Geochemical Tracers of Arctic Ocean Processes: a Study of Gallium, Barium, and Vanadium

The University of Southern Mississippi The Aquila Digital Community

Dissertations

Spring 2020

Geochemical Tracers of Arctic Ocean Processes: A Study of Gallium, Barium, and Vanadium

Laura M. Whitmore University of Southern Mississippi

Follow this and additional works at: https://aquila.usm.edu/dissertations

Part of the Biogeochemistry Commons, Environmental Chemistry Commons, Geochemistry Commons, and the Oceanography Commons

Recommended Citation Whitmore, Laura M., "Geochemical Tracers of Arctic Ocean Processes: A Study of Gallium, Barium, and Vanadium" (2020). Dissertations. 1780. https://aquila.usm.edu/dissertations/1780

This Dissertation is brought to you for free and open access by The Aquila Digital Community. It has been accepted for inclusion in Dissertations by an authorized administrator of The Aquila Digital Community. For more information, please contact [email protected]. GEOCHEMICAL TRACERS OF ARCTIC OCEAN PROCESSES: A STUDY OF GALLIUM, BARIUM, AND VANADIUM

Laura M. Whitmore

A Dissertation Submitted to the Graduate School, the College of Arts and Sciences and the School of Ocean Science and Engineering at The University of Southern Mississippi in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

Approved by:

Dr. Alan M. Shiller, Committee Chair Dr. Christopher T. Hayes Dr. Denis A. Wiesenburg Dr. Leila J. Hamdan Dr. Paul D. Quay

______Dr. Alan M. Shiller Dr. Jerry D. Wiggert Dr. Karen S. Coats Committee Chair Associate Director of Dean of the Graduate School School

May 2020

Laura M. Whitmore

2020

Published by the Graduate School

ABSTRACT

The Arctic Ocean is linked to the global oceans and climate through its connectivity with the North Atlantic Ocean and the regional thermohaline deep water formation sites. It’s also a region undergoing rapid environmental change. To inform the community of potential changes in geochemical and biogeochemical cycles, this dissertation addresses three dissolved geochemical tracers (gallium, barium, and vanadium) as indicators of Arctic Ocean processes. Gallium is tested as a replacement for nutrient-type tracers in an effort to deconvolve Pacific and Atlantic derived waters in the

Arctic Ocean basins. These water masses carry different heat and salt content and can influence sea ice melt, buoyancy, and deep water formation; thus, the accurate assignment and quantification of these waters is critical. It is shown that use of dissolved gallium yields a more realistic separation of these water types than is provided by the nutrient tracers. In contrast to gallium, dissolved barium and vanadium distributions are substantially modified by regional margin processes. Yet, the two elements differ in their behavior on the shelf: shelf processes create a benthic source of barium and a sink for vanadium. More specifically, particle scavenging coupled with reducing shelf sediments appear to remove vanadium from the water column. The source of barium is less clear, but, in part, particulate formation associated with biological activity likely shuttles barium from surface waters to shelf bottom waters where dissolution of the particulate barium is a source. The influences of these processes are observed throughout the upper water column of the western Arctic Ocean and, to some extent, Arctic Ocean deep waters. Furthermore, this work is pertinent to questions related to the net effect of marginal basin shelves on oceanic V & Ba cycling, their isotopic balance, and how

climate induced changes in shelf biogeochemical cycling will impact geochemical cycling.

iii

ACKNOWLEDGMENTS

Thank you to the members of my committee and my co-authors for their guidance and contributions to the topics presented in this dissertation. Dr. Alan Shiller has my deepest gratitude for his committed mentorship throughout my Ph.D. The members, past and present, of the Shiller and Hayes Labs have been instrumental in this work.

Especially Melissa Gilbert for her analytical expertise and time; Dr. Peng Ho for her breadth of trace element knowledge; and Dr. Virginie Sanial for her oceanographic knowledge and curiosity.

This research was supported by the National Science Foundation (OCE-1435312).

I owe many thanks to the captain and crew of the USCGC Healy (HLY1502) and to the scientific crew and team that has been working hard to understand the results. For their leadership through this effort, thank you to the cruise PIs: Bill Landing, Dave Kadko, and

Greg Cutter. Additionally, many members of the School of Ocean Science and

Engineering have supported my growth in academia, and they have my gratitude.

Thomas Kelly, who has discussed every part of this dissertation with me and has read and reread drafts upon drafts, is the gosh-darned best. He deserves an endless supply of my appreciation. I guess I’ll marry him.

DEDICATION

This work is dedicated to my parents, Elizabeth and Paul Whitmore, who have always encouraged my education. They inspired in me a curiosity, love, and care for nature around which my life has grown. Together they fostered a team of Whitmores in our family that has supported me through this process, from grinning to griping. My siblings, Leanne and Nick, have elevated me with hours of laughter. Mom and Dad, thanks for all the love.

TABLE OF CONTENTS

ABSTRACT ...... ii

ACKNOWLEDGMENTS ...... iv

DEDICATION ...... v

LIST OF TABLES ...... xi

LIST OF ILLUSTRATIONS ...... xii

LIST OF ABBREVIATIONS ...... 1

CHAPTER I - INTRODUCTION ...... 4

1.1 References ...... 8

CHAPTER II – GALLIUM: A NEW TRACER OF PACIFIC WATER IN THE ARCTIC

OCEAN ...... 10

2.1 Abstract ...... 10

2.2 Introduction ...... 11

2.3 Methods...... 16

2.3.1 Section Description ...... 16

2.3.2 Ancillary Hydrographic Data ...... 18

18 2.3.3 δ O-H2O Analytical Methods ...... 19

2.3.4 Ga Sample Collection and Analytical Methods ...... 19

2.3.5 Linear Mixing Model ...... 21

2.4 Results and Discussion ...... 22

2.4.1 Gallium Distribution ...... 22

2.4.2 Linear Mixing Model Results ...... 26

2.4.3 Inventories...... 31

2.4.4 Vertical Advection-Diffusion Model ...... 34

2.5 Conclusions ...... 37

2.6 Acknowledgements ...... 38

2.7 References ...... 39

CHAPTER III – ARCTIC OCEAN BARIUM BUDGET REVEALS LARGE MARGIN

INPUTS TO THE BASIN ...... 44

3.1 Abstract ...... 44

3.2 Introduction ...... 44

3.3 2015 Arctic GEOTRACES Sections ...... 48

3.3.1 Section Hydrography ...... 50

3.4 Methods...... 54

3.4.1 Sample Collection and Analysis ...... 54

3.4.1.1 Dissolved Barium Concentrations ...... 54

3.4.1.2 Dissolved Ba Isotopes ...... 57

3.4.1.3 Particulate Barium Concentration ...... 59

3.4.1.4 Ancillary Data ...... 59

3.4.2 Data Analysis ...... 60 vii

3.5 Results and Discussion ...... 61

3.5.1 Distribution of Ba in the Arctic Ocean ...... 61

3.5.2 Assessing Ba distributions relative to predicted Ba ...... 69

3.5.3 Barium Isotopes ...... 73

3.5.4 Basin-wide Ba Mass Balance...... 78

3.5.5 Supply of shelf-derived Ba to the Arctic Ocean basins ...... 84

3.5.6 Arctic Deep Water Ba signal ...... 86

3.5.7 Ba in the Canadian Arctic Archipelago ...... 87

3.6 Conclusions and Implications ...... 92

3.7 References ...... 93

CHAPTER IV – VANADIUM CYCLING IN THE WESTERN ARCTIC OCEAN IS

INFLUENCED BY SHELF-BASIN CONNECTIVITY ...... 103

4.1 Abstract ...... 103

4.2 Introduction ...... 104

4.3 Methods...... 107

4.3.1 Section Introduction and Geography ...... 107

4.3.2 Section Hydrography ...... 109

4.3.3 Dissolved Sample Collection and Analysis ...... 113

4.3.4 Particulate sample collection and analysis ...... 114

4.3.5 Data analysis...... 115 viii

4.4 Results ...... 116

4.4.1 Dissolved V (dV) distribution ...... 116

4.4.2 Particulate V (pV) distribution...... 119

4.5 Discussion ...... 120

4.5.1 Comparison of Arctic dV distribution to other ocean basins...... 120

4.5.2 The Pacific halocline & surface waters ...... 121

4.5.2.1 Influence of water-type mixing on dV distribution...... 122

4.5.2.2 Biological removal of dV...... 126

4.5.2.3 Chemically-driven particle interactions...... 131

4.5.2.4 Reducing shelf environment...... 138

4.5.2.5 Arctic Ocean V cycle...... 140

4.5.3 Deep water dV distribution ...... 141

4.5.4 Utility as a tracer ...... 142

4.6 Conclusions ...... 145

4.7 Acknowledgements ...... 146

4.8 References ...... 146

CHAPTER V – CONCLUSIONS ...... 162

5.1 References ...... 164

APPENDIX A – SUPPLEMENTAL TO CHAPTER II ...... 166

APPENDIX B – SUPPLEMENTARY MATERIAL TO CHAPTER III ...... 178 ix

APPENDIX C – SUPPLEMENTARY MATERIAL TO CHAPTER IV ...... 180

LIST OF TABLES

Table 2.1 Gallium reproducibility assessment ...... 21

Table 2.2 Endmember parameter values...... 22

Table 3.2 Barium endmember estimates (nmol kg-1)...... 69

Table 3.3 Estimated fluxes of barium from Arctic Ocean sources and sinks (in mol y-1) 82

Table 4.2 Mean dV concentrations of several water types discussed in the results...... 117

Table 4.3 Summary of literature V:P, V:Fe, and V:Mn ratios (mmol:mol) ...... 129

Table A.2 Summary of advection-diffusion model statistics...... 176

Table B.1 Standard Reference Material (SRM) summary and reproducibility analysis 178

LIST OF ILLUSTRATIONS

Figure 2.1 Nitrate-Phosphate relationship in the western Arctic Ocean...... 14

Figure 2.2 Regional and hydrographic description of the study area...... 18

Figure 2.3 Shelf distribution of dissolved gallium...... 23

Figure 2.4 Arctic Ocean dissolved gallium profiles...... 24

Figure 2.5 Comparison of Ga-derived water mass fraction and ANP-derived fractions. . 27

Figure 2.6 Sections of ANP-based and Ga-based water mass deconvolutions...... 29

Figure 2.7 Profiles of physical properties and Pacific fraction (ANP and Ga-derived). .. 30

Figure 2.8 Pacific water inventory (m)...... 32

Figure 3.1 Regional geography, hydrography, and station map...... 50

Figure 3.2 Intercalibration Comparison...... 57

Figure 3.3 Distributions of Dissolved Ba...... 63

Figure 3.4 Temperature-Salinity plots with dBa and pBaSSF z-axes...... 65

Figure 3.5 Particulate Ba of large and small size fractions...... 66

Figure 3.6 Shelf distribution of dBa and pBa...... 68

-1 Figure 3.7 Baanomaly (nmol kg ) for the western, central, and eastern Arctic Ocean...... 71

Figure 3.8 δ138Ba patterns in the western Arctic Ocean...... 75

Figure 3.9 δ138Ba distribution over the Bering and Chukchi Seas...... 78

Figure 3.10 Dissolved Ba distributions in the Canadian Arctic Archipelago...... 88

Figure 4.1 Arctic Ocean Geography & Section Description ...... 109

Figure 4.2 Water mass descriptions ...... 112

Figure 4.3 Dissolved and particulate V distributions...... 118

Figure 4.4 Global comparison of dV profiles...... 121

xii

Figure 4.5 Dissolved V versus salinity...... 123

Figure 4.6 TPD dV versus salinity...... 125

Figure 4.7 Dissolved and particulate V versus total chlorophyll a...... 127

Figure 4.8 Process associated labile pV distributions...... 130

Figure 4.9 TPD dV versus salinity...... 135

Figure 4.10 Dissolved V versus latitude...... 144

Distribution of Atlantic waters...... 167

Distribution of meteoric waters...... 168

Distribution of sea ice melt/formation waters...... 169

Comparison of the Yamamoto-Kawaii et al. (2008) nutrient method to the

Newton et al. (2013) nutrient method...... 170

Comparison of the Yamamoto-Kawaii et al. (2008) nutrient method to the Ga method (this study)...... 171

Salinity, temperature, and sigma advection-diffusion model results for

“Station 1” ...... 172

Salinity, temperature, and sigma advection-diffusion model results for

“Station 2” ...... 173

Advection-diffusion model results from the Pacific Fraction for “Station 1”

...... 174

Advection-diffusion model results from the Pacific Fraction for “Station 2”

...... 175

Ba:Ra ratio...... 178

Arctic Ocean 1994 & 2015 Station Map...... 179

xiii

Particulate V distributions (labile and nonlithogenic)...... 181

Dissolved and particulate trace elements versus total chlorophyll a and fucoxanthin...... 182

Dissolved and particulate V versus nutrients...... 183

Shelf profiles of dV and pV...... 184

Makarov Basin: full depth section profiles of process-associated pV distributions...... 185

Canada Basin: full depth section profiles of process-associated pV distributions...... 186

Pacific halocline: dV versus latitude...... 187

xiv

LIST OF ABBREVIATIONS

ACW Alaska Coastal Water

AFWM Annual flow weighted mean

AMOC Atlantic Meridional Overturning Circulation

ANP Arctic Nitrate-Phosphate tracer

BCO-DMO Biological and Chemical Oceanographic

… Data Management Office

BSBW Barents Sea Branch Water

CAA Canadian Arctic Archipelago

CCHDO CLIVAR and Carbon Hydrographic Data

… Office

CRDS Cavity Ring-down Spectroscopy dBa Dissolved barium dV Dissolved vanadium

EGC East Greenland Current

F Flux (elemental flux in nmol m-3) fAtl Fraction of Atlantic water in a sample fice Fraction of sea ice-derived water in a sample fmet Fraction of meteoric water in a sample fPac Fraction of Pacific water in a sample

FSBW Fram Strait Branch Water

GN01 2015 U.S. GEOTRACES Arctic Section

GN02 2015 CA GEOTRACES Arctic Section

… (Leg 1)

GN03 2015 CA GEOTRACES Arctic Section

… (Leg 2)

GN04 2015 European GEOTRACES Arctic

… Section

ICP-MS Inductively Coupled Plasma –

… Mass Spectrometry

J Volume flux of seawater (m3 y-1)

MC-ICP-MS Multiple Collector – Inductively Coupled

… Plasma – Mass Spectrometry

NADW North Atlantic Deep Water

ODF Oceanographic Data Facility

… (at the Scripps Institute for Oceanography) pBa Particulate barium pFe Particulate Iron

PH Pacific halocline

PML Polar mixed layer pMn Particulate manganese

POC Particulate organic carbon pV Particulate vanadium

RMSE Root mean square error

RSD Relative standard deviation

RSQ R-squared (r2)

S Salinity sBSW or SW Summer Bering Sea Water (summer water)

SGD Submarine Groundwater Discharge

SSR Sum of the Squared Residuals

T Temperature

TE Trace element

TPD Transpolar Drift

UAF University of Alaska Fairbanks

UCC Upper Continental Crust

UCSC University of California Santa Cruz

USCGC United States Coast Guard Cutter

USM University of Southern Mississippi

VSMOW Vienna Standard Mean Ocean Water

VUB Vrije Universiteit Brussel wBSW or WW Winter Bering Sea Water (winter water)

WHOI Woods Hole Oceanographic Institute

WOCE World Ocean Circulation Experiment

PML Polar mixed layer

CHAPTER I - INTRODUCTION

The Arctic Ocean is an important component of the ocean’s overturning circulation and the global climate system. Yet, because of its remoteness, ice cover, and extreme temperatures, it is understudied in comparison with other ocean basins. Waters exiting the Arctic Ocean contribute to Atlantic Meridional Overturning Circulation through deep water formation (Goosse et al., 1997), thereby linking the Arctic Ocean and the global oceanic heat, carbon, and elemental budgets. The formation of deep water in the North Atlantic Ocean occurs when surface waters cool, become denser, and eventually sink to depth. Changes within the Arctic, such as decreased surface salinity, can influence the rate of deep water formation and impact the overturning circulation.

The Arctic Ocean is a major source of freshwater to the North Atlantic Ocean because of low salinity of arctic source waters including river discharge, sea ice melt, and Pacific- derived seawater. Since the 1990s, observations of increased river discharge (Peterson,

2002), increased Pacific seawater flux through the Bering Strait (Woodgate et al., 2012), and decreased multi-year sea ice (Stroeve et al., 2007) have been made.

Within the Arctic Ocean, changes in the distribution and quantity of different water types can affect local biogeochemistry and heat. For example, sea ice melt can be driven from surficial heating or through the input of warmer seawater to the sea ice-ocean interface (Perovich et al., 2008; Polyakov et al., 2017), and, indeed, both mechanisms are at play. A decline in sea-ice has led to decreased albedo and greater absorption of radiant heat into the surface Arctic Ocean (Perovich and Polashenski, 2012). In the western

Arctic Ocean (north of Alaska), the surface ocean resiliency to sea-ice melt is supported by the distribution of cool Pacific-derived waters (Shimada et al., 2006). Yet any

warming of these waters or changes in the circulation patterns could drive erosion of the sea ice (Shimada et al., 2006; Polyakov et al., 2017) in the Arctic Ocean. Importantly, as a modifier of both surface salinity and heat, which drive the thermohaline circulation in the North Atlantic Ocean, changes in the Arctic Ocean may form important teleconnections to the climate in other regions.

Besides influencing salinity and heat, the distribution of water types (such as

Pacific-derived waters, Atlantic derived waters, river discharge, and sea ice melt or formation) can influence biogeochemistry and local ecology. Pacific-derived seawater tends to be more nutrient rich than Atlantic-derived seawater and the supply of nutrient rich waters influences the amount of primary productivity Arctic-wide (Nishino et al.,

2013). To understand how the distributions of waters influence local and regional processes, we must be able to accurately assign water masses and understand the biogeochemical cycles of potential tracers.

Trace element (TE) and isotopic research has been galvanized by the international

GEOTRACES program, which aims “to identify processes and quantify fluxes that control the distribution of key trace elements and isotopes in the ocean, and to establish the sensitivity of these distributions to changing environmental conditions”

(GEOTRACES Planning Group, 2006). Per this effort, the understanding of trace metal cycles has been bolstered due to TE connectivity to biological cycles and their utility in understanding physical regimes.

In 2015, three GEOTRACES-supported expeditions sampled Arctic waters. These cruises marked a significant moment in Arctic Ocean research, as no prior cruises had collected such a large amount of TE-associated data across the pan-Arctic region.

Utilizing samples from the 2015 GEOTRACES Arctic Ocean sections, this dissertation reports results from three dissolved trace element distributions: gallium (Ga), barium

(Ba), vanadium (V). Gallium and vanadium results are described for the 2015 U.S.

GEOTRACES Arctic Ocean section (GN01). In assessing barium distributions, data from the German-based (GN04) and Canadian-based cruises (GN02 & GN03) were included, and data from 2015 were compared to the 1994 Arctic Ocean Survey. By assessing the geochemical cycles (the sources and sinks) of these TEs, the distributions of TEs and water types within the Arctic Ocean are considered.

Herein, these tracers are used to address the distribution of water types (Pacific,

Atlantic, rivers and precipitation, and sea ice melt or formation). Furthermore, we consider the influence of the continental shelves on tracer distributions. Indeed, recent studies have demonstrated substantial shelf inputs of TEs, which had not been previously quantified (Kipp et al., 2018; Vieira et al., 2019). In Chapter 2, I confront the difficulty of discerning Pacific-derived seawater from Atlantic-derived seawater by employing Ga as a tracer. Previously used nutrient-based tracers have high uncertainties (e.g. Alkire et al.,

2015; Yamamoto-Kawai et al., 2008). Ga, which is conserved along isopycnals, yielded a potentially more realistic distinction between the two water masses than nutrient-based tracers.

In Chapter 3, the biogeochemical cycle of Ba in the Arctic Ocean is investigated.

Prior studies have applied dissolved Ba as a tracer of North American or Eurasian river expression in the surface Arctic Ocean (e.g., Guay and Kenison Falkner, 1997; Roeske et al., 2012). We consider the dissolved Ba budget to determine how well conservative mixing of water masses explains the upper 500 m of the water column. We determined

that a substantial dissolved Ba shelf source accounting for nearly 50% of the budget is required to satisfy the balance.

Margin processes recently have been realized to play an important role in ocean biogeochemical distributions (e.g., Charette et al., 2016). Chapter 3 demonstrates this for dissolved Ba in the Arctic and it is further realized for V. In Chapter 4, I address the role of Arctic Ocean margin processes on the dissolved V distribution. I observed that margin processes leave an imprint on the dissolved V distribution in the upper water column and deep waters.

In addition to the results described in the following chapters, my research has contributed to other GEOTRACES efforts led by collaborators. In Marsay et al. (2018), dissolved Ba, Ga, and V are described in sea ice and melt ponds. The dissolved Ba and V distributions are largely dominated by the mixing of seawater. Dissolved Ga is potentially influenced by atmospheric deposition in addition to the seawater source. Dissolved V in surface waters is further described in Kadko et al. (2019), wherein the mean residence time of V (and other TEs) with respect to atmospheric deposition is estimated as the product of the beryllium (7Be) flux and the TE:7Be ratio in aerosols. Charette et al. (in prep) investigates the connectivity of Eurasian rivers to the central Arctic Ocean via transport in the transpolar drift. Ba and V trend with the river distribution; however, V has clear indicators of removal in the estuary and over the shelf. Ga does not trend with the river signal, which possibly suggests that the dissolved Ga signal is not conserved in the estuary. Lastly, outside of the Arctic, Jenkins et al. (submitted) considers influence of the Loihi Seamount to the geochemical distributions of dissolved constituents in the

North Pacific Ocean.

Specific details of the studied tracer and of important oceanographic features are described within each chapter. Finally, the tracers will be considered together and expectations for future change will be addressed.

1.1 References

Alkire, M.B., Morison, J., Andersen, R., 2015. Variability in the meteoric water, sea-ice melt, and Pacific water contributions to the central Arctic Ocean, 2000-2014. J. Geophys. Res. Oceans 120, 1573–1598. https://doi.org/10.1002/2014JC010023 Charette, M.A., Kipp, L.E., Jensen, L.T., Dabrowski, J.S., Whitmore, L.M., submitted. The Transpolar Drift as a source of riverine and shelf-derived trace elements to the central Arctic Ocean. J. Geophys. Res. Oceans. Charette, M.A., Lam, P.J., Lohan, M.C., Kwon, E.Y., Hatje, V., Jeandel, C., Shiller, A.M., Cutter, G.A., Thomas, A., Boyd, P.W., Homoky, W.B., Milne, A., Thomas, H., Andersson, P.S., Porcelli, D., Tanaka, T., Geibert, W., Dehairs, F., Garcia- Orellana, J., 2016. Coastal ocean and shelf-sea biogeochemical cycling of trace elements and isotopes: lessons learned from GEOTRACES. Philos. Trans. R. Soc. Math. Phys. Eng. Sci. 374, 20160076. https://doi.org/10.1098/rsta.2016.0076 GEOTRACES Planning Group, 2006. GEOTRACES Science Plan. Goosse, H., Fichefet, T., Campin, J.-M., 1997. The effects of the water flow through the Canadian Archipelago in a global ice-ocean model. Geophys. Res. Lett. 24, 1507– 1510. https://doi.org/10.1029/97GL01352 Guay, C.K., Kenison Falkner, K., 1997. Barium as a tracer of Arctic halocline and river waters. Deep Sea Res. Part II Top. Stud. Oceanogr. 44, 1543–1569. https://doi.org/10.1016/S0967-0645(97)00066-0 Jenkins, W.J., Hatta, M., Fitzsimmons, J.N., Schlitzer, R., Lanning, N.T., Shiller, A.M., Buckley, N.R., German, C.R., Lott III, D.E., Weiss, G., Whitmore, L.M., Casciotti, K., Lam, P.J., Cutter, G.A., Cahill, K.L., submitted. An intermeidate- depth source of hydrothermal 3He and dissolved iron in the North Pacific. Earth Planet. Sci. Lett. Kadko, D., Aguilar-Islas, A., Bolt, C., Buck, C.S., Fitzsimmons, J.N., Jensen, L.T., Landing, W.M., Marsay, C.M., Rember, R., Shiller, A.M., Whitmore, L.M., Anderson, R.F., 2019. The residence times of trace elements determined in the surface Arctic Ocean during the 2015 US Arctic GEOTRACES expedition. Mar. Chem. 208, 56–69. https://doi.org/10.1016/j.marchem.2018.10.011 Kipp, L.E., Charette, M.A., Moore, W.S., Henderson, P.B., Rigor, I.G., 2018. Increased fluxes of shelf-derived materials to the central Arctic Ocean. Sci. Adv. 4, eaao1302. https://doi.org/10.1126/sciadv.aao1302 Marsay, C.M., Aguilar-Islas, A., Fitzsimmons, J.N., Hatta, M., Jensen, L.T., John, S.G., Kadko, D., Landing, W.M., Lanning, N.T., Morton, P.L., Pasqualini, A., Rauschenberg, S., Sherrell, R.M., Shiller, A.M., Twining, B.S., Whitmore, L.M., Zhang, R., Buck, C.S., 2018. Dissolved and particulate trace elements in late

summer Arctic melt ponds. Mar. Chem. 204, 70–85. https://doi.org/10.1016/j.marchem.2018.06.002 Nishino, S., Itoh, M., Williams, W.J., Semiletov, I., 2013. Shoaling of the nutricline with an increase in near-freezing temperature water in the Makarov Basin: shoaling of the Makarov Basin nutricline.. J. Geophys. Res. Oceans 118, 635–649. https://doi.org/10.1029/2012JC008234 Perovich, D.K., Polashenski, C., 2012. Albedo evolution of seasonal Arctic sea ice: albedo evolution of seasonal sea ice. Geophys. Res. Lett. 39, n/a-n/a. https://doi.org/10.1029/2012GL051432 Perovich, D.K., Richter-Menge, J.A., Jones, K.F., Light, B., 2008. Sunlight, water, and ice: Extreme Arctic sea ice melt during the summer of 2007. Geophys. Res. Lett. 35, L11501. https://doi.org/10.1029/2008GL034007 Peterson, B.J., 2002. Increasing River Discharge to the Arctic Ocean. Science 298, 2171– 2173. https://doi.org/10.1126/science.1077445 Polyakov, I.V., Pnyushkov, A.V., Alkire, M.B., Ashik, I.M., Baumann, T.M., Carmack, E.C., Goszczko, I., Guthrie, J., Ivanov, V.V., Kanzow, T., Krishfield, R., Kwok, R., Sundfjord, A., Morison, J., Rember, R., Yulin, A., 2017. Greater role for Atlantic inflows on sea-ice loss in the Eurasian Basin of the Arctic Ocean. Science 356, 285–291. https://doi.org/10.1126/science.aai8204 Roeske, T., Bauch, D., Rutgers van der Loeff, M., Rabe, B., 2012. Utility of dissolved barium in distinguishing North American from Eurasian runoff in the Arctic Ocean. Mar. Chem. 132–133, 1–14. https://doi.org/10.1016/j.marchem.2012.01.007 Shimada, K., Kamoshida, T., Itoh, M., Nishino, S., Carmack, E., McLaughlin, F., Zimmermann, S., Proshutinsky, A., 2006. Pacific Ocean inflow: Influence on catastrophic reduction of sea ice cover in the Arctic Ocean. Geophys. Res. Lett. 33, L08605. https://doi.org/10.1029/2005GL025624 Stroeve, J., Holland, M.M., Meier, W., Scambos, T., Serreze, M., 2007. Arctic sea ice decline: Faster than forecast: Arctic Ice loss - faster than forecast. Geophys. Res. Lett. 34. https://doi.org/10.1029/2007GL029703 Vieira, L.H., Achterberg, E.P., Scholten, J., Beck, A.J., Liebetrau, V., Mills, M.M., Arrigo, K.R., 2019. Benthic fluxes of trace metals in the Chukchi Sea and their transport into the Arctic Ocean. Mar. Chem. 208, 43–55. https://doi.org/10.1016/j.marchem.2018.11.001 Woodgate, R.A., Weingartner, T.J., Lindsay, R., 2012. Observed increases in Bering Strait oceanic fluxes from the Pacific to the Arctic from 2001 to 2011 and their impacts on the Arctic Ocean water column. Geophys. Res. Lett. 39. https://doi.org/10.1029/2012GL054092 Yamamoto-Kawai, M., McLaughlin, F.A., Carmack, E.C., Nishino, S., Shimada, K., 2008. Freshwater budget of the Canada Basin, Arctic Ocean, from salinity, δ 18 O, and nutrients. J. Geophys. Res. 113. https://doi.org/10.1029/2006JC003858

CHAPTER II – GALLIUM: A NEW TRACER OF PACIFIC WATER IN THE ARCTIC

OCEAN

Co-Authors: Angelica Pasqualini, Robert Newton, Alan Shiller

Submitted to: Journal of Geophysical Research: Oceans (Nov. 3, 2019)

2.1 Abstract

Determining the proportions of Atlantic and Pacific Ocean seawater entering the

Arctic Ocean is important both for understanding the mass balance of this basin as well as its contribution to formation of North Atlantic Deep Water, and because Atlantic waters carry a greater potential for global warming effects in the Arctic than do Pacific waters. To quantify the distribution and amount of Pacific and Atlantic-origin seawater in the western Arctic Ocean, we used dissolved Ga in a four-component linear endmember mixing model. Previously, nutrients, combined in their Redfield ratios, have been used to separate Pacific- and Atlantic-derived waters. These nutrient tracers are not conservative in practice, and there is a need to find quantities that are conserved, at least away from ocean boundaries. Dissolved Ga concentrations show measurable contrast between

Atlantic and Pacific source waters, shelf-influenced waters show little impact of shelf processes on the dissolved Ga distribution, and dissolved Ga in the Arctic basins is conserved along isopycnal surfaces. Thus, we explored the potential of Ga as a new parameter in Arctic source water deconvolution. The deconvolution of samples into their source water-masses using the Ga-informed mixing model was compared to that generated with the NO3:PO4 relationship. While distributions of the water masses were qualitatively similar, the Ga-based deconvolution predicted higher amounts of Pacific water at depths between about 150 and 300 m. The Ga-based decomposition yields a

smoother transition between the halocline and Atlantic layers, while nutrient-based solutions have sharper transitions. A 1-D advection-diffusion model was used to constrain estimates of the vertical diffusivity coefficient (kz). The Ga-based kz estimates agreed better with those from salinity and temperature than the nutrient method. The Ga- based approach implies greater vertical mixing between the Pacific and Atlantic waters.

2.2 Introduction

Buoyancy, or freshwater, in the Arctic Ocean is derived from inputs of river discharge, local precipitation, sea ice melt, and Pacific-derived seawater (since Pacific- derived seawater is less saline than Atlantic-derived seawater). The distribution of low salinity waters in the Arctic Ocean is important locally within the Arctic Ocean, and more broadly for its impacts on the regional and global circulation through export to the North

Atlantic. Within the Arctic Ocean, a freshwater-derived buoyancy anomaly helps stabilize all of the major boundary currents of the basin, including: the coastal currents along the Eurasian and North American boundaries, the shelf-slope currents of the

Atlantic Layer, and the topographically-guided “loop” currents (e.g., Rudels, 2018). In addition, freshwater content in the halocline insulates sea-ice from the warmer waters below and allows perennial sea-ice cover in the central Arctic Ocean (Steele & Boyd,

1998). As freshwater is exported from the Arctic Ocean, it enters the North Atlantic, where it has important regional impacts on the physical and ecological conditions of the

Nordic Seas (Torres-Valdés et al., 2013). By controlling the vertical stratification, Arctic

Ocean freshwater modulates the formation of deep water in the Nordic and Labrador Seas where the northern descending limb of the meridional overturning circulation is formed.

Over the last several decades, a great deal of progress has been made in tracking sea ice and river runoff through the Arctic marine system using a combination of in situ marine chemical sampling and satellite observations (e.g., Bauch et al., 2011; Guay &

Falkner, 1997; Jones & Anderson, 1986; Newton et al., 2013; Serreze et al., 2006).

However, the contribution of Pacific-sourced waters to buoyancy anomalies has been difficult to precisely identify, largely because Pacific water undergoes a range of biogeochemical transformations as and after it enters the Arctic Ocean.

While freshwater inputs and outputs to the Arctic Ocean must equal in the long run, recent studies have observed a pattern of freshwater accumulation followed by major freshwater export events (Belkin, 2004, Häkkinen, 1993; Hunkins & Whitehead, 1992).

Currently, the Arctic Ocean is in a period of freshwater accumulation (Haine et al., 2015).

This buildup can influence circulation patterns and, therefore, the distribution of water- mass components in the Arctic Ocean.

The Polar Mixed Layer (PML) overlays Pacific-derived waters with high nutrient concentrations and Atlantic-derived waters containing a substantial heat signal. The flux of heat and nutrients to the surface is regulated by the strength of stratification, which in the Arctic Ocean is primarily controlled by salinity. The strength and distribution of the halocline can influence the spatiotemporal extent of ice cover (e.g., Shaw et al., 2009) and euphotic zone nutrient concentrations, thereby influencing primary productivity and carbon export. Thus, making sensible projections of the impact of climate change on carbon cycling and freshwater exports from the Arctic Ocean requires accurate assignment of water mass sources.

To distinguish individual sources of salinity (and other geochemical constituents), samples in the Arctic Ocean are typically decomposed into four components: (1) meteoric water, (2) sea ice melt/formation, (3) Pacific-derived seawater, and (4) Atlantic-derived seawater. Ideally, conservative tracers with characteristics unique to each water type are applied in a linear mixing model in order to calculate the relative influence of individual sources given a sample’s composition. Oxygen isotopes (or sometimes total alkalinity) and salinity have been applied and are useful in distinguishing between meteoric, sea ice, and seawater components. However, a definitively conservative tracer for Atlantic and

Pacific-derived seawater has not yet been described.

Up to now, researchers have used nutrient-based tracers since nutrient concentrations are high in the Pacific inflow to the Arctic Ocean and low in Atlantic inflow (e.g., Ekwurzel et al., 2001; Jones et al., 1998; Newton et al., 2013, Wilson and

Wallace, 1990). To overcome the non-conservative nature of nutrient concentrations, nitrogen and phosphorus can be used simultaneously by combining them in their respective Redfield Ratios. The results have been useful qualitatively by identifying areas of Pacific, Atlantic and shelf-sea influence; however, they are of questionable quantitative value as significant errors appear as a result of processes that impart non-

Redfieldian ratios on the nutrients in the dissolved phase including nitrification, denitrification, phosphate dissolution, and exchange of oxygen at the ocean surface (e.g.,

Alkire et al., 2015, 2019; Bauch et al., 2011; Ekwurzel et al., 2001).

One commonly used nutrient tracer exploits an offset between the Atlantic

NO3:PO4 trend and the Pacific NO3:PO4 trend: Pacific waters are high in PO4 relative to

Atlantic waters (Figure 2.1). The distance a sample falls between the two trendlines

characterizes the amount of Pacific or Atlantic-derived water in the sample. In order to be quantitative, NO3 and PO4 must covary in a constant N:P ratio, which is not the case. In practice, the slopes of Atlantic water and Pacific water in N:P space differ significantly, but the impact of nitrogen fixation, nitrification and denitrification, especially over the

Arctic shelf seas, is likely large and uncontrolled in this tracer. In addition, the method requires knowledge of the end members, which are highly varied in space and may change with time (e.g., Alkire et al., 2019; Cooper et al., 1999).

Figure 2.1 Nitrate-Phosphate relationship in the western Arctic Ocean.

Nitrate-Phosphate trends for Atlantic-derived and Pacific-derived seawater. This study is compared to Jones et al., 1998 to demonstrate variability in the relationship at different times. Since the slopes of ‘Atlantic’ and ‘Pacific’ waters differ (Figure 2.1), the fraction of Pacific-derived water (fPac) in a sample depends on which connecting line one moves between the two end-members. It should be noted that results from using different

18 tracers (e.g., alkalinity instead of δ O-H2O) have yielded differences in Pacific water fraction greater than 10% (Ekwurzel et al., 2001 vs. Jones et al., 1998, reporting on the

1994 Arctic Ocean Survey). Alkire et al. (2015) documented differences of up to 60% in the fPac between nutrient tracer methods when the same end-members were applied.

Additionally, by varying endmembers within the method, there was a 13% median standard deviation for fPac. Generally, authors report errors between 10 % and 14 %, derived from endmember variation tests and perturbations in the slopes of the NO3:PO4 trend (Alkire et al., 2015; Bauch et al., 2011; Yamamoto-Kawai et al., 2008) even without including the uncertainty generated by biologically mediated modification of nutrient ratios. Uncertainty in the meteoric and sea ice fractions is consistently much less

(i.e. < 1% median standard deviation; Alkire et al., 2015).

Ultimately, the uncertainty in nutrient processes and internal cycling in the Arctic

Ocean limits the results of the nutrient tracer method to a qualitative identification of

Pacific water. These nutrient-based deconvolutions are able to report relative shifts to greater or lesser contribution, but the absolute level of that contribution cannot be quantitatively trusted. Because of the uncertainties, it is prudent to validate the outcome of the nutrient method with other conservative tracers.

Previous work has suggested that dissolved gallium (Ga) may be a useful tracer for deconvolving waters with sources in the Atlantic and Pacific Oceans (McAlister &

Orians, 2015). If this is the case, then the concentration of Ga could alleviate the need to use nutrient relationships to deconvolve water masses.

Here, we present the first trans-Arctic Ocean Ga sections describing data in the

Makarov and Canada basins of the Western Arctic Ocean. Our data agree with McAlister

& Orians' (2015) supposition that Ga could be a useful tracer in the Arctic Ocean.

Pacific-derived waters have low Ga concentrations (5 – 10 pmol kg-1) versus Atlantic-

derived waters (28 pmol kg-1). This contrast, in addition to Ga’s low reactivity relative

(residence time on the order of decades) to the residence times of shallow waters in the

Arctic (~ 10 years), is what drives Ga’s potential use as a tracer (Schlosser et al., 1999;

Shiller, 1998). We assessed the utility of Ga as a tracer for Pacific and Atlantic water by employing a linear end-member mixing model and comparing these results to the nutrient tracer method described in Newton et al. (2013). We applied a 1-D advection-diffusion model to assess the resulting vertical distributions of the Pacific fraction and compare them to expectations based on vertical gradients in temperature and salinity.

2.3 Methods

2.3.1 Section Description

We present and discuss freshwater components calculated from measurements of stable isotopes of water, salinity, gallium and nutrients along the 2015 U.S. Arctic Ocean

GEOTRACES section (GN01) in the upper 500 m of the water column. Samples were collected at 66 hydrocast stations (22 GEOTRACES and 44 CLIVAR repeat hydrography program) along two transects extending from the continental shelf to the North Pole roughly along longitudes 180°W and 150°W (Figure 2.2a). We define our Pacific endmembers using shelf data from the Bering Strait, as all Pacific waters are constricted through this location.

The western Arctic Ocean water column is characterized by the Polar Mixed

Layer (PML), the Pacific Halocline (PH), and Atlantic Halocline (AH) and Atlantic

Water (AW) (Figure 2.2b). The PML is composed largely of sea ice meltwater and meteoric water. Below the PML, cold, relatively fresh and nutrient rich Pacific water (S ~

32.5) sits above warm, saline AW (S ~ 34.8). Brine rejection can influence the water

column, although the exact influence is elusive (e.g., Bauch et al., 2011). Traditional brine rejection is the formation and subsequent sinking of hypersaline waters that are derived from ice formation. Salt rejection during sea ice formation can result in hypersaline brine channels in sea ice, which may be seasonally rejected to the surface ocean. Previous literature demonstrates the highest influence of brines above and below the PH (Bauch et al., 2011; Yamamoto-Kawai, 2005); although there is discussion that brines may drive deep water geochemical signals (e.g., Jones et al., 1995).

Figure 2.2 Regional and hydrographic description of the study area.

(A) Map of sampled stations with hydrograhic data (CLIVAR-only stations) as black triangles or hydrographic and trace metal data

(GEOTRACES stations) as blue circles. Geographic features are numbered and major surface circulation features are dictated on the map.

(B) Cross section of the Arctic Ocean (indicated on the map to the right) detailed with the rough distribution of major water masses.

Figure modified from Whitmore et al. (2019). 2.3.2 Ancillary Hydrographic Data

Ancillary hydrographic data for GN01 were produced by the Scripps Institution of Oceanography’s Oceanographic Data Facility (ODF), Shipboard Technical Support 18

group. Salinity samples collected from the rosette were analyzed onboard using a

Guildline 8400 salinometer (accuracy ca. ± 0.002). Nutrient concentrations were measured aboard the USCGC Healy within 1 – 4 hours from collection. Nutrient analyses were performed on a Seal Analytical continuous-flow Auto-Analyzer 3 (AA3) following the WOCE (World Ocean Circulation Experiment) standard techniques (Gordon et al.,

1993; Hager et al., 1972). Analytical methods used in 2015 were in accordance with the

GO-SHIP repeat hydrography manual (Hood et al., 2010). All hydrographic data are publicly available online through CLIVAR and Carbon Hydrographic Data Office

(CCHDO; Kadko et al., 2015) and through the Biological and Chemical Oceanographic

Data Management Office (BCO-DMO; Cutter et al., 2019a, 2019b).

