1  

A Mass Budget for Mercury and Methylmercury in the Arctic Ocean 1  

2  Anne L. Soerensena,b, c,*, Daniel J. Jacobb, Amina Schartupa,b, Jenny A. Fisherd, Igor 3  Lehnherre,f, Vincent L. St. Louise, Lars-Eric Heimbürgergh, Jeroen E. Sonkeg, David P. 4  Krabbenhofti, Elsie M. Sunderlanda,b 5   6   7   8  a Harvard T.H. Chan School of Public Health, Department of Environmental Health, Boston 9  MA 02215, USA 10  b Harvard University, John A. Paulson School of Engineering and Applied Sciences, 11  Cambridge MA, 02138, USA 12  c Stockholm University, Department of Environmental Science and Analytical Chemistry, 13  Stockholm, 11418, Sweden 14  d University of Wollongong, Centre for Atmospheric Chemistry, Wollongong NSW, 2522, 15  Australia 16  e University of Alberta, Department of Biological Sciences, Edmonton, Alberta, Canada T6G 17  2E9 18  f University of Toronto-Mississauga, Department of Geography, Mississauga, Ontario, 19  Canada L5L 1C6 20  g CNRS, Geoscience Environment Toulouse, Midi-Pyrenees Observatory, Toulouse, 31400, 21  France 22  h University of Bremen, Department of Geosciences, Bremen, 28334, Germany 23  i U.S. Geological Survey, Middleton, Wisconsin 53562, United States 24  

25  * corresponding author: anne.soerensen@aces.su.se 26  

27  

28  

  2  

Key points: 29  

• A Hg mass budget suggest the terrestrial environment is the largest net source to the 30  

Arctic Ocean 31  

• Atmospherically deposited MeHg, from oceanic Me2Hg evasion, is the largest source 32  

to the PML 33  

• Geochemical box modeling shows changes in Hg loading rapidly affect Arctic PML 34  

concentrations 35  

  3  

Abstract 36  

Elevated biological concentrations of methylmercury (MeHg), a bioaccumulative neurotoxin, 37  

are observed throughout the Arctic Ocean but major sources and degradation pathways in 38  

seawater are not well understood. Our mass budget calculations show that high total Hg in 39  

Arctic seawater relative to other basins reflect large freshwater inputs and sea ice cover that 40  

inhibits losses through evasion. We find that most net MeHg production (20 Mg a-1) occurs in 41  

the subsurface ocean (20-200 m). There, it is converted to dimethylmercury (Me2Hg: 17 Mg 42  

a-1), which diffuses to the polar mixed layer and evades to the atmosphere (15 Mg a-1). Me2Hg 43  

has an atmospheric lifetime of only a few hours and rapidly degrades back to MeHg. We 44  

postulate that most evaded Me2Hg is re-deposited as MeHg and that atmospheric deposition is 45  

the largest net MeHg source (8.5 Mg a-1) to the biologically productive surface ocean. 46  

Riverine MeHg accounts for an additional 15% of inputs to the surface ocean (2.5 Mg a-1). 47  

MeHg concentrations in the Arctic Ocean are elevated compared to lower latitude oceans. 48  

Increasing MeHg inputs from the Arctic terrestrial landscape are likely in the future given 49  

increasing freshwater discharges and permafrost melt. This may offset potential declines 50  

driven by increasing evasion from ice-free surface waters. Our simulations illustrate that for 51  

the most biologically relevant regions of the ocean, regulatory actions that decrease 52  

atmospheric Hg deposition have the capacity to rapidly affect aquatic Hg concentrations. 53  

54  

55  

Index terms: 4805 56  

  4  

Introduction 57  

Accumulation of methylmercury (MeHg) in Arctic biota is a major concern for the 58  

health of northern populations that consume large quantities of marine foods [AMAP, 2011a; 59  

