatures reveal details of 20th century warming at Bruce Plateau, Antarctic Peninsula. The Cryosphere 6(3), 675-686.

118

Appendix A

119

Maxwell Bay distal core multi-proxy analysis and interpretations (Milliken et al., 2009). Sedimentology from this study.

120

Proximal Maxwell Bay multi-proxy analysis and interpretations (Majewski et al., 2012). Sedimentology from this study.

121

Distal Andvord Bay multi-proxy analysis and interpretations (this study).

122

Proximal Andvord Bay multi-proxy analysis and interpretations (this study).

123

Beascochea Bay multi-proxy analysis and interpretations (Hardin, 2011).

124

Barilari Bay multi-proxy analysis and interpretations (Christ et al., 2014).

125

Lallemand Fjord multi-proxy analysis and interpretations (Shevenell et al., 1996; Taylor et al., 2001).

126

Neny Fjord multi-proxy analysis and interpretations (Allen et al., 2010). Sedimentology and geochemistry from this study.

127

Firth of Tay multi-proxy analysis and interpretations (Michalchuk et al., 2009; Majewski and Anderson, 2009; Foley, 2009).

128

Herbert Sound multi-proxy analysis and interpretations (see Chapter 2).

129

Ferrero Bay multi-proxy analysis and interpretations (see Chapter 3).