18 2.3.3 δ O-H2O Analytical Methods

18 16 Oxygen isotope ratios (H2 O/H2 O) were measured at Columbia University’s

Lamont Doherty Earth Observatory using a Picarro L2130-i Cavity Ring-Down

Spectroscopy (C.R.D.S.) analyzer following Walker et al. (2016). Oxygen isotope ratios

18 18 16 are reported as δ O, i.e. the per mil deviation of the H2 O/H2 O ratio from that of

Vienna Standard Mean Ocean Water (VSMOW-2) (Craig, 1961; Gat and Gonfiantini,

1981). Analytical precision is approximately ± 0.027 per mil (Pasqualini et al., 2017).

Stable isotopic analyses were performed on samples covering the entire water column at

GEOTRACES stations and the upper 500 m at U.S. Repeat Hydrography stations. This study only assesses the upper 500 m at all stations.

2.3.4 Ga Sample Collection and Analytical Methods

Following GEOTRACES standards (Cutter et al., 2014), dissolved Ga (referred to as ‘Ga’ throughout the chapter) samples were filtered (0.2 µm) into acid cleaned high-

density polyethylene bottles from GO-FLO bottles mounted on a trace metal clean rosette and subsequently acidified to 0.024 M HCl. Acidified seawater samples were prepared for analysis on an inductively coupled plasma mass spectrometer (ICP-MS) at the

University of Southern Mississippi (Center for Trace Analysis, ThermoFisher Element-

XR) by preconcentrating with magnesium hydroxide coprecipitation coupled with an isotope dilution approach (Ho et al., 2019; Shiller & Bairamadgi, 2006). Briefly, samples

(7 mL seawater) were spiked (spike volume variable with seawater Ga concentration) with an enriched isotope spike of 99.8% 71Ga (Oak Ridge National Laboratories).

Aqueous ammonium was added to the solution and the precipitate was collected and rinsed three times with a solution of ~0.1% NH4OH to remove barium, which can analytically interfere with 69Ga as doubly charged 138Ba. The precipitate was dissolved in

3% ultrapure nitric acid (0.47 M HNO3; Seastar Chemicals, Baseline) and analyzed on the ICP-MS for 69Ga, 71Ga, and 138Ba. Samples were introduced to the ICP-MS through a

PC3 Spray chamber (Elemental Scientific, Inc) and analyzed in low resolution.

Data are quality controlled with multiple approaches. Atlantic Ocean bulk surface and deep water samples (GS and GD, respectively) distributed from the 2008

GEOTRACES Intercalibration Cruise were analyzed during each run, the average concentration and precision from these repeat measurements is presented in Table 2.1.

Recovery was estimated for each run by spike enrichment of an in-house synthetic standard solution (Table 2.1). Intercalibration efforts were made in accordance to

GEOTRACES protocols. Ga data are available at BCO-DMO (Shiller, 2019).

Table 2.1 Gallium reproducibility assessment

Ga pmol/kg SD N GS 41.9 1.1 9 GD 32.8 1.4 13 % Recovery 96.8 5.3 11

2.3.5 Linear Mixing Model

The upper 500 m of the Arctic Ocean can be divided into four primary components: (1) Meteoric Water (fmet), (2) Sea-ice melt (fice), (3) Pacific-derived waters

(fPac), and (4) Atlantic-derived waters (fAtl). A linear mixing model approach was used to calculate the fraction of each water type in the upper 500 m of the water column (Eqns. 1

– 5). We applied salinity (S), δ18O, and a N:P-based tracer (which we call the Arctic

Nitrate-Phosphate, ANP) or Ga endmember values to calculate the fractions of each sample. The ANP value was calculated per Newton et al. (2013).

푓푚푒푡 + 푓𝑖푐푒 + 푓푃푎푐 + 푓퐴푡푙 = 1 (Eq. 2.1)

푓푚푒푡(푆푚푒푡) + 푓푆퐼푀(푆𝑖푐푒) + 푓푃푎푐(푆푃푎푐) + 푓퐴푡푙(푆퐴푡푙) = 푆표푏푠 (Eq. 2.2)

18 18 18 18 18 푓푚푒푡(δ 푂푚푒푡) + 푓푆퐼푀(δ 푂𝑖푐푒) + 푓푃푎푐(δ 푂푃푎푐) + 푓퐴푡푙(δ 푂퐴푡푙) = δ 푂표푏푠 (Eq 2.3)

푓푚푒푡(퐴푁푃푚푒푡) + 푓푆퐼푀(퐴푁푃𝑖푐푒) + 푓푃푎푐(퐴푁푃푃푎푐) + 푓퐴푡푙(퐴푁푃퐴푡푙) = 퐴푁푃표푏푠 (Eq. 2.4)

푓푚푒푡([퐺푎]푚푒푡) + 푓𝑖푐푒([퐺푎]𝑖푐푒) + 푓푃푎푐([퐺푎]푃푎푐) + 푓퐴푡푙([퐺푎]퐴푡푙) = [퐺푎]표푏푠 (Eq. 2.5)

The system is constrained to fractions between 0 and 1 for all parameters

(meteoric, sea ice melt, Atlantic, or Pacific), excepting fice which is allowed to be between -1 and 1. A negative fice reflects the extraction of freshwater to form ice and is considered evidence of brines formed with salt rejected from newly-formed sea ice.

When available, endmember values were determined from our data set; however, several values were taken from previous literature (Table 2.2). Our study was not able to 21

directly measure the meteoric endmember of Ga. Based on the concentrations in the Polar

Mixed Layer (roughly the upper 50 m of the water column), we determined the concentration of Ga in rivers would be low relative to the deeper waters. McAlister and

Orians (2015) indicated the riverine inputs from the Mackenzie River may be higher than the Pacific-derived seawater and, thus, higher than our suggested endmember. With the evidence in the central Arctic, we think a low Ga concentration is appropriate. We performed a sensitivity analysis by perturbing the meteoric Ga endmember between 0 and

14 pmol/kg; the variance in fmet was < 0.01%; fice varied within 1% and fPac and fAtl within

13%. This variability is similar to sensitivity tests conducted on nutrient-method endmember choices (Alkire et al., 2015). The uncertainty associated with the meteoric Ga endmember may be decreased with targeted studies assessing endmember geochemical properties.

Table 2.2 Endmember parameter values.

Salinity Water Mass δ18O [‰]† Ga [pmol kg-1] Arctic N:P (a) [g kg-1]† Atlantic Water 34.92 0.3 28 0 Pacific Water 32.50 -1.1 5 1 Meteoric Water 0 -20 2.3 0 Sea-Ice Meltwater 4 Surf. + 2.6 ‰ 2.3 Surface

(a) Pacific Water: [NO3] = 14 × [PO4] – 11; Atlantic Water: [NO3] = 17 × [PO4] – 2. † (b) (Newton et al., 2013 and references therein).

2.4 Results and Discussion

2.4.1 Gallium Distribution

The gallium distribution at isopycnal densities associated with the Pacific halocline indicate little modification of Ga (Figure 2.3a, b). The Pacific halocline (Ga =

7.4 ± 1.2 pmol kg-1; n = 41) is equivalent in concentration to the incoming Bering Strait waters (Ga 7.7 ± 1.1 pmol kg-1; n = 4). Notably, surface waters (< 100 m) at the Bering

Sea slope (i.e., Pacific-derived waters that have not had exposure to the shelf) are 4.4 ±

1.4 pmol kg-1; this indicates minor transformation of Pacific water as it transits northward. Other deviations from the mean Bering Strait waters occur at shelf station 66, where the concentration was 2.6 – 3.9 pmol kg-1 (Figure 2.3c). Additionally, shelf station

3 had low surface Ga (Ga = 4.4 pmol kg-1; z = 1 m) and higher concentrations at depth

(Ga = 9.7 ± 0.5 pmol kg-1; z > 15 m). Aside from station 3, shelf stations were homogeneous with depth (within the error of our measurement) (Figure 3c). The relative stability in Ga on the shelves and in the Pacific halocline is ideal in comparison to nutrient concentrations, which are impacted during transit over the shelves. The exchange of shelf nutrients on waters passing over the shelves could mean that even Atlantic- sourced waters passing over Eurasian shelves may develop “Pacific-like” nutrient signals.

Figure 2.3 Shelf distribution of dissolved gallium.

Shelf profiles of dissolved Ga (pmol kg-1), station locations and names are indicated on the map and by the legend. Dotted lines are indicative of the assigned Pacific endmember (5 pmol kg-1) and Atlantic endmember (28 pmol kg-1).

In the basin below the Pacific halocline, Ga increases into the Atlantic water layer

(Ga = 28 - 40 pmol kg-1; Figure 2.4a). Gallium concentration in the Western Arctic

Ocean below 500 m increases to roughly 30 pmol kg-1; however, the one station sampled in the Eurasian basin has much higher Ga below the depth of the Lomonosov ridge (Ga ~

40 pmol kg-1; Figure 2.4a).

Figure 2.4 Arctic Ocean dissolved gallium profiles.

(a) GN01 section (this study), different regions of the Arctic Ocean and shelf seas are identified by different color data points.

(b) Data from the GIPY14 cruise conducted during the international polar year and published in McAlister & Orians (2015). Colored data points indicate regions of the sections as indicated in the legend to the right and the map to the left. Prior studies agree that the deep ocean Ga distribution reflects the relatively unreactive behavior of this element (Shiller, 1998; Shiller & Bairamadgi, 2006).

However, McAlister & Orians (2012) demonstrated that Ga exhibited scavenged-type behavior in the Columbia River Plume. Our shelf section had a range of values from 2.6 –

10.0 pmol kg-1. This variability observed in Ga at our shelf stations may indicate a minor influence of particle scavenging over the continental shelves. Greater resolution over the

shelves is required to determine how significant the variability is in setting the basin Ga distribution. Recent work by Ho et al. (2019) described Ga and Al distributions in the equatorial Pacific; they concluded that the predominant non-conserved sources of Ga to their section were aeolian input and sediment resuspension. As the Arctic Ocean becomes more ice-free, the relative roles of these sources compared to oceanic inputs from

Atlantic and Pacific waters may become important. Kipp et al. (2018) suggested that an increasingly ice-free Arctic Ocean would result in potentially greater contributions from shelf sediments as exposure to wind effects increases.

Published open ocean dissolved Ga profiles and sections have been unable to discern an influence of scavenging; generally, open ocean profiles – including our upper water column western Arctic Ocean profiles – suggest little influence of scavenging-type behavior (i.e., particle reactivity or biological influence) on Ga concentration (Ho et al.,

2019; Shiller & Bairamadgi, 2006). For some elements, such as radium (Kipp et al.,

2018), the shelf can impart a large elemental signal and the transpolar drift (TPD) has been indicated as an important shelf-basin transport mechanism. Several GN01 stations are in TPD influenced waters (i.e., > 85°N; Figure 2.2a), but our Ga section does not indicate any specific shelf signal from these waters. Importantly, while our section enters

TPD-influenced surface waters, a section passing fully through TPD-derived waters would help assess the influence they have on the water column Ga distribution.

Nitrate and phosphate are higher in the Pacific-halocline than in the Bering Strait endmember because the shelves substantially modify nutrients. Furthermore, due to modifications over the shelf, the N:P slope is different between shelf-influenced Pacific- derived waters and Pacific waters before they interact with the shelves (e.g., Cooper et

al., 1999). Because of shelf modifications, the Pacific halocline is often discussed as

“shelf influenced” waters rather than truly Pacific-origin waters. In considering all the shelf stations sampled in the Bering and Chukchi Seas, the relative standard deviation

(RSD) of Ga is 31%, of NO3 is 140% and of PO4 is 61%. Our Ga distribution is less variable than nutrient concentrations, which supports its application as a conservative tracer.

2.4.2 Linear Mixing Model Results

In comparing the Ga-derived and ANP-derived water mass fractions, fmet and fice are roughly equivalent (Figure 2.5c, d). The two methods have minor discrepancies in fmet and fice, less than 0.05 and 0.02 respectively (where the discrepancy is determined by subtracting the fraction determined using the Ga-method from the ANP-method; fmet median discrepancy < 0.01 and fice median discrepancy < 0.01) (Figure 2.5 a, b). This is not surprising since these two fractions are largely determined by the S and δ18O distributions.

For fPac and fAtl, however, while both methods produce qualitatively similar spatial distributions, (Figure 2.6 e - h), quantitatively there can be a discrepancy of up to 0.66 in fPac or fAtl between the two methods. Between the nutrient and gallium methods, the median discrepancies of the fPac and fAtl are 0.11 and 0.10, respectively (Figure 2.6 i,j).

Figure 2.5 Comparison of Ga-derived water mass fraction and ANP-derived fractions.

(a) Fraction of Atlatnic (fAtl)

(b) Fraction of Pacific (fPac)

(d) Fraction of sea ice melt or formation (fice) The fractions of Pacific and Atlantic waters are roughly the inverse of each other, so in this discussion we focus on fPac (replicated figures for all other fractions are presented in the supplemental materials: Appendix Figures A1 – A3). In terms of fPac, the

ANP-derived water mass fraction under-predicted fPac relative to the Ga method. The

greatest differences between the two methods arise at depths between the core of the

Pacific halocline and the core of the Atlantic layer for stations north of 85°N (i.e., at stations north of the Pacific-Atlantic front).

The methods carry uncertainties in both the endmembers and the sample data.

Sample analyses generally have lower error than the endmember uncertainties; however,

Ga sample errors are often higher than those for nitrate or phosphate. To determine the sensitivity of the method, the endmembers for ANP and Ga were perturbed to their maximum and minimum expected. The sea ice and meteoric fractions consistently have low sensitivity to the endmember. However, the ANP and Ga method vary by about 10% for the Atlantic fraction. The ANP has lower variance within the Pacific fraction (5%) and the Ga method has greater variance within the Pacific fraction (~40%). The high sensitivity in the Ga method can be reduced by refining the endmember estimate.

Figure 2.6 Sections of ANP-based and Ga-based water mass deconvolutions.

The left column is the Makarov Basin transect (180°W) and the right column is the Canada Basin transect (150°W); the continued bathymetry, from 500 dbar to 4500 dbar, for each section is indicated in the bottom panel. (a, b) ANP distribution. (c, d) Dissolved Ga concentration (pmol kg-1). (e, f) Pacific fraction determined by applying the ANP tracer. (g, h) Pacific fraction determined by the Ga

tracer. (i, j) Distribution of the difference between the ANP and Ga-derived Pacific fractions. Contour lines are isopycnal surfaces (σθ) at 26.46 kg m-3 (core of the Pacific halocline; shallowest), 27.51 kg m-3 (core of the Lower halocline; middle), and 27.93 kg m-3 (core of the Fram Strait Branch Water; deepest). The profiles of the tracers (i.e., Ga or ANP), and the resulting fractions of Pacific and Atlantic water, suggest two contrasting mixing modes (Figure 2.7). The ANP-derived fPac transitions from high Pacific water to nearly zero Pacific water across a very narrow depth range, a feature indicative of minimal mixing between the water masses. The Ga- derived fPac transition between high and low Pacific content more gradually with depth, suggesting greater mixing between Atlantic and Pacific water-types than the ANP profile exhibits. Thus, understanding which tracer more reliably illustrates the relative distributions of the Atlantic and Pacific waters has implications relating to the exchange of materials and heat between the water types.

Figure 2.7 Profiles of physical properties and Pacific fraction (ANP and Ga-derived).

Profiles of the Pacific fraction from the (a) ANP and (b) Ga method. (c) Bottle salinity data. (d) CTD Temperature profile (at bottle depths). Solid lines are mean profiles determined using Java Ocean Atlas).

Previous literature has presented multiple approaches to determining the

“nutrient-tracer” and has outlined the variability between approaches (e.g., Alkire et al.,

2015). We chose not to replicate this in depth, as the uncertainties between different nutrient methods are well described. To discuss how different nutrient approaches may compare to the Ga-tracer method, we reproduced water mass deconvolution using the approach described in Yamamoto-Kawai et al. (2008) and compared it to the results using the ANP approach described above (originally detailed in Newton et al., 2013). The

Yamamoto-Kawai et al. (2008) approach differs from that of Newton et al. (2013) in the choice of slopes used to define the ‘Atlantic’ and ‘Pacific’ end members in N:P space.

The choice of slopes does make a quantitative difference in the division between Atlantic and Pacific water masses, but the difference is small relative to other sources of uncertainty and is not important to the current discussion. We note that in comparing the nutrient approaches (Yamamoto-Kawai et al., 2008 vs Newton et al., 2013) distributions are qualitatively similar – fPac are highest in the same regions, but still there are deviations of up to 0.6 in fPac between methods. Again, these differences have been described elsewhere and the important comparison in our study is that, regardless of nutrient methodology, the Ga approach systematically predicts a greater fPac in the transition zone than the nutrient approaches (Appendix Figure A4—A5).

2.4.3 Inventories

Geochemists and physical oceanographers have visualized the distribution of freshwater in the Arctic Ocean by summing the freshwater anomaly vertically in the water column, and showing a total ‘height’ of freshwater at each location (e.g., Ekwurzel et al., 2001; Serreze et al., 2006). The freshwater distribution is sufficient as a snapshot

of the current state of buoyancy forcing. However, to understand the mechanisms underlying this state, and to make reasonable forward projections as the Arctic Ocean evolves in time, it is necessary to estimate the sources and sinks of buoyancy—which starts with knowing the spatial distribution of the main sources of freshwater: meteoric water, sea-ice melt, and Pacific inflow. To visualize the differences between the ANP and Ga approaches, we summed the Pacific-sourced fraction, as estimated from each method, and plotted the totals as a function of latitude (Figure 2.8).

Figure 2.8 Pacific water inventory (m).

(a) Ga (blue) and ANP (black) inventories at stations in the western Arctic basin. (b) Difference in the Pacific water inventory as calculated from each method (ANP or Ga). The freshwater contribution from the Pacific fraction is greater in the Ga-derived approach. This increase in freshwater from fPac (relative to the ANP method) is balanced by a decrease in the meteoric freshwater component (Figure 2.5c). Near the continental slope (~72°N) the two methods nearly agree, while the difference between them increases approximately linearly to a maximum of about 7 m, at the northern end of the sections

(Figure 2.8b). To understand this pattern, recall that Ga is carried by the Atlantic fraction

and that the ANP value of Pacific water is set, by definition, by the values at the southern end of the transect, where shelf-modified Pacific inflow detrains from the Chukchi Shelf.

To try and understand whether the Ga method or ANP method is producing a more realistic estimate of the fPac distribution, we start first with a logical argument and then, in the next section, proceed to a modeling argument. In our logical approach, we will consider the implications of the ANP tracer being the more conservative parameter and then the implication of the Ga distribution being more conservative parameter.

First, we assume that the ANP tracer is relatively conservative and accurately estimates the Pacific water fraction throughout the transect: this scenario implies that Ga in the Atlantic water is most accurate at the southern end of the transect, where the

Atlantic water is farthest from its source, and most inaccurate at the northern end, closest to its source. This indicates that Ga underwent extreme scavenging between Fram Strait and the Lomonosov Ridge, and that Ga was added back into the water column between the ridge and the Chukchi shelf. Given the degree of scavenging required, we presume this is an unlikely scenario.

On the other hand, we can assume that the Ga tracer is conservative. In this scenario, nutrient addition/removal must be occurring. At the Chukchi continental slope, the two methods are in approximate agreement. As the Pacific-sourced water spreads northward, the increasing difference between the Ga and nutrient tracer implies that nitrogen and phosphorus are added to the system at non-Redfieldian ratios, with greater nitrogen input relative to phosphorus (i.e., N:P ratios greater than 16:1). McAlister &

Orians (2015) posited that differences between N* (where N* = NO3 – 13.6 × PO4 + 11) and Ga distributions may provide evidence of N-fixation at the surface; indeed, there is a

growing body of literature providing evidence of diazotrophy in the Arctic Ocean (Blais et al., 2012; Fernández-Méndez et al., 2016; Harding et al., 2018; Sipler et al., 2017).

However, the greatest difference in the ANP and Ga water mass fractions occurs at depths between 150 and 300 m where the N* and Ga profiles are similar. It is unclear if

N-fixation at the surface translates to the deeper waters in which we observe the greatest difference or if there is other in-situ nutrient cycling that should be considered.

Additionally, besides the assumption and requirement of Redfieldian ratios, there are inherent uncertainties in the ANP and other nutrient methods that could drive incorrect estimations of the Pacific and Atlantic fractions. Specifically, there is no unbiased or a priori way to draw the mixing line between Atlantic and Pacific N:P regressions. Thus, the nutrient method has an inherent uncertainty in fPac and fAtl, since different mixing lines lead to differing estimates of those fractions. Importantly, Ga scavenging would be unusual away from sedimentary interfaces and there is no independent evidence of significant Ga addition over the Amerasian Basin. Thus, through this discussion, we favor the Ga-supported fPac estimates.

2.4.4 Vertical Advection-Diffusion Model

Since ANP- and Ga-based deconvolutions imply different degrees of diapycnal mixing, it is instructive to use each to constrain the vertical mixing coefficient in a one- dimensional, vertical mixing model. The model is a 2-parameter advection-diffusion model with a small, background upwelling, constant vertical diffusion and fixed boundary conditions above (in the core of the Pacific halocline, ~100 m) and below (in the core of the Atlantic Layer ~300 m). The essential limitation of the model (its main divergence from the real Arctic Ocean situation) is that the vertical gradients are assumed

to be a result of purely vertical processes. We know that lateral spreading along isopycnals is important in ventilating the upper and intermediate layers of the Arctic

Ocean (e.g., Steele & Boyd, 1998). Therefore, we present these results with the understanding that the model’s vertical eddy diffusion subsumes the results of a combination of three-dimensional processes.

Away from the continental shelves and slopes, over the deep Arctic Ocean the shapes of property profiles fall into two types, those north and those south of 85°N. We therefore modeled the northern and southern stations along the GN01 track separately.

Gallium data were too sparse between the Pacific halocline and Atlantic layer to capture inflection points at each station, so we gridded the data vertically and created mean profiles for those two groups of stations (Figure 2.7). The model ran with a simple forward Euler numerical integration for 10 years with a 12-hour time step, with upper and lower boundary values held fixed at the mean observed values throughout the run. The upwelling rate for all runs was fixed at 0.7 × 10−7 m s−1, but we adjusted the model so that the median of the simulated values was always at the same depth as the median of the observations. The model was run iteratively with vertical eddy diffusivity

−5 −5 −6 −6 −7 2 −1 coefficients (kz) of 1 × 10 , 5 × 10 , 1 × 10 , 5 × 10 and 1 × 10 m s . We compared the model results to the mean profile of the observations, assessing the goodness of fit by applying a linear regression and then determining the R-squared

(RSQ), the squared residuals (SSR), and the root mean squared error (RMSE) (Appendix

Table A1). We applied this process to temperature and salinity from the CTD traces and to the Pacific water fractions as determined by the Ga- and ANP-based decomposition techniques. For temperature and salinity, the model results indicate a vertical diffusivity

coefficient of roughly 5 × 10−6 m2 s−1 for stations both north and south of 85°N.

Previous studies have indicated diffusivity coefficients of roughly 1 × 10−6 m2 s−1 (e.g.,

Nguyen et al., 2009; Wallace et al., 1987; Zhang & Steele, 2007); we therefore consider our results reasonable for the purposes of our study.

South of 85°N both fPac_ANP and fPac_Ga profiles yielded eddy diffusivity coefficients equivalent to those determined from salinity and temperature. As expected from visual inspection of the Pacific component profiles (Figure 2.7), fPac_ANP profiles were matched best by lower diffusivity coefficients north of 85°N than fPac_Ga. North of

85°N, the diffusivity coefficients derived from the ANP-method were lower than the best fits for temperature or salinity, which aligned closely with the Ga-derived coefficients.

Thus, while the Pacific water profiles from both methods are within the range of plausible diapycnal rates, the Ga method is a better match, overall, to those rates implied from the basic physical parameters, salinity and heat.

We note, in addition, that one of the differences between the two methods is that the gallium-based decomposition extends the Pacific-influenced upper halocline farther toward the Eurasian side of the Arctic Ocean than the ANP-based analysis does. This method’s contrast is even more severe when comparing with a decomposition of the same samples using the NO parameter (where NO = 9 × NO3 + O2; Alkire et al., 2019). Thus, even along isopycnal surfaces, Ga-based distributions imply a more gradual transition from Pacific to Atlantic source waters. That is, overall Ga implies a more dispersive mixing regime over the central Arctic Ocean basins than the nutrient-based deconvolution.

2.5 Conclusions

The use of a nutrient tracer to deconvolve Pacific and Atlantic derived waters in the Arctic Ocean is problematic. Nutrient tracer methods have significant disagreements between specific approaches and, when the same approach is applied, slight changes in the Redfield relationship between years and cruises increases uncertainty in the determined Pacific inventories (e.g., Alkire et al., 2019; Alkire et al., 2015, Yamamoto-

Kawai et al., 2008).

The application of Ga as a tracer exposes uncertainty associated with the assumed quasi-conservative behavior of nutrient tracers in an Arctic Ocean linear endmember mixing model for Atlantic and Pacific waters. The Ga distribution indicates minor variability resulting from shelf exposure, especially when compared to nutrients (e.g.,

Cooper et al., 1999). In applying Ga as a conservative-type tracer in place of nutrients, the resulting water-source distributions are effectively equivalent for fraction of meteoric water and fraction of sea ice melt. However, the Ga method often predicts a greater fPac

(and lower fraction of Atlantic waters) relative to historically used nutrient methods.

The spatial distribution of fPac_Ga – fPac_ANP reveals that the greatest difference of fPac occurs in the northernmost region of the western Arctic Ocean (near the North Pole).

This regional inconsistency is also observed in our simplified vertical advection-diffusion model. In assessing vertical mixing processes, we find that the Ga method is more reflective of the salinity and temperature distributions for regions north of the Pacific-

Atlantic front. For regions south of the Pacific-Atlantic front, the two methods have slightly better agreement.

Our study has demonstrated that relative to the nutrient approach, Ga-derived water mass deconvolution implies more dispersive mixing between the Atlantic and

Pacific water types. The Ga method thus implicates the potential for greater heat transfer and nutrients between the water masses than the ANP method. In assuming that the Ga tracer is accurate, this implies an addition of nitrate or removal of phosphate independent of Redfieldian processes.

Future studies should include more samples in the Eurasian basin which would enable the assessment of the transpolar drift influence and of Eurasian shelf influences, since the GN01 transect only had one station in the Eurasian Basin (and that station was heavily influenced by TPD surface waters). Additionally, a multiple tracer technique (i.e., using an overdetermined system) integrating information from the suite of tracers that programs like GEOTRACES have generated may be a useful tool.

2.6 Acknowledgements

This research was supported by the National Science Foundation [OCE-1435312

(AMS), OCE-1436666 (RN)]. Thank you to the USCGC Healy crew and the scientific leadership—Bill Landing, Dave Kadko, and Greg Cutter—for their support in a successful expedition. Further thanks are extended to Melissa Gilbert for ICP-MS support at the University of Southern Mississippi’s Center for Trace Analysis. Data used in this study are available at the Biological and Chemical Oceanography Data Management

Office (DOI: 10.1575/1912/bco-dmo.772645.1) and the EarthChem Library (DOI:

10.1594/IEDA/100633).

2.7 References

Alkire, M. B., Morison, J., & Andersen, R. (2015). Variability in the meteoric water, sea- ice melt, and Pacific water contributions to the central Arctic Ocean, 2000-2014. Journal of Geophysical Research: Oceans, 120(3), 1573–1598. https://doi.org/10.1002/2014JC010023 Alkire, M. B., Rember, R., & Polyakov, I. (2019). Discrepancy in the Identification of the Atlantic/Pacific Front in the Central Arctic Ocean: NO Versus Nutrient Relationships. Geophysical Research Letters, 46(7). https://doi.org/10.1029/2018GL081837 Bauch, D., van der Loeff, M. R., Andersen, N., Torres-Valdes, S., Bakker, K., & Abrahamsen, E. P. (2011). Origin of freshwater and polynya water in the Arctic Ocean halocline in summer 2007. Progress in Oceanography, 91(4), 482–495. https://doi.org/10.1016/j.pocean.2011.07.017 Blais, M., Tremblay, J.-É., Jungblut, A. D., Gagnon, J., Martin, J., Thaler, M., & Lovejoy, C. (2012). Nitrogen fixation and identification of potential diazotrophs in the Canadian Arctic: N 2 fixation in the Arctic Ocean. Global Biogeochemical Cycles, 26(3). https://doi.org/10.1029/2011GB004096 Cooper, L. W., Cota, G. F., Pomeroy, L. R., Grebmeier, J. M., & Whitledge, T. E. (1999). Modification of NO, PO, and NO/PO during flow across the Bering and Chukchi shelves: Implications for use as Arctic water mass tracers. Journal of Geophysical Research: Oceans, 104(C4), 7827–7836. https://doi.org/10.1029/1999JC900010 Craig, H. (1961). Isotopic Variations in Meteoric Waters. Science, New Series, 133(3465), 1702–1703. Cutter, G., Andersson, P. S., Codispoti, L. A., Croot, P., Francois, R., Lohan, M., et al. (2014). Sampling and Sample-handling Protocols for GEOTRACES Cruises, Version 2. http://www.geotraces.org/library-88/scientific- publications/reports/169-sampling-and-sample-handling-protocols-for-geotraces- cruises Cutter, G., Kadko, D., & Landing, W. (2019a). Bottle data from the CTD-ODF carousel on the GEOTRACES Arctic Section cruise (HLY1502) from August to October 2015 (U.S. GEOTRACES Arctic project). Biological and Chemical Oceanography Data Management Office (BCO-DMO)., Dataset version 2019-07- 29. https://doi.org/10.1575/1912/bco-dmo.646825.4 Cutter, G., Kadko, D., & Landing, W. (2019b). Bottle data from the GEOTRACES Clean Carousel sampling system (GTC) on the Arctic Section cruise (HLY1502) from August to October 2015 (U.S. GEOTRACES Arctic project). Dataset Version 2019-07-29. https://doi.org/10.1575/1912/bco-dmo.647259.4 Ekwurzel, B., Schlosser, P., Mortlock, R. A., Fairbanks, R. G., & Swift, J. H. (2001). River runoff, sea ice meltwater, and Pacific water distribution and mean residence times in the Arctic Ocean. Journal of Geophysical Research: Oceans, 106(C5), 9075–9092. https://doi.org/10.1029/1999JC000024 Fernández-Méndez, M., Turk-Kubo, K. A., Buttigieg, P. L., Rapp, J. Z., Krumpen, T., Zehr, J. P., & Boetius, A. (2016). Diazotroph Diversity in the Sea Ice, Melt

Ponds, and Surface Waters of the Eurasian Basin of the Central Arctic Ocean. Frontiers in Microbiology, 7, 1884. https://doi.org/10.3389/fmicb.2016.01884 Gat, J. R., & Gonfiantini, R. (Eds.). (1981). Stable isotope hydrology: deuterium and oxygen-18 in the water cycle. Vienna, Italy: IAEA. Gordon, L. I., Jennings, J. C., Ross, A. A., & Krest, J. M. (n.d.). A Suggested Protocol for Continuous Flow Automated Analysis of Seawater Nutrients (Phosphate, Nitrate, Nitrite and Silicic Acid) in the WOCE Hydrographic Program and the Joint Global Ocean Fluxes Study, 56. Guay, C. K., & Kenison Falkner, K. (1997). Barium as a tracer of Arctic halocline and river waters. Deep Sea Research Part II: Topical Studies in Oceanography, 44(8), 1543–1569. https://doi.org/10.1016/S0967-0645(97)00066-0 Hager, S. W., Atlas, E. L., Gordon, L. I., Mantyla, A. W., & Park, P. K. (1972). A comparison at sea of manual an autoanalyzer analyses of phosphate, nitrate, and silicate. Limnology and Oceanography, 17(6), 931–937. https://doi.org/10.4319/lo.1972.17.6.0931 Haine, T. W. N., Curry, B., Gerdes, R., Hansen, E., Karcher, M., Lee, C., et al. (2015). Arctic freshwater export: Status, mechanisms, and prospects. Global and Planetary Change, 125, 13–35. https://doi.org/10.1016/j.gloplacha.2014.11.013 Häkkinen, S. (1993). An Arctic source for the great salinity anomaly: A simulation of the Arctic ice-ocean system for 1955–1975. Journal of Geophysical Research, 98(C9), 16397. https://doi.org/10.1029/93JC01504 Harding, K., Turk-Kubo, K. A., Sipler, R. E., Mills, M. M., Bronk, D. A., & Zehr, J. P. (2018). Symbiotic unicellular cyanobacteria fix nitrogen in the Arctic Ocean. Proceedings of the National Academy of Sciences, 115(52), 13371–13375. https://doi.org/10.1073/pnas.1813658115 Ho, P., Resing, J. A., & Shiller, A. M. (2019). Processes controlling the distribution of dissolved Al and Ga along the U.S. GEOTRACES East Pacific Zonal Transect (GP16). Deep Sea Research Part I: Oceanographic Research Papers, 147, 128– 145. https://doi.org/10.1016/j.dsr.2019.04.009 Hood, E. M., Sabine, C. L., & Sloyan, B. M. (Eds.). (2010). GO-SHIP repeat hydrography manual: A collection of expert reports and guidelines. IOCCP Report Number 14, ICPO Publication Series Number 134. Retrieved from http://www.go-ship.org/HydroMan.html Hunkins, K., & Whitehead, J. A. (1992). Laboratory Simulation of Exchange Through Fram Strait. Journal of Geophysical Research, 97(C7), 11299–11321. https://doi.org/10.1029/92JC00735 Jones, E. P., & Anderson, L. G. (1986). On the origin of the chemical properties of the Arctic Ocean halocline. Journal of Geophysical Research, 91(C9), 10759. https://doi.org/10.1029/JC091iC09p10759 Jones, E. Peter, Anderson, L. G., & Swift, J. H. (1998). Distribution of Atlantic and Pacific waters in the upper Arctic Ocean: Implications for circulation. Geophysical Research Letters, 25(6), 765–768. https://doi.org/10.1029/98GL00464

Jones, E.P., Rudels, B., & Anderson, L. G. (1995). Deep waters of the Arctic Ocean: origins and circulation. Deep Sea Research Part I: Oceanographic Research Papers, 42(5), 737–760. https://doi.org/10.1016/0967-0637(95)00013-V Kadko, D., Smethie, W., Ho, D., Millero, F., Key, R., McNichol, A., et al. (2015). ARC01. 33HQ20150809_hy1.csv. CLIVAR and Carbon Hydrographic Data Office, La Jolla, CA, USA. https://doi.org/10.7942/C2MW25 Kipp, L. E., Charette, M. A., Moore, W. S., Henderson, P. B., & Rigor, I. G. (2018). Increased fluxes of shelf-derived materials to the central Arctic Ocean. Science Advances, 4(1), eaao1302. https://doi.org/10.1126/sciadv.aao1302 McAlister, J., & Orians, K. (2012). Calculation of river-seawater endmembers and differential trace metal scavenging in the Columbia River plume. Estuarine, Coastal and Shelf Science, 99, 31–41. https://doi.org/10.1016/j.ecss.2011.12.013 McAlister, J. A., & Orians, K. J. (2015). Dissolved gallium in the Beaufort Sea of the Western Arctic Ocean: A GEOTRACES cruise in the International Polar Year. Marine Chemistry, 177, 101–109. https://doi.org/10.1016/j.marchem.2015.05.007 Newton, R., Schlosser, P., Mortlock, R., Swift, J., & MacDonald, R. (2013). Canadian Basin freshwater sources and changes: Results from the 2005 Arctic Ocean Section: AOS 2005 freshwater sources and changes. Journal of Geophysical Research: Oceans, 118(4), 2133–2154. https://doi.org/10.1002/jgrc.20101 Nguyen, A. T., Menemenlis, D., & Kwok, R. (2009). Improved modeling of the Arctic halocline with a subgrid-scale brine rejection parameterization. Journal of Geophysical Research: Oceans, 114(C11). https://doi.org/10.1029/2008JC005121 Pasqualini, A., Schlosser, P., Newton, R., & Koffman, T. N. (2017). U.S. GEOTRACES Arctic Section Ocean Water Hydrogen and Oxygen Stable Isotope Analyses, (Version 1.0). http://dx.doi.org/10.1594/ieda/100633 Rudels, B. (2018). Arctic Ocean Circulation. In Reference Module in Earth Systems and Environmental Sciences. Elsevier. https://doi.org/10.1016/B978-0-12-409548- 9.11209-6 Schlosser, P., Bayer, R., Bönisch, G., Cooper, L. W., Ekwurzel, B., Jenkins, W. J., et al. (1999). Pathways and mean residence times of dissolved pollutants in the ocean derived from transient tracers and stable isotopes. Science of The Total Environment, 237–238, 15–30. https://doi.org/10.1016/S0048-9697(99)00121-7 Serreze, M. C., Barrett, A. P., Slater, A. G., Woodgate, R. A., Aagaard, K., Lammers, R. B., et al. (2006). The large-scale freshwater cycle of the Arctic. Journal of Geophysical Research, 111(C11). https://doi.org/10.1029/2005JC003424 Shaw, W. J., Stanton, T. P., McPhee, M. G., Morison, J. H., & Martinson, D. G. (2009). Role of the upper ocean in the energy budget of Arctic sea ice during SHEBA. Journal of Geophysical Research, 114(C6), C06012. https://doi.org/10.1029/2008JC004991 Shiller, A. (2019). Dissolved Ba, Cd, Cu, Ga, Mn, Ni, and V concentration data from the US GEOTRACES Arctic Expeditions (GN01, HLY1502) from August to October 2015. Biological and Chemical Oceanography Data Management Office (BCO- DMO), Dataset version 2019-07-09. https://doi.org/10.1575/1912/bco- dmo.772645.1

Shiller, A. M. (1998). Dissolved gallium in the Atlantic Ocean. Marine Chemistry, 61(1– 2), 87–99. https://doi.org/10.1016/S0304-4203(98)00009-7 Shiller, A. M., & Bairamadgi, G. R. (2006). Dissolved gallium in the northwest Pacific and the south and central Atlantic Oceans: Implications for aeolian Fe input and a reconsideration of profiles. Geochemistry, Geophysics, Geosystems, 7, Q08M09. https://doi.org/10.1029/2005GC001118 Sipler, R. E., Gong, D., Baer, S. E., Sanderson, M. P., Roberts, Q. N., Mulholland, M. R., & Bronk, D. A. (2017). Preliminary estimates of the contribution of Arctic nitrogen fixation to the global nitrogen budget: Arctic nitrogen fixation. Limnology and Oceanography Letters, 2(5), 159–166. https://doi.org/10.1002/lol2.10046 Steele, M., & Boyd, T. (1998). Retreat of the cold halocline layer in the Arctic Ocean. Journal of Geophysical Research: Oceans, 103(C5), 10419–10435. https://doi.org/10.1029/98JC00580 Torres-Valdés, S., Tsubouchi, T., Bacon, S., Naveira-Garabato, A. C., Sanders, R., McLaughlin, F. A., et al. (2013). Export of nutrients from the Arctic Ocean: Arctic Ocean nutrient exports. Journal of Geophysical Research: Oceans, 118(4), 1625–1644. https://doi.org/10.1002/jgrc.20063 Walker, S. A., Azetsu-Scott, K., Normandeau, C., Kelley, D. E., Friedrich, R., Newton, R., et al. (2016). Oxygen isotope measurements of seawater ( H2 18 O/ H2 16 O): A comparison of cavity ring-down spectroscopy (CRDS) and isotope ratio mass spectrometry (IRMS): CRDS oxygen isotope measurements in seawater. Limnology and Oceanography: Methods, 14(1), 31–38. https://doi.org/10.1002/lom3.10067 Wallace, D. W. R., Moore, R. M., & Jones, E. P. (1987). Ventilation of the Arctic Ocean cold halocline: rates of diapycnal and isopycnal transport, oxygen utilization and primary production inferred using chlorofluoromethane distributions. Deep Sea Research Part A. Oceanographic Research Papers, 34(12), 1957–1979. https://doi.org/10.1016/0198-0149(87)90093-8 Whitmore, L. M., Morton, P. L., Twining, B. S., & Shiller, A. M. (2019). Vanadium cycling in the Western Arctic Ocean is influenced by shelf-basin connectivity. Marine Chemistry, 216, 103701. https://doi.org/10.1016/j.marchem.2019.103701 Yamamoto-Kawai, M., McLaughlin, F. A., Carmack, E. C., Nishino, S., & Shimada, K. (2008). Freshwater budget of the Canada Basin, Arctic Ocean, from salinity, δ 18 O, and nutrients. Journal of Geophysical Research, 113(C1). https://doi.org/10.1029/2006JC003858 Yamamoto-Kawai, Michiyo. (2005). Freshwater and brine behaviors in the Arctic Ocean deduced from historical data of δ 18 O and alkalinity (1929–2002 A.D.). Journal of Geophysical Research, 110(C10). https://doi.org/10.1029/2004JC002793 Zhang, J., & Steele, M. (2007). Effect of vertical mixing on the Atlantic Water layer circulation in the Arctic Ocean. Journal of Geophysical Research: Oceans, 112(C4). https://doi.org/10.1029/2006JC003732 Sipler, R. E., Gong, D., Baer, S. E., Sanderson, M. P., Roberts, Q. N., Mulholland, M. R., & Bronk, D. A. (2017). Preliminary estimates of the contribution of Arctic nitrogen fixation to the global nitrogen budget: Arctic nitrogen fixation.