Riget et al., 2011b]. High MeHg concentrations in Arctic fish and marine mammals have been 60  

attributed to a combination of enhanced mercury (Hg) deposition at high latitudes and 61  

enhanced biomagnification due to climate-driven shifts in ocean biogeochemistry and trophic 62  

structure [Braune et al., 2015; Lehnherr, 2014; Loseto et al., 2008; Stern et al., 2012]. Prior 63  

work has characterized total mercury (Hg) inputs and reservoirs in the Arctic Ocean but has 64  

not quantified the link to bioavailable MeHg [Fisher et al., 2013; Kirk et al., 2012; Outridge 65  

et al., 2008; Zhang et al., 2015]. Here we use all available data to construct Arctic Ocean 66  

mass budgets for Hg and MeHg and gain insight into processes driving MeHg bioavailability 67  

for biota. 68  

In marine ecosystems, MeHg is produced from divalent inorganic Hg (HgII) by 69  

bacteria in benthic sediment and in the marine water column [Lehnherr, 2014]. It is also 70  

produced in freshwater ecosystems and wetlands and transported by rivers into the surface 71  

ocean [Nagorski et al., 2014]. Lehnherr et al. [Lehnherr et al., 2011] measured water column 72  

methylation at the subsurface chlorophyll maximum and the oxycline at multiple stations in 73  

the Canadian Arctic Archipelago and concluded that this is likely the major MeHg source to 74  

polar marine waters. Levels of dimethylmercury (Me2Hg) measured in Arctic surface 75  

seawater are unusually high compared to other oceans [Bowman et al., 2014; Horvat et al., 76  

2003]. Me2Hg is mainly produced from MeHg in the ocean as shown by incubation 77  

experiments in Arctic Coastal waters [Lehnherr et al., 2011]. Recent measurements have 78  

identified evaded Me2Hg from the Arctic surface ocean and subsequent atmospheric MeHg 79  

deposition as a substantial source to ocean surface waters and adjacent terrestrial ecosystems 80  

[Baya et al., 2015; Lin and Pehkonen, 1999; St Pierre et al., 2015]. 81  

  5  

Concentrations of total Hg are elevated in Arctic surface seawater relative to other 82  

basins [Heimburger et al., 2015; Mason et al., 2012]. Sea ice cover, an extensive continental 83  

shelf, large freshwater inputs, and salinity driven stratification of the surface mixed layer 84  

distinguish the Arctic Ocean from the mid-latitude oceans. Springtime releases of bromine 85  

result in rapid oxidation of gaseous elemental Hg (Hg0) to the water-soluble divalent species 86  

(HgII) and atmospheric Hg depletion events (AMDEs) [Schroeder et al., 1998; Steffen et al., 87  

2015]. AMDEs were originally hypothesized to drive large increases in oceanic Hg inputs, 88  

but subsequent work has indicated that much of the deposited HgII is reemitted back to the 89  

atmosphere after photochemical reduction in the surface ocean and snowpack [AMAP, 2011a; 90  

Kirk et al., 2006; Lalonde et al., 2002]. Rivers are a large source of total Hg to the Arctic 91  

Ocean, but prior modeling work has not evaluated their importance for MeHg cycling [Amos 92  

et al., 2014; Dastoor and Durnford, 2014; Fisher et al., 2012; Zhang et al., 2015]. 93  

Here we develop a five-box geochemical model for the Arctic Ocean parameterized 94  

using all available Hg and MeHg measurements. We force the model with changes in sea ice 95  

and anthropogenic Hg deposition from 1850 to 2010. We use this analysis to gain insight into 96  

critical processes affecting MeHg concentrations in Arctic Ocean biota and timescales of 97  

response associated with changes in external Hg sources. 98  

Model Overview 99  

Our five-box geochemical model for the Arctic Ocean includes compartments 100  

representing: (1) the low-salinity upper 20 m of the water column known as the polar mixed 101  

layer (PML); (2) the subsurface ocean that extends to the bottom on the shelf and to 200 m 102  

depth in the Central Arctic Basin and Baffin Bay; (3) the deep ocean below 200 m to the sea 103  

floor in the Central Basin and Baffin Bay; (4/5) active sediment layers on the shelf (2 cm) and 104  

in the Central Basin (1 cm). The active sediment layer depth characterizes benthic sediment 105  

  6  

that interacts with the water column through bioturbation/resuspension and chemical diffusion 106  