Limnology and Oceanography Letters, 2(5), 159–166. https://doi.org/10.1002/lol2.10046 Steele, M., & Boyd, T. (1998). Retreat of the cold halocline layer in the Arctic Ocean. Journal of Geophysical Research: Oceans, 103(C5), 10419–10435. https://doi.org/10.1029/98JC00580 Torres-Valdés, S., Tsubouchi, T., Bacon, S., Naveira-Garabato, A. C., Sanders, R., McLaughlin, F. A., et al. (2013). Export of nutrients from the Arctic Ocean: Arctic Ocean nutrient exports. Journal of Geophysical Research: Oceans, 118(4), 1625–1644. https://doi.org/10.1002/jgrc.20063 Walker, S. A., Azetsu-Scott, K., Normandeau, C., Kelley, D. E., Friedrich, R., Newton, R., et al. (2016). Oxygen isotope measurements of seawater ( H2 18 O/ H2 16 O): A comparison of cavity ring-down spectroscopy (CRDS) and isotope ratio mass spectrometry (IRMS): CRDS oxygen isotope measurements in seawater. Limnology and Oceanography: Methods, 14(1), 31–38. https://doi.org/10.1002/lom3.10067 Wallace, D. W. R., Moore, R. M., & Jones, E. P. (1987). Ventilation of the Arctic Ocean cold halocline: rates of diapycnal and isopycnal transport, oxygen utilization and primary production inferred using chlorofluoromethane distributions. Deep Sea Research Part A. Oceanographic Research Papers, 34(12), 1957–1979. https://doi.org/10.1016/0198-0149(87)90093-8 Whitmore, L. M., Morton, P. L., Twining, B. S., & Shiller, A. M. (2019). Vanadium cycling in the Western Arctic Ocean is influenced by shelf-basin connectivity. Marine Chemistry, 216, 103701. https://doi.org/10.1016/j.marchem.2019.103701 Yamamoto-Kawai, M., McLaughlin, F. A., Carmack, E. C., Nishino, S., & Shimada, K. (2008). Freshwater budget of the Canada Basin, Arctic Ocean, from salinity, δ 18 O, and nutrients. Journal of Geophysical Research, 113(C1). https://doi.org/10.1029/2006JC003858 Yamamoto-Kawai, Michiyo. (2005). Freshwater and brine behaviors in the Arctic Ocean deduced from historical data of δ 18 O and alkalinity (1929–2002 A.D.). Journal of Geophysical Research, 110(C10). https://doi.org/10.1029/2004JC002793 Zhang, J., & Steele, M. (2007). Effect of vertical mixing on the Atlantic Water layer circulation in the Arctic Ocean. Journal of Geophysical Research: Oceans, 112(C4). https://doi.org/10.1029/2006JC003732

CHAPTER III – ARCTIC OCEAN BARIUM BUDGET REVEALS LARGE MARGIN

INPUTS TO THE BASIN

3.1 Abstract

This chapter addresses the sources and sinks of dissolved Ba in the Arctic Ocean.

Using a pan-Arctic data set from the 2015 Arctic GEOTRACES expeditions, a box model approach is used to determine the relative influence of Pacific-derived seawater, Atlantic- derived seawater, and riverine discharge on the Arctic Ocean Ba budget. Under the assumption of steady state, the results indicate only 50% of the Ba budget is accounted for. We ascribe the additional 50% to the inputs from shelf sediments. Such a large shelf term indicates that the basin Ba budget is influenced by margin exchange in addition to redistribution due to particle formation, as previously described.

Additionally, we consider the influence of the margins on the Arctic Ocean outflow through the Canadian Arctic Archipelago. Although a shallow shelf region, the

Canadian Arctic Archipelago does not impart a large margin signal and the distribution of dissolved Ba (dBa) distribution is predominantly influenced by mixing of Arctic Ocean waters with Baffin Bay waters. The spatial variability of margin inputs cannot be resolved in this study. Although, by assessing Ba distributions, it is evident that shelves throughout the Arctic Ocean region must contribute substantially different amounts of dBa to Arctic Ocean.

3.2 Introduction

Arctic Ocean waters contribute to the formation of North Atlantic Deep Water

(NADW). Thus, processes in the Arctic Ocean can influence heat, carbon, and elemental distributions globally as waters are circulated via the Atlantic Meridional Overturning

Circulation (AMOC). As the region continues to be influenced by climate, determining current tracer distributions to develop a quasi-baseline is imperative. Dissolved barium

(dBa) has a history of research in the Arctic Ocean. As a trace element that is not contaminant-prone, it has been sampled on several Arctic Ocean sections (e.g.,

Abrahamsen et al., 2009; Guay and Kenison Falkner, 1997; Guay et al., 2009; Taylor,

2003). In terms of historical measurements, dBa offers a unique opportunity to assess the distributions through time. However, a better understanding of the sources, sinks, and internal processes influencing the Ba distribution in the Arctic Ocean must be acquired to appropriately consider the available data.

The distribution of dBa is influenced by circulation, particulate Ba (pBa) formation, and inputs from the seafloor (e.g., Jacquet et al., 2005; Pyle et al., 2018). In consideration of pBa water column dynamics, pBa has been applied as a proxy for productivity and carbon export (e.g., Dehairs et al., 1980; Dymond et al., 1992; Eagle et al., 2003). This proxy was developed following an early study that described a correlation between barite flux and particulate organic carbon flux (Dymond et al., 1992). The mechanisms driving the relationship remain unclear; although, the most widely accepted explanation entails removal of Ba in surface waters associated with biological productivity. Barium becomes associated with particulate organic carbon in the surface ocean, and, as the particles sink and are remineralized, microenvironments that favor barite formation form (Cardinal et al., 2005; Chow and Goldberg, 1960; Dehairs et al.,

1980; Ganeshram et al., 2003; Martinez-Ruiz et al., 2019). Since barite is fairly insoluble, pBa will largely escape dissolution before sinking to the seafloor when ballasted with sinking organic matter. In this regard, barite flux from the surface is likely a function of

barium concentration in the surface, export production, and the remineralization of sinking materials. Thus, while the mechanisms are not well defined, there is a relationship between Ba and C that may be used to infer productivity. In the central

Arctic Ocean, productivity is low relative to other ocean basins and modern measurements of export are limited (Honjo et al., 2010 and references therein).

Incorporating Ba as a tool to estimate export must be done with consideration.

Nonetheless, it would add a tool to a region where direct measurements of export via sediment traps has been limited.

Additionally, several studies have assessed the viability of Ba as a tracer of North

American and Eurasian river input into the Arctic Ocean (Abrahamsen et al., 2009;

Alkire et al., 2015; Guay and Falkner, 1997; Guay et al., 2009; Roeske et al., 2012).

Roughly 10% of the world’s river discharge, including several major rivers, enters the

Arctic marine system (Milliman and Farnsworth, 2013). The surface Arctic Ocean has an almost-estuarine quality with a relatively fresh surface and a strong halocline. These fresh surface waters exit the Arctic and can influence the convective formation of deep waters.

While salinity can be used to estimate the freshwater budget in the central Arctic

(Carmack et al., 2016), the study of individual rivers and where they are expressed in the

Arctic marine system has been difficult due to limitations of riverine tracers. Oxygen isotopes have a unique signature for Arctic meteoric water, but they are similar across the entire Arctic region (Östlund and Hut, 1984). High concentrations of dBa in Arctic rivers relative to seawater identified Ba as a possible tracer of rivers. Furthermore, a substantial difference between North American and Eurasian river Ba concentrations suggests dBa’s application as a tracer distinguishing continental freshwater sources (Guay and Falkner,

1997). The components of the freshwater budget in the Arctic Ocean must be quantified and monitored to predict how climate variability will influence downstream processes, such as the formation and composition of NADW (Yang et al., 2016).

Previous Arctic Ocean Ba work has suggested that much of the Arctic Ocean riverine component is derived from Eurasian Rivers (Guay et al. 2009). However, studies applying Ba as a freshwater tracer acknowledge that non-conservative processes, such as particle formation, may influence its distribution (e.g., Abrahamsen et al., 2009; Guay et al., 2009; Roeske et al., 2012a). Nonetheless, few studies have been able to describe the non-conservative behavior of dissolved Ba in the Arctic (Hendry et al., 2018; Roeske et al., 2012a; Taylor et al., 2003; Thomas et al., 2011), which limits its utility to a predominantly qualitative description of freshwater sources. Thomas et al. (2011) and

Hendry et al. (2018) assessed dissolved Ba distributions in the Amundsen Gulf and north of Svalbard, respectively. These studies came to similar conclusions: that biological Ba precipitation seasonally influences Ba in surface waters (i.e., < 50 m). Herein, we expand the scope of previous studies to a pan-Arctic perspective and discuss several approaches to assessing non-conservative Ba sources and sinks in the Arctic marine system.

Beyond biological pBa formation, sea ice dynamics and margin influences must be considered among potential non-conservative sources and sinks in the Arctic Ocean

Ba cycle. Margins in general have been identified as an important source of trace metals to the oceans (Jeandel et al., 2011). Due to its broad continental shelves and the expected increase in shelf fluxes due to the decline in sea ice coverage, the Arctic Ocean margins must be considered as an especially important trace metal source to the region (Charette et al., 2016; Kipp et al., 2018). We expect that Arctic Ocean margins would be a source

of Ba, as previous studies elsewhere have indicated a flux of Ba out of margin sediments via diffusion or advective submarine groundwater discharge (e.g., McManus et al., 1994;

Shaw et al., 1998).

To address the issues outlined above, we have assembled a pan-Arctic data set utilizing 2015 Arctic GEOTRACES sections to assess the Arctic Ba cycle, with an eye toward margin influences. We incorporate dissolved barium isotopic data and particulate

Ba data as additional tools to describe the dBa distributions. Finally, we compare this study to the 1994 Arctic Ocean Survey.

3.3 2015 Arctic GEOTRACES Sections

Four oceanographic expeditions were conducted between July and October 2015 that spanned the Western and Eastern sectors of the Arctic Ocean, and included shelf areas such as the Bering Sea, Barents Sea, and Canadian Arctic Archipelago (Figure 3.1).

The cruises participated in the international GEOTRACES program in an effort to characterize trace elements and their isotopes in the Arctic Ocean. Cruises departed from the United States (GN01: Aug. 9 – Oct 12, 2015), Norway (GN04: Aug. 17 – Oct. 14,

2015), and Canada (GN02 & GN03: July 10 – Aug. 20, 2015 & Sept. 4 – Oct. 1, 2015, respectively) and are referred to by their GEOTRACES cruise ID (GN0#) throughout the text.

GN01 transited through the Bering and Chukchi Seas and completed two transects: one in the Makarov Basin (180°W) and another in the Canada Basin (150°W).

GN02 and GN03 completed surveys through the Canadian Arctic Archipelago, with a primary transect from the Canada Basin of the western Arctic Ocean through Baffin Bay to the Labrador Sea (Figure 3.1). GN02 and GN03 also conducted two high resolution

cross sections at critical gateways: across the Canada Basin/Amundsen Gulf sill and another across Lancaster Sound (see Section 3.5.7, Figure 3.10A). GN04 conducted a transect from the Barents Sea to the North Pole roughly along the 30°E longitudinal line

(due to ice conditions, there are longitudinal variations in the transect). Near the North

Pole, GN04 conducted a high resolution transect perpendicular to 135°E. Each cruise shared a crossover station and data from these stations were analyzed for quality control and intercalibration purposes.

Throughout the manuscript, we will refer to the key regions as the “western Arctic

Ocean”, “central Arctic Ocean”, “eastern Arctic Ocean” and “Canadian Arctic

Archipelago (CAA).” The eastern and western Arctic Ocean are the Eurasian and North

American sides of the Lomonosov Ridge, respectively. The central Arctic Ocean is the region north of 85°N, which—during the 2015 expeditions—was influenced by

Transpolar Drift waters (see section 2.1 for further discussion on regional hydrography).

Figure 3.1 Regional geography, hydrography, and station map.

A) Local geographic features and predominant surface circulation. B) Station map for the 2015 GEOTRACES expeditions.

U.S. GEOTRACES (GN01) are black diamonds, European GEOTRACES (GN04) are red diamonds, and Canadian

GEOTRACES (GN02 and GN03) are orange diamonds. C) Regional hydrographic features. 3.3.1 Section Hydrography

Seawater enters the Arctic Ocean through the Bering Strait (Pacific-derived waters), the Fram Strait and the Barents Sea (Atlantic-derived). Substantial freshwater inputs to the regions sustain low salinity waters in the Arctic Ocean’s mixed layer (Polar Mixed

Layer; PML). While only a minor fraction of the world’s ocean volume (< 3%), the

Arctic Ocean receives ~11% of global river discharge (e.g., McClelland et al., 2012;

Milliman and Farnsworth, 2013). River discharge and seasonal sea ice melt input contribute to the PML basin-wide. The combination of these inputs, plus Pacific-derived 50

seawater (S ~ 32.5), yield a relatively fresh halocline expressed in the western Arctic

Ocean (Pacific Halocline; PH; Figure 3.1B). Pacific-derived waters enter the Arctic

Ocean via the Bering Strait and are exposed to shallow shelf conditions before entering the Arctic Ocean basins where they make up the PH. During shelf exposure, geochemical and physical properties undergo modifications due to exchange with shelf sediments, seasonal brine formation and sea ice melt, and particulate interactions and biological activity (e.g., Fransson et al., 2001; Gong and Pickart, 2016; Whitmore et al., 2019).

Indeed, transformations of physical properties such as temperature (T) and salinity (S) result in the formation of warm, fresh summer water (SW) and cold, salty winter water

(WW) that contribute to the PH (Gong and Pickart, 2016; Weingartner et al., 1998).

Alaska Coastal Water (ACW) may also be expressed in the PH; these waters are warm waters that originate in the Gulf of Alaska and flow along the coastline into the Arctic

Ocean. Compared to SW, ACW is warmer and fresher (e.g., Steele et al., 2004).

Throughout the discussion, we do not distinguish warm PH waters between ACW and

SW. The vertical and spatial extent of the PH is variable and responsive to atmospheric conditions (e.g., Steele et al., 2004) and have a residence time of roughly 15 years

(considering estimates from Kipp et al., 2019; Schlosser et al., 1999).

Pacific-derived waters are less saline and less dense than Atlantic derived waters and thus overlay saltier, warmer Atlantic-derived “intermediate waters”. Two primary

Atlantic-derived water masses circulate in the Arctic Ocean with residence times of 20 –

30 years (Kipp et al., 2019; Schlosser et al., 1999): Barents Sea Branch Water (BSBW) and Fram Strait Branch Water (FSBW). Barents Sea Branch Waters are exposed to the shelves before entering the basins; the density of these waters increases through cooling

and upon entering the basins they form a layer that flows beneath the FSBW (Rudels,

2018). Below the intermediate waters (>1500 m), Arctic Deep Water circulates within each basin. The Lomonosov Ridge restricts flow between the eastern Arctic Ocean basins and the western Arctic Ocean basins. In the western Arctic Ocean, the Canada and

Makarov Basins are further divided by the Alpha-Mendeleyev Ridge. As there are few outflow sites for deep waters, these waters have long residence times of ~200—500 years

(Kipp et al., 2019; Schlosser et al., 1999; Tanhua et al., 2009). Notably, the western

Arctic Ocean basins have older waters than eastern Arctic Ocean basins. Arctic Deep

Water geochemical signals may be influenced by near-slope mixing processes and brines

(Bauch et al., 1995; Middag et al., 2009; Roeske et al., 2012b; Rudels and Quadfasel,

1991).

Waters exiting the Arctic Ocean leave through the Fram Strait and through the

Canadian Arctic Archipelago (CAA; Rudels, 2018). Waters leaving the Arctic Ocean via the Fram Strait flow along the eastern coast of Greenland in the East Greenland Current

(EGC). Our study has no samples from this region; see Bourke et al. (1987) and Michel et al. (2015) for further discussion on this pathway. Net volume fluxes out of the Fram

Strait and the CAA (via Davis Strait) are roughly equivalent (~2 Sv; Beszczynska-Möller et al., 2012). All waters entering the CAA must transit through one of several relatively shallow straits (< 500 m) before ultimately entering the Labrador Sea (e.g., McLaughlin et al., 2004; Melling, 2000). Importantly, the waters transiting through the Canadian

Arctic Archipelago can influence circulation and chemical distributions in the North

Atlantic (e.g., Macdonald et al., 2000; Wadley and Bigg, 2002; Yang et al., 2016).

Similar to Pacific-derived waters that transit over shelves, seawater that passes through

the CAA may be modified by processes occurring during that transit. For example, sediment exchange, biological activity, riverine input, and sea ice melt and formation may influence the geochemical composition of CAA waters. Numerous small rivers discharge into the CAA; while these rivers do not impact the central Arctic Basin, they may contribute significantly to geochemical signals observed within the Archipelago.

The estimated combined discharge of all CAA rivers is about 10% of the total river discharge into the Arctic (Alkire et al., 2017; Haine et al., 2015); this assessment is imprecise because few of the rivers draining into the CAA are gauged.

The main flow of seawater through the CAA is eastward and southward toward

Baffin Bay and the Labrador Sea. Difficulty in interpreting CAA geochemical distributions is highlighted by the complexity of the currents and pathways present throughout the region. Arctic Ocean seawater enters the CAA through four main pathways: (1) Amundsen Gulf, (2) McClure Strait, (3) Queen Elizabeth Islands, and (4)

Nares Strait (see Section 3.5.7; Figure 3.10A). Notably, waters within the Amundsen

Gulf may not contribute largely to bulk flow through the CAA. Lanos (2009) indicated that the prevailing circulation was into the gulf on the northern side of the sill, and out of the gulf on the southern side of the Amundsen Sill. Regarding other pathways, the flow is generally east and southward; however, the straits are wide enough for counter currents to form along the coastlines (see Section 3.5.7, Figure 3.10B) (LeBlond, 1980; McLaughlin et al., 2004). Furthermore, the region is tidally influenced and winds play a role setting the surface currents (McLaughlin et al., 2004; Peterson et al., 2012). For this study, we focus on waters entering the Amundsen Gulf and waters entering via the McClure Strait passing through the Lancaster Sound.

3.4 Methods

3.4.1 Sample Collection and Analysis

For all cruises, dBa samples were filtered and collected into acid cleaned containers from a trace metal clean rosette following GEOTRACES protocols (Cutter et al., 2014). Specifics to each rosette can be accessed via the cruise reports

(http://www.geotraces.org/library-88/scientific-publications/cruise-reports). Additionally, large and small fraction (> 51 µm & 1 - 51 µm) particulate barium (pBa) samples were collected via McLane Research in situ pumps (WTS-LV) during the GN01 section

(following Cutter et al., 2014). Pump casts were set up as described in Xiang and Lam

(submitted). Briefly, filter holders on the McLane pumps were prepared for two flow paths (quartz fiber “QMA” and polyethersulfone “Supor” flow paths) with 142 mm- diameter filter holders. Each path housed a “pre-filter” (51 µm polyester mesh; Sefar 07-

51/33). Following the prefilter, the “QMA” path had paired 1.0 µm quartz fiber filters

(Whatman QMA) that had been pre-combusted at 450°C for 4 hours. The “Supor” path had paired 0.8 µm polyethersulfone (Pall Supor800) filters. At basin stations (GN01), dBa was collected from the clean rosette which conducted two casts with a total of 23 depths (one overlapping depth). Particulate samples were typically collected from two pump casts for a total of 16 depths; at three stations, three casts were conducted for a total of 24 depths. In comparing the dBa to pBa, sample depths are often not a match.

3.4.1.1 Dissolved Barium Concentrations

Samples from GN01 were analysed at the Center for Trace Analysis (University of Southern Mississippi; USM; Shiller, 2019). Samples from GN02/GN03 were analysed at Vrije Universiteit Brussel (VUB) and GN04 samples were analysed at the University

of Alaska, Fairbanks (UAF; Rember, 2018). All samples were acidified prior to analysis via inductively coupled plasma mass spectrometry (ICP-MS) and concentrations were determined by isotope dilution (Klinkhammer and Chan, 1990). Specifics to each facility are detailed below.

At USM, dBa was determined using an ICP-MS (ThermoFisher Element XR) in low resolution; samples were introduced with a PC3 spray chamber (Elemental

Scientific). Prior to analysis, samples were acidified to 0.024 M HCl (Fisher Optima). In preparation for analysis, following isotope dilution methods, samples were diluted 30- fold with ultra-pure water and spiked with enriched 135Ba solution (Oak Ridge National

Laboratory) to a target 138/135Ba ratio between 0.5 and 1. Standards and GEOTRACES reference samples (GS & GD, distributed from the 2008 GEOTRACES Intercalibration

Cruise) were analysed in every run for reproducibility, which was within < 2% RSD

(Supplemental Table 1).

At VUB, a volume of 0.25 ml of sample was pipetted into an acid cleaned 15 mL polyethylene tube and acidified with 0.15 ml concentrated ultra-pure nitric acid to ensure the stability of Ba measurements. This acidified sub-sample was spiked with 0.15 ml of a

135Ba-spike solution yielding a 138Ba/ 135Ba ratio between 0.7 and 1 to minimize error propagation (Klinkenberg et al., 1996; Webster, 1960). Subsequently, the sample was diluted 30-fold with 7 ml Milli‐Q grade water to reduce salt content to less than 0.2%.

Quantities of sample, spike and dilution water were assessed gravimetrically. The same procedure was employed to prepare blanks (Milli‐Q grade water) and reference waters:

SLRS-5 & SLRS-3 (National Research Council Canada; Ba concentrations = 14.0 ± 0.5

µg L-1 and 13.4 ± 0.6 µg L-1, respectively) and ‘OMP’ seawater (Mediterranean seawater

prepared at Observatoire Midi Pyrénées, Toulouse, France; Ba concentration = 10.4 ± 0.2

µg L-1). Isotope ratios were measured with a sector-ﬁeld inductively-coupled plasma mass spectrometry (SF-ICP-MS; Element 2, Thermo Finnigan). Reproducibility of our method is within <1% (RSD) as tested on repeat preparations of SLRS-5.

Dissolved barium from each cruise and laboratory were intercalibrated using samples from crossover stations, which are stations where two cruises occupied the same location. GN01 and GN02/GN03 had a crossover station in the Canada Basin; however, due to rough sea state, GN01 was not able to completely sample the planned station. To accommodate for this, we compared GN01 data at two stations to GN02/GN03 data at nearby stations (Figure 3.2). The results from the comparison indicate that the data agree well at depths below 500 m (Figure 3.2). GN01 and GN04 had a crossover station at the

North Pole. The data indicate a 5—8% offset (mean 7%). We conducted our interpretive analysis below three ways: using the original data, using the GN04 data plus 7%

(correcting the GN04 to the GN01), and using the GN01, GN02, and GN03 data minus

7% (correcting these cruises to the GN04 data). Regardless, the interpretations of the analyses we conducted were not sensitive to the offset.

Figure 3.2 Intercalibration Comparison.

Intercalibration stations were met for GN01-GN04 (#1 on map) and GN01-GN02 (#2, #3 on map). The map to the right indicates the location of the station. Station profiles of dBa for each cruise/station are in nmol kg-1. GN01 data re black points, GN02/GN03 are orange, and GN04 data are red points. 3.4.1.2 Dissolved Ba Isotopes

Dissolved Ba isotope measurements were made on a subset of the GN01 samples at the NIRVANA Labs at Woods Hole Oceanographic Institution. Methods followed those described by Bates et al. (2017). A double spike containing roughly equal parts

135Ba and 136Ba was added to 5 mL of seawater to achieve a spike- to sample-derived Ba

(molar) ratio of between 1—2. Following equilibration, Ba was preconcentrated by co- precipitating with CaCO3 via addition of 350 uL of 1 M Na2CO3 solution. The resulting precipitate was dissolved in 6 M HCl, dried, and reconstituted in 250 µL of 2 M HCl in preparation for ion-exchange chromatography. Chromatographic protocols follow those described by Horner et al. (2015); matrix elements and Ba were eluted (and isobaric

REE's retained) using 2 M HCl and 2 M HNO3, respectively, and the process subsequently repeated. Following chromatographic separation, samples were dried and 57

reconstituted in 0.5 M HNO3 for isotopic analysis. The volume was adjusted to achieve a natural (i.e., non spike-derived) Ba concentration of ~20 ng/mL.

Barium-isotopic analyses were performed using a ThermoFinnigan Neptune multiple collector ICP-MS situated at the WHOI Plasma Facility. Sample solutions were aspirated at 140 uL/min with ~1 L/min Ar through a PFA micro-concentric nebulizer

(Elemental Scientific) and desolvated in an Aridus II (CETAC). The resultant aerosol

+ was introduced into the MC-ICP-MS and admixed with 3—5 mL/min N2 to reduce BaO formation (Miyazaki et al., 2014). Analyses were performed in static mode by simultaneously monitoring baseline-corrected ion currents corresponding to m/z 131 (Xe;

L3), 135 (Ba; L1), 136 (Xe, Ba, Ce; center cup), 137 (Ba; H1), 138 (Ba, Ce, La; H2), 139

(La; H3), and 140 (Ce; H4) for 30 integrations, each ~4.2 s in duration. (Detector baselines were measured by deflecting the ion beam and measuring intensities for 30 s prior to each analysis.) Data reduction was performed using the three-dimensional geometric interpretation of the double spike problem (Siebert et al., 2001) whereby

138/135, 137/135, and 136/135 correspond to the x-, y-, and z-axes, respectively. Sample isotopic composition were solved iteratively—with additional nested loops for isobaric corrections—and reported relative to the nearest four bracketing measurements of NIST standard reference material 3104a in the delta-notation:

138⁄134 138 Basample δ BaNIST (‰) = 138⁄134 − 1 (Eqn. 1) BaNIST

All samples were analyzed between 2—8 times (median n = 4). Reported values represent the weighted mean of n measurements, whereby the weightings were assigned according to the inverse square of the corresponding measurement uncertainty.

Uncertainties are reported as the greater of either the weighted uncertainty for n measurements (± 2 SE, standard error), or our long-term precision of ± 0.03 permil

(± 2 SD, standard deviation; Horner et al., 2015).

3.4.1.3 Particulate Barium Concentration

Particulate barium concentrations were obtained via a refluxing digestion method

(Cullen and Sherrell, 1999; Ohnemus et al., 2014; Planquette and Sherrell, 2012).

Briefly, the filter was placed onto the wall of a 15 mL flat-bottom screw-cap Savillex vial to avoid immersion. The digestion includes a 4-h refluxing at 110 ˚C with an ultrapure

TM (ARISTAR® or Optima grade) 50% HNO3/10% HF (v/v) mixture and drying down of the acid mixture. By ICP-MS (Thermo Scientific Element XR) at the UCSC Plasma

Analytical Facility, final pBa sample solutions were analyzed in low resolution in low resolution. Indium (1 ppb) was used as an internal standard for ICP-MS analysis.

The lithogenic and non-lithogenic components are considered; where the non- lithogenic fraction represents authigenically formed barite. This fraction is determined by adjusting the observed concentrations by the terrigenous Ba:Al ratio (Eqn 2). The terrigenous Ba:Al ratios is determined from upper continental crust (UCC) values reported by Taylor and McLennan (1995).

퐵푎푈퐶퐶 퐵푎푛표푛푙𝑖푡ℎ표푔푒푛𝑖푐 = 퐵푎표푏푠 − (퐴푙표푏푠 × ) (Eqn. 2) 퐴푙푈퐶퐶

3.4.1.4 Ancillary Data

Ancillary data, such as salinity and temperature, were retrieved from public databases when possible, including BCO-DMO for GN01 (Cutter et al., 2019) and

PANGAEA for GN04 (Ober et al., 2016). Water mass fractions for the central Arctic

Ocean were determined using a four-component linear mixing model. The four- 59

component mixing model uses salinity (S), water oxygen isotopic composition (δ18O), and nitrate and phosphate to determine the fraction of Atlantic, Pacific, meteoric, or sea- ice derived waters in each sample. This method is outlined in greater detail in other literature (see: Bauch et al., 2011; Newton et al., 2013).

3.4.2 Data Analysis

The three cruises cover a large spatial region of the Arctic Ocean. To visualize the distributions, we’ve assembled transects that may incorporate samples from more than one cruise. We defined two transects in the Arctic basins and one through the Canadian

Arctic Archipelago. Section A includes stations in the Bering and Chukchi Seas, the

Makarov Basin (along the Alpha-Mendeleyev Ridge) and into the Eurasian Basin (Figure

3.3A). Section B progresses from the Beaufort Sea shelf-break, through the Canada and

Eurasian Basins, and onto the Barents Sea Shelf (Figure 3.3B). Section C progresses from the Canada Basin, through the Canadian Arctic Archipelago, through Baffin Bay and ends in Baffin Bay and ends south of Baffin Bay (Figure 3.3C). Section plots were generated using weighted-average gridding in Ocean Data View 5.1.5 (Schlitzer, 2018).

The saturation state has been employed in the discussion of particle distributions.

Generally, the surface ocean is undersaturated with respect to barite (BaSO4) (Monnin et al., 1999); thus the formation of barite is unlikely in the absence of microenvironments rich in Ba. In considering ocean distributions, it is helpful to at least consider the barium saturation index (SI). The SI was calculated following Rushdi et al. (2000) and Millero

(1982). The SI is a function of dissolved Ba molality [Ba], sulfate molality [SO4], and the stoichiometric solubility product (Ksp) (Eqn. 3). The solubility product is a function of temperature (T), salinity (S), and pressure (p) (Eqn. 12 in Rushdi et al., 2000).

[퐵푎]×[푆푂 ] 푆퐼 = 4 (Eqn. 3) 퐾푠푝

3.5 Results and Discussion

3.5.1 Distribution of Ba in the Arctic Ocean

Sections A and B showed similar features in the basins. The western basin was characterized by high dBa (> 60 nmol kg-1) in surface waters down to roughly 250 – 350

-3 m (σθ ≈ 26 -27 kg m ) (Figure 3.3A, B). These high Ba waters were associated with the

PML and the Pacific Halocline (PH). Within the halocline, waters associated with WW and SW had differing dBa concentration with the winter-associated waters yielding higher concentrations of dBa (Figure 3.4A). Below the halocline waters dBa concentration decreased, which was associated with the presence of Atlantic-derived water masses. In the Eurasian Arctic Ocean, dBa in the surface waters (< 50 m) was much lower (~40 nmol kg-1) than western Arctic Ocean surface waters, reaching as low as ~32.5 nmol kg-1. The Eurasian Arctic Ocean profile was more typical of global oceanic

Ba profiles: lowest in the surface and increasing with depth. Sites with low surface concentrations of dBa in the eastern Arctic Ocean region (~32.5 nmol kg-1; Figure 3.3A at ~ 4000 km), and over the Bering Sea shelf (~10—15 nmol kg-1; see figures 3.3B at ~

600 km and Figure 3.6A), are likely indicative of biological and/or particle scavenging.

In comparing the deep basins, below 2000 m, dBa was higher in the western

Arctic Ocean basins; this may relate to the greater age of the water in those deep western basins (i.e., Canada and Makarov Basins) and/or the supply of particulate barite from the surface ocean. The PH has been reported to direct shelf-derived particles over the western

Arctic Ocean (Aguilar-Islas et al., 2013; Kondo et al., 2016), thereby indicating that supply of particulate barite originating over the shelves may contribute to higher particle 61

fluxes to the deep western Arctic Ocean. Additionally, the waters of the western Arctic

Ocean basins are reported to have longer residence times than the eastern Arctic Ocean basins (Kipp et al., 2019; Schlosser et al., 1999). Surface waters of the western Arctic

Ocean were supersaturated with respect to barite (Figure 3.3A, B), while solubility of Ba with respect to barite is near or under saturation in the eastern Arctic Ocean. All central

Arctic Ocean deep basin samples were undersaturated with respect to barite (Figure 3.3).

Figure 3.3 Distributions of Dissolved Ba.

A) dBa in nmol kg-1 (A1) and dBa saturation index (SI) (A2) of Section A. B) dBa (B1) and dBa SI of Section B. C) dBa (C1) and dBa SI (C2) of Section C. Hashed areas indicate the

background where no data was available. Location of each section is indicated in the map to the right of the panels.

Western Arctic Ocean (GN01; Sections A and B) pBa distributions support previous conclusions of shelf-derived particles in the PH (Aguilar-Islas et al., 2013;

Kondo et al., 2016). Basin areas have total pBa concentrations up to roughly 1 nM; the small size fraction is generally an order of magnitude greater than the large size fractions.

Maximum pBa concentrations are observed in the upper 500 m of the water column and are highest near the continental slope (Figure 3.5). There is a correlation between the pBa max at each station and the core of the PH. This is qualitatively demonstrated by the pBa maximum being associated within PH depth waters (Figure 3.4B). In other ocean basins, the pBa maximum is observed in mesopelagic depth waters (Bishop, 1988). The formation of pBa is associated with particulate organic carbon. We expect the formation of pBa in the western Arctic Ocean basins to be predominantly associated with laterally sourced particulate organic carbon due to low in-situ primary productivity and large shelf particulate organic carbon inputs (Honjo et al., 2010; Xiang & Lam, submitted).

Figure 3.4 Temperature-Salinity plots with dBa and pBaSSF z-axes.

A) dBa (nmol kg-1), circled regions indicate Bering/Chukchi Sea Summer Waters (SW) and Bering/Chukchi Sea Winter

Waters (WW). B) pBa (nmol L-1), circled regions indicate SW and WW. For both dBa and pBa, WW tend to have higher

concentrations than SW. Shelf concentrations of pBa are up to 16 nM; the highest concentrations are observed near the bottom (Figure 3.6B). This region is largely characterized by lithogenic pBa, whereas the basins are largely nonlithogenic pBa (Figure 3.5 A2, B2). Although, the absolute nonlithogenic pBa on the shelves is largely higher than the basin nonlithogenic pBa signals. Generally, authigenically formed pBa particles are classified in the small size fraction (Lam and Marchal, 2015); we suspect authigenically formed pBa over the shelves is advected offshore with shelf bottom waters. While there may be a small amount of local in situ pBa production in the Arctic Ocean basins, the western Arctic

Ocean pBa distribution is overprinted by the lateral inputs.

Figure 3.5 Particulate Ba of large and small size fractions.

A) Large ( > 51 µm) and small size fraction particle distributions (A1) and the respective percent contribution of nonlithogenic barium (A2) for Section A and Section B (B1, B2). Particulate

data were available for fewer stations; the stations where data was available are in light green in the section maps for each panel. Note, the z-axis in panels A1 and B1 are logarithmically

scaled.

Particulate Ba flux has been utilized to predict particulate organic carbon (POC) flux (Dehairs et al., 1980; Dymond et al., 1992). This approach assumes a 1-D vertical model of particle sinking. However, in the western Arctic Ocean, this approach is complicated by the overprinting of lateral fluxes. While we observed predominantly lateral pBa signatures in the western Arctic Ocean, POC and δ13C-POC distributions indicated a combination of lateral and vertical inputs (Xiang and Lam, submitted). The combination of sources complicates the use of pBa in the western Arctic Ocean as an analogue for vertical POC export.

Figure 3.6 Shelf distribution of dBa and pBa.

Dissolved Ba (A) generally has higher concentrations than the Bering Sea slope station (Station 1) and appears well mixed in the water column. However, station 2 exhibited a large depletion in the surface and very high concentrations near the bottom. Particulate

Ba (B) consistently increases with depth, compared to shelf concentrations the Bering Sea slope station (Station 1) have very low concentrations. In the following sections, we consider the influence of various Ba sources and sinks on the distribution of Ba in the Arctic Ocean. We then consider the possible

consequences of Ba sources and sinks in the Arctic Ocean and its outflow pathways on the communication of Arctic Ocean geochemical properties to the North Atlantic.

3.5.2 Assessing Ba distributions relative to predicted Ba

We utilized water mass deconvolutions determined by a four-component linear endmember mixing model and Ba endmember concentrations to predict the conservative distribution of dBa (Bapred).

퐵푎푝푟푒푑 = 퐵푎푚푒푡푓푚푒푡 + 퐵푎푆퐼푀푓푆퐼푀 + 퐵푎푃푎푐푓푃푎푐 + 퐵푎퐴푡푙푓퐴푡푙 (Eqn. 4)

The four components are identified in the subscripts of Equation 4: meteoric (met, representative of riverine component and precipitation), sea ice melt/formation (SIM),

Pacific-derived waters (Pac) and Atlantic-derived waters (Atl). The dBa of each endmember is weighted by the fraction (f) of the component to determine the predicted concentration of dBa (Bapred) in each sample. Barium endmembers for these components have been widely discussed in previous literature (Abrahamsen et al., 2009; Guay and

Falkner, 1997; Guay et al., 2009; Taylor, 2003).

Minimum Best Estimate Maximum

1 Bamet 90 130 190

2 BaSIM 2 6.5 11

BaPac 56.2 57.3 58.4

1 BaAtl 40 43 46

Table 3.2 Barium endmember estimates (nmol kg-1).

1 Guay et al., 2009; 2 Marsay et al., 2018 Considering the prior literature and available data, we determined a minimum, best-estimate, and maximum Ba endmember concentration for each component (Table

3.2). For the meteoric Ba, we combine annual flow weighted means (AFWM) of the major rivers to determine an Arctic-wide estimate as well as consider the effects of estuarine processes. An average of the AFWMs from the seven major rivers represents our maximum estimate (190 nmol kg-1) (Holmes et al., 2018). Although it does not incorporate estuarine processes, it equally weights the contribution of each river to the central basin. North American rivers have less expression in the central basins and have very high Ba concentrations (> 300 nmol kg-1) (Guay and Falkner, 1997); thus, the mean of all AFWMs would bias the river Ba estimate high. Our “best-guess” estimate (130 nmol kg-1) is from Guay et al. (2009) and considers both the AFWMs and previous estimates of the effective river endmember (i.e., includes estuarine processes). Our minimum estimate is an average of Eurasian river AFWMs; this is low because it does not include estuarine processes or any influence from North American rivers. The sea ice endmember (BaSIM) is estimated as the mean of sea-ice Ba concentrations collected during the GN01 expedition (Marsay et al., 2018). The minimum and maximum sea-ice estimates are plus or minus one standard deviation. The BaPac estimate is determined from GN01 samples in the Bering Strait, plus or minus one standard deviation. Lastly, the

BaAtl estimate is from Guay et al. (2009).

퐵푎푎푛표푚푎푙푦 = 퐵푎표푏푠 − 퐵푎푝푟푒푑 (Eqn. 5)

We then determined the Baanomaly (Eqn. 5); where the anomaly is roughly 0, we consider Ba behavior to be conservative. Ba excess (Baanomaly > 0) indicates observed concentrations higher than predicted; thus, we would expect there to be an additional source of Ba that was not included in the four-component model. Alternatively, deficits

(Baanomaly < 0) indicate a removal term that should be considered. Under all scenarios

(minimum, best-estimate, maximum), there is an excess in the region of the Pacific- derived halocline in the western Arctic Ocean (Figure 3.7). Additionally, there is occasionally a surface deficit, which indicates non-conservative removal of dBa in the surface mixed layer. The greatest deficits were observed at shelf-break stations in the western Arctic Ocean. Note, the deconvolution of water masses was not conducted for stations over the shelves. We may expect even greater deficits in these regions, but, due to the hydrography of the region, it is not appropriate to apply the same water mass deconvolution.

-1 Figure 3.7 Baanomaly (nmol kg ) for the western, central, and eastern Arctic Ocean.

The green dashed line represents no anomaly (i.e., conservative behavior). The gray lines are the mean profiles of each region for the minimum, “best-guess”, and maximum endmember estimate. The map to the right depicts the stations categorized into each region, where blue dots are the western Arctic Ocean, black dots are the central Arctic Ocean, and red dots are the eastern Arctic Ocean. Compared to the western Arctic Ocean, the central Arctic Ocean (> 85°N) had a more consistent surface deficit, although there were stations with both negative and positive Baanomaly (Figure 3.7). In the central Arctic Ocean, the deficit was typically 71

associated with the surface 75 m and excesses (positive anomalies) were at a maximum in the sub-surface. The excesses are consistent regionally with the northernmost extent of the PH. The surface deficits pose an interesting conundrum: are the deficits occurring in situ or laterally transported from shelf waters? In the central Arctic Ocean, TPD waters— which originate from the East Siberian and Laptev Seas—most strongly influence the surface 50 m and have both a shelf and riverine component (Charette et al., in prep; Kipp et al., 2018). Dissolved Ba is correlated with the meteoric fraction in these waters

(Charette et al., in prep) and suggests a reasonable riverine endmember (i.e., in the range of possible Eurasian river endmembers). However, Laptev Sea dBa and fmet distributions indicated a decoupling of the dBa distribution from the meteoric water distribution;

Roeske et al. (2012b) attribute the decoupling of surface deficits to particle formation, and bottom water excesses to particle dissolution. Our central Arctic Ocean deficit may indicate that, at the time TPD waters were advected from the shelves, the net effect of shelf processes on the dBa distribution in shelf surface waters was removal.

In the eastern Arctic Ocean, there was not an excess in the Baanomaly. Additionally, in considering our range of endmembers, the shape of the Baanomaly profile is reversed at the maximum endmember estimate. This analysis in the eastern Arctic Ocean was most sensitive to the range of meteoric endmember Ba concentrations. Our best-estimate of

Baanomaly shows the surface waters of the eastern Arctic Ocean clustering around zero but may have slight deficits (Figure 3.4). Deficits in this region are likely driven by pBa formation associated with primary productivity. Indeed, in the Barents Sea region,

Hendry et al. (2018) reported Ba deficits in the water column throughout winter and spring, but not in summer. In the spring, they attribute the deficits to Ba uptake associated

with under-ice biological activity. In the winter, biological indicators are low and thus would not be a substantial sink of Ba (Hendry et al., 2018). Our samples, collected in late summer, have only minor deficits in the surface. This suggests that, compared to end member error, the influence of biological activity was minimal in the eastern Arctic

Ocean during the GN04 section.

Hendry et al. (2018) and Thomas et al. (2011) predominantly ascribe surface dBa deficits to seasonal bio-activity in the surface and associated barite formation. Indeed, the association between warm Pacific halocline waters (SW) and decreased Ba relative to winter-associated Pacific halocline waters (WW) does indicate that seasonality may influence the Ba distribution (Figure 3.4A). Evidence has suggested that the shelves may be a large source or sink of several tracers (Jensen et al., 2019; Kadko et al., 2019; Kipp et al., 2018; Kondo et al., 2016; Marsay et al., 2018; Whitmore et al., 2019), as such, we hypothesize that the excess observed in the Pacific halocline is derived from a shelf flux of dBa.

In comparing the regions, the western Arctic Ocean is heavily influenced by a shelf source that is incorporated into Pacific-derived waters. The eastern Arctic Ocean does not appear to have an equivalent shelf flux term. The Barents Sea is an expansive margin, but deeper than the western Arctic Ocean shelf area. The behavior of these two regions may be the result of a different reservoir of shelf-derived Ba or varying supplies of the same source. The shelf source of Ba will be considered further in Section 4.5.

3.5.3 Barium Isotopes

In this section we consider dBa isotope signatures as another means to assess the relative influence of source (i.e., conservative mixing) or local processes (e.g., particle

formation) on the dBa distribution. Recent literature on Ba isotopes has indicated that

δ138Ba distributions are largely influenced by ocean circulation (Bates et al., 2017;

Horner et al., 2015; Hsieh and Henderson, 2017). Oceanic profiles are typically isotopically enriched in the surface and depleted at depth (Figure 3.8). Such a profile is compatible with incorporation of isotopically light Ba in the surface (i.e., biological incorporation or barite formation) and remineralization at depth. However, studies have demonstrated that Rayleigh-type models (open and closed), which model the fractionation of δ138Ba by biological or chemical processes, cannot solely describe the observed vertical distributions (Bates et al., 2017; Hsieh and Henderson, 2017). Thus, circulation has been indicated as the likely main control on the observed ocean distribution of δ138Ba. The shape of the δ138Ba profile in the western Arctic Ocean differs from observations made in other ocean basins: the surface is isotopically light and δ138Ba increases to a maximum near 500 m. Below 500 m, δ138Ba becomes lighter and the value stabilizes below 2000 m (Figure 3.8A). Interestingly, despite a different vertical profile shape, the local Arctic Ocean δ138Ba and inverse dBa relationship are generally similar to the global pattern (Figure 3.8B).

Figure 3.8 δ138Ba patterns in the western Arctic Ocean.

A) The vertical distribution of Arctic Ocean data (blue dots) and global data (gray dots). The depth axis is logarithmically

scaled to expand the surface range. Solid blue and dashed black lines represent example Arctic Ocean, North Atlantic

Ocean, and North Pacific Ocean profiles.

B) The δ138Ba versus 1/dBa pattern. Arctic Ocean data from this study are in blue dots, and global data are gray. The upper water column of the Arctic Ocean basins has large lateral advective fluxes that influence dissolved and particulate distributions (e.g., Aguilar-Islas et al.,

2013; Rudels, 2018). Due to these lateral inputs, we suspect that the δ138Ba distribution in the upper water column may reflect conservative mixing of Arctic Ocean water types.

The conservative mixing of δ138Ba in the Arctic Ocean can be predicted by the weighting the δ138Ba in each component by the endmember concentration of Ba in that component and the fraction of the component present in the sample (Eqn. 6). The denominator,

138 [Ba]pred, is defined in equation 4; additionally, throughout equation 6, δ Ba is identified as δ in an effort to improve readability.

[(훿푚푒푡×[퐵푎]푚푒푡×푓푚푒푡)+(훿푎푡푙×[퐵푎]푎푡푙×푓푎푡푙)+(훿푝푎푐×[퐵푎]푝푎푐×푓푝푎푐)+(훿푠푖푚×[퐵푎]푠푖푚×푓푠푖푚)] 훿푝푟푒푑 = (Eqn. 6) [퐵푎]푝푟푒푑 75

The fractions of each water mass component (meteoric, Atlantic, Pacific, and sea ice) and the dBa endmembers are described in detail in Section 4.2. Endmember δ138Ba has not been previously determined for this region; thus, we assessed the available literature to determine a range of reasonable endmember concentrations for each

138 component. We expect δ Bamet to be near 0 ‰; mean isotopic composition of upper continental crust samples is δ138Ba = 0.00 ± 0.03‰ (Charbonnier et al., 2018; Nan et al.,

2018). Additionally supporting this expectation: riverine samples from the Rio de la Plata

(which drains to the South Atlantic Ocean on eastern coastline of South America near

Buenos Aires, Argentina) range between -0.06 ‰ and 0.11 ‰ ( δ138Ba ; Hsieh and

Henderson, 2017). Although, a recent study in the Mackenzie River Basin indicates that shales with high dBa also have higher isotopic ratios. Thus, we may expect higher δ138Ba concentrations entering the Arctic Ocean (Charbonnier et al., 2019).

The Pacific endmember, as a mean of previously published data from the surface

200 m of North Pacific stations, is 0.61 ± 0.02 ‰ (n = 10). Our study encapsulated the surface waters (< 55 m) of a station at the Bering Sea Slope and we suspect these waters may be more representative of the water entering the Arctic Ocean than the data from the central North Pacific basin; at this station our data range from 0.42 – 0.50 ‰ (n = 3). We assessed the Atlantic endmember similarly to the Pacific; previously published surface (<

200 m) data in the North Atlantic indicate a mean δ138Ba of 0.53 ± 0.03 ‰ (n = 10).

To determine if conservative endmember mixing could explain the vertical profile in the upper 500 m, we ran an optimization procedure in R (“optim”; R Core Team, 2018) where our cost was defined as Σ|(퐵푎푝푟푒푑 − 퐵푎표푏푠)|. The optimization routine iteratively tests possible endmembers and returns endmember values where the cost function is the

lowest. We determined Bapred by assuming the isotopic contributions from ice were negligible. This is not unreasonable, as dBa in sea ice is low (~ 5 nM) and the fraction of sea ice in our samples was also low (median = -0.01; n = 28). Indeed, at [Ba]SIM = 5 nmol

-1 138 kg the cost is insensitive to a range of potential δ BaSIM endmembers (i.e., cost varies

138 by less than 2% for δ BaSIM between -1 ‰ and 1 ‰). Our endmembers were thus

138 138 138 determined to be: δ Bamet = 0.24 ‰, δ Baatl = 0.55 ‰, and δ Bapac = 0.37 ‰).

138 138 In considering these values, δ Bamet = 0.24 ‰ and δ Baatl = 0.55 ‰ are

138 reasonable. However, δ Bapac = 0.37 ‰ is lower than expected. With the endmember assumptions we’ve made (i.e., sea ice melt component is negligible), this indicates there is either modification to one of our endmember terms or there is an additional and substantial source term. The shelf δ138Ba distribution has a decrease of δ138Ba over the

Chukchi Shelf, with the lowest values in shelf bottom waters and in Pacific halocline depth waters (50 – 150 m in the Arctic Ocean basin; Figure 3.9). This persistence of lower δ138Ba in shelf bottom waters and Pacific halocline waters supports our hypothesis from the previous section that there is a shelf contribution of Ba.

Figure 3.9 δ138Ba distribution over the Bering and Chukchi Seas.

A) The shelf section depicting δ138Ba on the z-axis, the map to the right identifies the section with a blue line.

B) Individual shelf station profiles, including two profiles from the Beaufort Sea that are not included in panel A. The dashed

line references surface waters in the Bering Sea (Station 1) 3.5.4 Basin-wide Ba Mass Balance

138 Our analysis of both the Baanomaly and δ Ba suggest an additional dBa source to the Arctic Ocean as well as a potential sink in surface waters. In this section we quantify the advective fluxes (sources and sinks) of Ba to determine the magnitude of the net non- conservative components. Ba in the Arctic Ocean has advective sources from rivers,

Pacific-derived sea water, and Atlantic-derived sea water. Sea ice may be a source or a sink of materials, and its role in the Ba cycle in the Arctic Ocean is unclear. Other sinks may include particle interactions and transport out of the system. By assuming the system is in steady-state, the subtraction of sinks from sources equals zero (Eqn. 7).

⏟퐹푟𝑖푣푒푟푠 + 퐹 푝푎푐𝑖푓𝑖푐 + 퐹 푎푡푙푎푛푡𝑖푐 + 퐹 𝑖푐푒 . 𝑖푛 + 퐹 푠ℎ 푒푙푓 ⏟− 퐹푝푎푟푡𝑖푐푙푒푠 − 퐹 𝑖푐푒 .표푢푡 − 퐹푡푟푎푛푠푝표푟푡 .표푢푡 = 0 (Eqn. 7) 푠표푢푟푐푒푠 푠𝑖푛푘푠 where F represents the flux of Ba from sources (rivers, Pacific-derived waters, Atlantic- derived waters, ice and shelf contributions) and sinks (particles, ice, and circulation out of the system). Comparable to Kipp et al. (2018), we assess the fluxes of Ba from these 78

sources and sinks in the surface 500 m of the water column (all fluxes have units of mol y-1). An issue with this approach is that it does not account for the spatial heterogeneity of the Arctic Ocean water column and treats all regions of the Arctic Ocean as homogeneous in terms of Ba distribution and residence time. However, in considering the same box as Kipp et al. (2018), we are able to more directly compare our results to the previous literature.

Furthermore, we are unable to make flux estimates of a few of the mass balance

138 terms, including Fparticles, Fice.in, and Fice.out. However, the Baanomaly and δ Ba endmember tests were insensitive to a range of sea ice concentrations; thus, we expect Fice.in and

Fice.out are minor components. Additionally, the distribution of large and small size fraction pBa indicates that the highest pBa concentrations are largely advected from the shelves into the central basin (see section 4.1). Thus, the in situ formation of pBa may be minor. With this in mind, we consider the “net non-conservative” term to be the sum of

Fshelf, Fparticles, Fice.in, and Fice.out (Eqn. 8). Since we expect Fparticles, Fice.in, and Fice.out to all be small, the largest contribution to this term likely will be the shelf component.

퐹푠ℎ푒푙푓 + 퐹𝑖푐푒.𝑖푛 − 퐹푝푎푟푡𝑖푐푙푒푠 − 퐹𝑖푐푒.표푢푡 = 퐹푟𝑖푣푒푟푠 + 퐹푝푎푐𝑖푓𝑖푐 + 퐹푎푡푙푎푛푡𝑖푐 − 퐹푡푟푎푛푠푝표푟푡.표푢푡 (Eqn. 8)

Generally, the mass balance follows the radium (Ra) budget in Kipp et al. (2018); although Kipp et al. (2018) solved for the shelf flux and our budget determines the net non-conservative flux. Fluxes of Ba from rivers, Pacific seawater, and Atlantic seawater were solved using the following form:

퐹푠표푢푟푐푒 = [퐵푎]푠표푢푟푐푒 × 퐽푠표푢푟푐푒 (Eqn. 9) where, Fsource represents the flux of Ba from rivers, Pacific seawater, or Atlantic seawater

-1 in mol y . Fsource is determined as the product of the endmember concentration of Ba in

-3 that source ([Ba]source as nmol m ) and the volume flux (J) of that source into the surface

500 m (m3 y-1). The dBa endmembers for rivers, Pacific seawater, and Atlantic seawater are as described in section 4.2.

The fluxes of the terms are derived from the literature and largely summarized in

Kipp et al. (2018). The Pacific volume flux, as published by Woodgate et al. (2012) and measured in the Bering Strait in 2011, is (3.5 ± 0.3) × 1013 m3 y-1; the minimum and maximum estimates from this term are defined by the mean plus or minus one standard deviation. The Atlantic flux, a more difficult term to quantify because of the broad pathways by which it enters the Arctic Ocean, is estimated at (2.1 ± 0.1) × 1014 m3 y-1

(Beszczynska-Möller et al., 2012). This value is the average net flux into the Fram Strait, and thus it is not perfectly representative of the volume entering the system. Additionally,

Atlantic waters entering the upper 500 m Arctic Ocean water column are a combination of Fram Strait and Barents Sea-derived waters, but likely don’t account for 100% of either of those components. We follow Kipp et al., 2018 in the choice of our “best guess”

Atlantic flux for consistency. However, we use only the net Fram Strait flux (plus or minus one standard deviation) as opposed to using the range of fluxes presented available for the Fram Strait and Barents Sea branches in determining the minimum and maximum

(Beszczynska-Möller et al., 2012; Rudels, 2015).

The river flux term was determined from Haine et al. (2015) using data between

12 3 -1 2000 and 2010; Jrivers equals 4.2 ± 0.4 × 10 m y . Lastly, to compare the 2015 cruise to the 1994 Arctic Ocean Survey, we modified the terms to be more representative of the

12 3 -1 1990s. Haine et al. (2015) reported Jrivers of 3.9 ± 0.4 × 10 m y between 1980 and

2000. Woodgate et al. (2012) reported Pacific fluxes through the Bering Strait of 2.2 ±

0.3) × 1013 m3 y-1. Given the uncertainty in our original Atlantic flux term and a lack of estimates available specific to that decade, we apply the same fluxes as the 2015 mass balance.

To determine the flux of barium out of the system (Ftransport.out) we determined an average dBa inventory for the upper 500 m of the Arctic Ocean basin. Here, we trapezoidally integrated the surface 500 m of each station where the bottom depth was >

1000 m (yields a station Ba inventory in nmol m-2; see Figure 3.1B for a reference to the

1000 m isobath and the stations within it). Station inventories were averaged and then multiplied by the area of Arctic Ocean (where the bottom depth is > 1000 m) to determine an Arctic-wide inventory (0.023 ± 0.003 mol Ba). The flux of Ba out of the system is then the inventory divided by the residence time of waters in the surface 500 m.

The residence time of waters in the surface 500 m is variable (~ 1 – 30 years), but in treating the surface 500 m homogeneously we use only one residence time (10 years). To determine the minimum and maximum shelf terms, we calculated the balance with the maximum source terms and minimum sink terms (minimum shelf input) and the minimum source terms and maximum sink terms (maximum shelf input).

Table 3.3 Estimated fluxes of barium from Arctic Ocean sources and sinks (in mol y-1)

% Of Total Sinks Year Minimum Flux Best Estimate Maximum flux (“Best-Estimate”)

Sinks Transport Out 2015 2.0 × 1010 2.2 × 1010 2.5 × 1010 100 1994 2.1 × 1010 2.2 × 1010 2.3 × 1010 100 Sources Pacific Advection 2015 1.8 × 109 2.0 × 109 2.2 × 109 9 1994 1.1 × 109 1.3 × 109 1.5 × 109 6 Atlantic Advection 2015 7.8 × 109 9.0 × 109 1.0 × 1010 40 1994 7.8 × 109 9.0 × 109 1.0 × 1010 41 Rivers 2015 3.4 × 108 5.5 × 108 8.8 × 108 2 1994 3.2 × 108 5.1 × 108 8.2 × 108 2 Net 2015 6.3 × 109 1.1 × 1010 1.5 × 1010 48 Non-conservative 8.7 × 109 1.1 × 1010 1.4 × 1010 (Fshelf, Fice.in, Fice.out, 1994 51 Fparticles) Shelf 5.7 × 1010 9.1 × 1010 1.9 × 1010 2015 41 (Ba:Ra-derived) 1994 NA NA NA NA

In assessing the described fluxes in and out of the Arctic marine system, the sources only account for ~52% (range: 40 - 68%) of the ascribed sink, indicating that net non-conservative sources are roughly 48% of the budget (Table 3.3). The net non- conservative term is the sum of sources and sinks derived from the shelf, particle formation or dissolution, and sea ice related input or removal.

This box model approach is sensitive to the endmember terms and fluxes. Our model is most sensitive to the residence time of waters and the inventory, which set the fluxes of barium out of the system (Ftransport.out). A 15% variation in the residence time or the inventory results in roughly a 30% variation in Fshelf. The model is also sensitive to the Atlantic term, where 15% variation in the endmember or volume flux results in a 10% variation to Fshelf. 82

We independently calculated Fshelf by utilizing the Ba:Ra ratio over the Chukchi shelf and the Ra shelf flux (FRa.shelf, Eqn 4).

퐵푎 퐹 = × 퐹 (Eqn. 10) 푠ℎ푒푙푓 푅푎 푅푎.푠ℎ푒푙푓

Kipp et al. (2018) estimated that 80% of the Ra budget was derived from the shelves; thus, we expect a substantial part of the Ba budget to be associated with the shelf sediment diffusion. We utilized the Bering Sea Ba:Ra ratio observed on the shelves

(Appendix Figure B1) and estimated a shelf flux of Ba accounting for 41% (range: 23 –

97%) of the inputs relative to sinks, which effectively closes the mass balance. To summarize, our flux balance approach indicates ~50% of the Ba budget must come from an additional source (i.e., not Pacific Ocean, Atlantic Ocean, or river discharge), which is likely the continental margins. Similarly, the Ba:Ra ratio, suggested ~40% of the Ba budget is derived from the margins. Importantly, because of the integrated nature of our flux balance approach, the method may not realize the intricacies of each profile.

Recent literature has suggested that, as Arctic Ocean shelves become more frequently ice-free, shelf fluxes to the Arctic Ocean will increase (Charette et al., in prep;

Kipp et al., 2018). Broadly, this is difficult to assess due to the paucity of historic geochemical trace metal measurements in the Arctic. However, Ba measurements in the

Arctic began in the early 1990s. We compare the GN01 cruise to the 1994 Arctic Ocean

Survey described in Guay and Falkner (1997); the 1994 expedition replicates many of the stations in both the GN01 and GN04 transects (see Appendix Figure B2 for a station map). In comparing the data, there is no evidence of major changes in the dBa endmember concentrations; however, riverine and Pacific fluxes were adjusted to reflect conditions prior to 2000 (Haine et al., 2015; Woodgate et al., 2012).We estimated that the 83

net non-conservative component of the budget was 51% (range: 45 - 56%). Our data, within the errors of the method, do not reflect an increase in the net non-conservative flux of Ba between 1994 and 2015. Interestingly, this contrasts with the results of Kipp et al.

(2018) for the Ra budget assessed between 2007 and 2015. However, between 1994 and

2015, the Ba mass balance was most sensitive to changes in volume fluxes. Alternatively, the Ra budget appears to be more sensitive to the change in concentration. Thus, it is possible that the Ba model is not sensitive enough to capture any change in shelf fluxes due to only minor changes in concentration.

3.5.5 Supply of shelf-derived Ba to the Arctic Ocean basins

We’ve demonstrated that the shelf accounts for roughly ~50% of the dBa budget by two means. First, a box model assuming steady state balanced fluxes of Pacific,

Atlantic, and river inputs against the transport of Ba out of the system by circulation. The difference between the sinks and sources indicate a shelf flux of Ba accounting for 45—

56% of the budget. Second, we utilized estimate of shelf Ra flux and the Bering and

Chukchi Sea Ba:Ra ratio to determine Ba shelf flux, this estimate accounted for ~41% of the Ba budget.

Such a large flux of Ba from the shelf sediments does require a consideration of the source of Ba in the shelf sediments. Previous studies have indicated possible mechanisms for sediment-derived Ba fluxes. One such mode is advective submarine groundwater discharge (SGD), which has high dBa concentrations and could be a source of Ba to the Arctic marine system. Few studies have demonstrated large advective SGD fluxes in the Arctic shelf system; considering the geologic and geophysical variability in the pan-Arctic domain, it is not surprising that studies in various regions identify varying

inputs of SGD in the Arctic Ocean region (Charkin et al., 2017; Lecher, 2017; Lecher et al., 2016). Importantly, this work has yet to quantify how SGD inputs may influence local biogeochemistry. A consensus among the community is that as permafrost thaws SGD fluxes will increase (Lecher, 2017 and references therein).

Alternatively, work on the California continental margin suggested that a substantial amount of pBa raining from the water column is regenerated at the sediment- water interface (Colbert and McManus, 2005; McManus et al., 1994). The shelf-derived

Ba signal in the Arctic Ocean is predominantly expressed in the PH, which has a residence time of roughly a decade. Seasonal biological behavior potentially drives differences in the SW and WW concentrations of dBa in the PH. Furthermore, the dissolution of shelf pBa could be a source of dBa. We expect pBa to have lighter isotopic composition than the dissolved water it was formed from (Horner et al., 2015; von

Allmen et al., 2010); thus, dissolution of pBa may decrease the isotopic signature of the seawater. We observed that Pacific-derived waters in the western Arctic Ocean have a lighter isotopic composition than Pacific waters in the Bering Sea. Dissolution of marine barite may explain this distribution, and indeed, Arctic Ocean margin sediments often have high biogenic barium content, especially in association with the ice edge (Nurnberg et al., 1996). However, continental sources of dBa also have isotopically light endmembers (i.e., rivers are expected to be around 0 ‰) and cannot be ruled out in our study. It is possible that terrigenous sources, such as SGD or terrigenous particles, could produce a similar decrease in the dBa isotopic composition. Hydrothermal sources of Ba are expected to be negligible in this region of the Arctic Ocean. There is evidence of Ba inputs from hydrothermal sites (Eickmann et al., 2014); however, hydrothermal inputs to

the Arctic Ocean have only been described for the Eastern Arctic Ocean and influence waters below 1000 m (Klunder et al., 2012; Edmonds et al., 2003).

3.5.6 Arctic Deep Water Ba signal

Western Arctic Ocean deep waters have higher dBa concentrations than the eastern Arctic Ocean deep waters (Figure 3.3A, B). Baffin Bay deep waters have the highest dBa concentrations (Figure 3.3C). Roeske et al. (2012b) indicated that Arctic

Ocean deep waters have clear indications of particle dissolution, and that the western

Arctic Ocean deep waters are characterized by dissolution of shelf-derived particles.

Particle dissolution is also a likely culprit for high dBa concentrations in the deep waters of Baffin Bay. This region is potentially less subject to lateral shelf inputs, and may have a larger contribution of in situ pBa formation associated with biological activity (Lalande et al., 2009). The comparatively large POC fluxes in the northern region of Baffin Bay

(Honjo et al., 2010; Lalande et al., 2009)—and, thus, concomitant export of pBa—paired with the long residence time of Baffin Bay deep waters (Bourke and Paquette, 1991), could account for the elevated dBa concentrations by dissolution of pBa.

In addition to observing an increase of dBa in the western Arctic Ocean deep waters, δ138Ba decreases to roughly 0.46 ± 0.01 ‰ (n = 4) (Figure 3.8A). In other ocean basins, the deep δ138Ba signature decreases as well and is attributed to the mixing of isotopically lighter water masses. The decline in the deep Arctic Ocean may occur by two pathways: (1) waters enter the deep basin at 0.46 ‰ and there is no subsequent modification or (2) waters enter at the expected Atlantic endmember (0.55 ‰) and are modified to 0.46 ‰. Without additional data that has higher resolution vertically and regionally, assessing the likelihood of the first option is difficult.

The second option requires a source of low δ138Ba. Two possible sources include shelf derived brines and particle dissolution. Previous literature has indicated that shelf derived brines may contribute to Arctic deep water geochemical signals (e.g., Jones et al.,

1995). The concentrations of dBa in sea ice are low (e.g., Marsay et al., 2018) and trend with salinity. The low concentration in sea ice indicates that dBa is generally not incorporated into the ice matrix and we would expect very little fractionation of the isotopes in seawater. The formations of brines over the shelf would thus preserve the isotopic signature of shelf waters (roughly 0.37 ‰, predicted from the PH endmember estimate); were these brines to enter the deep Arctic Ocean basins, they may lighten the

δ138Ba signature. Roeske et al., (2012b) indicated that dissolution of shelf-derived particles explained the increase in dBa in western Arctic Ocean deep waters; thus, it seems likely that the δ138Ba decrease could also be supported by this mechanism.

3.5.7 Ba in the Canadian Arctic Archipelago

Throughout the previous sections, we have determined that the regional shelves alter the composition (concentration and isotopes) of the incoming Pacific derived waters.

The downstream connectivity and implications of the shelf source of dBa, and potentially other tracers, is important to consider. In the following section, we investigate a suite of

Canadian Arctic Archipelago data to assess the dBa cycle on these outflow shelves.

There is potential for the CAA shelves to influence the Ba distribution. In addition to potential sedimentary sources of Ba over the shelves, there are substantial river inputs to the region. Previous studies have indicated Arctic Ocean-type geochemical signatures throughout the region (Jones et al., 2003). To consider the influence of the CAA on dBa distributions, the GN02/GN03 cruises conducted a transect through a primary

passageway of the CAA: from McClure Strait to Lancaster Sound and into Baffin Bay

(Section C; Figure 3.3C). In addition, samples were collected at several sites throughout the CAA that are potentially influenced by the continent. Generally, the distributions we observed agree with those described in Thomas et al. (2011).

Figure 3.10 Dissolved Ba distributions in the Canadian Arctic Archipelago.

A) A geographic map of the Canadian Arctic Archipelago (CAA), with features numbered and labeled. Sections drawn for the

region are identified by purple, blue, and green lines. B) Predominant circulation pathways for the area. C) A cross section

of the Amundsen Gulf, depicting dBa on the z-axis, thatched areas of the panel indicate regions with no data. D) A cross

section of the Lancaster Sound depicting dba on the z-axis. E) Dissolved Ba—Salinity trends for the Amundsen Gulf cross

section. F) Dissolved Ba—Salinity trends for stations along Section C (purple line) between the McClure Sound (#3) and

Lancaster Sound (#6).

First, we address a cross section of the Amundsen Gulf sill. This cross section captures waters flowing between the Amundsen Gulf and the Canada Basin of the western Arctic Ocean. The Amundsen Gulf is directly to the east of the Beaufort Sea and the Mackenzie River outflow. As North America’s most prominent discharge basin to the

Arctic, this region may help elucidate where the influence of the Mackenzie River is observed. Mackenzie River outflow is largely directed to the east, although depending on atmospheric forcing it may reach the shelf break and interior Canada Basin (e.g.,

Carmack et al., 2016; Fichot et al., 2013). Importantly, the influence of the Mackenzie

River has not been well quantified in the Arctic Ocean or the CAA due to the complexities in regional circulation and in source and sink terms for geochemical tracers

(e.g., Roeske et al., 2012a). Yamamoto-Kawai et al. (2010) demonstrated that, during the summer-fall seasons of 2002, Mackenzie River waters were evident in the coastal surface waters of the Amundsen Gulf, but no further east. One station in the Amundsen Sill cross section (on the southern shelf) had a Mackenzie River-like influence. In the cross section of the Amundsen Gulf sill, we observed a generally Arctic-like dBa distribution: the upper water column has a Pacific-halocline type dBa signature (~65 nmol kg-1) and the deeper waters (sill depth ~ 400 m) have lower concentrations (~40 nmol kg-1) (Figure

3.10C). Additionally, in this region dBa concentrations decline with decreasing salinity

(Figure 3.10E). This likely demonstrates the influence of sea-ice melt as the trend generally agrees with a low Ba, zero salinity endmember (Figure 3.10E). The Amundsen sill cross section reflects the predominant western Arctic Ocean/PH source of waters entering the Amundsen Gulf (Lanos, 2009).

Seawater also enters the CAA through the McClure Strait; this pathway is roughly

Section C (Figure 3.3 C1). West of Barrow Strait, the dBa distribution shows the influence of the PH; these waters are typified by high dBa concentrations (~65 nmol kg-1) at densities associated with the PH. Near the Barrow Strait, isopycnals associated with

-3 the PH shoal (σθ < 27.5 kg m ) and dBa concentrations decrease concomitantly by roughly 10 nmol kg-1. This decrease could be driven by a few mechanisms: (1) sea ice melt, (2) dilution by local rivers, (3) particle formation in surface waters or (4) dilution with low Ba seawater. The decrease in concentration occurs at salinities between 32 and

33, and is not correlated to decreasing salinity, which excludes sea ice melt or river discharges as drivers of the Ba decrease (Figure 3.10F). Particle formation may remove dBa from these waters, but we do not have sufficient data to determine this; it seems unlikely, since a removal of 10 nmol kg-1 in the region of the Barrow Strait would have to occur rapidly to be result in this distribution. If particles were at play, we may expect a more gradual decline in the dBa distribution throughout the region.

We suspect the decrease is driven by mixing with “Baffin Bay-derived” waters.

Across Lancaster Sound (Figure 3.10D), the Ba distribution reveals the influence of at least two water types: high dBa (~55 nmol kg-1) and low dBa (~45 nmol kg-1). Across the surface at all Lancaster Sound stations, dBa is roughly 55 nmol kg-1. Below the surface layer, the stations on the northern side of Lancaster Sound decrease to concentrations around ~45 nmol kg-1, while the stations on the south side tend to remain around 55 nmol kg-1. These observations are consistent with the local circulation wherein Baffin Bay- derived waters flow westward on the northern side of the Sound and CAA-derived waters flow eastward on the southern side of the Sound (Figure 3.10B). Prinsenberg et al. (2009)

described current measurements on the north and south shores of Lancaster Sound and noted that the southern shore consistently had eastward flow. However, the north shore had much greater variability in both magnitude and direction. In the summer, currents were often to the west and in winter they were often to the east (Prinsenberg et al., 2009).

The Baffin Bay-derived waters have the potential to erode the PH signal near Barrow

Strait, since they are lower dBa and warmer. However, given the seasonal variability in flow patterns, it is possible that at another time of year, the PH signal may reach further through the CAA. Flux of dBa across Lancaster Sound, as the product of the Lancaster

Sound cross-sectional mean (± 1 SD) and the range of net volume fluxes through

Lancaster Sound into Baffin Bay (0.7 ± 0.3 Sv from Prinsenberg et al., 2009) is 1.1 × 109 mol y-1 (5.6 × 108—1.7 × 109 mol y-1). This generally agrees with Thomas et al. (2011) for Ba fluxes through the Lancaster Sound. The Lancaster Sound Ba mean we derived

(48.5 ± 3.8 nM) is from five stations across the Lancaster Sound cross section, whereas the mean reported by Thomas et al. (2011) is from one station (51.2 nM). Both of these

Lancaster Sound means are smaller than the CAA dBa mean reported and utilized by

Taylor et al., (2003) (62.4 ± 5.1 nM).

In addition to the Lancaster Sound pathway to Baffin Bay, Pacific halocline waters reach Baffin Bay through additional channels, including Nares Strait. In general,

Baffin Bay Atlantic-origin waters may enter through Nares Strait (northern Baffin Bay), or through the northward circulation of Atlantic waters through Davis Strait (southern

Baffin Bay). Arctic-origin waters flow southward into Baffin Bay through Nares Strait and circulate on the western coast of Baffin Bay (e.g., Tang et al., 2004). Waters within

Baffin Bay are influenced by Arctic Ocean sourced waters, CAA sourced waters,

Atlantic-derived waters, local runoff, and seasonal biological processes (e.g., Michel et al., 2015 and references therein). The section through Baffin Bay (Figure 3.3 C) depicts high (~ 55 nmol kg-1) concentration in surface waters all the way to Davis Strait. Below

-3 -1 ~100 m (σθ ~ 27 kg m ) concentrations decline to 40—45 nmol kg . In Baffin Bay, the concentrations increase again below the depth of the CAA and Davis Strait sills (~1000 m). The concentrations in Baffin Bay deep waters are higher than observed in any other

Arctic region, reaching ~105 nmol kg-1 (see section 4.6 for further discussion).

South of Baffin Bay, in the Labrador Sea, surface dBa concentrations are low

(much more “Atlantic-like”) and a direct signal of Baffin Bay waters dBa signals is not evident (Figure 3.3C). This suggests that there is drawdown of the surface dBa or that the locations sampled do not capture the outflow of Davis Strait.

3.6 Conclusions and Implications

The large non-conservative net dBa source term (~50%), ascribed in this study to shelf inputs, certainly complicates the application of dBa as a conservative tracer of river sources. Previous studies have noted issues with the use dBa for this purpose due to non- conservative behavior (Hendry et al., 2018; Roeske et al., 2012a). Roeske et al. (2012a) observed substantial depletions and excesses of dBa over the Laptev Sea region where the riverine fraction was consistently between 15-20%. They suggested that the formation of pBa in surface waters redistributed the Ba to shelf sediments where the particles then dissolved. Our box model analysis indicates that there is a substantial source term that cannot be accounted for solely by redistribution of dBa in the surface 500 m of the water column. Diagenesis of terrigenous particles or/and submarine groundwater are potential sources of Ba to margin sediments.

Such a large margin source certainly confirms the importance of continental margins to basin geochemical distributions in the Arctic Ocean; furthermore, it affirms the need to constrain margin inputs at regional scales. Due to the connectivity of the

Arctic Ocean to the North Atlantic Ocean through the Canadian Arctic Archipelago, there is a distinct need to understand how waters passing through the shallow archipelago region are modified. In terms of dBa, we observed the broad influence of Baffin Bay waters mixing with western Arctic Ocean waters in the eastern region of the archipelago

(i.e., Barrow Strait to Lancaster Sound). This suggests that the transformations that waters undergo in Baffin Bay are necessary to quantitatively define, as they set the signature of waters leaving through Davis Strait.

In continuing this research, a study of the Ba cycle over several Arctic Ocean shelves will be critical in defining the origin of the shelf Ba signal in the Arctic Ocean basin. Coupling water column observations with pore water conditions would improve our understanding of the shelf-basin interplay. Furthermore, constraining the seasonal properties of waters leaving the Arctic Ocean through the Davis Strait would help identify the variability in geochemical signals entering the Labrador Sea.

3.7 References

Abrahamsen, E.P., Meredith, M.P., Falkner, K.K., Torres-Valdes, S., Leng, M.J., Alkire, M.B., Bacon, S., Laxon, S.W., Polyakov, I., Ivanov, V., 2009. Tracer-derived freshwater composition of the Siberian continental shelf and slope following the extreme Arctic summer of 2007: Siberian freshwater. Geophys. Res. Lett. 36, L07602. https://doi.org/10.1029/2009GL037341 Aguilar-Islas, A.M., Rember, R., Nishino, S., Kikuchi, T., Itoh, M., 2013. Partitioning and lateral transport of iron to the Canada Basin. Polar Sci. 7, 82–99. https://doi.org/10.1016/j.polar.2012.11.001 Alkire, M.B., Jacobson, A.D., Lehn, G.O., Macdonald, R.W., Rossi, M.W., 2017. On the geochemical heterogeneity of rivers draining into the straits and channels of the

Canadian Arctic Archipelago: Canadian Arctic rivers. J. Geophys. Res. Biogeosciences 122, 2527–2547. https://doi.org/10.1002/2016JG003723 Alkire, M.B., Morison, J., Andersen, R., 2015. Variability in the meteoric water, sea-ice melt, and Pacific water contributions to the central Arctic Ocean, 2000-2014. J. Geophys. Res. Oceans 120, 1573–1598. https://doi.org/10.1002/2014JC010023 Bates, S.L., Hendry, K.R., Pryer, H.V., Kinsley, C.W., Pyle, K.M., Woodward, E.M.S., Horner, T.J., 2017. Barium isotopes reveal role of ocean circulation on barium cycling in the Atlantic. Geochim. Cosmochim. Acta 204, 286–299. https://doi.org/10.1016/j.gca.2017.01.043 Bauch, D., Schlosser, P., Fairbanks, R., 1995. Freshwater balance and the sources of deep 18 and bottom waters in the Arctic Ocean inferred from the distribution of H2 O. Prog. Oceanogr. 35, 53–80. Bauch, D., van der Loeff, M.R., Andersen, N., Torres-Valdes, S., Bakker, K., Abrahamsen, E.P., 2011. Origin of freshwater and polynya water in the Arctic Ocean halocline in summer 2007. Prog. Oceanogr. 91, 482–495. https://doi.org/10.1016/j.pocean.2011.07.017 Beszczynska-Möller, A., Fahrbach, E., Schauer, U., Hansen, E., 2012. Variability in Atlantic water temperature and transport at the entrance to the Arctic Ocean, 1997–2010. ICES J. Mar. Sci. 69, 852–863. https://doi.org/10.1093/icesjms/fss056 Bishop, J.K.B., 1988. The barite-opal-organic carbon association in oceanic particulate matter. Nature 332. Bourke, R.H., Newton, J.L., Paquette, R.G., Tunnicliffe, M.D., 1987. Circulation and water masses of the East Greenland shelf. J. Geophys. Res. 92, 6729. https://doi.org/10.1029/JC092iC07p06729 Bourke, R.H., Paquette, R.G., 1991. Formation of Baffin Bay Bottom and Deep Waters, in: Elsevier Oceanography Series. Elsevier, pp. 135–155. https://doi.org/10.1016/S0422-9894(08)70065-5 Cardinal, D., Savoye, N., Trull, T.W., André, L., Kopczynska, E.E., Dehairs, F., 2005. Variations of carbon remineralisation in the Southern Ocean illustrated by the Baxs proxy. Deep Sea Res. Part Oceanogr. Res. Pap. 52, 355–370. https://doi.org/10.1016/j.dsr.2004.10.002 Carmack, E.C., Yamamoto-Kawai, M., Haine, T.W.N., Bacon, S., Bluhm, B.A., Lique, C., Melling, H., Polyakov, I.V., Straneo, F., Timmermans, M.-L., Williams, W.J., 2016. Freshwater and its role in the Arctic Marine System: Sources, disposition, storage, export, and physical and biogeochemical consequences in the Arctic and global oceans: Freswater and the Arctic Marine System. J. Geophys. Res. Biogeosciences 121, 675–717. https://doi.org/10.1002/2015JG003140 Charbonnier, Q., Bouchez, J., Gaillardet, J., 2019. Black shale weathering in Mackenzie Basin: constraints from barium isotopes. Charbonnier, Q., Moynier, F., Bouchez, J., 2018. Barium isotope cosmochemistry and geochemistry. Sci. Bull. 63, 385–394. https://doi.org/10.1016/j.scib.2018.01.018 Charette, M.A., Kipp, L.E., Jensen, L.T., Dabrowski, J.S., Whitmore, L.M., in prep. The Transpolar Drift as a source of riverine and shelf-derived trace elements to the central Arctic Ocean. J. Geophys. Res. Oceans.