[Clough et al., 1997; Kuzyk et al., 2013; Trefry et al., 2014]. 107  

We estimate external inputs and exchanges/loss of Hg species (HgII, Hg0, MeHg, 108  

Me2Hg) from each compartment based on a comprehensive review of all measured 109  

concentrations and fluxes (Tables 1-4). Mass budgets for each species are used to derive first-110  

order rate coefficients and create a set of coupled first-order differential equations for changes 111  

in chemical mass over time following Sunderland et al. [Sunderland et al., 2010]. Processes 112  

in the model include: (1) external inputs from atmospheric deposition, rivers, erosion, snow 113  

and ice melt, and other oceans, (2) advective transport by seawater circulation, ice transport, 114  

and settling of suspended particulate matter, (3) diffusive transport through the water column 115  

and at the sediment-water and air-sea interfaces, and (4) chemical transformations through 116  

inorganic Hg redox reactions and methylation, and degradation of MeHg and Me2Hg. Details 117  

of mass budget calculations are provided in the Tables S3-S9. 118  

Hydrologic and solids budget 119  

Hydrologic and solids budgets used to characterize advective transport of Hg species 120  

are provided in the supplemental information (Figure S1). The hydrologic budget is based on 121  

seawater inflow and outflow from the North Atlantic and North Pacific, river discharge, 122  

precipitation and evapotranspiration [Beszczynska-Moller et al., 2012; Curry et al., 2011; 123  

Panteleev et al., 2006; Serreze et al., 2006; Smedsrud et al., 2010]. Internal circulation is 124  

based on Alfimov et al. [Alfimov et al., 2006]. 125  

We adapted the solids budget from Rachold et al. [Rachold et al., 2004] to include 126  

enhanced productivity with reduced sea ice cover in recent years [Hill et al., 2013]. 127  

Productivity (550 Tg C a-1) is distributed equally between the surface layer and mid-depth Chl 128  

max [Arrigo et al., 2011]. Settling of suspended particles is based on the fraction of solids 129  

remaining following remineralization at each depth in the water column [Cai et al., 2010; 130  

  7  

Moran et al., 1997; Rachold et al., 2004]. We do not estimate solids resuspension from 131  

benthic sediment due to limited data. Burial rates for benthic sediment are based on shelf and 132  

deep Central Arctic Basin measurements (Table 3). 133  

Mercury reservoirs 134  

We synthesized all data collected since 2004 on Hg species concentrations in seawater 135  

(Figure 1), sediment, ice, and biota to estimate reservoirs of Hg species in the Arctic Ocean 136  

(Tables 1-2). Volumes used to calculate reservoirs are provided in Table 3. Maximum 137  

reservoirs of MeHg in primary producers and zooplankton occur in the Arctic summer and are 138  

estimated from peak biomass, reservoirs of carbon (g C m-2) and Hg concentrations (Table 3). 139  

For phytoplankton and bacteria we used Hg concentrations in seston (<200 µm) as a proxy 140  

(Table 2) and for zooplankton (>200 µm) we estimate mean MeHg burden from Arctic 141  

observations (Table S2). 142  

Mercury fluxes 143  

Atmospheric Hg fluxes include direct deposition to ice-free ocean surface waters, a 144  

delayed pulse of inputs from seasonal meltwater, and evasion of Hg0 and Me2Hg from the 145  

surface ocean. We adopt the recent Hg0 evasion estimate of Zhang et al. [Zhang et al., 2015] 146  

that matches observational constraints from atmospheric monitoring data. Evasion of Me2Hg 147  

from ice-free surface waters of 15 Mg a-1 is based on the concentration gradient between the 148  

PML (Table 1) and marine boundary layer (Table 2), and mean wind speed over the Arctic 149  

Ocean (Table 3). Additional details are provided in the SI (Table S9). 150  

Atmospheric inputs of inorganic Hg to the ice-free surface ocean (Table 3) are taken 151  

from recent modeling using the GEOS-Chem global Hg model for consistency with the 152  

evasion flux [Fisher et al., 2013; Zhang et al., 2015]. Me2Hg has an atmospheric lifetime of 153  

only a few hours against decomposition to MeHg [Lin and Pehkonen, 1999], while the 154  

atmospheric lifetime of MeHg against photo-degradation is longer (2-5 days) [Bittrich et al., 155  