Charette, M.A., Lam, P.J., Lohan, M.C., Kwon, E.Y., Hatje, V., Jeandel, C., Shiller, A.M., Cutter, G.A., Thomas, A., Boyd, P.W., Homoky, W.B., Milne, A., Thomas, H., Andersson, P.S., Porcelli, D., Tanaka, T., Geibert, W., Dehairs, F., Garcia- Orellana, J., 2016. Coastal ocean and shelf-sea biogeochemical cycling of trace elements and isotopes: lessons learned from GEOTRACES. Philos. Trans. R. Soc. Math. Phys. Eng. Sci. 374, 20160076. https://doi.org/10.1098/rsta.2016.0076 Charkin, A.N., Rutgers van der Loeff, M., Shakhova, N.E., Gustafsson, Ö., Dudarev, O.V., Cherepnev, M.S., Salyuk, A.N., Koshurnikov, A.V., Spivak, E.A., Gunar, A.Y., Ruban, A.S., Semiletov, I.P., 2017. Discovery and characterization of submarine groundwater discharge in the Siberian Arctic seas: a case study in the Buor-Khaya Gulf, Laptev Sea. The Cryosphere 11, 2305–2327. https://doi.org/10.5194/tc-11-2305-2017 Chow, T.J., Goldberg, E.D., 1960. On the marine geochemistry of barium. Geochim. Cosmochim. Acta 20, 192–198. https://doi.org/10.1016/0016-7037(60)90073-9 Colbert, D., McManus, J., 2005. Importance of seasonal variability and coastal processes on estuarine manganese and barium cycling in a Pacific Northwest estuary. Cont. Shelf Res. 25, 1395–1414. https://doi.org/10.1016/j.csr.2005.02.003 Cullen, J.T., Sherrell, R.M., 1999. Techniques for determination of trace metals in small samples of size-fractionated particulate matter: phytoplankton metals off central California. Mar. Chem. 67, 233–247. https://doi.org/10.1016/S0304- 4203(99)00060-2 Cutter, G., Andersson, P.S., Codispoti, L.A., Croot, P., Francois, R., Lohan, M., Obata, Hajime., Rutgers van der Loeff, M., 2014. Sampling and Sample-handling Protocols for GEOTRACES Cruises Version 2. http://www.geotraces.org/library- 88/scientific-publications/reports/169-sampling-and-sample-handling-protocols- for-geotraces-cruises Cutter, G., Kadko, D., Landing, W., 2019. Bottle data from the GEOTRACES Clean Carousel sampling system (GTC) on the Arctic Section cruise (HLY1502) from August to October 2015 (U.S. GEOTRACES Arctic project). Dataset Version 2019-07-29. https://doi.org/10.1575/1912/bco-dmo.647259.4 Dehairs, F., Chesselet, R., Jedwab, J., 1980. Discrete suspended particles of barite and the barium cycle in the open ocean. Earth Planet. Sci. Lett. 49, 528–550. https://doi.org/10.1016/0012-821X(80)90094-1 Dymond, J., Suess, E., Lyle, M., 1992. Barium in deep-sea sediment: a geochemical proxy for paleoproductivity 19. Eagle, M., Paytan, A., Arrigo, K.R., van Dijken, G., Murray, R.W., 2003. A comparison between excess barium and barite as indicators of carbon export. Paleoceanography 18, 1021. https://doi.org/10.1029/2002PA000793 Edmonds, H.N., Michael, P.J., Baker, E.T., Connelly, D.P., Snow, J.E., Langmuir, C.H., Dick, H.J.B., Mühe, R., German, C.R., Graham, D.W., 2003. Discovery of abundant hydrothermal venting on the ultraslow-spreading Gakkel ridge in the Arctic Ocean. Nature 421, 252–256. https://doi.org/10.1038/nature01351 Eickmann, B., Thorseth, I.H., Peters, M., Strauss, H., Bröcker, M., Pedersen, R.B., 2014. Barite in hydrothermal environments as a recorder of subseafloor processes: a

multiple-isotope study from the Loki’s Castle vent field. Geobiology 12, 308– 321. https://doi.org/10.1111/gbi.12086 Fichot, C.G., Kaiser, K., Hooker, S.B., Amon, R.M.W., Babin, M., Belanger, S., Walker, S.A., Benner, R., 2013. Pan-Arctic distributions of continental runoff in the Arctic Ocean. Sci. Rep. 3, 1053. https://doi.org/10.1038/srep01053 Fransson, A., Chierici, M., Anderson, L.G., Bussmann, I., Kattner, G., Peter Jones, E., Swift, J.H., 2001. The importance of shelf processes for the modification of chemical constituents in the waters of the Eurasian Arctic Ocean: Implication for carbon fluxes. Cont. Shelf Res. 21, 225–242. https://doi.org/10.1016/S0278- 4343(00)00088-1 Ganeshram, R.S., François, R., Commeau, J., Brown-Leger, S.L., 2003. An experimental investigation of barite formation in seawater. Geochim. Cosmochim. Acta 67, 2599–2605. https://doi.org/10.1016/S0016-7037(03)00164-9 Gong, D., Pickart, R.S., 2016. Early summer water mass transformation in the eastern Chukchi Sea. Deep Sea Res. Part II Top. Stud. Oceanogr. 130, 43–55. https://doi.org/10.1016/j.dsr2.2016.04.015 Guay, C.K., Kenison Falkner, K., 1997. Barium as a tracer of Arctic halocline and river waters. Deep Sea Res. Part II Top. Stud. Oceanogr. 44, 1543–1569. https://doi.org/10.1016/S0967-0645(97)00066-0 Guay, C.K.H., McLaughlin, F.A., Yamamoto-Kawai, M., 2009. Differentiating fluvial components of upper Canada Basin waters on the basis of measurements of dissolved barium combined with other physical and chemical tracers. J. Geophys. Res. 114. https://doi.org/10.1029/2008JC005099 Haine, T.W.N., Curry, B., Gerdes, R., Hansen, E., Karcher, M., Lee, C., Rudels, B., Spreen, G., de Steur, L., Stewart, K.D., Woodgate, R., 2015. Arctic freshwater export: Status, mechanisms, and prospects. Glob. Planet. Change 125, 13–35. https://doi.org/10.1016/j.gloplacha.2014.11.013 Hendry, K.R., Pyle, K.M., Barney Butler, G., Cooper, A., Fransson, A., Chierici, M., Leng, M.J., Meyer, A., Dodd, P.A., 2018. Spatiotemporal Variability of Barium in Arctic Sea-Ice and Seawater. J. Geophys. Res. Oceans 123, 3507–3522. https://doi.org/10.1029/2017JC013668 Holmes, R.M., McClelland, J.W., Tank, S.E., Spencer, R.G.M., Shiklomanov, A.I., 2018. Arctic Great Rivers Observatory. Water Quality Dataset, Version 20181010. Honjo, S., Krishfield, R.A., Eglinton, T.I., Manganini, S.J., Kemp, J.N., Doherty, K., Hwang, J., McKee, T.K., Takizawa, T., 2010. Biological pump processes in the cryopelagic and hemipelagic Arctic Ocean: Canada Basin and Chukchi Rise. Prog. Oceanogr. 85, 137–170. https://doi.org/10.1016/j.pocean.2010.02.009 Horner, T.J., Kinsley, C.W., Nielsen, S.G., 2015. Barium-isotopic fractionation in seawater mediated by barite cycling and oceanic circulation. Earth Planet. Sci. Lett. 430, 511–522. https://doi.org/10.1016/j.epsl.2015.07.027 Hsieh, Y.-T., Henderson, G.M., 2017. Barium stable isotopes in the global ocean: Tracer of Ba inputs and utilization. Earth Planet. Sci. Lett. 473, 269–278. https://doi.org/10.1016/j.epsl.2017.06.024

Jacquet, S.H.M., Dehairs, F., Cardinal, D., Navez, J., Delille, B., 2005. Barium distribution across the Southern Ocean frontal system in the Crozet–Kerguelen Basin. Mar. Chem. 95, 149–162. https://doi.org/10.1016/j.marchem.2004.09.002 Jeandel, C., Peucker-Ehrenbrink, B., Jones, M.T., Pearce, C.R., Oelkers, E.H., Godderis, Y., Lacan, F., Aumont, O., Arsouze, T., 2011. Ocean margins: The missing term in oceanic element budgets? Eos Trans. Am. Geophys. Union 92, 217–224. https://doi.org/10.1029/2011EO260001 Jensen, L.T., Wyatt, N.J., Twining, B.S., Rauschenberg, S., Landing, W.M., Sherrell, R.M., Fitzsimmons, J.N., 2019. Biogeochemical Cycling of Dissolved Zinc in the Western Arctic (Arctic GEOTRACES GN01). Glob. Biogeochem. Cycles 33, 343–369. https://doi.org/10.1029/2018GB005975 Jones, E.P., Rudels, B., Anderson, L.G., 1995. Deep waters of the Arctic Ocean: origins and circulation. Deep Sea Res. Part Oceanogr. Res. Pap. 42, 737–760. https://doi.org/10.1016/0967-0637(95)00013-V Jones, E.P., Swift, J.H., Anderson, L.G., Lipizer, M., Civitarese, G., Falkner, K.K., Kattner, G., McLaughlin, F., 2003. Tracing Pacific water in the North Atlantic Ocean. J. Geophys. Res. Oceans 108. https://doi.org/10.1029/2001JC001141 Kadko, D., Aguilar-Islas, A., Bolt, C., Buck, C.S., Fitzsimmons, J.N., Jensen, L.T., Landing, W.M., Marsay, C.M., Rember, R., Shiller, A.M., Whitmore, L.M., Anderson, R.F., 2019. The residence times of trace elements determined in the surface Arctic Ocean during the 2015 US Arctic GEOTRACES expedition. Mar. Chem. 208, 56–69. https://doi.org/10.1016/j.marchem.2018.10.011 Kipp, L.E., Charette, M.A., Moore, W.S., Henderson, P.B., Rigor, I.G., 2018. Increased fluxes of shelf-derived materials to the central Arctic Ocean. Sci. Adv. 4, eaao1302. https://doi.org/10.1126/sciadv.aao1302 Kipp, L.E., Kadko, D.C., Pickart, R.S., Henderson, P.B., Moore, W.S., Charette, M.A., 2019. Shelf‐basin interactions and water mass residence times in the Western Arctic Ocean: insights provided by radium isotopes. J. Geophys. Res. Oceans. https://doi.org/10.1029/2019JC014988 Klinkenberg, H., Van Borm, W., Souren, F., 1996. A theoretical adaptation of the classical isotope dilution technique for practical routine analytical determinations by means of inductively coupled plasma mass spectrometry. Spectrochim. Acta Part B At. Spectrosc. 51, 139–153. https://doi.org/10.1016/0584-8547(95)01386-5 Klinkhammer, G.P., Chan, L.H., 1990. Determination of barium in marine waters by isotope dilution inductively coupled plasma mass spectrometry. Anal. Chim. Acta 232, 323–329. Klunder, M.B., Laan, P., Middag, R., de Baar, H.J.W., Bakker, K., 2012. Dissolved iron in the Arctic Ocean: Important role of hydrothermal sources, shelf input and scavenging removal: Dissoved iron in the Arctic Ocean. J. Geophys. Res. Oceans 117, C04014. https://doi.org/10.1029/2011JC007135 Kondo, Y., Obata, Hajime., Hioki, N., Ooki, A., Nishino, S., Kikuchi, T., Kuma, K., 2016. Transport of trace metals (Mn, Fe, Ni, Zn and Cd) in the western Arctic Ocean (Chukchi Sea and Canada Basin) in late summer 2012. Deep Sea Res. Part Oceanogr. Res. Pap. 116, 236–252. https://doi.org/10.1016/j.dsr.2016.08.010

Lalande, C., Forest, A., Barber, D.G., Gratton, Y., Fortier, L., 2009. Variability in the annual cycle of vertical particulate organic carbon export on Arctic shelves: Contrasting the Laptev Sea, Northern Baffin Bay and the Beaufort Sea. Cont. Shelf Res. 29, 2157–2165. https://doi.org/10.1016/j.csr.2009.08.009 Lam, P.J., Marchal, O., 2015. Insights into Particle Cycling from Thorium and Particle Data. Annu. Rev. Mar. Sci. 7, 159–184. https://doi.org/10.1146/annurev-marine- 010814-015623 Lanos, R., 2009. Circulation regionale, masses d’eau, cycles d’evolution et transports entre la mer de Beaufort et le golfe d’Amundsen (Ph.D. Thesis). University du Quebec, Quebec, Canada. LeBlond, P.H., 1980. On the Surface Circulation in Some Channels of the Canadian Arctic Archipelago. ARCTIC 33, 189–197. https://doi.org/10.14430/arctic2554 Lecher, A., 2017. Groundwater Discharge in the Arctic: A Review of Studies and Implications for Biogeochemistry. Hydrology 4, 41. https://doi.org/10.3390/hydrology4030041 Lecher, A.L., Kessler, J., Sparrow, K., Garcia-Tigreros Kodovska, F., Dimova, N., Murray, J., Tulaczyk, S., Paytan, A., 2016. Methane transport through submarine groundwater discharge to the North Pacific and Arctic Ocean at two Alaskan sites: SGD methane transport. Limnol. Oceanogr. 61, S344–S355. https://doi.org/10.1002/lno.10118 Macdonald, R.W., Barrie, L.A., Bidleman, T.F., Diamond, M.L., Gregor, D.J., Semkin, R.G., Strachan, W.M.J., Li, Y.F., Wania, F., Alaee, M., Alexeeva, L.B., Backus, S.M., Bailey, R., Bewers, J.M., Gobeil, C., Halsall, C.J., Harner, T., Hoff, J.T., Jantunen, L.M.M., Lockhart, W.L., Mackay, D., Muir, D.C.G., Pudykiewicz, J., Reimer, K.J., Smith, J.N., Stern, G.A., Schroeder, W.H., Wagemann, R., Yunker, M.B., 2000. Contaminants in the Canadian Arctic: 5 years of progress in understanding sources, occurrence and pathways. Sci. Total Environ. 254, 93– 234. https://doi.org/10.1016/S0048-9697(00)00434-4 Marsay, C.M., Aguilar-Islas, A., Fitzsimmons, J.N., Hatta, M., Jensen, L.T., John, S.G., Kadko, D., Landing, W.M., Lanning, N.T., Morton, P.L., Pasqualini, A., Rauschenberg, S., Sherrell, R.M., Shiller, A.M., Twining, B.S., Whitmore, L.M., Zhang, R., Buck, C.S., 2018. Dissolved and particulate trace elements in late summer Arctic melt ponds. Mar. Chem. 204, 70–85. https://doi.org/10.1016/j.marchem.2018.06.002 Martinez-Ruiz, F., Paytan, A., Gonzalez-Muñoz, M.T., Jroundi, F., Abad, M.M., Lam, P.J., Bishop, J.K.B., Horner, T.J., Morton, P.L., Kastner, M., 2019. Barite formation in the ocean: Origin of amorphous and crystalline precipitates. Chem. Geol. 511, 441–451. https://doi.org/10.1016/j.chemgeo.2018.09.011 McClelland, J.W., Holmes, R.M., Dunton, K.H., Macdonald, R.W., 2012. The Arctic Ocean Estuary. Estuaries Coasts 35, 353–368. https://doi.org/10.1007/s12237- 010-9357-3 McLaughlin, F.A., Carmack, E.C., Ingram, R.G., Williams, W.J., Michel, C., 2004. Chapter 31. Oceanography of the Northwest Passage (26,P), in: The Global Coastal OCean, Interdisciplinary Regional Studes and Syntheses. Harvard university Press, pp. 1211–1242.

McManus, J., Berelson, W.M., Klinkhammer, G.P., Kilgore, T.E., Hammond, D.E., 1994. Remobilization of barium in continental margin sediments. Geochim. Cosmochim. Acta 58, 4899–4907. https://doi.org/10.1016/0016-7037(94)90220-8 Melling, H., 2000. Exchanges of freshwater through the shallow straits of the North American Arctic, in: The Freshwater Budget of the Arctic Ocean, NATO Science Series. Kluwer Academic Publishers, The Netherlands, pp. 479–502. Michel, C., Hamilton, J., Hansen, E., Barber, D., Reigstad, M., Iacozza, J., Seuthe, L., Niemi, A., 2015. Arctic Ocean outflow shelves in the changing Arctic: A review and perspectives. Prog. Oceanogr. 139, 66–88. https://doi.org/10.1016/j.pocean.2015.08.007 Middag, R., de Baar, H.J.W., Laan, P., Bakker, K., 2009. Dissolved aluminium and the silicon cycle in the Arctic Ocean. Mar. Chem. 115, 176–195. https://doi.org/10.1016/j.marchem.2009.08.002 Millero, F.J., 1982. The effect of pressure on the solubility of minerals in water and seawater. Geochim. Cosmochim. Acta 46, 11–22. https://doi.org/10.1016/0016- 7037(82)90286-1 Milliman, J.D., Farnsworth, K.L., 2013. River Discharge to the Coastal Ocean: a global synthesis. Cambridge University Press, Cambridge, United Kingdom. Miyazaki, T., Kimura, J.-I., Chang, Q., 2014. Analysis of stable isotope ratios of Ba by double-spike standard-sample bracketing using multiple-collector inductively coupled plasma mass spectrometry. J. Anal. At. Spectrom. 29, 483. https://doi.org/10.1039/c3ja50311a Monnin, C., Jeandel, C., Cattaldo, T., Dehairs, F., 1999. The marine barite saturation state of the world’s oceans. Mar. Chem. 65, 253–261. https://doi.org/10.1016/S0304-4203(99)00016-X Nan, X.-Y., Yu, H.-M., Rudnick, R.L., Gaschnig, R.M., Xu, J., Li, W.-Y., Zhang, Q., Jin, Z.-D., Li, X.-H., Huang, F., 2018. Barium isotopic composition of the upper continental crust. Geochim. Cosmochim. Acta 233, 33–49. https://doi.org/10.1016/j.gca.2018.05.004 Newton, R., Schlosser, P., Mortlock, R., Swift, J., MacDonald, R., 2013. Canadian Basin freshwater sources and changes: Results from the 2005 Arctic Ocean Section: AOS 2005 freshwater sources and changes. J. Geophys. Res. Oceans 118, 2133– 2154. https://doi.org/10.1002/jgrc.20101 Nurnberg, D., 1996. Biogenic barium and opal in shallow Eurasian shelf sediments in relation to the pelagic Arctic Ocean environment. Rep Polar Res 212, 96–118. Ober, S., Rijkenberg, M.J.A., Gerringa, L.J.A., 2016. Physical oceanography measured with ultra clean CTD/Water sampler-system during POLARSTERN cruise PS94 (ARK-XXIX/3). R. Neth. Inst. Sea Res. Texel. https://doi.org/10.1594/PANGAEA.859560 Ohnemus, D.C., Auro, M.E., Sherrell, R.M., Lagerström, M., Morton, P.L., Twining, B.S., Rauschenberg, S., Lam, P.J., 2014. Laboratory intercomparison of marine particulate digestions including Piranha: a novel chemical method for dissolution of polyethersulfone filters. Limnol. Oceanogr. Methods 12, 530–547. https://doi.org/10.4319/lom.2014.12.530

Östlund, H.G., Hut, G., 1984. Arctic Ocean water mass balance from isotope data. J. Geophys. Res. 89, 6373. https://doi.org/10.1029/JC089iC04p06373 Peterson, I., Hamilton, J., Prinsenberg, S., Pettipas, R., 2012. Wind-forcing of volume transport through Lancaster Sound: transport through Lancaster Sound. J. Geophys. Res. Oceans 117. https://doi.org/10.1029/2012JC008140 Planquette, H., Sherrell, R.M., 2012. Sampling for particulate trace element determination using water sampling bottles: methodology and comparison to in situ pumps: Particulate trace element sampling. Limnol. Oceanogr. Methods 10, 367–388. https://doi.org/10.4319/lom.2012.10.367 Prinsenberg, S., Hamilton, J., Peterson, I., Pettipas, R., 2009. Observing and interpreting the seasonal variability of the oceanographic fluxes passing through Lancaster Sound of the Canadian Arctic Archipelago, in: Nihoul, J.C.J., Kostianoy, A.G. (Eds.), Influence of Climate Change on the Changing Arctic and Sub-Arctic Conditions. Springer Netherlands, Dordrecht, pp. 125–143. https://doi.org/10.1007/978-1-4020-9460-6_10 Pyle, K.M., Hendry, K.R., Sherrell, R.M., Legge, O., Hind, A.J., Bakker, D., Venables, H., Meredith, M.P., 2018. Oceanic fronts control the distribution of dissolved barium in the Southern Ocean. Mar. Chem. 204, 95–106. https://doi.org/10.1016/j.marchem.2018.07.002 R Core Team, 2018. R: A language and environment for statistical computing. R foundation for Statistical computing, Vienna, Austria. Rember, R., 2018. Dissolved barium measured on water bottle samples during POLARSTERN cruise PS94 (ARK-XXIX/3) to the central Arctic ocean in 2015. PANGAEA. https://doi.org/10.1594/PANGAEA.896022 Roeske, T., Bauch, D., Rutgers van der Loeff, M., Rabe, B., 2012a. Utility of dissolved barium in distinguishing North American from Eurasian runoff in the Arctic Ocean. Mar. Chem. 132–133, 1–14. https://doi.org/10.1016/j.marchem.2012.01.007 Roeske, T., Rutgers van der Loeff, M., Middag, R., Bakker, K., 2012b. Deep water circulation and composition in the Arctic Ocean by dissolved barium, aluminium and silicate. Mar. Chem. 132–133, 56–67. https://doi.org/10.1016/j.marchem.2012.02.001 Rudels, B., 2018. Arctic Ocean Circulation, in: Reference Module in Earth Systems and Environmental Sciences. Elsevier. https://doi.org/10.1016/B978-0-12-409548- 9.11209-6 Rudels, B., 2015. Arctic Ocean circulation, processes and water masses: A description of observations and ideas with focus on the period prior to the International Polar Year 2007–2009. Prog. Oceanogr. 132, 22–67. https://doi.org/10.1016/j.pocean.2013.11.006 Rudels, B., Quadfasel, D., 1991. Convection and deep water formation in the Arctic Ocean-Greenland Sea System. J. Mar. Syst. 2, 435–450. https://doi.org/10.1016/0924-7963(91)90045-V Rushdi, A.I., McManus, J., Collier, R.W., 2000. Marine barite and celestite saturation in seawater. Mar. Chem. 69, 19–31. https://doi.org/10.1016/S0304-4203(99)00089-4 Schlitzer, R., 2018. Ocean Data View 5.1.5 [WWW Document]. URL https://odv.awi.de

100

Schlosser, P., Bayer, R., Bönisch, G., Cooper, L.W., Ekwurzel, B., Jenkins, W.J., Khatiwala, S., Pfirman, S., Smethie, W.M., 1999. Pathways and mean residence times of dissolved pollutants in the ocean derived from transient tracers and stable isotopes. Sci. Total Environ. 237–238, 15–30. https://doi.org/10.1016/S0048- 9697(99)00121-7 Shaw, T.J., Moore, W.S., Kloepfer, J., Sochaski, M.A., 1998. The flux of barium to the coastal waters of the southeastern USA: the importance of submarine groundwater discharge. Geochim. Cosmochim. Acta 62, 3047–3054. https://doi.org/10.1016/S0016-7037(98)00218-X Shiller, A., 2019. Dissolved Ba, Cd, Cu, Ga, Mn, Ni, and V concentration data from the US GEOTRACES Arctic Expeditions (GN01, HLY1502) from August to October 2015. Biol. Chem. Oceanogr. Data Manag. Off. BCO-DMO Dataset version 2019-07-09. https://doi.org/10.1575/1912/bco-dmo.772645.1 Siebert, C., Nägler, T.F., Kramers, J.D., 2001. Determination of molybdenum isotope fractionation by double-spike multicollector inductively coupled plasma mass spectrometry. Geochem. Geophys. Geosystems 2, 2000GC00124. https://doi.org/10.1029/2000GC000124 Steele, M., Morison, J., Ermold, W., Rigor, I., Ortmeyer, M., Shimada, K., 2004. Circulation of summer Pacific halocline water in the Arctic Ocean. J. Geophys. Res. 109, C02027. https://doi.org/10.1029/2003JC002009 Tang, C.C.L., Ross, C.K., Yao, T., Petrie, B., DeTracey, B.M., Dunlap, E., 2004. The circulation, water masses and sea-ice of Baffin Bay. Prog. Oceanogr. 63, 183– 228. https://doi.org/10.1016/j.pocean.2004.09.005 Tanhua, T., Jones, E.P., Jeansson, E., Jutterström, S., Smethie, W.M., Wallace, D.W.R., Anderson, L.G., 2009. Ventilation of the Arctic Ocean: Mean ages and inventories of anthropogenic CO 2 and CFC-11. J. Geophys. Res. 114, C01002. https://doi.org/10.1029/2008JC004868 Taylor, J.R., Falkner, K.K., Schauer, U., Meredith, M., 2003. Quantitative considerations of dissolved barium as a tracer in the Arctic Ocean. J. Geophys. Res. 108, 3374. https://doi.org/10.1029/2002JC001635 Taylor, S.R., McLennan, S.M., 1995. The geochemical evolution of the continental crust. Rev. Geophys. 33, 241. https://doi.org/10.1029/95RG00262 Thomas, H., Shadwick, E., Dehairs, F., Lansard, B., Mucci, A., Navez, J., Gratton, Y., Prowe, F., Chierici, M., Fransson, A., Papakyriakou, T.N., Sternberg, E., Miller, L.A., Tremblay, J.-É., Monnin, C., 2011. Barium and carbon fluxes in the Canadian Arctic Archipelago. J. Geophys. Res. 116, C00G08. https://doi.org/10.1029/2011JC007120 von Allmen, K., Böttcher, M.E., Samankassou, E., Nägler, T.F., 2010. Barium isotope fractionation in the global barium cycle: First evidence from barium minerals and precipitation experiments. Chem. Geol. 277, 70–77. https://doi.org/10.1016/j.chemgeo.2010.07.011 Wadley, M.R., Bigg, G.R., 2002. Impact of flow through the Canadian Archipelago and Bering Strait on the North Atlantic and Arctic circulation: An ocean modelling study. Q. J. R. Meteorol. Soc. 128, 2187–2203. https://doi.org/10.1256/qj.00.35

101

Webster, R.K., 1960. Mass spectrometric isotope dilution analysis, in: Methods in Geochemistry. Interscience, London, pp. 202–246. Weingartner, T.J., Cavalieri, D.J., Aagaard, K., Sasaki, Y., 1998. Circulation, dense water formation, and outflow on the northeast Chukchi Shelf. J. Geophys. Res. Oceans 103, 7647–7661. https://doi.org/10.1029/98JC00374 Whitmore, L.M., Morton, P.L., Twining, B.S., Shiller, A.M., 2019. Vanadium cycling in the Western Arctic Ocean is influenced by shelf-basin connectivity. Mar. Chem. 216, 103701. https://doi.org/10.1016/j.marchem.2019.103701 Woodgate, R.A., Weingartner, T.J., Lindsay, R., 2012. Observed increases in Bering Strait oceanic fluxes from the Pacific to the Arctic from 2001 to 2011 and their impacts on the Arctic Ocean water column. Geophys. Res. Lett. 39. https://doi.org/10.1029/2012GL054092 Xiang, Y., Lam, P.J., submitted. Size-fractionated marine suspended particle dynamics in the Western Arctic Ocean from the U.S. GEOTRACES Arctic cruise (GN01). Yamamoto-Kawai, M., Carmack, E.C., McLaughlin, F.A., Falkner, K.K., 2010. Oxygen isotope ratio, barium and salinity in waters around the North American coast from the Pacific to the Atlantic: Implications for freshwater sources to the Arctic throughflow. J. Mar. Res. 68, 97–117. https://doi.org/10.1357/002224010793078988 Yang, Q., Dixon, T.H., Myers, P.G., Bonin, J., Chambers, D., van den Broeke, M.R., Ribergaard, M.H., Mortensen, J., 2016. Recent increases in Arctic freshwater flux affects Labrador Sea convection and Atlantic overturning circulation. Nat. Commun. 7, 10525. https://doi.org/10.1038/ncomms10525

102

CHAPTER IV – VANADIUM CYCLING IN THE WESTERN ARCTIC OCEAN IS

INFLUENCED BY SHELF-BASIN CONNECTIVITY

Co-Authors: Peter Morton, Benjamin Twining, Alan Shiller

Originally published in Marine Chemistry, reference:

L.M Whitmore, P.L. Morton, B.S. Twining, A.M Shiller. (2019). Vanadium cycling in the Western Arctic Ocean is influenced by shelf-basin connectivity. Marine Chemistry,

216: 103701. DOI: 10.1016/j.marchem.2019.103701

4.1 Abstract

Water in the western Arctic Ocean tends to show lower dissolved vanadium concentrations than profiles observed elsewhere in the open ocean. Dissolved V in

Pacific-derived basin waters was depleted by approximately 15 - 30% from the effective

Pacific Ocean endmember. The depletion originates on western Arctic shelves and is not a result of mixing with a water mass with low V. While biological uptake may account for some of the V removal from the water column, adsorption onto particulate Fe is likely the dominant factor in removing V from shelf waters to the sediments. Once in the sediments, reduction should result in sequestering the V while Fe (and Mn) can be remobilized. A similar Fe-shuttling mechanism for V was previously described for the

Peru margin (Scholz et al., 2011). Off the shelves, particulate Mn concentrations often exceed particulate Fe concentrations and thus may exert greater control on the V distribution in basin waters. Nonetheless, particulate V concentrations are much lower in basin waters and dissolved V thus behaves largely conservatively away from the shelf environment. Dissolved V concentrations in Atlantic-derived and Arctic deep waters were as much as 5 nmol/kg lower than those observed in deep waters of other ocean

103

basins. The uniformity in deep water dissolved V between the sampled basins suggests that slow removal of V from the deep basins is probably not a factor in the deep water depletion. Vanadium-depleted incoming Atlantic waters (i.e., the source of Arctic deep waters) and/or removal of vanadium from incoming waters that pass over the shelves probably accounts for the deep water dissolved V depletion. Overall, our results demonstrate the utility of the V distribution as an additional tool to help understand the

Arctic marine system. Furthermore, our work is pertinent to questions related to the net effect of marginal basin shelves on oceanic vanadium cycling, its isotopic balance, and how climate-induced changes in shelf biogeochemical cycling will impact vanadium cycling.

4.2 Introduction

The Arctic Ocean plays a vital role in global climate and global ocean dynamics.

As Atlantic waters enter the Arctic they cool and sink, a critical process in the ocean’s meridional overturning circulation. The sinking waters export materials from the surface to the ocean’s interior where they can be preserved for long periods of time. Waters entering the Arctic basins are modified through shelf interactions such as sediment-water exchange, local biological uptake and remineralization, and riverine inputs. Previous research has shown that the extent and distribution of these shelf modifications may impact biogeochemical cycling (Charette et al., 2016; Fransson et al., 2001; Macdonald and Gobeil, 2012). Furthermore, the global connectivity of shelf-basin systems is currently an important study area since there is evidence that inputs from ocean margins may be large enough to account for imbalances in oceanic elemental budgets (Charette et al., 2016; Jeandel et al., 2011). Jakobsson (2002) indicates that the shelves make up

104

slightly more than half the area of the Arctic Ocean and this large shelf-to-basin ratio makes it an ideal region to study the role of ocean margins on basin geochemical signatures and cycling. Currently, the extent to which the shelves impart modifications on

Arctic water masses, the depth of that influence, and the mechanisms driving modification remain poorly described.

There is further uncertainty regarding the influence of shelf-derived waters on

Arctic deep waters partly because the processes driving geochemical signatures in Arctic deep waters are poorly understood. The Arctic Ocean basin is delineated by several topographic features. The Lomonosov Ridge divides the region into two primary basins

(the Western Arctic Basin and the Eurasian Basin), both of which are further divided by

18 other underwater ridges (Figure 4.1). Bauch et al. (1995) argued that the different H2 O isotopic compositions of the Western Arctic Basin and Eurasian Basin deep waters reflected differences in shelf water contributions to the basins’ deep waters. Supporting the role of shelves in setting the deep water geochemistry, Roeske et al. (2012) concluded that deep water Ba concentrations were associated with Ba dissolution from shelf derived particles. Additionally, Klunder et al. (2012) indicated that the patterns in dissolved Fe in the Arctic Ocean were very basin specific and that basins with high dissolved Fe concentrations may have more resuspension at the slopes or downslope convective mixing slope interactions relative to basins with lower dissolved Fe concentrations.

Further constraints are necessary to identify the relative contributions of shelf and slope processes in influencing Arctic deep water signatures.

Previous work has demonstrated that vanadium (V) cycling is linked to shelf environment and sediment processes, such as reactions associated with redox state (Beck

105

et al., 2010; Reckhardt et al., 2017; Scholz et al., 2011; Shiller and Mao, 1999). Redox driven V removal has been observed and described (Emerson and Huested, 1991;

Reckhardt et al., 2017) but has not yet been connected to the open ocean V cycle. In pelagic environments, a slight surface water dissolved V (dV) depletion is observed (e.g.,

Collier, 1984; Ho et al., 2018). Shiller and Mao (1999) suggested that the open ocean surface depletion of dV might be derived from advected shelf waters with low dV concentrations.

In contrast, open ocean studies ascribe the surface dV depletion to biological uptake and export (Collier, 1984; Klein et al., 2013; Middelburg et al., 1988). Indeed, some studies have demonstrated biological uptake of V (Crans et al., 2004; Osterholz et al., 2014). Vanadium may be biologically incorporated and is associated with enzymes such as V-haloperoxidases or V-nitrogenases, although V-nitrogenases have not yet been observed in the marine environment (Crans et al., 2004 and references therein). Klein et al. (2013) concluded that biological V removal associated with diatom blooms was a significant removal mechanism for dV locally and accumulation of V by diazotrophs has been reported (Nuester et al., 2012). Additionally, V is hyperaccumulated by some ascidians, although the physiological role isn’t well understood (Crans et al., 2004 and references therein).

Beyond redox processes and biological uptake, in the oxic marine environment dV can be scavenged onto surfaces of particles such as iron (Fe) or manganese (Mn) oxyhydroxides (Bauer et al., 2017; Brinza et al., 2008; Koschinsky and Hein, 2003;

Trefry and Metz, 1989). Particle interactions could be an important removal mechanism

106

for dV; oxides formed in the water column sorb dV and carry it to the sediments where it may be reduced to a less soluble species (Scholz et al., 2017, 2011).

Since dV has been shown to be significantly modified by shelf-associated processes (Reckhardt et al., 2017; Scholz et al., 2011; Wehrli and Stumm, 1989), basin- wide dV distributions may provide a constraint on the extent to which a water mass has been modified by the shelf. This is potentially useful for predicting nutrient stoichiometry, source chemistry, and valuable for monitoring biogeochemical changes in the Arctic Ocean due to climate change. While Shiller and Mao (1999) suggested the use of dV depletions as a tracer of shelf-derived waters, understanding the impact of this process on dV distributions in the Arctic Ocean requires better knowledge of the dV input and removal processes over the shelves and deep basins.

This study describes the first dV transect in the Arctic Ocean using samples collected during the 2015 U.S. GEOTRACES Arctic Section (GN01). This is an important step in determining the utility of dV as a tracer of shelf influences. Herein we discuss (1) the mechanisms influencing dV distribution in the basins and on the shelves,

(2) the extent of shelf modification of dV, and (3) the role of shelf waters in affecting

Arctic deep water composition. Our Arctic Ocean dV distributions are discussed in both local and global contexts.

4.3 Methods

4.3.1 Section Introduction and Geography

The 2015 US GEOTRACES Arctic Section (GN01) was conducted from August

9, 2015 to October 12, 2015 aboard the USCGC Healy. The cruise was comprised of two sections in the Western Arctic Ocean, one in the Makarov Basin and the other in the

107

Canada Basin. The cruise was operated by U.S. GEOTRACES in conjunction with

CLIVAR. Trace metal samples described in this study (i.e., dissolved V, particulate V, particulate Fe, and particulate Mn) were collected by GEOTRACES scientists at stations identified in Figure 4.1; data are accessible at the Biological and Chemical Oceanography

Data Management Office (BCO-DMO) database (Shiller, 2019; Twining et al., 2019).

Each GEOTRACES basin station had a deep and a shallow cast that overlapped at 500 m, for a total of 23 depths. Data from the CLIVAR contingent was used to increase resolution for ancillary parameters such as nutrients, salinity and temperature (at 0.5° latitude resolution in the basins). Pigment samples were collected by the GEOTRACES contingent from shallow depths (<100 m) at GEOTRACES stations and results were provided by cruise leadership. Ancillary data used in this study (e.g., nutrients, salinity, temperature) are located in the BCO-DMO database (Landing et al., 2016, 2017).

The western Arctic Ocean is divided from the Eurasian sector by the Lomonosov

Ridge. Our study was predominantly in the Western Arctic Ocean, which is further divided into the Canada and Makarov Basins by the Alpha-Mendeleev Ridge (Figure

4.1). These basins have a mean depth of approximately 4000 m (Aagaard, 1981). The large shelf seas of the Arctic Ocean also influence the distribution and composition of waters in the Arctic Ocean Basins. Our study includes samples in the Bering and Chukchi

Seas, but basin waters described in our study may also be influenced by the East Siberian

Sea or the Beaufort Sea.

108

Figure 4.1 Arctic Ocean Geography & Section Description

(A) Map of GEOTRACES stations (blue dots) during the 2015 GN01 cruise. Numbers denote important geographic features. Gray arrows indicate major surface circulation. Dashed and solid light gray lines indicate bathymetry.

(B) Cross section of the Arctic Ocean, with major water masses denoted. Geographic features are numbered as referenced in panel A.

The side map indicates the location of the cross section in the basin

4.3.2 Section Hydrography

Circulation of surface waters in the western Arctic basins is influenced by atmospheric oscillations (Proshutinsky and Johnson, 1997; Steele, 2004). The Beaufort 109

Gyre and the Transpolar Drift (TPD) are the dominant surface circulation features in the

GN01 section (Figure 4.1). The TPD is a major surface current that bisects the Arctic

Ocean. It generally originates in the East Siberian Arctic Sea and flows across the central

Arctic toward the Fram Strait. In the GN01 section, at latitudes > 85°N and depths < 50 m the presence of the Transpolar Drift (TPD) has been reported (Kipp et al., 2018).

Water column structure in the Arctic basins is complicated by transformation of seawater by brines formed when seawater freezes and by additional inputs of freshwater from rivers and sea ice melt. The presence of water types with different temperatures and salinities leads to a layering of water masses (Figure 4.1B) across a large density gradient, predominantly driven by the salinity structure. In this study, water masses were defined primarily using criteria described in Steele et al. (2004) and Talley et al. (2011), and by assessing the T-S plot of our study region with respect to previous literature

(Figure 4.2A).

The surface waters (the Polar Mixed Layer) of the Arctic Ocean are freshened by seasonal sea ice melt and river discharge; freezing throughout winter alters the composition of this layer via the formation of brines (Talley et al., 2011). In this study,

3 we describe surface waters as water above the σθ = 24 kg/m isopycnal (i.e., above the

Pacific halocline).

The Pacific halocline is comprised of waters that have entered the Arctic Ocean through the Bering Strait. Chemical properties of these waters get strongly modified by the shelf (Fransson et al., 2001; Vieira et al., 2019). The Pacific halocline is often discussed in terms of three water types, summer Bering Sea Water (sBSW), winter

Bering Sea Water (wBSW), and Alaska Coastal Water (ACW) (Steele, 2004). In this

110

study, we refer to the Pacific halocline as one unit unless otherwise indicated, meaning that where there is the presence of a warm halocline layer (sBSW or ACW) it is not discussed separately from wBSW. As such, the Pacific halocline is defined as 24 kg/m3 <

3 σθ < 27 kg/m . Where we discuss the influence of sBSW versus wBSW, sBSW is defined

3 3 3 3 as 24 kg/m < σθ < 26 kg/m and wBSW is 26 kg/m < σθ < 27 kg/m (Figure 4.2B, C).

Notably, the lateral extent of the Pacific halocline is variable and it is typically constrained to the Western Arctic basins (Alkire et al., 2015; Steele, 2004). By considering nutrient and dissolved trace element distributions, we identified the furthest extent of the Pacific halocline to be roughly at 85°N for the GN01 transect.

3 Directly below the Pacific halocline is the Atlantic halocline (27 km/m < σθ <

27.6 kg/m3). North Atlantic waters enter the Arctic Ocean via the Fram Strait (sill depth ≈

2500 m) and the Barents Sea (shelf depth ≈ 200 m) (Rudels, 2019; von Appen et al.,

2015). Barents Sea Branch Waters are modified and become more dense than Fram Strait

Branch Waters; in the central Arctic basins, Fram Strait Branch Waters are directly below

3 3 the Atlantic halocline (27.6 kg/m < σθ < 27.95 kg/m ) and Barents Sea Branch Waters extend from below Fram Strait Branch Waters to roughly the sill depth (~ 2000 m) of the

3 3 Lomonosov ridge (27.95 kg/m < σθ < 28.02 kg/m ; Mauldin et al., 2010; Smethie et al.,

111

2000).

Figure 4.2 Water mass descriptions

A) Temperature-salinity plot for the GN01 Section. The gray contours represent isopycnals (σθ). Thick contour lines are the

isopycnal boundaries between water masses: (1) Polar Mixed Layer<24 σθ, (2) Pacific Halocline=24–27 σθ, (3) Atlantic

Halocline=27–27.6 σθ, and (4) Atlantic Waters>27.6 σθ.

B) Shelf and Makarov Basin transect for potential density with the water mass isopycnal separations denoted in contours;

stations used are noted in green on the map to the right of the section plot.

C) Canada Basin transect for potential density with the water mass isopycnal separations denoted in contours; stations used

are noted in green on the map to the right of the section plot. Both GEOTRACES station data and CLIVAR station data

were used to generate all the figures in this panel. Arctic Deep Waters are roughly defined as waters below the sill depth of the

Lomonosov Ridge (Figure 4.1B). The distinct formation mechanisms of Arctic Deep

Water remain unclear. Studies have indicated that the chemical signature of Arctic deep waters may be representative of interactions with the shelf and slope delivered by downslope convective mixing (Bauch et al., 1995; Middag et al., 2009; Roeske et al.,

2012; Rudels and Quadfasel, 1991).

112

4.3.3 Dissolved Sample Collection and Analysis

Water column samples were collected from GO-FLO (General Oceanics) bottles on a trace metal clean rosette (Cutter et al., 2014). The surface cast collected the shallowest sample at 20 m to avoid contamination from the ship. Additional surface samples were then collected at 1 m over the edge of a small boat. Where sea ice was stable, water samples were collected at 1 m, 5 m and 20 m through a hole in the ice.

Small boat and under-ice samples were collected through Teflon coated Tygon tubing using a trace metal clean pump (IWAKI, model WMD-30LFY-115). All samples were passed through pre-cleaned 0.2 µm filters (Acropak-200 or Supor; Pall Corp.) and collected into 125 mL, pre-cleaned (1.2 M HCl, rinsed with ultrapure water) high density polyethylene (HDPE) bottles; small boat and under-ice samples were first collected into large acid-washed carboys and subsampled into 125 mL bottles. Samples were acidified to 0.024 M HCl (Fisher Optima) within 1 – 3 months after sample collection and were stored at room temperature.

Concentrations of dV were determined using a ThermoFisher Element XR inductively coupled plasma mass spectrometer (ICP-MS) introduced with a PC3 spray chamber (Elemental Scientific). Samples were prepared for quantification using an isotope dilution method with preconcentration and matrix removal on a seaFAST system

(Elemental Scientific, Inc.) following Ho et al. (2018). An enriched vanadium spike (50V

= 44.3%; Oak Ridge National Laboratories) was added to each sample at a volume determined from the geometric mean of the natural isotope ratio and the enriched spike isotope ratio. Approx. 14 mL of seawater sample was extracted on the seaFAST and eluted with 1 mL of 1.2 M ultrapure HNO3. Extracted samples were introduced into the

113

ICP-MS using a PC3 spray chamber (Elemental Scientific, Inc.) to obtain the intensities of 50V and 51V. Additionally, 47Ti and 52Cr were monitored to correct for any 50Ti or 50Cr isobaric interference on 50V; the correction was generally <1%. Reproducibility of the method was determined by assessing repeat measurements of U.S. GEOTRACES intercalibration samples (GS and GD) (Appendix Table C1); replicate measurements of

GS and GD between runs had relative standard deviations of 1.8% and 1.4%, respectively

(Appendix Table C1). Furthermore, GS and GD measurements were intercalibrated with those by Ho et al. (2018) and are comparable within the error of our measurements.

4.3.4 Particulate sample collection and analysis

Particulate samples were collected from the same trace metal clean Go-Flo bottles as the dissolved samples, by filtering an average of ~6 L of seawater through an acid- washed Supor polyethersulfone 0.45 µm filter (25 mm or 47 mm) held in a Swinnex or

Advantec filter holder (Twining et al., 2015). Process blanks were also generated by passing 1 – 2 L of filtered (0.2 µm) seawater through unused, acid-cleaned Supor filters.

Sample and process blank filters were analyzed for total concentrations by digesting half of the filter in acid-cleaned PFA vials (Savillex) with 2 mL of a solution of 4 M HCl, 4 M

HNO3 and 4 M HF (all Optima or double-distilled grade) and heating at 110°C for four hours. The digest solution was transferred to a second PFA vial, to which was added 60

µL of 18.4 M H2SO4 (Optima) and 20 µL of 9.8 M H2O2 (Optima) and taken to dryness.

The final residue was redissolved in 0.32 M HNO3 (Optima or double-distilled grade) and a 10 ppb In spike added to account for matrix effects and instrumental drift during ICP-

MS analysis. Full details of the method are described elsewhere (Marsay et al., 2018;

Ohnemus et al., 2014; Twining et al., 2015). Labile element concentrations were

114

determined by leaching the complementary filter half with acetic acid and hydroxylamine-hydrochloride (Berger et al., 2008), and blank corrected using leached process blank filters. Blanks and reference material recoveries are reported in Appendix

Table C2.

4.3.5 Data analysis.

Trends between dV or pV and environmental parameters were regressed using type I or type II reduced major axis linear regression and reported with slope, correlation coefficients, and two-sided p-values in the supplementary material (Appendix Table C3).

Shelf and basin conditions often have separate trends and are discussed separately in the text.

Averages used to describe dV of water masses in the basins were made using water column linear interpolations binned at 1 m intervals and averaging values between isopycnals outlined in section 4.3.2 (Figure 4.2). Alternatively, averages from the shelf stations were made using the discrete samples from each station.

Our discussion of particulate trace metals is divided into three primary components: (1) total particulate concentrations, (2) labile particulate concentrations, and

(3) non-lithogenic particulate concentrations. Total particulate and labile particulate concentrations were determined with the methods above (section 2.4). The lithogenic fraction was determined by applying crustal element/Al ratios (Rudnick and Gao, 2014) and calculating the lithogenic element of interest from the total pAl concentration; the assumption being made is that all pAl is lithogenic in origin. We calculate the non- lithogenic fraction by subtracting the calculated lithogenic fraction from the total particle concentration. In principle, the non-lithogenic and labile fractions determine the non-

115

refractory component of particulate matter: biological material, readily soluble authigenic particles (e.g., oxyhydroxide phases), and/or surface-adsorbed metals (Berger et al., 2008;

Ohnemus and Lam, 2015). We thus expect the non-lithogenic & labile particle concentrations to be roughly equivalent. Differences in these two fractions may arise if the global crustal ratios are not representative of our region or if the Berger leach method did not mobilize all of the non-refractory material when particle loading is high (e.g., at shelf stations).

Section figures were made using a weighted average gridding method in Ocean

Data View 5.1.5 (Schlitzer, 2018).

4.4 Results

4.4.1 Dissolved V (dV) distribution

The occupied shelf stations (stations 2 – 6, 66) exhibited variability in dV concentrations ranging between 26 and 32 nmol/kg (Figure 4.3A, B). The surface waters of the Bering Sea slope station had a dV of 31.5 ± 0.4 nmol/kg (station 1). Continental shelf stations 2 and 3, also in the Bering Sea, had dV equal to 26.1 ± 0.9 and 27.5 ± 1.5, respectively. Despite the lower dV at stations 2 and 3, the Bering Strait station (station 4; dV = 31.4 ± 0.6 nmol/kg) had dV similar to the Bering Sea slope station (station 1)

(Table 4.1). Stations 5 and 6 had concentrations similar to the Bering Strait station

(station 4) and station 66, which was nearer the shelf break, had concentrations similar to stations 2 and 3.

In the occupied Arctic Ocean basin stations, dV had a concentration range from

3 13 – 22 nmol/kg in surface waters (σ θ < 24 kg/m ), with the lowest concentrations in

TPD-derived waters (> 85°N, < 50 m) (Stations 30 – 43; Figure 4.3A, B); the

116

concentrations observed in TPD-associated surface waters were roughly 10 nmol/kg less than basin surface water observations from south of 85°N. All observed surface water (σ θ

3 < 24 kg/m ) concentrations were low relative to the deep waters (29.9 ± 1.1 nmol/kg; σ θ

> 27.9 kg/m3) and had a 25% – 57% depletion with respect to the deep waters. Dissolved

V concentrations increased with depth, and relative to surface waters, dV was slightly higher in the Pacific halocline (dV = 25.4 ± 1.2 nmol/kg; Table 4.1, Figure 4.3). In

Atlantic and Arctic deep waters, dV is fairly constant with an average of 29.9 ± 1.1 nmol/kg (determined using samples below the core of the Atlantic water layer: σ θ > 27.9 kg/m3 for FSBW as defined by Smethie et al., 2000).

Table 4.2 Mean dV concentrations of several water types discussed in the results.

Region Station dV (nmol/kg) mean SD n Bering Sea Slope, surface 1 31.5 0.4 4 Shelf (Bering Sea) 2 26.1 0.9 4 Shelf (Bering Sea) 3 27.5 1.5 4 Shelf (Bering Strait)* 4 31.4 0.6 4 Shelf (Chukchi Sea) 5 30.7 1.0 3 Shelf (Chukchi Sea) 6 30.4 0.5 4 Shelf (Beaufort Sea) 66 26.0 0.6 4 10 - 30, Pacific Halocline† 25.4 1.2 11 48 - 60 Atlantic/Deep Arctic 14 - 57 29.9 1.1 161 Waters * Bering Strait Endmember (effective Pacific endmember)

† An average of each station’s Pacific halocline mean, determined using interpolated profiles.

117

Figure 4.3 Dissolved and particulate V distributions.

(A, B): Dissolved V distributions for (A) the shelf region and Makarov Basin and (B) the Canada Basin.

(C, D): Total particulate V distributions for (C) the shelf region and Makarov Basin and (D) the Canada Basin (the color scale is logarithmic). The location of the transect is indicated in green on the maps below the sections and station numbers are indicated above the sections. The contours are isopycnals delineating water masses as described in Figure 4.2. In panels A & C, caution should be used interpreting the gradient between Station 6 and 10, as there are few data points to inform the gridding. 118

4.4.2 Particulate V (pV) distribution

Particulate trace metals in the Arctic have been minimally sampled and to our knowledge this is the first report of pV in the Arctic. Total pV concentrations were up to

5 orders of magnitude larger on the shelf than in the basins (Shelf/Slope pV: 35 – 11000 pmol/kg (median = 370 pmol/kg); Basin pV: 4 – 292 pmol/kg (median = 27 pmol/kg).

Shelf pV profiles generally had lower concentrations in the surface than at depth

(Stations 2-6, 66; Figure 4.3C, D). Dissolved V concentrations were roughly 3 orders of magnitude greater than pV concentrations. However, when comparing dV and pV distributions, one should remember that the pV concentrations represent the standing stock of suspended particulate material concentrations and not the settling flux. Thus, the magnitudes of changes in dV and pV are not directly comparable (as discussed for Cd and P cycling in Knauer and Martin, 1981), but comparison of the vertical profiles of the two phases are useful for investigating potential sources and internal transformations such as biological uptake and redox processes.

In the basins, vertical profiles of pV consistently showed a subsurface maximum, with decreasing concentrations towards the surface and at depth (Figure 4.3C, D). The basin pV maximum was observed around the 26.6 kg/m3 isopycnal in a density range which encapsulated winter Bering Sea Water (wBSW) of the Pacific halocline and the

3 Atlantic halocline (σ θ ≈ 26 – 27.3 kg/m ). Additionally, elevated pV concentrations were observed along the continental slopes with the highest concentrations on the shelves. The spatial trends observed between the non-lithogenic, labile, and total pV were similar

(Appendix Figure C1).

119

The non-lithogenic fraction of pV represented 47% – 83% (median = 63%) of the total pV on the shelves and 43% – 100% (median = 80%) in the basins. The labile fraction of pV represented 15% – 100% (median = 36%) of total pV on the shelves and

5% - 100% (median = 64%) in the basins. Such a large labile pool indicates a substantial amount of the pV distribution is associated with in situ formation of pV and/or represents a form of pV that may be bioavailable.

4.5 Discussion

4.5.1 Comparison of Arctic dV distribution to other ocean basins.

Outside of the Arctic Ocean, typical oceanic profiles of dV are slightly depleted at the surface and increase with depth to a concentration of 32.5 – 35 nmol/kg. Surface dV is usually depleted by approximately 10% relative to deep waters, although some observations indicate more substantial surface depletions attributed to biological activity

(Collier, 1984; Ho et al., 2018; Klein et al., 2013). The surface depletion we observed in

3 the Western Arctic Ocean was 25 – 57%. The deep water (σ θ > 27.87 kg/m ) dV in the

Western Arctic Ocean basins (29.9 ± 1.1 nmol/kg) is up to 5 nmol/kg lower than previous observations in the Pacific and North Atlantic Oceans (i.e., Collier, 1984; Ho et al., 2018;

Klein et al., 2013; Middelburg et al., 1988; Figure 4.4).

The differences in dV distribution compared to the Pacific and Atlantic Oceans may stem from the unique features of the Arctic Ocean, including high riverine input and shelf exposure relative to basin area. Herein, we suggest that the dV structure in the

Arctic basins reflects the source water type and the extent to which that water has been modified on the shelf.

120

Figure 4.4 Global comparison of dV profiles.

Comparison of Arctic dV data (this study) to previously published Pacific and Atlantic V data: (1) Collier, 1984; (2) Ho et al., 2018;

(3) Middelburg et al., 1988; (4) This study. 4.5.2 The Pacific halocline & surface waters

All water entering the Arctic from the Pacific is channeled through the Bering

Strait. This concentration is our effective Pacific endmember (dV = 31.4 ± 0.6 nmol/kg;

Station 4, Table 4.1). However, between the Bering Strait and the basins there was a significant decrease in dV (Figure 4.3A, B). By comparing dV in the Bering Strait endmember to dV in the Pacific halocline (i.e. along the same isopycnal range, σ θ = 24 –

27 kg/m3), we observed a 5 – 9 nmol/kg decrease in dV, or 17 – 30% removal of dV between the effective Pacific endmember and Pacific halocline waters (ΔV).

Additionally, an extremely low near surface dV signal north of 85°N (Stations 30

– 43; Figure 4.3A, B) provides further evidence for shelf influence on the dV distribution in the basins (dV = 11.9 – 21.6 nmol/kg). The surface waters north of 85°N are associated with the TPD, which has been demonstrated to carry a significant shelf-derived radium

121

signal (≤ 50 m; Kipp et al., 2018). In addition to the shelf-derived radium signal, a higher fraction of river water (~20%) was also reported in the TPD waters (Kipp et al., 2018).

There are four plausible mechanisms that may drive the observed low dV signals in the basin which will be assessed in the following sections: (1) mixing with low dV water types (i.e., dilution; section 4.2.1), (2) dV removal associated with biological activity (section 4.2.2), (3) inorganic particle scavenging of dV (section 4.2.3), and (4) influences of reducing shelf sediments (section 4.2.4).

4.5.2.1 Influence of water-type mixing on dV distribution.

Conservative mixing of different water types can influence the distribution of dV.

In this section we aim to separate potential dV depletion due to mixing of lower dV waters from depletion due to non-conservative processes. In the Arctic basin there are several water types with different endmember dV concentrations including sea ice melt, riverine input, Pacific-derived waters, and Atlantic-derived waters. Both sea ice melt and riverine inputs are characterized by low salinity and low dV relative to the Pacific- derived or Atlantic-derived seawater (Marsay et al., 2018; Peterson et al., 2016).

Given the strong covariance between salinity and dissolved V (dV; Figure 4.5), we took a two-pronged approach to assess the relative contribution of these low salinity endmembers to the observed dV distributions. Both approaches utilize a two endmember mixing model (Eq. 1 – 3), where f indicates the fraction of freshwater or saline water,

[dV] indicates the concentration of dissolved V, and S indicates salinity in the observed fresh or saline endmembers.

1 = ffresh + fsaline (Eq. 4.1)

[dV]obs= [dV]fresh ⋅ ffresh + [dV]saline ⋅ fsaline (Eq. 4.2)

[S]obs= [S]fresh ⋅ ffresh + [S]saline ⋅ fsaline (Eq. 4.3) 122

The first approach is a salinity normalization, which assumes that the freshwater fraction has [S]fresh = 0 and [dV]fresh = 0. In doing this, we can rearrange the equation to solve for [dV]saline (Eq. 4). We consider these assumptions reasonable because the main sources of freshwater to the Arctic Ocean (sea ice melt and river discharge) have low concentrations of dV (Marsay et al., 2018; Peterson et al., 2016).

Figure 4.5 Dissolved V versus salinity.

All samples were used in plotting this figure. Original data (dVobs) are filled with the density color bar and outlined in black and salinity normalized data (dVsaline) are gray circles. The circled regions and numbered regions indicate: (1) Surface waters south of

85°N, (2) Surface waters north of 85°N (i.e., TPD-influenced), (3) Bering Sea and Strait (i.e., roughly the Bering Sea endmember), and (4) The Pacific halocline. The largest correction between the observed and the normalized (group 1) contains surface waters south of 85°N (i.e., group 1) where many samples are influenced by sea ice melt in the marginal ice zone. We account for the influence of conservative mixing by normalizing our observed data (dVobs, Sobs) to a reference salinity (Ssaline) of 35 (Eq. 4; Figure 4.5). We recognize that this is higher than Pacific seawater (S ≈ 32.5) and slightly higher than Atlantic waters

(S ≈ 34.9), so all samples will have a slight positive offset.

[dV]obs [dV] = × Ssaline (Eq. 4.4) saline Sobs

123

In this approach, salinity-normalized dissolved V (dVsaline) preserves non- conservative changes in dV while removing the influence of conservative mixing with low dV, low salinity endmembers such as ice melt or river discharge. Stations that were within the marginal ice zone – the transitional zone between open ocean and sea ice covered seas – had surface samples that were significantly influenced by dilution

(Stations 8 – 14 and 51 – 54; Figures 4.3 and 4.5). This result is expected as the dV in ice samples collected during the GN01 cruise was 0.3 – 4.4 nM (Marsay et al., 2018).

We reevaluate the change in V (ΔV) between the Bering Sea endmember and the

Pacific-halocline using the salinity normalized dV by subtracting the average dVsaline of the Pacific halocline at each station from the average dVsaline from the Bering Strait; this resulting nonconservative ΔV (ΔVsaline) yielded a difference of 5 – 10 nmol/kg, equivalent to a 17% – 33% removal of dV. This range is the same as the ΔV derived from the un-normalized data, indicating that conservative mixing of our Pacific endmember with a low-salinity, low-dV water mass does not account for the depletion of dV in the

Pacific halocline relative to the Bering Strait inflow. Thus, an additional mechanism(s) is required to account for the observed ΔV.

Our second approach uses data from TPD-influenced waters in the central Arctic.

In these waters, we had a range of salinities from 27 – 33. We made no initial assumption on what the fresh dV endmember would be and instead used the dV and salinity data to inform our endmember. In plotting dV versus salinity along this section, a strong linear correlation is apparent (r2 = 0.89; p < 0.001; (Figure 4.6A). Using this correlation, we extrapolated back to the zero-salinity endmember and determined the apparent endmember dVfresh to be approximately -24 nmol/kg (Figure 4.6B, see figure for 95%

124

confidence interval). The negative result implies significant removal in the estuarine and/or shelf environment.

Figure 4.6 TPD dV versus salinity.

(A) Type I least squares regression for observed TPD dV data.

(B) Regression extrapolated to the zero-salinity endmember. The gray band in both panels represents the 95% confidence interval.

Negative dV at the zero salinity endmember indicates there must be dV removal processes between the river source and the basin. The effect of dilution from low-salinity, low-dV endmembers such as river water or sea ice melt therefore does not account for the ΔV observed in the Pacific halocline or the low dV waters in the TPD. With that in mind, we investigated other potential removal

125

mechanisms, including biological removal, dissolved-particulate interactions, and redox reactivity, discussed in detail in the following sections.

4.5.2.2 Biological removal of dV

It has been proposed that the surface ocean depletion of dV is, to some degree, the result of biological uptake (e.g., Collier, 1984; Middelburg et al., 1988). This biological uptake may be unintentional due to the chemical similarity of vanadate and phosphate.

However, a direct biological need for V has also been demonstrated in marine systems.

For example, V is utilized in V-haloperoxidases (Crans et al., 2004), which are found in a variety of organisms including some diatoms. This might explain the observations of

Klein et al. (2013) and Osterholz et al. (2014) who demonstrated scavenging removal of

V associated with diatoms. Elevated V in North Atlantic Trichodesmium supports a role for V in diazotrophs (Nuester et al., 2012), though V nitrogenases have so far only been identified in soil organisms. Interestingly, recent studies have demonstrated N2-fixation is occurring Arctic-wide (Blais et al., 2012; Fernández-Méndez et al., 2016; Harding et al.,

2018; Sipler et al., 2017); thus, V uptake by diazotrophs may facilitate removal of dV in the Arctic water column. To evaluate whether biological processes might significantly influence the V cycle in the Arctic, we used several approaches comparing our dV and pV data with ancillary pigment and nutrient data.

To investigate the relationships with biological activity, we first compared dV and pV to total chlorophyll-a and to fucoxanthin (an accessory pigment found in diatoms)

(Figure 4.7, Appendix Figure C2). We found weak correlations for dV or log-transformed pV and log-transformed pigments (i.e., chlorophyll-a or fucoxanthin; r2 < 0.5 for any tested combination of pigments and dV or pV; Appendix Table C3). Significant p-values

126

were determined in consideration of all data or shelf-only regressions; basin-only regressions had insignificant p-values (Appendix Table C3). In a system with substantial biological control, we expect pV and pigments to trend positively and dV and pigments to trend negatively. Our observations indicate a generally positive trend between pV and pigments, as expected (Figure 4.7B); however, we also observed a positive trend between dV and pigments which is opposite of the expectation (Figure 4.7A, Appendix Table C3).

This suggests that while there may be an association between biological activity and V, it does not drive the dV distribution.

Figure 4.7 Dissolved and particulate V versus total chlorophyll a.

(A) Dissolved V concentration plotted against total chlorophyll a. (B) Labile pV versus total chlorophyll a. Circles indicate data from shelf stations and squares indicate data from basin stations. Similarly, linear relationships between dV or pV and dissolved nutrients were not observed in this study (Appendix Figure C3). This may be due to the preferential regeneration of nutrients in shelf sediments relative to V: vanadium delivered to the sediments (by particle scavenging or biological uptake and settling) is potentially retained under reducing conditions commonly found in sediments (e.g., Breit and Wanty, 1991; see section 4.2.4), whereas nutrients may diffuse back into the water column during

127

remineralization. Cid et al. (2012) compared bioactive trace element and major nutrient relationships in the Chukchi and Beaufort Seas and concluded that trace metal distributions (excepting Cd) were not dominated by biological uptake and remineralization. While V was not included in their study, the dV-nutrient relationships from our Arctic samples are consistent with the conclusions of Cid et al. (2012).

Relationships between dissolved V and nutrients may not be representative of biological uptake ratios. To account for this, we utilized a biological ratio of V:P estimated from the literature and our labile particulate phosphorus (pP) concentrations to predict the biological pV (pVBIO) signal: 푝푉퐵퐼푂 = 푉⁄푃 ∗ 푝푃퐿푎푏𝑖푙푒, where V/P is the molar ratio of these elements in biological materials. Our approach assumes all the observed labile pP was biological in origin. Published particulate V:P ratios range from 0.018 – 63 mmol:mol (Table 4.2). We applied a V:P ratio of 0.5 mmol:mol, which is roughly the median of the range reported for the mixed layer in the Equatorial Pacific (0.17 – 2.6 mmol:mol; Ohnemus et al., 2017) and slightly higher than the range of values reported by

Klein et al. (2013) for surface waters of the North Atlantic (0.076 – 0.38 mmol:mol;

Table 4.2). Higher literature estimates are usually condition-specific (e.g., in an oxygen deficient zone: Ho et al., 2018; Ohnemus et al., 2017). The highest literature V:P estimates are from Trichodesmium in P-limited conditions in the North Atlantic (63 mmol:mol; Nuester et al., 2012) or under stationary phase conditions in isolated culture samples (10 mmol:mol; Osterholz et al., 2014). These two situations are unlikely to be representative of either the bulk Arctic planktonic community or the P-replete conditions of the western Arctic Ocean and thus we discount them as appropriate estimates.

128

The resulting distribution of pVBIO suggests that pVBIO can only account for a substantial fraction of labile pV in surface waters, < 50 m depth in the basin and < 15 m over the shelves (Figure 4.8C, D). A few samples over-predict pVBIO (i.e. pVBIO > labile pV), which indicates that our estimate of biological V:P may be an overestimate at some stations (i.e., surface waters at the Bering Sea Slope – station 1, on the shelf – station 3, and in the marginal ice zone – station 14). Furthermore, we note that the relationship between labile pV and labile pP, while positive, is not strongly correlated on the shelf or in the basin (r2 < 0.2; p > 0.09) (Figure 4.9A, B). In this comparison, we again assumed that all of the observed labile pP was biogenic. We interpret the weak correlation as evidence that the observed pV signal is not strongly driven by biological uptake.

Table 4.3 Summary of literature V:P, V:Fe, and V:Mn ratios (mmol:mol)

Publication V:P V:Fe V:Mn Ohnemus et al., 2017* 0.17 - 4.6 4.15 - Ho et al., 2018* 4 5.3 - Osterholz et al., 2014 10 - - Nuester et al., 2017 41 - 63 - - Bauer et al., 2017 - - 0.3 Klein et al., 2013 0.076 - 0.38 - - 0.018 - Tang & Morel, 2006 0.020 - - Tovar-Sanchez & Sanudo-Wilhelmy, 2011 5 - 11.4 - - Feely et al., 1998 - 2.3 - 4.4 - Our Study (labile, nonlithogenic) - †3.6, 7.6 ††8.8, 12.5 † Shelf only data.

†† Basin-only data

* V:P and V:Fe were determined using subsets of large oceanic sections. V:Fe was determined using hydrothermally-impacted waters with the assumption that all particulate Fe was derived from the oxidation of hydrothermal Fe. In contrast, V:P ratios were assessed after correction for lithogenic and Fe-oxide contributions in waters expected to have a dominant biological profile, such as the surface mixed layer.

129

Figure 4.8 Process associated labile pV distributions.

(A, B) The distribution of labile pV in the surface 500 m (the color scale is logarithmic). The relative amounts of labile pV (%) that are associated with (C, D) biological particles, (E, F) iron oxide-derived particles, and (G, H) manganese oxide-derived particles. The

Makarov and Canada Basin sections are the left panel and right panel, respectively; the color bar for the z variable is located to the far right of each pair and the section location can be identified in the maps at the base of the figure. The contours on each panel are isopycnals delineating water masses as described in Figure 4.2. Full depth section profiles are provided in the supplementary material

(Appendix Figure C5a, b). 130

Our investigation of various biological proxies and dV and pV distributions suggest that V removal by biological uptake is not the dominant mechanism of dV removal in the Western Arctic Ocean. In our results (section 3.1), we noted that stations

2, 3, and 66 had lower dV concentrations than other shelf stations (Table 4.1); we postulate that this may be due to a stripping of dV in the surface resulting from bloom conditions or resuspension of sediments. Particle profiles at these stations have low particle concentrations in the surface and high particle concentrations at depth (Appendix

Figure C4). This profile shape may be indicative of post-bloom conditions at stations 2 and 3 and/or resuspension at the seafloor.

We conclude that while biological uptake may contribute to the removal of dV in near surface waters, it does not appear to have a large influence below the surface water layer; therefore, biological uptake cannot explain the observed depletion in the Pacific halocline of 8 – 9 nmol/kg (ΔV). Notably, this assessment would be improved by further constraints on cellular V:P ratios.

4.5.2.3 Chemically-driven particle interactions.

Generally speaking, Arctic basins have low particle concentrations, including both POC and particulate trace elements, compared to other ocean basins (Honjo et al.,

2010; O’Brien et al., 2013). Nonetheless, lateral transport of particulate trace metals from the shelf into the basin has been described for some trace metals (Aguilar-Islas et al.,

2013; Kondo et al., 2016; Macdonald and Gobeil, 2012).

We can hypothesize that pV distributions in the Arctic Ocean would be similar to pFe or pMn due to scavenging of dV onto oxyhydroxide surfaces. We suspect that particle scavenging may lead to a substantial sink of dV when coupled with reducing

131

sediment conditions whereby V may be further immobilized (Scholz et al., 2011, 2017).

Under oxic conditions, dV species predominantly exist in the V(V) oxidation state, although small amounts of V(IV) may be stabilized by organic complexes (Huang et al.,

2015 and references therein). While V(V) has not been demonstrated to be strongly complexed with organics in the marine environment, various studies have demonstrated scavenging of V(V) by ligand exchange onto Fe and Mn oxyhydroxides (e.g., Bauer et al., 2017; Brinza et al., 2008; Wehrli and Stumm, 1989). The dominant association of V is likely to depend on the relative concentrations of pFe and pMn (as suggested in Bauer et al., 2017). Partition coefficients can be of a similar order of magnitude for Fe and Mn oxyhydroxides, roughly 2,000 to 90,000 L kg-1 for Fe oxyhydroxides and ~60,000 L kg-1 for Mn oxyhydroxides (Huang et al., 2015 and references therein; Takematsu et al., 1985)

Similarly, Scholz et al. (2017) described Mn as a dominant scavenger in the surface ocean above the oxygen minimum zone, but Fe oxides predominated within the oxygen minimum zone. In our section, pFe is higher on the shelves than pMn, and thus pFe likely has more control on V cycling than pMn. However, in the basins pMn is often greater than pFe and, thus, pMn may have a greater relative influence on V cycling in the basins than it does on the shelves. That said, because pFe is much greater on the shelves than either pFe or pMn in the basins, scavenging of V onto oxyhydroxide surfaces is likely far more important on the shelves than in basin waters.

To test the influence of authigenic marine oxides on pV (designated as

“pVCHEM”), we determine the component of labile pV due to Fe oxides (pVCHEM, Fe) and

Mn oxides (pVCHEM, Mn) by utilizing particulate V:Fe and V:Mn ratios from the literature

(Eq. 5, 6).

132

푉 푝푉 = × 푝퐹푒 (Eq. 5) 퐶퐻퐸푀,퐹푒 퐹푒 퐿푎푏𝑖푙푒 푉 푝푉 = × 푝푀푛 (Eq. 6) 퐶퐻퐸푀,푀푛 푀푛 퐿푎푏𝑖푙푒

We applied the V:Fe ratio reported in Ohnemus et al. (2017) (Table 4.2), which was determined by assessing hydrothermal plume samples and assuming all the pFe was sourced from oxides formed when hydrothermal waters entered the oxic marine environment. Other studies have also assessed the V:Fe ratio in hydrothermal plume- associated samples (Table 4.2; Ho et al., 2018; Feely, 1998) and reported V:Fe ratios similar to Ohnemus et al. (2017). Both Ho et al. (2018) and Ohnemus et al. (2017) discuss possible limitations of applying a hydrothermal V:Fe ratio specifically to particle interactions elsewhere. However, upper water column V:Fe ratios specific to the Arctic shelf Fe-oxyhydroxide phase have not been described, therefore application of V:Fe ratios from hydrothermal systems is our most reasonable option.

The estimated pVCHEM, Fe accounts for all of the observed labile pV over the shelves (Figure 4.8E, F). In contrast to the shelf region, the labile pV in the basins cannot be entirely accounted for by pVCHEM, Fe (Figure 4.8E, F). Notably, pVCHEM, Fe exceeded

100% of the measured labile pV in 13% of water column samples; of the samples exceeding 100%, a range from 102% – 165% was observed – the majority of which are located on the continental shelf. Considering pVCHEM, Fe determined from non-lithogenic pFe, roughly 60 – 70% of non-lithogenic pV can be accounted for using the V:Fe relationship. In the case of non-lithogenic pVCHEM, Fe, only one water column sample exceeded 100% of the observed non-lithogenic pV, this sample was at 1 m and collected directly under the sea ice; it was greater than 100 percent for the labile fraction as well.

133

The difference in non-lithogenic and labile pVCHEM, Fe may be attributed to methodological concerns mentioned in section 2.5. Even though the fraction of pV that non-lithogenic pVCHEM, Fe accounts for is less than observed for labile data, the distribution of pVCHEM, Fe determined from non-lithogenic pFe has the same distribution of as pVCHEM, Fe determined using labile pFe. Thus, for our purposes in assessing the relative role of Fe-oxide interactions on dV, either result is appropriate.

We further examined the relationship between pV and pFe using non-lithogenic and labile fractions in property-property space. Both non-lithogenic and labile pV and

2 pFe exhibit tight positive correlations in the shelf regions (r nonlithogenic = 0.96, p < 0.001;

2 r labile = 0.98, p < 0.001; Figure 4.9C). The slopes of these relationships are 7.3 and 3.6 for the non-lithogenic and labile fractions, respectively. This is a broader range than reported in other studies, but encompasses the ratios presented in the literature (Table

4.2). The relationships in the basin are more poorly correlated than the shelf trends

2 2 (r nonlithogenic = 0.42, p < 0.001; r labile = 0.15, p < 0.001; Figure 4.9D). We suspect this is due to the low concentrations of non-lithogenic and labile pFe observed in the basins. We note that slope stations add scatter to the basin and slope trends, which may be due to unique conditions at each slope station, such as surface sea ice melt (at marginal ice zone stations 8 and 10) or resuspension along the shelf break.

134

Figure 4.9 TPD dV versus salinity.

Particulate V to particulate-element trends (pP, pFe, or pMn). Data are divided into shelf-only (right column) and basin-only (left column) scatter plots. Because of large regional differences in particle abundance, axis ranges between the shelf and basin panels are variable. Slope data are plotted on both shelf-only and basin-only plots and marked with triangles. A,B) The pV:pP relationship, which is positive but not strongly correlated. C,D) The pV:pFe relationship, which is strongly correlated (see Appendix Table C3 in the supplemental for more statistical information). E,F) The pV:pMn trend: basin-only data and the labile shelf data yield significant relationships. In all panels, dashed lines indicate the trendline of a type II major axis linear regression, while the colored bands indicate the 95% confidence interval determined from a jackknife procedure (n = 1000).

135

Vanadium may also be adsorbed by Mn oxyhydroxides. Thus, we also considered the relationship of labile and non-lithogenic pV to pMn fractions. Our analyses suggest that pMn-V interactions might be important in the basins where pFe concentrations are often lower than pMn concentrations. We examined the pV and pMn relationship in property-property space using the non-lithogenic and labile fractions (Figure 4.9E, F).

Shelf-only and slope-only trends show strong correlations in the labile fraction; the shelf samples have a steep slope and basin samples have a comparatively lower slope (shelf:

2 2 r labile = 0.70, p < 0.001; basin: r labile = 0.81, p < 0.001). On the shelf, the pV:pMn correlation might not be causative; the increasing pV is likely driven by the much larger pFe concentrations, as discussed above. The trends in both the shelf and the basin are

2 weaker for the non-lithogenic fractions (shelf: r nonlithogenic = 0.27, p = 0.006; basin:

2 r nonlithogenic = 0.56, p < 0.001). Possible explanations for the differences in labile and non- lithogenic data were discussed in section 2.5. For the basin samples, where pMn is generally greater than pFe, we tentatively interpret the pV:pMn slopes as representative of an adsorptive pV:pMn relationship.

Interestingly, the slopes derived from our labile and non-lithogenic basin pV:pMn observations are 30 – 40 times greater than the relationship reported in Bauer et al. (2017)

(Table 4.2). This may suggest that our data are only correlative, and that pV may not be driven by pMn. However, in the Baltic Sea environment studied by Bauer et al.

(2017) pMn concentrations can be much higher than those in our Arctic section and dV is significantly lower than in the Arctic water column. Bauer et al. (2017) described the relationship between pV and pMn in an environment where Mn is effectively trapped at the oxic-anoxic interface in the water column, whereas the relationships we observed in

136

the Arctic Ocean are defined in an oxic water column. In the suboxic zone of the Black

Sea, Yiğiterhan et al. (2011) observed similar non-lithogenic pV/pMn ratios as Bauer et al. (2017) but could not resolve whether the Mn or Fe oxyhydroxides were the controlling phase. Thus, the different pV:pMn ratios in the Baltic and Black Seas versus the Arctic

Ocean are not unreasonable. We thereby chose to assess our system using the stoichiometry observed in our system; further investigation would be required to determine the reasons for the large difference in pV:pMn stoichiometry between studies.

We calculated the pVCHEM,Mn by multiplying labile or non-lithogenic pMn with our V:Mn ratio (Table 4.2). The distribution of percent pVCHEM,Mn indicates that the abundance of pMn in the basins, away from the continental slopes could be responsible for a significant portion of the basin pV distribution (Figure 4.8G, H; see also Appendix

Figure C5B for full depth sections). However, on the shelf and slope, pVCHEM,Mn is far less important, which is in agreement with the greater amount of pFe than pMn in the shelf/slope environment and our analysis above of pVCHEM, Fe.

Overall, the tight correlation between pFe and pV in our data is evidence that the distribution of pV, and thus dV, is largely influenced by scavenging of V to Fe oxides.

This relationship predominates on the shelves where pFe concentrations are highest and are also much greater than pMn concentrations. In contrast, the relationship between pV and pMn in the basins and the distribution of percent pVCHEM,Mn indicates that pMn may assume a more dominant role on the pV cycle in those waters where pMn is greater than pFe. Notably, because particle concentrations in the basins are so much lower than on the shelves, the relative influence of particle scavenging on dV in the basins is also much less, resulting in a largely conservative dV distribution in shallow basin waters (Sec. 5.1).

137

4.5.2.4 Reducing shelf environment.

The final mechanism for dV removal is reduction at the sediment-water interface or within shelf sediments (Joung and Shiller, 2016; Scholz et al., 2011; Shiller and Mao,

1999). Under reducing conditions, V(V) is reduced to V(IV) and sometimes V(III). V(IV) and V(III) species tend to rapidly partition to solid phases (Huang et al., 2015; Wehrli and Stumm, 1989). Given the previously described particle associations between V and

Fe or Mn oxyhydroxides (section 4.2.3), particle-bound V delivered to the sediments has the potential to be further immobilized in the sediments (Scholz et al., 2017, 2011).

Scholz et al. (2011) indicate that the formation of authigenic V in sediments requires highly reducing sediments. Indeed, analysis of various continental margin sediments suggests that V only becomes enriched in sediments under strongly reducing conditions such as where an oxygen deficient zone impinges on the bottom and Fe reduction occurs at the sediment-water interface (Morford and Emerson, 1999). Where there was slight oxygen penetration into the sediments (< 1 cm), Morford and Emerson (1999) found release of V from margin sediments. V depletion has been observed in hypoxic waters of the Louisiana Shelf (Joung and Shiller, 2016; Shiller and Mao, 1999) and seasonally variable V depletion and enrichment was reported for waters in the Mississippi Sound and Bight where hypoxia is less extensive (Ho et al., 2019).

Hardison et al. (2017) reported only slight oxygen penetration (5 – 8 mm) in sediments from the north-eastern Chukchi Sea and Brüchert et al. (2018) found similarly slight oxygen penetration in sediments of the East Siberian Shelf. Anderson et al. (2017) suggested that nitrate depletion (relative to phosphate) in the East Siberian Sea was indicative of hypoxia. Additionally, Kvitek et al. (1998) note the possibility of anoxic

138

brine filled ice gouges in an Arctic embayment. Thus, while sediments were not examined as a part of this study, it is reasonable to expect conditions favorable for immobilization of V on some Arctic shelves.

From the GN01 section, there are further indicators that the shelf sediments are reducing enough to remove dV and that sediment-water exchange may be a V sink on the shelf. For example, under the reducing conditions wherein dV can be removed from solution, pMn is released to the dissolved phase (Wang and Sañudo-Wilhelmy, 2008).

High dMn concentrations at GN01 shelf stations indicate that the shelf is a source of dMn

(Jensen et al., 2018; Shiller, 2019) and, thus, also likely a sink for V. Kondo et al. (2016) also observed elevated dMn in bottom waters of the outer shelf, which indicates consistency between years and across the spatial area of the Chukchi Shelf. Additionally, dissolved methane was elevated in some shelf bottom waters during the GN01 cruise

(Whitmore and Shiller, 2016), indicating a sediment source of dissolved methane

(although several other methane sources are possible). The δ13C signature of methane in samples collected from the Bering and Chukchi shelves in 2012 was indicative of a microbial source (Kudo et al., 2018) which implies highly reducing shelf sediment conditions. Thus, while we cannot quantitatively ascribe sediment reduction as a dV removal mechanism, it is likely that reducing shelf conditions play an important role in removing dV from Arctic shelf waters.

Off the Peru margin, Ho et al. (2018) did not observe the sort of shelf influence we see in the Arctic, despite the operation of an Fe shuttling mechanism there (Scholz et al., 2011). The reason for this difference is probably physiographic: as noted in our

139

introduction, shelves comprise half the area of the Arctic Ocean; in contrast, the Peru margin is narrow, especially when compared to the expanse of the Pacific Ocean.

4.5.2.5 Arctic Ocean V cycle.

In this consideration of mechanisms responsible for dV removal, we demonstrated that mixing of endmembers does not account for low dV observed in the waters of the basins. Biology likely has a significant role in near surface waters, but a minor role below that; even on the shelves, pVBIO accounts for little of the labile pV below the very surface. However, pFe, which is elevated on the shelves, does have a strong relationship with pV, suggesting that Fe oxides exert a control on the V cycle. Particulate V trended significantly with pMn only in the basins where pFe was very low. The large amount of particles on the shelf compared to the basins suggests the role of the shelf in setting basin geochemistry is substantial.

We suggest the same paired sorption-reduction mechanism that Scholz et al.

(2011, 2017) described for V removal in the Peruvian margin is occurring over the

Bering and Chukchi shelves and likely other Arctic shelves. Delivery of V to shelf sediments in our region is likely driven by V scavenged by authigenic marine oxides

(predominantly Fe oxides) and to a minor degree by biologically-associated V.