  8  

2011]. This allows MeHg to be scavenged in precipitation and dry deposited to the ocean and 156  

adjacent terrestrial ecosystems before degradation. We base atmospheric MeHg deposition 157  

(8.5 Mg a-1) on the evaded Me2Hg flux and the seasonal fraction of the surface ocean that is 158  

ice-free (Table 3; Table S3). Measured MeHg concentrations in Arctic precipitation (snow, 159  

rain) account for <0.5 Mg a-1 of this flux (Table 4) while the dry deposition flux, calculated 160  

using the HgII dry deposition velocity for water surfaces as a proxy for MeHg [Baya et al., 161  

2015] (Table 3), accounts for between 1 and 5 Mg a-1. Fog deposition provides another 162  

plausible pathway for enhanced deposition of water-soluble gases (including Hg) in the Arctic 163  

summer [Evenset et al., 2004; Poulain et al., 2007; Rice and Chernyak, 1997]. Presently no 164  

direct measurements of MeHg concentrations in Arctic fog droplets are available. However, at 165  

lower latitudes concentrations (17±19 pM) up to two orders of magnitude higher than oceanic 166  

precipitation have been observed [Weiss-Penzias et al., 2012]. 167  

We use a mid-point estimate of inorganic Hg releases from snowmelt from various 168  

models that range between 20 and 50 Mg a-1 [Dastoor and Durnford, 2014; Fisher et al., 169  

2012]. For MeHg, the snowmelt flux is based on the volume of snow deposited to the ice-170  

covered surface ocean and mean concentrations in aged snow and snow melt (Table 4). Hg 171  

and MeHg inputs from melting of multi-year sea ice are taken from Beattie et al. [Beattie et 172  

al., 2014]. 173  

We use inputs of inorganic Hg from Arctic rivers based on terrestrial dissolved 174  

organic carbon (DOC) inputs to the ocean from the modeling work of Dastoor and Durnford 175  

[Dastoor and Durnford, 2014]. This estimate agrees well with the upper bound from a 176  

compilation of empirical measurements from Amos et al. [Amos et al., 2014] and recent work 177  

by Zhang et al. [Zhang et al., 2015]. Inputs of MeHg from rivers are calculated from the 178  

observed fraction of total Hg (1-10%, Table 4). Coastal erosion of inorganic Hg is based on 179  

the solids budget (Figure S1A) and Hg concentrations in eroding material (61-114 ng g-1) 180  

  9  

from Leitch et al. [Leitch, 2006]. MeHg concentrations are assumed to be negligible in 181  

eroding material. 182  

Advective transport of Hg species with inflowing/outflowing seawater and ice is based 183  

on flow volumes and measured concentrations at the appropriate flow depths (Figure S1B). 184  

Internal transport of Hg species associated with seawater flow among boxes is based on the 185  

hydrologic budget (Figure S1A). The surface ocean is strongly stratified [Macdonald et al., 186  

2005], so we only consider vertical Hg fluxes in the PML associated with settling of 187  

suspended particles and diffusive transport. We estimate settling rates and burial of Hg 188  

species based on the solids budget and Hg concentration data from Table 2. 189  

Within the water column, the eddy diffusion flux is based on concentration gradients 190  

between the surface ocean (10 m), the average mid-depth peak (140 m) and the deep ocean 191  

(300 m) [Heimburger et al., 2015], and a mean diffusivity of 1 x 10-4 m2 s-1(range 0.5 -2.3 x 192  

10-4) [Joos et al., 1997; Shaw and Stanton, 2014; Strode et al., 2010]. For benthic sediment 193  

we use data from North Atlantic estuarine and shelf regions to estimate partitioning of HgII 194  

and MeHg between the dissolved and solid phases (log Kd: Hg = 3.5; MeHg = 2.7) [Hollweg 195  

et al., 2010; Schartup et al., 2015a; Sunderland et al., 2006]. Dissolved concentrations are 196  

used to calculate upper and lower bounds for diffusion of Hg species at the sediment-water 197  

interface [Sunderland et al., 2010] (Table S6). 198  

Chemical transformations 199  

We specify the seawater oxidation rate following the parameterization of Soerensen et 200  

al. [Soerensen et al., 2010] and constrain the reduction rate using the measured Hg0:HgII 201  

fraction in PML and subsurface seawater along with Hg0 evasion loss. Internal production and 202  

decomposition of MeHg and Me2Hg in the water column is based on a compilation of rate 203  

constants measured in seawater (Table S1). We calculate a base methylation rate in the deep 204  

ocean using experimental measurements from dark filtered seawater [Monperrus et al., 2007]. 205  