Particulate V may subsequently be immobilized within the shelf sediments due to its reduction to less soluble forms. Low dV shelf-influenced waters are advected into the central Arctic basins, whereby a surface depletion is observed. If our suggested mechanism is correct, the extensive shelf V removal might make the Arctic Ocean one place where water column V isotope fraction (Wu et al., 2019) could be observed.

140

Shiller and Mao (1999) suggested that expansion of anthropogenic shelf hypoxia might increase the shelf contribution to dV removal from seawater. Importantly, our

Arctic Ocean data suggest that shelf removal of dV could be a non-trivial aspect of the V cycle, especially in marginal seas with large shelf regions. Furthermore, if climate change factors such as declining sea ice contribute to changing biogeochemical cycling in the

Arctic (e.g., Anderson et al., 2017), then the Arctic Ocean dV distribution may be an indicator of the impact of that change.

4.5.3 Deep water dV distribution

The deep and bottom water dV concentrations we observed in the Arctic Ocean

(dV = 30.1 ± 0.9 where z > 2000 m) are lower than reports from other deep ocean basins

(Collier, 1984; Ho et al., 2018; Middelburg et al., 1988). There are only a few mechanisms by which Arctic Deep Water (ADW) could have diminished dV relative to other ocean basins: source waters of ADW could have low dV to begin with; the source waters could be modified during their transformation into ADW; or, there could be slow removal of dV during the long residence time of ADW. Since the deep waters of each basin we sampled (Makarov, Canada, and Eurasian) have the same dV concentration, despite having different apparent ages (e.g., Schlosser et al., 1997), slow removal of dV is unlikely to be a significant factor. Supporting this is the observation that in the global ocean deep water dV only slightly increases between the Atlantic and Pacific Ocean (Ho et al., 2018; Middelburg et al., 1988); therefore, slow scavenging removal is unlikely a major factor in the dV distribution.

Jones et al. (1995) indicate three main sources of Arctic Deep Water: a) continental slope density flows comprised mainly of Atlantic and intermediate layer

141

waters but triggered by shelf brines, b) Barents Sea Branch Waters (BSBW), and c) Fram

Strait Branch Waters (FSBW). They suggest that the density flows are the most important component while FSBW is the least important component. Regardless of the relative magnitudes of these sources, it would seem that, except for FSBW, the sources of ADW have experienced substantial shelf interactions and thus likely would have diminished dV relative to their external source, such as we observed in the TPD and elsewhere in shallow Arctic waters.

With regard to the external source of ADW, most of this is from the North

Atlantic which, as noted previously, enters the Arctic Ocean via the Fram Strait and the

Barents Sea, having sill depths of roughly 2500 m and 200 m, respectively (Rudels, 2019; von Appen et al., 2015). In shallow North Atlantic waters, Klein et al. (2013) observed a range of surface dV in the North Atlantic of 12 – 32 nmol/kg (5 – 10 m water depth) and

Middelburg et al. (1988) observed surface dV of 31.7 – 35.2 nmol/kg (< 100 m depth).

Because of the sill depths we should also consider deeper waters that enter the Arctic, which according to Middelburg et al. (1988) are roughly 32.5 nmol/kg.

It thus seems likely that ADW has diminished dV due partly to the slightly low dV of inflowing shallow North Atlantic waters but also due to additional removal of dV as these waters interact with the shelves. To better assess the ADW dV signature, dissolved V data from the Eurasian side of the Arctic would be required to further characterize deep water distributions and assess endmember values.

4.5.4 Utility as a tracer

Geochemical tracers can be used to inform us about the mechanisms driving water mass modifications on the shelves and the connectivity between the shelf and the basin.

142

Additionally, because Arctic waters are a key driver of global circulation processes, understanding geochemical distributions in the Arctic marine system can help us further utilize the distributions of trace elements and their isotopes globally.

One of the most prominent features in the western Arctic Basins is the Pacific halocline, formed from a mixture of lower-salinity sources such as sea ice melt, river drainage, and Pacific water. Pacific halocline waters are often traced with a combined nutrient tracer, which is only quasi-conservative, in part due to the non-conservative behavior of the nutrients (e.g., Alkire et al., 2015; Newton et al., 2013). Although there are concerns with the quantitative results of the nutrient tracer, there are few other tracers that have been described that can deconvolve these water masses. Thus, there is a demand for tracers that can elucidate the Arctic marine system more clearly. Our V data reveal a distribution wherein in the Pacific halocline has less dV than Atlantic waters below and less than Pacific waters entering the Arctic. This distribution reveals the extent of shelf modification on Arctic water masses from the incoming Pacific and Atlantic waters.

We note that interactions of V with particles predominantly influence dV on the shelves, but Fe and Mn oxyhydroxides have the potential to alter dV distribution in the basins. To assess if our basin distribution behaves conservatively, as is often the case with dV in oceanic waters below the euphotic zone (e.g., Morris, 1975), we examined the dV associated with each water mass as a function of location in the basin. In this case, distance from the margin is a rough proxy for age (i.e., exposure to potential non- conservative processes) (Figure 4.10).

143

Figure 4.10 Dissolved V versus latitude.

Dissolved V (nmol/kg) in the GN01 Section, dV concentrations presented are averages at each station for the respective water masses.

The gray shaded region represents the shelf break and the red shaded region represents stations north of the pacific halocline (which also have TPD-influenced surface waters). Using average dV for each water mass at each station, we determined that dV in the polar mixed layer, the Atlantic halocline, and Arctic deep waters basin did not significantly change with distance from the shelf (Figure 4.10). We note that the Pacific halocline dV had a poor correlation with latitude (r2 = 0.37; p = 0.05). However, the

2 normalized dVsaline showed a stronger correlation (r = 0.57; p = 0.01) (Appendix Table

C3; Appendix Figure C6). In the Makarov Basin and Canada Basin we used data between

75°N and 85°N and data between 73°N and 85°N, respectively (i.e., the shelf break and the northern-most limit of the Pacific halocline). These correlations suggest that, in general, behavior of dV in the basins is conservative. However, in the Pacific halocline there may be some non-conservative behavior.

Our data lend support to interactions with reducing environments and scavenging by Fe-oxyhydroxides as primary drivers for V removal on the shelves. While the conditions that impart depleted dV signals are highly dynamic and may not be conducive

144

to separating endmembers, it appears that dV falls into a suite of tools that can help unify our understanding of the Arctic marine system.

4.6 Conclusions

In the western Arctic Ocean basins, the vertical distribution of dissolved vanadium (dV) shows a surface depletion of 24 – 49% with respect to deep water dV.

The depleted surface waters are associated with water mass features (i.e., the Pacific halocline, Polar Mixed Layer, and Transpolar Drift), which indicates that lateral transport from shelf waters is important in setting the dV signature in the basins. A consideration of processes that could influence the dV distribution suggests that scavenging of V by Fe oxyhydroxides and delivery to the sediments is the dominant mechanism by which V is removed from the water column. To a minor degree, biological removal in the surface layer and scavenging associated with Mn oxyhydroxides may also contribute to the water column removal of dV. The dV depletion is likely facilitated by an Fe shuttling mechanism (Scholz et al., 2011, 2017) whereby Fe oxyhydroxides delivered to Arctic shelf sediments are reductively dissolved in the sediments. The dissolved Fe diffusing from the sediments is ultimately re-oxidized and precipitated in the oxic shelf water column and thus further removes dV by repeating the scavenging process. Evidence suggests that the shelf sediments are likely reducing enough to reduce and immobilize the shuttled V within them. Vanadium-depleted shelf waters are subsequently advected into the basins, where dV behaves roughly conservatively.

Deep basin waters in the Western Arctic Ocean are significantly lower in dV concentration than other ocean basins. This may be indicative of sources of water to the deep basin and further study is needed to determine the extent to which the depletion

145

results from direct endmember contributions or substantial margin influence on incoming deep Arctic source waters.

Further studies of V in the water column and sediments of the Arctic basin and on the regional continental shelves will help better constrain the processes, influence, and extent of shelf modification on the basin dV signature. Arctic Ocean waters leave the

Arctic predominantly via the Fram and Davis Straits and contribute to formation of intermediate and deep waters in the North Atlantic. The study of tracers in these water masses must include the Arctic as an important site of modification to tracer signals. Our

Arctic work also raises important questions as to the net effect of marginal basin shelves on oceanic vanadium cycling, its isotopic balance, and how climate-induced changes in shelf biogeochemical cycling will impact vanadium cycling.

4.7 Acknowledgements

This research was supported by National Science Foundation [OCE-1436312

(AMS), OCE-1435862 (BST), OCE-1436019 (PLM)]. We extend thanks to Melissa

Gilbert for her analytical ICP-MS work at the University of Southern Mississippi’s

Center for Trace Analysis and Angelique White for the pigment analyses. We also thank the USCGC Healy’s crew and the scientific leadership, Bill Landing, Greg Cutter, and

Dave Kadko, for their support of a successful expedition.

4.8 References

Aagaard, K., 1981. On the deep circulation in the Arctic Ocean. Deep Sea Res. 28A, 251–268. Aguilar-Islas, A.M., Rember, R., Nishino, S., Kikuchi, T., Itoh, M., 2013. Partitioning and lateral transport of iron to the Canada Basin. Polar Sci. 7, 82–99. https://doi.org/10.1016/j.polar.2012.11.001

146

Alkire, M.B., Morison, J., Andersen, R., 2015. Variability in the meteoric water, sea-ice melt, and Pacific water contributions to the central Arctic Ocean, 2000-2014. J. Geophys. Res. Oceans 120, 1573–1598. https://doi.org/10.1002/2014JC010023 Anderson, L.G., Björk, G., Holby, O., Jutterström, S., Mörth, C.M., O'Regan, M., Pearce, C., Semiletov, I., Stranne, C., Stöven, T., Tanhua, T., Ulfsbo, A., Jakobsson, M., 2017. Shelf–Basin interaction along the East Siberian Sea. Ocean Sci. 13, 349–363. https://doi.org/10.5194/os-13-349-2017 Bauch, D., Schlosser, P., Fairbanks, R., 1995. Freshwater balance and the sources of deep 18 and bottom waters in the Arctic Ocean inferred from the distribution of H2 O. Prog. Oceanogr. 35, 53–80. Bauer, S., Blomqvist, S., Ingri, J., 2017. Distribution of dissolved and suspended particulate molybdenum, vanadium, and tungsten in the Baltic Sea. Mar. Chem. 196, 135–147. https://doi.org/10.1016/j.marchem.2017.08.010 Beck, A.J., Cochran, J.K., Sañudo-Wilhelmy, S.A., 2010. The distribution and speciation of dissolved trace metals in a shallow subterranean estuary. Mar. Chem. 121, 145–156. https://doi.org/10.1016/j.marchem.2010.04.003 Berger, C.J.M., Lippiatt, S.M., Lawrence, M.G., Bruland, K.W., 2008. Application of a chemical leach technique for estimating labile particulate aluminum, iron, and manganese in the Columbia River plume and coastal waters off Oregon and Washington. J. Geophys. Res. 113, C00B01. https://doi.org/10.1029/2007JC004703 Blais, M., Tremblay, J.-É., Jungblut, A.D., Gagnon, J., Martin, J., Thaler, M., Lovejoy, C., 2012. Nitrogen fixation and identification of potential diazotrophs in the Canadian Arctic: N 2 fixation in the Arctic Ocean. Glob. Biogeochem. Cycles 26. https://doi.org/10.1029/2011GB004096 Breit, G.N., Wanty, R.B., 1991. Vanadium accumulation in carbonaceous rocks: A review of geochemical controls during deposition and diagenesis. Chem. Geol. 91, 83–97. https://doi.org/10.1016/0009-2541(91)90083-4 Brinza, L., Benning, L.G., Statham, P.J., 2008. Adsorption studies of Mo and V onto ferrihydrite. Mineral. Mag. 72, 385–388. https://doi.org/10.1180/minmag.2008.072.1.385 Brüchert, V., Bröder, L., Sawicka, J.E., Tesi, T., Joye, S.P., Sun, X., Semiletov, I.P., Samarkin, V.A., 2018. Carbon mineralization in Laptev and East Siberian sea shelf and slope sediment. Biogeosciences 15, 471–490. https://doi.org/10.5194/bg-15-471-2018 Charette, M.A., Lam, P.J., Lohan, M.C., Kwon, E.Y., Hatje, V., Jeandel, C., Shiller, A.M., Cutter, G.A., Thomas, A., Boyd, P.W., Homoky, W.B., Milne, A., Thomas, H., Andersson, P.S., Porcelli, D., Tanaka, T., Geibert, W., Dehairs, F., Garcia- Orellana, J., 2016. Coastal ocean and shelf-sea biogeochemical cycling of trace elements and isotopes: lessons learned from GEOTRACES. Philos. Trans. R. Soc. Math. Phys. Eng. Sci. 374, 20160076. https://doi.org/10.1098/rsta.2016.0076 Cid, A.P., Nakatsuka, S., Sohrin, Y., 2012. Stoichiometry among bioactive trace metals in the Chukchi and Beaufort Seas. J. Oceanogr. 68, 985–1001. https://doi.org/10.1007/s10872-012-0150-8

147

Collier, R.W., 1984. Particulate and dissolved vanadium in the North Pacific Ocean. Nature 390, 441–444. Crans, D.C., Smee, J.J., Gaidamauskas, E., Yang, L., 2004. The Chemistry and Biochemistry of Vanadium and the Biological Activities Exerted by Vanadium Compounds. Chem. Rev. 104, 849–902. Cutter, G., Andersson, P.S., Codispoti, L.A., Croot, P., Francois, R., Lohan, M., Obata, Hajime., Rutgers van der Loeff, M., 2014. Sampling and Sample-handling Protocols for GEOTRACES Cruises Version 2. http://www.geotraces.org/library- 88/scientific-publications/reports/169-sampling-and-sample-handling-protocols- for-geotraces-cruises Emerson, S.R., Huested, S.S., 1991. Ocean anoxia and the concentrations of molybdenum and vanadium in seawater. Mar. Chem. 34, 177–196. https://doi.org/10.1016/0304-4203(91)90002-E Fernández-Méndez, M., Turk-Kubo, K.A., Buttigieg, P.L., Rapp, J.Z., Krumpen, T., Zehr, J.P., Boetius, A., 2016. Diazotroph Diversity in the Sea Ice, Melt Ponds, and Surface Waters of the Eurasian Basin of the Central Arctic Ocean. Front. Microbiol. 7, 1884. https://doi.org/10.3389/fmicb.2016.01884 Fransson, A., Chierici, M., Anderson, L.G., Bussmann, I., Kattner, G., Peter Jones, E., Swift, J.H., 2001. The importance of shelf processes for the modification of chemical constituents in the waters of the Eurasian Arctic Ocean: Implication for carbon fluxes. Cont. Shelf Res. 21, 225–242. https://doi.org/10.1016/S0278- 4343(00)00088-1 Harding, K., Turk-Kubo, K.A., Sipler, R.E., Mills, M.M., Bronk, D.A., Zehr, J.P., 2018. Symbiotic unicellular cyanobacteria fix nitrogen in the Arctic Ocean. Proc. Natl. Acad. Sci. 115, 13371–13375. https://doi.org/10.1073/pnas.1813658115 Ho, P., Lee, J.-M., Heller, M.I., Lam, P.J., Shiller, A.M., 2018. The distribution of dissolved and particulate Mo and V along the U.S. GEOTRACES East Pacific Zonal Transect (GP16): The roles of oxides and biogenic particles in their distributions in the oxygen deficient zone and the hydrothermal plume. Mar. Chem. 201, 242–255. https://doi.org/10.1016/j.marchem.2017.12.003 Ho, P., Shim, M.J., Howden, S.D., Shiller, A.M., 2019. Temporal and spatial distributions of nutrients and trace elements (Ba, Cs, Cr, Fe, Mn, Mo, U, V and Re) in Mississippi coastal waters: Influence of hypoxia, submarine groundwater discharge, and episodic events. Cont. Shelf Res. 175, 53–69. https://doi.org/10.1016/j.csr.2019.01.013 Honjo, S., Krishfield, R.A., Eglinton, T.I., Manganini, S.J., Kemp, J.N., Doherty, K., Hwang, J., McKee, T.K., Takizawa, T., 2010. Biological pump processes in the cryopelagic and hemipelagic Arctic Ocean: Canada Basin and Chukchi Rise. Prog. Oceanogr. 85, 137–170. https://doi.org/10.1016/j.pocean.2010.02.009 Huang, J.-H., Huang, F., Evans, L., Glasauer, S., 2015. Vanadium: Global (bio)geochemistry. Chem. Geol. 417, 68–89. https://doi.org/10.1016/j.chemgeo.2015.09.019 Jakobsson, M., 2002. Hypsometry and volume of the Arctic Ocean and its constituent seas. Geochem. Geophys. Geosystems 3, 1–18. https://doi.org/10.1029/2001GC000302

148

Jeandel, C., Peucker-Ehrenbrink, B., Jones, M.T., Pearce, C.R., Oelkers, E.H., Godderis, Y., Lacan, F., Aumont, O., Arsouze, T., 2011. Ocean margins: The missing term in oceanic element budgets? Eos Trans. Am. Geophys. Union 92, 217–224. https://doi.org/10.1029/2011EO260001 Jensen, L., Cullen, J.T., Ball, G.T., Sherrell, R.M., Fitzsimmons, J.N., 2018. CT13A-04: Dissolved Fe and Mn along US Arctic GEOTRACES GN01: Effects of scavenging in intermediate and deep waters. Jones, E.P., Rudels, B., Anderson, L.G., 1995. Deep waters of the Arctic Ocean: origins and circulation. Deep Sea Res. Part Oceanogr. Res. Pap. 42, 737–760. https://doi.org/10.1016/0967-0637(95)00013-V Joung, D., Shiller, A.M., 2016. Temporal and spatial variations of dissolved and colloidal trace elements in Louisiana Shelf waters. Mar. Chem. 181, 25–43. https://doi.org/10.1016/j.marchem.2016.03.003 Kipp, L.E., Charette, M.A., Moore, W.S., Henderson, P.B., Rigor, I.G., 2018. Increased fluxes of shelf-derived materials to the central Arctic Ocean. Sci. Adv. 4, eaao1302. https://doi.org/10.1126/sciadv.aao1302 Klein, N.J., Beck, A.J., Hutchins, D.A., Sañudo-Wilhelmy, S.A., 2013. Regression modeling of the North East Atlantic Spring Bloom suggests previously unrecognized biological roles for V and Mo. Front. Microbiol. 4, 45. https://doi.org/10.3389/fmicb.2013.00045 Klunder, M.B., Laan, P., Middag, R., de Baar, H.J.W., Bakker, K., 2012. Dissolved iron in the Arctic Ocean: Important role of hydrothermal sources, shelf input and scavenging removal: Dissoved iron in the Arctic Ocean. J. Geophys. Res. Oceans 117, C04014. https://doi.org/10.1029/2011JC007135 Knauer, G.A., Martin, J.H., 1981. Phosphorus-cadmium cycling in northeast Pacific waters. J. Mar. Res. Kondo, Y., Obata, Hajime., Hioki, N., Ooki, A., Nishino, S., Kikuchi, T., Kuma, K., 2016. Transport of trace metals (Mn, Fe, Ni, Zn and Cd) in the western Arctic Ocean (Chukchi Sea and Canada Basin) in late summer 2012. Deep Sea Res. Part Oceanogr. Res. Pap. 116, 236–252. https://doi.org/10.1016/j.dsr.2016.08.010 Koschinsky, A., Hein, J.R., 2003. Uptake of elements from seawater by ferromanganese crusts: solid-phase associations and seawater speciation. Mar. Geol. 198, 331– 351. https://doi.org/10.1016/S0025-3227(03)00122-1 Kudo, K., Yamada, K., Toyoda, S., Yoshida, N., Sasano, D., Kosugi, N., Ishii, M., Yoshikawa, H., Murata, A., Uchida, H., Nishino, S., 2018. Spatial distribution of dissolved methane and its source in the western Arctic Ocean. J. Oceanogr. 74, 305–317. https://doi.org/10.1007/s10872-017-0460-y Kvitek, R.G., Conlan, K.E., Iampietro, P.J., 1998. Black pools of death: hypoxic, brine- filled ice gouge depressions become lethal traps for benthic organisms in a shallow Arctic embayment. Mar. Ecol. Prog. Ser. 162, 1–10. https://doi.org/10.3354/meps162001 Landing, W., Cutter, G., Kadko, D., 2017. CTD-GTC bottle data from GEOTRACES- Arctic Section cruise HLY1502, August to October 2015 (U.S. GEOTRACES Arctic project) [WWW Document]. Biol. Chem. Oceanogr. Data Manag. Off. BCO-DMO. URL http://lod.bco-dmo.org/id/dataset/647259

149

Landing, W., Cutter, G., Kadko, D., 2016. CTD-ODF bottles from GEOTRACES-Arctic Section cruise HLY1502 [WWW Document]. Biol. Chem. Oceanogr. Data Manag. Off. BCO-DMO. URL http://lod.bco-dmo.org/id/dataset/646825 Macdonald, R.W., Gobeil, C., 2012. Manganese sources and sinks in the Arctic ocean with reference to periodic enrichments in basin sediments. Aquat. Geochem. 18, 565–591. https://doi.org/10.1007/s10498-011-9149-9 Marsay, C.M., Aguilar-Islas, A., Fitzsimmons, J.N., Hatta, M., Jensen, L.T., John, S.G., Kadko, D., Landing, W.M., Lanning, N.T., Morton, P.L., Pasqualini, A., Rauschenberg, S., Sherrell, R.M., Shiller, A.M., Twining, B.S., Whitmore, L.M., Zhang, R., Buck, C.S., 2018. Dissolved and particulate trace elements in late summer Arctic melt ponds. Mar. Chem. 204, 70–85. https://doi.org/10.1016/j.marchem.2018.06.002 Mauldin, A., Schlosser, P., Newton, R., Smethie, W.M., Bayer, R., Rhein, M., Jones, E.P., 2010. The velocity and mixing time scale of the Arctic Ocean Boundary Current estimated with transient tracers. J. Geophys. Res. 115, C08002. https://doi.org/10.1029/2009JC005965 Middag, R., de Baar, H.J.W., Laan, P., Bakker, K., 2009. Dissolved aluminium and the silicon cycle in the Arctic Ocean. Mar. Chem. 115, 176–195. https://doi.org/10.1016/j.marchem.2009.08.002 Middelburg, J.J., Hoede, D., Van Der Sloot, H.A., Van Der Weijden, C.H., Wijkstra, J., 1988. Arsenic, antimony and vanadium in the North Atlantic Ocean. Geochim. Cosmochim. Acta 52, 2871–2878. https://doi.org/10.1016/0016-7037(88)90154-8 Morford, J.L., Emerson, S., 1999. The geochemistry of redox sensitive trace metals in sediments. Geochim. Cosmochim. Acta 63, 1735–1750. https://doi.org/10.1016/S0016-7037(99)00126-X Morris, A.W., 1975. Dissolved molybdenum and vanadium in the northeast Atlantic Ocean. Deep Sea Res. Oceanogr. Abstr. 22, 49–54. https://doi.org/10.1016/0011- 7471(75)90018-2 Newton, R., Schlosser, P., Mortlock, R., Swift, J., MacDonald, R., 2013. Canadian Basin freshwater sources and changes: Results from the 2005 Arctic Ocean Section: AOS 2005 freshwater sources and changes. J. Geophys. Res. Oceans 118, 2133– 2154. https://doi.org/10.1002/jgrc.20101 Nuester, J., Vogt, S., Newville, M., Kustka, A.B., Twining, B.S., 2012. The Unique Biogeochemical Signature of the Marine Diazotroph Trichodesmium. Front. Microbiol. 3, 150. https://doi.org/10.3389/fmicb.2012.00150 O’Brien, M.C., Melling, H., Pedersen, T.F., Macdonald, R.W., 2013. The role of eddies on particle flux in the Canada Basin of the Arctic Ocean. Deep Sea Res. Part Oceanogr. Res. Pap. 71, 1–20. https://doi.org/10.1016/j.dsr.2012.10.004 Ohnemus, D.C., Auro, M.E., Sherrell, R.M., Lagerström, M., Morton, P.L., Twining, B.S., Rauschenberg, S., Lam, P.J., 2014. Laboratory intercomparison of marine particulate digestions including Piranha: a novel chemical method for dissolution of polyethersulfone filters. Limnol. Oceanogr. Methods 12, 530–547. https://doi.org/10.4319/lom.2014.12.530

150

Ohnemus, D.C., Lam, P.J., 2015. Cycling of lithogenic marine particles in the US GEOTRACES North Atlantic transect. Deep Sea Res. Part II Top. Stud. Oceanogr. 116, 283–302. https://doi.org/10.1016/j.dsr2.2014.11.019 Ohnemus, D.C., Rauschenberg, S., Cutter, G.A., Fitzsimmons, J.N., Sherrell, R.M., Twining, B.S., 2017. Elevated trace metal content of prokaryotic communities associated with marine oxygen deficient zones: Elevated trace metals in ODZ prokaryotes. Limnol. Oceanogr. 62, 3–25. https://doi.org/10.1002/lno.10363 Osterholz, H., Simon, H., Beck, M., Maerz, J., Rackebrandt, S., Brumsack, H.-J., Feudel, U., Simon, M., 2014. Impact of diatom growth on trace metal dynamics (Mn, Mo, V, U). J. Sea Res. 87, 35–45. https://doi.org/10.1016/j.seares.2013.09.009 Peterson, B.J., Holmes, R.M., McClelland, J.W., Amon, R., Brabets, T., et al, 2016. PARTNERS Project Arctic River Biogeochemical Data. Arct. Data Cent. https://doi.org/doi:10.18739/A2166T Proshutinsky, A.Y., Johnson, M.A., 1997. Two circulation regimes of the wind-driven Arctic Ocean. J. Geophys. Res. Oceans 102, 12493–12514. https://doi.org/10.1029/97JC00738 Reckhardt, A., Beck, M., Greskowiak, J., Schnetger, B., Böttcher, M.E., Gehre, M., Brumsack, H.-J., 2017. Cycling of redox-sensitive elements in a sandy subterranean estuary of the southern North Sea. Mar. Chem. 188, 6–17. https://doi.org/10.1016/j.marchem.2016.11.003 Roeske, T., Rutgers van der Loeff, M., Middag, R., Bakker, K., 2012. Deep water circulation and composition in the Arctic Ocean by dissolved barium, aluminium and silicate. Mar. Chem. 132–133, 56–67. https://doi.org/10.1016/j.marchem.2012.02.001 Rudels, B., 2019. Arctic Ocean Circulation, in: Encyclopedia of Ocean Sciences. Elsevier, pp. 262–277. https://doi.org/10.1016/B978-0-12-409548-9.11209-6 Rudels, B., Quadfasel, D., 1991. Convection and deep water formation in the Arctic Ocean-Greenland Sea System. J. Mar. Syst. 2, 435–450. https://doi.org/10.1016/0924-7963(91)90045-V Rudnick, R.L., Gao, S., 2014. 4.1 - Composition of the Continental Crust, in: Treatise on Geochemistry. Elsevier, Oxford, pp. 1–51. Schlitzer, R., 2018. Ocean Data View 5.1.5 [WWW Document]. URL https://odv.awi.de Schlosser, P., Kromer, B., Ekwurzel, B., Bönisch, G., McNichol, A., Schneider, R., von Reden, K., Östlund, H.G., Swift, J.H., 1997. The first trans-Arctic 14C section: comparison of the mean ages of the deep waters in the Eurasian and Canadian basins of the Arctic Ocean. Nucl. Instrum. Methods Phys. Res. Sect. B Beam Interact. Mater. At. 123, 431–437. https://doi.org/10.1016/S0168-583X(96)00677- 5 Scholz, F., Hensen, C., Noffke, A., Rohde, A., Liebetrau, V., Wallmann, K., 2011. Early diagenesis of redox-sensitive trace metals in the Peru upwelling area – response to ENSO-related oxygen fluctuations in the water column. Geochim. Cosmochim. Acta 75, 7257–7276. https://doi.org/10.1016/j.gca.2011.08.007 Scholz, F., Siebert, C., Dale, A.W., Frank, M., 2017. Intense molybdenum accumulation in sediments underneath a nitrogenous water column and implications for the reconstruction of paleo-redox conditions based on molybdenum isotopes.

151

Geochim. Cosmochim. Acta 213, 400–417. https://doi.org/10.1016/j.gca.2017.06.048 Shiller, A., 2019. Dissolved Ba, Cd, Cu, Ga, Mn, Ni, and V concentration data from the US GEOTRACES Arctic Expeditions (GN01, HLY1502) from August to October 2015. Biol. Chem. Oceanogr. Data Manag. Off. BCO-DMO Dataset version 2019-07-09. https://doi.org/10.1575/1912/bco-dmo.772645.1 Shiller, A.M., Mao, L., 1999. Dissolved vanadium on the Louisiana Shelf: Effect of oxygen depletion. Cont. Shelf Res. 1007–1020. Sipler, R.E., Gong, D., Baer, S.E., Sanderson, M.P., Roberts, Q.N., Mulholland, M.R., Bronk, D.A., 2017. Preliminary estimates of the contribution of Arctic nitrogen fixation to the global nitrogen budget: Arctic nitrogen fixation. Limnol. Oceanogr. Lett. 2, 159–166. https://doi.org/10.1002/lol2.10046 Smethie, W.M., Schlosser, P., Bönisch, G., Hopkins, T.S., 2000. Renewal and circulation of intermediate waters in the Canadian Basin observed on the SCICEX 96 cruise. J. Geophys. Res. Oceans 105, 1105–1121. https://doi.org/10.1029/1999JC900233 Steele, M., 2004. Circulation of summer Pacific halocline water in the Arctic Ocean. J. Geophys. Res. 109. https://doi.org/10.1029/2003JC002009 Takematsu, N., Sato, Y., Okabe, S., Nakayama, E., 1985. The partition of vanadium and molybdenum between manganese oxides and sea water. Geochim. Cosmochim. Acta 49, 2395–2399. https://doi.org/10.1016/0016-7037(85)90239-X Talley, L.D., Pickard, G.L., Emery, W.J., Swift, J.H., 2011. Arctic Ocean and Nordic Seas, in: Descriptive Physical Oceanography. Elsevier, pp. 401–436. https://doi.org/10.1016/B978-0-7506-4552-2.10012-5 Trefry, J.H., Metz, S., 1989. Role of hydrothermal precipitates in the geochemical cycling of vanadium. Nature 342, 531–533. [dataset] Twining, B., Morton, P., Salters, V., 2019. Trace element concentrations (labile and total measurements) in particles collected with GO-Flo bottles and analyzed with ICP-MS from the US GEOTRACES Arctic cruise (HLY1502; GN01) from August to October 2015. Biol. Chem. Oceanogr. Data Manag. Off. BCO-DMO Dataset version 2019-07-02. https://doi.org/10.1575/1912/bco-dmo.771474.2 Twining, B.S., Rauschenberg, S., Morton, P.L., Ohnemus, D.C., Lam, P.J., 2015. Comparison of particulate trace element concentrations in the North Atlantic Ocean as determined with discrete bottle sampling and in situ pumping. Deep-Sea Res. Part II Top. Stud. Oceanogr. 116, 273–282. https://doi.org/10.1016/j.dsr2.2014.11.005 Vieira, L.H., Achterberg, E.P., Scholten, J., Beck, A.J., Liebetrau, V., Mills, M.M., Arrigo, K.R., 2019. Benthic fluxes of trace metals in the Chukchi Sea and their transport into the Arctic Ocean. Mar. Chem. 208, 43–55. https://doi.org/10.1016/j.marchem.2018.11.001 von Appen, W.-J., Schauer, U., Somavilla, R., Bauerfeind, E., Beszczynska-Möller, A., 2015. Exchange of warming deep waters across Fram Strait. Deep Sea Res. Part Oceanogr. Res. Pap. 103, 86–100. https://doi.org/10.1016/j.dsr.2015.06.003 Wang, D., Sañudo-Wilhelmy, S.A., 2008. Development of an analytical protocol for the determination of V(IV) and V(V) in seawater: Application to coastal

152

environments. Mar. Chem. 112, 72–80. https://doi.org/10.1016/j.marchem.2008.05.011 Wehrli, B., Stumm, W., 1989. Vanadyl in natural waters: Adsorption and hydrolysis promote oxygenation. Geochim. Cosmochim. Acta 53, 69–77. Whitmore, L.M., Shiller, A.M., 2016. Distribution and flux of methane in the 2015 U.S. GEOTRACES Arctic Section. Goldschmidt Abstr. 3408. Wu, F., Owens, J.D., Huang, T., Sarafian, A., Huang, K.-F., Sen, I.S., Horner, T.J., Blusztajn, J., Morton, P., Nielsen, S.G., 2019. Vanadium isotope composition of seawater. Geochim. Cosmochim. Acta 244, 403–415. https://doi.org/10.1016/j.gca.2018.10.010 Abrahamsen, E.P., Meredith, M.P., Falkner, K.K., Torres-Valdes, S., Leng, M.J., Alkire, M.B., Bacon, S., Laxon, S.W., Polyakov, I., Ivanov, V., 2009. Tracer-derived freshwater composition of the Siberian continental shelf and slope following the extreme Arctic summer of 2007: Siberian freshwater. Geophys. Res. Lett. 36, n/a- n/a. https://doi.org/10.1029/2009GL037341 Aguilar-Islas, A.M., Rember, R., Nishino, S., Kikuchi, T., Itoh, M., 2013. Partitioning and lateral transport of iron to the Canada Basin. Polar Sci. 7, 82–99. https://doi.org/10.1016/j.polar.2012.11.001 Alkire, M.B., Jacobson, A.D., Lehn, G.O., Macdonald, R.W., Rossi, M.W., 2017. On the geochemical heterogeneity of rivers draining into the straits and channels of the Canadian Arctic Archipelago: Canadian Arctic rivers. J. Geophys. Res. Biogeosciences 122, 2527–2547. https://doi.org/10.1002/2016JG003723 Alkire, M.B., Morison, J., Andersen, R., 2015. Variability in the meteoric water, sea-ice melt, and Pacific water contributions to the central Arctic Ocean, 2000-2014. J. Geophys. Res. Oceans 120, 1573–1598. https://doi.org/10.1002/2014JC010023 Bates, S.L., Hendry, K.R., Pryer, H.V., Kinsley, C.W., Pyle, K.M., Woodward, E.M.S., Horner, T.J., 2017. Barium isotopes reveal role of ocean circulation on barium cycling in the Atlantic. Geochim. Cosmochim. Acta 204, 286–299. https://doi.org/10.1016/j.gca.2017.01.043 Bauch, D., Schlosser, P., Fairbanks, R., 1995. Freshwater balance and the sources of deep 18 and bottom waters in the Arctic Ocean inferred from the distribution of H2 O. Prog. Oceanogr. 35, 53–80. Bauch, D., van der Loeff, M.R., Andersen, N., Torres-Valdes, S., Bakker, K., Abrahamsen, E.P., 2011. Origin of freshwater and polynya water in the Arctic Ocean halocline in summer 2007. Prog. Oceanogr. 91, 482–495. https://doi.org/10.1016/j.pocean.2011.07.017 Beszczynska-Möller, A., Fahrbach, E., Schauer, U., Hansen, E., 2012. Variability in Atlantic water temperature and transport at the entrance to the Arctic Ocean, 1997–2010. ICES J. Mar. Sci. 69, 852–863. https://doi.org/10.1093/icesjms/fss056 Bishop, J.K.B., 1988. The barite-opal-organic carbon association in oceanic particulate matter. Nature 332. Bourke, R.H., Newton, J.L., Paquette, R.G., Tunnicliffe, M.D., 1987. Circulation and water masses of the East Greenland shelf. J. Geophys. Res. 92, 6729. https://doi.org/10.1029/JC092iC07p06729

153

Bourke, R.H., Paquette, R.G., 1991. Formation of Baffin Bay Bottom and Deep Waters, in: Elsevier Oceanography Series. Elsevier, pp. 135–155. https://doi.org/10.1016/S0422-9894(08)70065-5 Cardinal, D., Savoye, N., Trull, T.W., André, L., Kopczynska, E.E., Dehairs, F., 2005. Variations of carbon remineralisation in the Southern Ocean illustrated by the Baxs proxy. Deep Sea Res. Part Oceanogr. Res. Pap. 52, 355–370. https://doi.org/10.1016/j.dsr.2004.10.002 Carmack, E.C., Yamamoto-Kawai, M., Haine, T.W.N., Bacon, S., Bluhm, B.A., Lique, C., Melling, H., Polyakov, I.V., Straneo, F., Timmermans, M.-L., Williams, W.J., 2016. Freshwater and its role in the Arctic Marine System: Sources, disposition, storage, export, and physical and biogeochemical consequences in the Arctic and global oceans: Freswater and the Arctic Marine System. J. Geophys. Res. Biogeosciences 121, 675–717. https://doi.org/10.1002/2015JG003140 Charbonnier, Q., Bouchez, J., Gaillardet, J., 2019. Black shale weathering in Mackenzie Basin: constraints from barium isotopes. Charbonnier, Q., Moynier, F., Bouchez, J., 2018. Barium isotope cosmochemistry and geochemistry. Sci. Bull. 63, 385–394. https://doi.org/10.1016/j.scib.2018.01.018 Charette, M.A., Kipp, L.E., Jensen, L.T., Dabrowski, J.S., Whitmore, L.M., in prep. The Transpolar Drift as a source of riverine and shelf-derived trace elements to the central Arctic Ocean. J. Geophys. Res. Oceans. Charette, M.A., Lam, P.J., Lohan, M.C., Kwon, E.Y., Hatje, V., Jeandel, C., Shiller, A.M., Cutter, G.A., Thomas, A., Boyd, P.W., Homoky, W.B., Milne, A., Thomas, H., Andersson, P.S., Porcelli, D., Tanaka, T., Geibert, W., Dehairs, F., Garcia- Orellana, J., 2016. Coastal ocean and shelf-sea biogeochemical cycling of trace elements and isotopes: lessons learned from GEOTRACES. Philos. Trans. R. Soc. Math. Phys. Eng. Sci. 374, 20160076. https://doi.org/10.1098/rsta.2016.0076 Charkin, A.N., Rutgers van der Loeff, M., Shakhova, N.E., Gustafsson, Ö., Dudarev, O.V., Cherepnev, M.S., Salyuk, A.N., Koshurnikov, A.V., Spivak, E.A., Gunar, A.Y., Ruban, A.S., Semiletov, I.P., 2017. Discovery and characterization of submarine groundwater discharge in the Siberian Arctic seas: a case study in the Buor-Khaya Gulf, Laptev Sea. The Cryosphere 11, 2305–2327. https://doi.org/10.5194/tc-11-2305-2017 Chow, T.J., Goldberg, E.D., 1960. On the marine geochemistry of barium. Geochim. Cosmochim. Acta 20, 192–198. https://doi.org/10.1016/0016-7037(60)90073-9 Colbert, D., McManus, J., 2005. Importance of seasonal variability and coastal processes on estuarine manganese and barium cycling in a Pacific Northwest estuary. Cont. Shelf Res. 25, 1395–1414. https://doi.org/10.1016/j.csr.2005.02.003 Cullen, J.T., Sherrell, R.M., 1999. Techniques for determination of trace metals in small samples of size-fractionated particulate matter: phytoplankton metals off central California. Mar. Chem. 67, 233–247. https://doi.org/10.1016/S0304- 4203(99)00060-2 Cutter, G., Andersson, P.S., Codispoti, L.A., Croot, P., Francois, R., Lohan, M., Obata, Hajime., Rutgers van der Loeff, M., 2014. Sampling and Sample-handling Protocols for GEOTRACES Cruises Version 2. http://www.geotraces.org/library-

154

88/scientific-publications/reports/169-sampling-and-sample-handling-protocols- for-geotraces-cruises Cutter, G., Kadko, D., Landing, W., 2019. Bottle data from the GEOTRACES Clean Carousel sampling system (GTC) on the Arctic Section cruise (HLY1502) from August to October 2015 (U.S. GEOTRACES Arctic project). Dataset Version 2019-07-29. https://doi.org/10.1575/1912/bco-dmo.647259.4 Dehairs, F., Chesselet, R., Jedwab, J., 1980. Discrete suspended particles of barite and the barium cycle in the open ocean. Earth Planet. Sci. Lett. 49, 528–550. https://doi.org/10.1016/0012-821X(80)90094-1 Dymond, J., Suess, E., Lyle, M., 1992. Barium in deep-sea sediment: a geochemical proxy for paleoproductivity 19. Eagle, M., Paytan, A., Arrigo, K.R., van Dijken, G., Murray, R.W., 2003. A comparison between excess barium and barite as indicators of carbon export: BARITE A PROXY FOR EXPORT PRODUCTION. Paleoceanography 18, n/a-n/a. https://doi.org/10.1029/2002PA000793 Fichot, C.G., Kaiser, K., Hooker, S.B., Amon, R.M.W., Babin, M., Belanger, S., Walker, S.A., Benner, R., 2013. Pan-Arctic distributions of continental runoff in the Arctic Ocean. Sci. Rep. 6. Fransson, A., Chierici, M., Anderson, L.G., Bussmann, I., Kattner, G., Peter Jones, E., Swift, J.H., 2001. The importance of shelf processes for the modification of chemical constituents in the waters of the Eurasian Arctic Ocean: Implication for carbon fluxes. Cont. Shelf Res. 21, 225–242. https://doi.org/10.1016/S0278- 4343(00)00088-1 Ganeshram, R.S., François, R., Commeau, J., Brown-Leger, S.L., 2003. An experimental investigation of barite formation in seawater. Geochim. Cosmochim. Acta 67, 2599–2605. https://doi.org/10.1016/S0016-7037(03)00164-9 Gong, D., Pickart, R.S., 2016. Early summer water mass transformation in the eastern Chukchi Sea. Deep Sea Res. Part II Top. Stud. Oceanogr. 130, 43–55. https://doi.org/10.1016/j.dsr2.2016.04.015 Guay, C.K., Kenison Falkner, K., 1997. Barium as a tracer of Arctic halocline and river waters. Deep Sea Res. Part II Top. Stud. Oceanogr. 44, 1543–1569. https://doi.org/10.1016/S0967-0645(97)00066-0 Guay, C.K.H., McLaughlin, F.A., Yamamoto-Kawai, M., 2009. Differentiating fluvial components of upper Canada Basin waters on the basis of measurements of dissolved barium combined with other physical and chemical tracers. J. Geophys. Res. 114. https://doi.org/10.1029/2008JC005099 Haine, T.W.N., Curry, B., Gerdes, R., Hansen, E., Karcher, M., Lee, C., Rudels, B., Spreen, G., de Steur, L., Stewart, K.D., Woodgate, R., 2015. Arctic freshwater export: Status, mechanisms, and prospects. Glob. Planet. Change 125, 13–35. https://doi.org/10.1016/j.gloplacha.2014.11.013 Hendry, K.R., Pyle, K.M., Barney Butler, G., Cooper, A., Fransson, A., Chierici, M., Leng, M.J., Meyer, A., Dodd, P.A., 2018. Spatiotemporal Variability of Barium in Arctic Sea-Ice and Seawater. J. Geophys. Res. Oceans 123, 3507–3522. https://doi.org/10.1029/2017JC013668

155

Holmes, R.M., McClelland, J.W., Tank, S.E., Spencer, R.G.M., Shiklomanov, A.I., 2018. Arctic Great Rivers Observatory. Water Quality Dataset, Version 20181010. Honjo, S., Krishfield, R.A., Eglinton, T.I., Manganini, S.J., Kemp, J.N., Doherty, K., Hwang, J., McKee, T.K., Takizawa, T., 2010. Biological pump processes in the cryopelagic and hemipelagic Arctic Ocean: Canada Basin and Chukchi Rise. Prog. Oceanogr. 85, 137–170. https://doi.org/10.1016/j.pocean.2010.02.009 Horner, T.J., Kinsley, C.W., Nielsen, S.G., 2015. Barium-isotopic fractionation in seawater mediated by barite cycling and oceanic circulation. Earth Planet. Sci. Lett. 430, 511–522. https://doi.org/10.1016/j.epsl.2015.07.027 Hsieh, Y.-T., Henderson, G.M., 2017. Barium stable isotopes in the global ocean: Tracer of Ba inputs and utilization. Earth Planet. Sci. Lett. 473, 269–278. https://doi.org/10.1016/j.epsl.2017.06.024 Jacquet, S.H.M., Dehairs, F., Cardinal, D., Navez, J., Delille, B., 2005. Barium distribution across the Southern Ocean frontal system in the Crozet–Kerguelen Basin. Mar. Chem. 95, 149–162. https://doi.org/10.1016/j.marchem.2004.09.002 Jeandel, C., Peucker-Ehrenbrink, B., Jones, M.T., Pearce, C.R., Oelkers, E.H., Godderis, Y., Lacan, F., Aumont, O., Arsouze, T., 2011. Ocean margins: The missing term in oceanic element budgets? Eos Trans. Am. Geophys. Union 92, 217–224. https://doi.org/10.1029/2011EO260001 Jensen, L.T., Wyatt, N.J., Twining, B.S., Rauschenberg, S., Landing, W.M., Sherrell, R.M., Fitzsimmons, J.N., 2019. Biogeochemical Cycling of Dissolved Zinc in the Western Arctic (Arctic GEOTRACES GN01). Glob. Biogeochem. Cycles 33, 343–369. https://doi.org/10.1029/2018GB005975 Jones, E.P., Rudels, B., Anderson, L.G., 1995. Deep waters of the Arctic Ocean: origins and circulation. Deep Sea Res. Part Oceanogr. Res. Pap. 42, 737–760. https://doi.org/10.1016/0967-0637(95)00013-V Jones, E.P., Swift, J.H., Anderson, L.G., Lipizer, M., Civitarese, G., Falkner, K.K., Kattner, G., McLaughlin, F., 2003. Tracing Pacific water in the North Atlantic Ocean. J. Geophys. Res. Oceans 108. https://doi.org/10.1029/2001JC001141 Kadko, D., Aguilar-Islas, A., Bolt, C., Buck, C.S., Fitzsimmons, J.N., Jensen, L.T., Landing, W.M., Marsay, C.M., Rember, R., Shiller, A.M., Whitmore, L.M., Anderson, R.F., 2019. The residence times of trace elements determined in the surface Arctic Ocean during the 2015 US Arctic GEOTRACES expedition. Mar. Chem. 208, 56–69. https://doi.org/10.1016/j.marchem.2018.10.011 Kipp, L.E., Charette, M.A., Moore, W.S., Henderson, P.B., Rigor, I.G., 2018. Increased fluxes of shelf-derived materials to the central Arctic Ocean. Sci. Adv. 4, eaao1302. https://doi.org/10.1126/sciadv.aao1302 Kipp, L.E., Kadko, D.C., Pickart, R.S., Henderson, P.B., Moore, W.S., Charette, M.A., 2019. Shelf‐basin interactions and water mass residence times in the Western Arctic Ocean: insights provided by radium isotopes. J. Geophys. Res. Oceans. https://doi.org/10.1029/2019JC014988 Klinkenberg, H., Van Borm, W., Souren, F., 1996. A theoretical adaptation of the classical isotope dilution technique for practical routine analytical determinations by means of inductively coupled plasma mass spectrometry. Spectrochim. Acta Part B At. Spectrosc. 51, 139–153. https://doi.org/10.1016/0584-8547(95)01386-5

156

Klinkhammer, G.P., Chan, L.H., 1990. Determination of barium in marine waters by isotope dilution inductively coupled plasma mass spectrometry. Anal. Chim. Acta 232, 323–329. Kondo, Y., Obata, Hajime., Hioki, N., Ooki, A., Nishino, S., Kikuchi, T., Kuma, K., 2016. Transport of trace metals (Mn, Fe, Ni, Zn and Cd) in the western Arctic Ocean (Chukchi Sea and Canada Basin) in late summer 2012. Deep Sea Res. Part Oceanogr. Res. Pap. 116, 236–252. https://doi.org/10.1016/j.dsr.2016.08.010 Lalande, C., Forest, A., Barber, D.G., Gratton, Y., Fortier, L., 2009. Variability in the annual cycle of vertical particulate organic carbon export on Arctic shelves: Contrasting the Laptev Sea, Northern Baffin Bay and the Beaufort Sea. Cont. Shelf Res. 29, 2157–2165. https://doi.org/10.1016/j.csr.2009.08.009 Lam, P.J., Marchal, O., 2015. Insights into Particle Cycling from Thorium and Particle Data. Annu. Rev. Mar. Sci. 7, 159–184. https://doi.org/10.1146/annurev-marine- 010814-015623 Lanos, R., 2009. Circulation regionale, masses d’eau, cycles d’evolution et transports entre la mer de Beaufort et le golfe d’Amundsen (Ph.D. Thesis). University du Quebec, Quebec, Canada. LeBlond, P.H., 1980. On the Surface Circulation in Some Channels of the Canadian Arctic Archipelago. ARCTIC 33, 189–197. https://doi.org/10.14430/arctic2554 Lecher, A., 2017. Groundwater Discharge in the Arctic: A Review of Studies and Implications for Biogeochemistry. Hydrology 4, 41. https://doi.org/10.3390/hydrology4030041 Lecher, A.L., Kessler, J., Sparrow, K., Garcia-Tigreros Kodovska, F., Dimova, N., Murray, J., Tulaczyk, S., Paytan, A., 2016. Methane transport through submarine groundwater discharge to the North Pacific and Arctic Ocean at two Alaskan sites: SGD methane transport. Limnol. Oceanogr. 61, S344–S355. https://doi.org/10.1002/lno.10118 Macdonald, R.W., Barrie, L.A., Bidleman, T.F., Diamond, M.L., Gregor, D.J., Semkin, R.G., Strachan, W.M.J., Li, Y.F., Wania, F., Alaee, M., Alexeeva, L.B., Backus, S.M., Bailey, R., Bewers, J.M., Gobeil, C., Halsall, C.J., Harner, T., Hoff, J.T., Jantunen, L.M.M., Lockhart, W.L., Mackay, D., Muir, D.C.G., Pudykiewicz, J., Reimer, K.J., Smith, J.N., Stern, G.A., Schroeder, W.H., Wagemann, R., Yunker, M.B., 2000. Contaminants in the Canadian Arctic: 5 years of progress in understanding sources, occurrence and pathways. Sci. Total Environ. 254, 93– 234. https://doi.org/10.1016/S0048-9697(00)00434-4 Marsay, C.M., Aguilar-Islas, A., Fitzsimmons, J.N., Hatta, M., Jensen, L.T., John, S.G., Kadko, D., Landing, W.M., Lanning, N.T., Morton, P.L., Pasqualini, A., Rauschenberg, S., Sherrell, R.M., Shiller, A.M., Twining, B.S., Whitmore, L.M., Zhang, R., Buck, C.S., 2018. Dissolved and particulate trace elements in late summer Arctic melt ponds. Mar. Chem. 204, 70–85. https://doi.org/10.1016/j.marchem.2018.06.002 Martinez-Ruiz, F., Paytan, A., Gonzalez-Muñoz, M.T., Jroundi, F., Abad, M.M., Lam, P.J., Bishop, J.K.B., Horner, T.J., Morton, P.L., Kastner, M., 2019. Barite formation in the ocean: Origin of amorphous and crystalline precipitates. Chem. Geol. 511, 441–451. https://doi.org/10.1016/j.chemgeo.2018.09.011

157

McClelland, J.W., Holmes, R.M., Dunton, K.H., Macdonald, R.W., 2012. The Arctic Ocean Estuary. Estuaries Coasts 35, 353–368. https://doi.org/10.1007/s12237- 010-9357-3 McLaughlin, F.A., Carmack, E.C., Ingram, R.G., Williams, W.J., Michel, C., 2004. Chapter 31. Oceanography of the Northwest Passage (26,P), in: The Global Coastal OCean, Interdisciplinary Regional Studes and Syntheses. Harvard university Press, pp. 1211–1242. McManus, J., Berelson, W.M., Klinkhammer, G.P., Kilgore, T.E., Hammond, D.E., 1994. Remobilization of barium in continental margin sediments. Geochim. Cosmochim. Acta 58, 4899–4907. https://doi.org/10.1016/0016-7037(94)90220-8 Melling, H., 2000. Exchanges of freshwater through the shallow straits of the North American Arctic, in: The Freshwater Budget of the Arctic Ocean, NATO Science Series. Kluwer Academic Publishers, The Netherlands, pp. 479–502. Michel, C., Hamilton, J., Hansen, E., Barber, D., Reigstad, M., Iacozza, J., Seuthe, L., Niemi, A., 2015. Arctic Ocean outflow shelves in the changing Arctic: A review and perspectives. Prog. Oceanogr. 139, 66–88. https://doi.org/10.1016/j.pocean.2015.08.007 Middag, R., de Baar, H.J.W., Laan, P., Bakker, K., 2009. Dissolved aluminium and the silicon cycle in the Arctic Ocean. Mar. Chem. 115, 176–195. https://doi.org/10.1016/j.marchem.2009.08.002 Millero, F.J., 1982. The effect of pressure on the solubility of minerals in water and seawater. Geochim. Cosmochim. Acta 46, 11–22. https://doi.org/10.1016/0016- 7037(82)90286-1 Milliman, J.D., Farnsworth, K.L., 2013. River Discharge to the Coastal Ocean: a global synthesis. Cambridge University Press, Cambridge, United Kingdom. Miyazaki, T., Kimura, J.-I., Chang, Q., 2014. Analysis of stable isotope ratios of Ba by double-spike standard-sample bracketing using multiple-collector inductively coupled plasma mass spectrometry. J. Anal. At. Spectrom. 29, 483. https://doi.org/10.1039/c3ja50311a Monnin, C., Jeandel, C., Cattaldo, T., Dehairs, F., 1999. The marine barite saturation state of the world’s oceans. Mar. Chem. 65, 253–261. https://doi.org/10.1016/S0304-4203(99)00016-X Nan, X.-Y., Yu, H.-M., Rudnick, R.L., Gaschnig, R.M., Xu, J., Li, W.-Y., Zhang, Q., Jin, Z.-D., Li, X.-H., Huang, F., 2018. Barium isotopic composition of the upper continental crust. Geochim. Cosmochim. Acta 233, 33–49. https://doi.org/10.1016/j.gca.2018.05.004 Newton, R., Schlosser, P., Mortlock, R., Swift, J., MacDonald, R., 2013. Canadian Basin freshwater sources and changes: Results from the 2005 Arctic Ocean Section: AOS 2005 freshwater sources and changes. J. Geophys. Res. Oceans 118, 2133– 2154. https://doi.org/10.1002/jgrc.20101 Ober, S., Rijkenberg, M.J.A., Gerringa, L.J.A., 2016. Physical oceanography measured with ultra clean CTD/Water sampler-system during POLARSTERN cruise PS94 (ARK-XXIX/3). R. Neth. Inst. Sea Res. Texel. https://doi.org/10.1594/PANGAEA.859560

158

Ohnemus, D.C., Auro, M.E., Sherrell, R.M., Lagerström, M., Morton, P.L., Twining, B.S., Rauschenberg, S., Lam, P.J., 2014. Laboratory intercomparison of marine particulate digestions including Piranha: a novel chemical method for dissolution of polyethersulfone filters. Limnol. Oceanogr. Methods 12, 530–547. https://doi.org/10.4319/lom.2014.12.530 Östlund, H.G., Hut, G., 1984. Arctic Ocean water mass balance from isotope data. J. Geophys. Res. 89, 6373. https://doi.org/10.1029/JC089iC04p06373 Peterson, I., Hamilton, J., Prinsenberg, S., Pettipas, R., 2012. Wind-forcing of volume transport through Lancaster Sound: TRANSPORT THROUGH LANCASTER SOUND. J. Geophys. Res. Oceans 117, n/a-n/a. https://doi.org/10.1029/2012JC008140 Planquette, H., Sherrell, R.M., 2012. Sampling for particulate trace element determination using water sampling bottles: methodology and comparison to in situ pumps: Particulate trace element sampling. Limnol. Oceanogr. Methods 10, 367–388. https://doi.org/10.4319/lom.2012.10.367 Prinsenberg, S., Hamilton, J., Peterson, I., Pettipas, R., 2009. Observing and interpreting the seasonal variability of the oceanographic fluxes passing through Lancaster Sound of the Canadian Arctic Archipelago, in: Nihoul, J.C.J., Kostianoy, A.G. (Eds.), Influence of Climate Change on the Changing Arctic and Sub-Arctic Conditions. Springer Netherlands, Dordrecht, pp. 125–143. https://doi.org/10.1007/978-1-4020-9460-6_10 Pyle, K.M., Hendry, K.R., Sherrell, R.M., Legge, O., Hind, A.J., Bakker, D., Venables, H., Meredith, M.P., 2018. Oceanic fronts control the distribution of dissolved barium in the Southern Ocean. Mar. Chem. 204, 95–106. https://doi.org/10.1016/j.marchem.2018.07.002 R Core Team, 2018. R: A language and environment for statistical computing. R foundation for Statistical computing, Vienna, Austria. Rember, R., 2018. Dissolved barium measured on water bottle samples during POLARSTERN cruise PS94 (ARK-XXIX/3) to the central Arctic ocean in 2015. PANGAEA. https://doi.org/10.1594/PANGAEA.896022 Roeske, T., Bauch, D., Rutgers van der Loeff, M., Rabe, B., 2012a. Utility of dissolved barium in distinguishing North American from Eurasian runoff in the Arctic Ocean. Mar. Chem. 132–133, 1–14. https://doi.org/10.1016/j.marchem.2012.01.007 Roeske, T., Rutgers van der Loeff, M., Middag, R., Bakker, K., 2012b. Deep water circulation and composition in the Arctic Ocean by dissolved barium, aluminium and silicate. Mar. Chem. 132–133, 56–67. https://doi.org/10.1016/j.marchem.2012.02.001 Rudels, B., 2018. Arctic Ocean Circulation, in: Reference Module in Earth Systems and Environmental Sciences. Elsevier. https://doi.org/10.1016/B978-0-12-409548- 9.11209-6 Rudels, B., 2015. Arctic Ocean circulation, processes and water masses: A description of observations and ideas with focus on the period prior to the International Polar Year 2007–2009. Prog. Oceanogr. 132, 22–67. https://doi.org/10.1016/j.pocean.2013.11.006

159

Rudels, B., Quadfasel, D., 1991. Convection and deep water formation in the Arctic Ocean-Greenland Sea System. J. Mar. Syst. 2, 435–450. https://doi.org/10.1016/0924-7963(91)90045-V Rushdi, A.I., McManus, J., Collier, R.W., 2000. Marine barite and celestite saturation in seawater. Mar. Chem. 69, 19–31. https://doi.org/10.1016/S0304-4203(99)00089-4 Schlitzer, R., 2018. Ocean Data View 5.1.5 [WWW Document]. URL https://odv.awi.de Schlosser, P., Bayer, R., Bönisch, G., Cooper, L.W., Ekwurzel, B., Jenkins, W.J., Khatiwala, S., Pfirman, S., Smethie, W.M., 1999. Pathways and mean residence times of dissolved pollutants in the ocean derived from transient tracers and stable isotopes. Sci. Total Environ. 237–238, 15–30. https://doi.org/10.1016/S0048- 9697(99)00121-7 Shaw, T.J., Moore, W.S., Kloepfer, J., Sochaski, M.A., 1998. The flux of barium to the coastal waters of the southeastern USA: the importance of submarine groundwater discharge. Geochim. Cosmochim. Acta 62, 3047–3054. https://doi.org/10.1016/S0016-7037(98)00218-X Shiller, A., 2019. Dissolved Ba, Cd, Cu, Ga, Mn, Ni, and V concentration data from the US GEOTRACES Arctic Expeditions (GN01, HLY1502) from August to October 2015. Biol. Chem. Oceanogr. Data Manag. Off. BCO-DMO Dataset version 2019-07-09. https://doi.org/10.1575/1912/bco-dmo.772645.1 Siebert, C., Nägler, T.F., Kramers, J.D., 2001. Determination of molybdenum isotope fractionation by double-spike multicollector inductively coupled plasma mass spectrometry: TECHNICAL BRIEF. Geochem. Geophys. Geosystems 2, n/a-n/a. https://doi.org/10.1029/2000GC000124 Steele, M., Morison, J., Ermold, W., Rigor, I., Ortmeyer, M., Shimada, K., 2004. Circulation of summer Pacific halocline water in the Arctic Ocean. J. Geophys. Res. 109, C02027. https://doi.org/10.1029/2003JC002009 Tang, C.C.L., Ross, C.K., Yao, T., Petrie, B., DeTracey, B.M., Dunlap, E., 2004. The circulation, water masses and sea-ice of Baffin Bay. Prog. Oceanogr. 63, 183– 228. https://doi.org/10.1016/j.pocean.2004.09.005 Tanhua, T., Jones, E.P., Jeansson, E., Jutterström, S., Smethie, W.M., Wallace, D.W.R., Anderson, L.G., 2009. Ventilation of the Arctic Ocean: Mean ages and inventories of anthropogenic CO 2 and CFC-11. J. Geophys. Res. 114, C01002. https://doi.org/10.1029/2008JC004868 Taylor, J.R., Falkner, K.K., Schauer, U., Meredith, M., 2003. Quantitative considerations of dissolved barium as a tracer in the Arctic Ocean. J. Geophys. Res. 108, 3374. https://doi.org/10.1029/2002JC001635 Taylor, S.R., McLennan, S.M., 1995. The geochemical evolution of the continental crust. Rev. Geophys. 33, 241. https://doi.org/10.1029/95RG00262 Thomas, H., Shadwick, E., Dehairs, F., Lansard, B., Mucci, A., Navez, J., Gratton, Y., Prowe, F., Chierici, M., Fransson, A., Papakyriakou, T.N., Sternberg, E., Miller, L.A., Tremblay, J.-É., Monnin, C., 2011. Barium and carbon fluxes in the Canadian Arctic Archipelago. J. Geophys. Res. 116, C00G08. https://doi.org/10.1029/2011JC007120 von Allmen, K., Böttcher, M.E., Samankassou, E., Nägler, T.F., 2010. Barium isotope fractionation in the global barium cycle: First evidence from barium minerals and

160

precipitation experiments. Chem. Geol. 277, 70–77. https://doi.org/10.1016/j.chemgeo.2010.07.011 Wadley, M.R., Bigg, G.R., 2002. Impact of flow through the Canadian Archipelago and Bering Strait on the North Atlantic and Arctic circulation: An ocean modelling study. Q. J. R. Meteorol. Soc. 128, 2187–2203. https://doi.org/10.1256/qj.00.35 Webster, R.K., 1960. Mass spectrometric isotope dilution analysis, in: Methods in Geochemistry. Interscience, London, pp. 202–246. Weingartner, T.J., Cavalieri, D.J., Aagaard, K., Sasaki, Y., 1998. Circulation, dense water formation, and outflow on the northeast Chukchi Shelf. J. Geophys. Res. Oceans 103, 7647–7661. https://doi.org/10.1029/98JC00374 Whitmore, L.M., Morton, P.L., Twining, B.S., Shiller, A.M., 2019. Vanadium cycling in the Western Arctic Ocean is influenced by shelf-basin connectivity. Mar. Chem. 216, 103701. https://doi.org/10.1016/j.marchem.2019.103701 Woodgate, R.A., Weingartner, T.J., Lindsay, R., 2012. Observed increases in Bering Strait oceanic fluxes from the Pacific to the Arctic from 2001 to 2011 and their impacts on the Arctic Ocean water column. Geophys. Res. Lett. 39. https://doi.org/10.1029/2012GL054092 Xiang, Y., Lam, P.J., submitted. Sixe-fractionated marine suspended particle dynamics in the Western Arctic Ocean from the U.S. GEOTRACES Arctic cruise (GN01). Yamamoto-Kawai, M., Carmack, E.C., McLaughlin, F.A., Falkner, K.K., 2010. Oxygen isotope ratio, barium and salinity in waters around the North American coast from the Pacific to the Atlantic: Implications for freshwater sources to the Arctic throughflow. J. Mar. Res. 68, 97–117. https://doi.org/10.1357/002224010793078988 Yang, Q., Dixon, T.H., Myers, P.G., Bonin, J., Chambers, D., van den Broeke, M.R., Ribergaard, M.H., Mortensen, J., 2016. Recent increases in Arctic freshwater flux affects Labrador Sea convection and Atlantic overturning circulation. Nat. Commun. 7, 10525. https://doi.org/10.1038/ncomms10525

161

CHAPTER V – CONCLUSIONS

This dissertation has considered the geochemical cycles of three trace elements— gallium, barium, and vanadium—in the Arctic Ocean. Gallium was used as a tracer of

Atlantic sourced waters in the Arctic Ocean, in part, because of the concentration difference between Atlantic seawater (~30 nmol kg-1) and Pacific water (~5 nmol kg-1).

Ga concentrations over the shelves suggested little modification of the Pacific endmember and, in the basins, Ga was conserved along isopycnals. Furthermore, by applying a 1-D vertical advection-diffusion model, it was determined that the Ga tracer reflected more realistic vertical distributions of Pacific water than the nutrient tracer. The application of Ga is therefore preferred over the use of less conservative nutrient-type tracers in the Arctic Ocean when performing water mass deconvolutions. Charette et al.

(in prep) addressed the distribution of Ga in Transpolar Drift waters and indicates that there is likely substantial removal of Ga from riverine waters in the estuary. Ho et al.

(2019) also indicated a source of Ga to Pacific Ocean deep waters, in part, from sediment resuspension. More continuous exposure of the shallow Arctic Shelves, due to decreased sea ice cover, may change the amount of sediment resuspension over shallow shelves, thereby influencing Ga distributions. Future studies with Ga in the Arctic Ocean should consider a pan-Arctic perspective and test the utility of Ga as a conservative tracer in other regional environments—such as the Canadian Arctic Archipelago.

Declining sea ice cover also has potential to influence Ba and V distributions since both elements have modifications imparted by the margins. Generally, it has been suggested that warming climate will drive increases in shelf-derived fluxes (e.g., Charette et al., in prep; Kipp et al., 2018; Lecher, 2017).

162

This study demonstrated that the dissolved Ba distribution has a large “non- conservative” source, contributing roughly 50% of the dBa budget. Fluxes of dissolved

Ba from shelf sediments largely accounts for the “non-conservative” source and inputs from sea ice or in situ particle formation are expected to be small. This source must be acknowledged if Ba is to be applied as a tracer of North American and Eurasian rivers.

Indeed, dissolution of particulate Ba formed on the shelf and advected into the basins has the potential to drive dissolved Ba distributions in the western Arctic Ocean. However, the exact mechanisms contributing to the shelf source of dBa remain unclear. To improve upon our understanding of the Ba distribution in the Arctic, studies must target the mechanisms of particle formation and dissolution and the source of Ba in shelf bottom waters. A comparison of shelf environments, such as those over the Barents Sea, Chukchi

Sea, and East Siberian Seas, may help tease apart potential mechanisms. Specifically, analysis of speciation, isotopes, and phase association to particulate material can be useful to identify the types of processes modifying elemental constituents.

In contrast to dBa, the shelf was determined to be a sink of V due to particle scavenging supported by iron oxide shuttling. This mechanism was proposed for the

Peruvian margin (Scholz et al., 2017, 2011); compared to the Peruvian margin, a much larger removal of V from the water column is observed. Indeed, unlike the Peruvian margin, low dV concentrations resulting from shelf exposure are observed widely in the surface waters of the western Arctic Ocean. These low concentrations of dV can be used to qualitatively describe shelf influenced waters. In consideration of declining sea ice and potential increases in sediment resuspension, there are two possible consequences of such environmental change. First, more rapid recycling of iron oxides may increase the

163

amount of scavenged V. However, more permanent removal of V from the water column requires reducing shelf sediments. Increased resuspension may support more oxic conditions in shelf bottom waters, simultaneously reducing the Fe flux from the sediments and decreasing the amount of V retained in the sediments. Coupled water column-sediment interface studies on the Arctic shelves must be conducted to identify the mechanisms controlling the dV distribution.

In considering all three trace elements, process studies may be useful to identify and separate the role of in situ processes from advective ones. Furthermore, the properties of Arctic outflow waters, such as those through Davis Strait and the Fram Strait, must be described in a time series context. Such a study could identify the potential that outflow regions have on modifying signatures developed in the Arctic Ocean interior and determine the fluxes of these TEs into the North Atlantic.

5.1 References

Charette, M.A., Kipp, L.E., Jensen, L.T., Dabrowski, J.S., Whitmore, L.M., in prep. The

Transpolar Drift as a source of riverine and shelf-derived trace elements to the

central Arctic Ocean. J. Geophys. Res. Oceans.

Ho, P., Resing, J.A., Shiller, A.M., 2019. Processes controlling the distribution of

dissolved Al and Ga along the U.S. GEOTRACES East Pacific Zonal Transect

(GP16). Deep Sea Res. Part Oceanogr. Res. Pap. 147, 128–145.

https://doi.org/10.1016/j.dsr.2019.04.009

Kipp, L.E., Charette, M.A., Moore, W.S., Henderson, P.B., Rigor, I.G., 2018. Increased

fluxes of shelf-derived materials to the central Arctic Ocean. Sci. Adv. 4,

eaao1302. https://doi.org/10.1126/sciadv.aao1302

164

Lecher, A., 2017. Groundwater Discharge in the Arctic: A Review of Studies and

Implications for Biogeochemistry. Hydrology 4, 41.

https://doi.org/10.3390/hydrology4030041

Scholz, F., Hensen, C., Noffke, A., Rohde, A., Liebetrau, V., Wallmann, K., 2011. Early

diagenesis of redox-sensitive trace metals in the Peru upwelling area – response to

ENSO-related oxygen fluctuations in the water column. Geochim. Cosmochim.

Acta 75, 7257–7276. https://doi.org/10.1016/j.gca.2011.08.007

Scholz, F., Siebert, C., Dale, A.W., Frank, M., 2017. Intense molybdenum accumulation

in sediments underneath a nitrogenous water column and implications for the

reconstruction of paleo-redox conditions based on molybdenum isotopes.

Geochim. Cosmochim. Acta 213, 400–417.

https://doi.org/10.1016/j.gca.2017.06.048

165

APPENDIX A – SUPPLEMENTAL TO CHAPTER II

Within this supplementary data, we include additional figures that add to the sections discussed within the manuscript. Furthermore, we include a summary of the statistical tests conducted in the advection diffusion model addressed in the manuscript.

Figures A1 – A3 depict the section results of the water mass deconvolution for the

Atlantic fraction (fAtl), meteoric fraction (fmet) and sea ice melt/formation fraction (fice).

The Pacific fraction is presented and discussed thoroughly in Chapter II Section 2.4.2

(Figure 2.6). The Atlantic fraction roughly mirrors the discussion of the Pacific fraction.

The meteoric and sea ice fractions are very similar between methods (within 0.05; i.e.,

5%).

166

Distribution of Atlantic waters.

The left column contains figures from the Makarov Basin (180°W) transect and the right column has figures for the Canada Basin

(150°W) transect. The bottom panel in each column depicts the respective bathymentry. (a,b) fAtl determined by the ANP method.

(c,d) fAtl determined by the Ga method. (e,f) The difference between fAtl (ANP) and fAtl (Ga)

167

Distribution of meteoric waters.

The left column contains figures from the Makarov Basin (180 W) transect and the right column has figures for the Canada Basin (150

W) transect. The bottom panel in each column represents the respective bathymetry. (a, b) fmet determined by the ANP method. (c, d) fmet determined by the Ga method. (e, f) The difference between fmet (ANP) and fmet (Ga).

168

Distribution of sea ice melt/formation waters.

The left column contains figures from the Makarov Basin (180 W) transect and the right column has figures for the Canada Basin

(150 W) transect. The bottom panel in each column represents the respective bathymetry. (a, b) fice determined by the ANP method.

(c, d) fice determined by the Ga method. (e, f) The difference between fice (ANP) and fice (Ga).

Figures A4 and A5 compare the ANP and Ga method to that of Yamamoto-Kawai et al. (2008; YK). The nutrient methods (ANP & YK) are roughly 1:1 for all parameters.

The Ga method predicts a greater Pacific fraction than the YK method; this is discussed in Section 2.4.2 of the manuscript.

169

Comparison of the Yamamoto-Kawaii et al. (2008) nutrient method to the

Newton et al. (2013) nutrient method.

Data are generally 1:1 (gray dashed line), but have some scatter. (a) Fraction of Atlantic. (b) Fraction of Pacific. (c) Fraction of meteoric. (d) Fraction of sea ice melt/formation.

170

Comparison of the Yamamoto-Kawaii et al. (2008) nutrient method to the Ga method (this study).

Data have similar patterns as described in the main text of the article (Figure 2.5) in comparing the ANP (Newton et al., 2013) method to the Ga method. (a) Fraction of Atlantic and (b) Pacific deviate substantially from the 1:1 line (gray dashed line), but the (c) meteoric and (d) sea ice fractions are generally 1:1. The following figures A6—A9 depict the results of the advection-diffusion modelling effort. “Station 1” is the ID given to the mean of all profiles north of 85°N and

“Station 2” is the ID given to the mean of all profiles south of 85°N (excluding shelf stations). The model assumes low upwelling; because the model simplifies the system we 171

adjusted the model so that the mean of the modeled profile and the mean of the actual profile (for the parameter on the x-axis) were at the same depth. The statistical results from comparing the modeled profiles to the observed are presented in Table S1. See section 3.4 of the manuscript for further discussion.

Salinity, temperature, and sigma advection-diffusion model results for

“Station 1”

Model results (thin black line) compared to observed data (thick black line) before adjustments (left panel) and after the median of the model is matched to the median of the data (right panel)

172

Salinity, temperature, and sigma advection-diffusion model results for

“Station 2”

Model results (thin black line) compared to observed data (thick black line) before adjustments (left panel) and after the median of the model is matched to the median of the data (right panel).

173

Advection-diffusion model results from the Pacific Fraction for “Station 1”

Model results (thin black line) compared to observed data (blue points) before adjustments (left panel) and after the median of the model is matched to the median of the data (right panel). The thick black line is the 1 m interpolated profile.

174

Advection-diffusion model results from the Pacific Fraction for “Station 2”

175

Table A.2 Summary of advection-diffusion model statistics. a. Physical Features > 85N

2 kz Parameter r SSR RMSE 1E-07 Salinity 0.56 16.6504 0.2893 5E-07 Salinity 0.73 7.8773 0.1990 1E-06 Salinity 0.81 4.5661 0.1511 5E-06 Salinity 0.94 1.6183 0.0889 1E-05 Salinity 0.91 2.5415 0.1092 1E-07 sigma 0.62 7.1242 0.1911 5E-07 sigma 0.78 3.0158 0.1244 1E-06 sigma 0.85 1.6842 0.0927 5E-06 sigma 0.94 1.1242 0.0748 1E-05 sigma 0.88 1.8202 0.0933 1E-07 Temp 0.80 38.1646 0.3795 5E-07 Temp 0.89 18.6330 0.2652 1E-06 Temp 0.94 8.9158 0.1831 5E-06 Temp 0.99 0.8950 0.0575 1E-05 Temp 0.93 4.5557 0.1278 b. Physical Features < 85N

2 kz Parameter r SSR RMSE 1E-07 Salinity 0.68 22.1305 0.3037 5E-07 Salinity 0.81 11.0923 0.2150 1E-06 Salinity 0.89 5.6794 0.1535 5E-06 Salinity 0.99 0.7009 0.0534 1E-05 Salinity 0.92 2.4189 0.0974 1E-07 sigma 0.69 11.8398 0.2235 5E-07 sigma 0.83 5.6743 0.1547 1E-06 sigma 0.90 2.7706 0.1079 5E-06 sigma 0.98 0.5169 0.0461 1E-05 sigma 0.91 1.6247 0.0803 1E-07 Temp 0.77 34.7823 0.3506 5E-07 Temp 0.86 18.7996 0.2577 1E-06 Temp 0.91 10.3967 0.1913 5E-06 Temp 1.00 0.1571 0.0233 1E-05 Temp 0.95 2.5037 0.0917

176

Table A1 (continued). c. Water Mass Distributions > 85N

2 kz Parameter r SSR RMSE 1E-07 fPac_ANP 0.90 0.0072 0.0347 5E-07 fPac_ANP 0.97 0.0014 0.0154 1E-06 fPac_ANP 0.98 0.0009 0.0123 5E-06 fPac_ANP 0.86 0.0057 0.0307 1E-05 fPac_ANP 0.74 0.0069 0.0340 1E-07 fPac_Ga 0.54 0.1500 0.1464 5E-07 fPac_Ga 0.71 0.0840 0.1095 1E-06 fPac_Ga 0.81 0.0501 0.0846 5E-06 fPac_Ga 0.96 0.0086 0.0351 1E-05 fPac_Ga 0.94 0.0068 0.0311 d. Water Mass Distributions < 85N

2 kz Parameter r SSR RMSE 1E-07 fPac_ANP 0.80 0.1411 0.1252 5E-07 fPac_ANP 0.90 0.0652 0.0851 1E-06 fPac_ANP 0.94 0.0330 0.0605 5E-06 fPac_ANP 0.99 0.0030 0.0183 1E-05 fPac_ANP 0.93 0.0130 0.0380 1E-07 fPac_Ga 0.61 0.2387 0.1629 5E-07 fPac_Ga 0.76 0.1240 0.1174 1E-06 fPac_Ga 0.83 0.0896 0.0946 5E-06 fPac_Ga 0.99 0.0051 0.0225 1E-05 fPac_Ga 0.99 0.0029 0.0172 Model-derived profiles were compared to observed profiles and assessed by considering the r2, sum of the squared residuals, and root mean squared error. In most cases all the statistics are in agreement. The “best match” is indicated in the table by highlighted green rows, if there was disagreement among the statistics, it is indicated in yellow. a) results of physical parameters (salinity, temperature, and sigma) at latitudes > 85°N. b) results of physical parameters (salinity, temperature, and sigma) at latitudes < 85°N. c) results of water mass (Pacific fraction) analysis at latitudes > 85°N. d) results of water mass (Pacific fraction) analysis at latitudes < 85°N.

177

APPENDIX B – SUPPLEMENTARY MATERIAL TO CHAPTER III

Table B.1 Standard Reference Material (SRM) summary and reproducibility analysis

GS GD SLRS-5 SLRS-3 OMP USM mean ± sd 44.3 ± 0.8 54.1 ± 0.9 NA NA NA RSD 1.80% 1.70% NA NA NA n 12 12 NA NA NA VUB mean ± sd NA NA 99.6 ± 3.6 95.3 ± 4.3 74.0 ± 1.4 RSD NA NA <1% NA NA n NA NA NA NA NA UAF mean ± sd NA NA NA NA NA RSD NA NA NA NA NA n NA NA NA NA NA

Ba:Ra ratio.

Ba:Ra trends and ratio determined by Bering and Chukchi Sea shelf data.

178

Arctic Ocean 1994 & 2015 Station Map.

To conduct the Ba mass balance we used central Arctic Ocean stations; generally, stations above ~75 N with a bottom depth < 1000 m

(dashed black bathymetric line) were considered as inside our box.

179

APPENDIX C – SUPPLEMENTARY MATERIAL TO CHAPTER IV

Co-Authors: Peter Morton, Benjamin Twining, Alan Shiller

Originally published in Marine Chemistry, reference:

L.M Whitmore, P.L. Morton, B.S. Twining, A.M Shiller. (2019). Vanadium

cycling in the Western Arctic Ocean is influenced by shelf-basin connectivity.

Marine Chemistry, 216: 103701. DOI: 10.1016/j.marchem.2019.103701

180

Particulate V distributions (labile and nonlithogenic).

Makarov and Canada Basin transects for (A, B) labile pV and (C, D) nonlithogenic pV. Station numbers are listed above the plots, the color bar to the right applies to both the Makarov and Canada Basin transect for a given variable and the maps below reference the transect for the section plots above it.

181

Dissolved and particulate trace elements versus total chlorophyll a and fucoxanthin.

(A) Dissolved V concentration plotted against fucoxanthin. (B) Labile pV versus fucoxanthin. (C) Labile pFe versus total chlorophyll a. (D) Labile pMn versus total chlorophyll a. (E) Labile pFe versus fucoxanthin. (F) Labile pMn versus fucoxanthin. Circles indicate data from shelf stations and squares indicate data from basin stations. A generally positive trend emerges, however there are no distinct trends in consideration of shelf-only and basin-only data.

182

Dissolved and particulate V versus nutrients.

(A) dV and (B) labile pV (logarithmic) plotted against Nitrogen (Nitrate + Nitrite). (C) dV and (D) labile pV (logarithmic) versus phosphate. (E) dV and (F) labile pV (logarithmic) versus silicate. Shelf data are in black and basin data are in orange. Generally, there is high scatter and no discernable trend, although there may be a positive relationship between dV and nitrogen.

183

Shelf profiles of dV and pV.

(A) dV profiles. (B) pV profiles. Map panel depicts location of the plotted stations. Stations 2, 3, and 66 have low dV, indicating heightened removal of dV relative to the expected endmember. Stations 2, 3, and 6 have high bottom water pV, which may suggest resuspension, post-bloom conditions, or active diffusion of reduced waters at the sediment-water interface – thereby increasing authigenic oxide formation in the water column.

184

Makarov Basin: full depth section profiles of process-associated pV distributions.

(A) Labile pV distribution. (B) Percent pVBIO relative to labile pV. (C) Percent pVCHEM, Fe relative to labile pV. (D) Percent pVCHEM, Mn relative to labile pV. Transect is depicted in the map below and station numbers are above panels A and B. 185

Canada Basin: full depth section profiles of process-associated pV distributions.

186

Pacific halocline: dV versus latitude.

Black data are original data and blue are normalized dV, shaded bands represent the associated 95% confidence bounds

187