  10  

To account for the influence of enhanced microbial activity in the subsurface ocean (20-200 206  

meters) on methylation, we scale the base rate by monthly primary production. Our average 207  

methylation rate correspond to the upper range of experimentally measured values in summer 208  

at mid-latitudes (0.017 d-1) [Monperrus et al., 2007], and our calculated fall value (0.007 d-1) 209  

agrees with rates measured in the Canadian Arctic Archipelago [Lehnherr et al., 2011] (Table 210  

S1). No methylation rates have been measured in the Arctic Ocean PML and we therefore 211  

estimate the net biotic methylation needed to close the PML budget for MeHg. This is 212  

reasonable given significant, but low, net methylation recently reported in oxic marine waters 213  

of the sub-Arctic [Schartup et al., 2015b] and detectable methylation measured at the base of 214  

the PML in the Canadian Arctic Archipelago [Lehnherr et al., 2011]. 215  

Production of Me2Hg from MeHg in seawater is taken from direct rate 216  

measurements [Lehnherr et al., 2011]. Photolytic MeHg degradation to HgII in seawater is 217  

based on the mean of published rates (Table S1) and averaged summer photosynthetically 218  

active radiation (PAR) within the photic zone of ice free surface waters and melt ponds 219  

[Porter et al., 2011; Serreze and Barry, 2005] (Table 3). Parsing the photodemethylation rate 220  

into specific UV-A, UV-B and PAR fractions based on Arctic freshwater studies results in a 221  

lower degradation flux [Lehnherr and St Louis, 2009]. Biotic degradation/decomposition of 222  

Me2Hg to MeHg is based on data from the deep Pacific Ocean [Mason et al., 1995]. We 223  

neglect photodegradation of Me2Hg in the PML as Me2Hg is not readily photodegraded in the 224  

ocean [Black et al., 2009]. We constrain the pool of MeHg available for demethylation using 225  

the measured ratio of MeHg:HgII in seawater. 226  

Methylation in benthic sediment (0.03 d-1) is based on measurements from the North 227  

Atlantic shelf and estuaries [Heyes et al., 2006; Hollweg et al., 2010]. The benthic sediment 228  

demethylation rate is constrained using the dissolved fraction of total Hg as MeHg in 229  

porewater (Table S6). 230  

  11  

231  

Results and discussion 232  

Total Hg Budget 233  

Figure 2 shows that more than half of the Hg reservoir in the Arctic Ocean is 234  

contained in the active layers of benthic and shelf sediment (3900 Mg) followed by an 235  

additional 35% in the deep ocean and only 7% in the upper ocean. We estimate the total Hg 236  

reservoir in Arctic seawater at 2870 Mg, with a lifetime against losses of 13 years. Although 237  

the age of total Hg in the deep water (>200 m) is longer (45 years) it is still less than central 238  

Arctic basin tracer ages, that can exceed several hundred years, because the reservoir also 239  

includes Baffin Bay and upper intermediate waters. Evasion of Hg0 and Me2Hg (100 Mg a-1) 240  

are the major removal pathways from the water column, accounting for 44% of total losses 241  

(Figure 3). Seawater outflow to the North Atlantic accounts for an additional 86 Mg a-1 242  

removal of total Hg (38%) and sedimentation of suspended solids is relatively smaller at 38 243  

Mg a-1 (17% total losses). 244  

Figure 3 shows that the terrestrial environment accounts for more than a third of total 245  

Hg inputs to the water column (erosion and rivers: 83 Mg a-1), as does the atmosphere (direct 246  

deposition and meltwater: 76 Mg a-1). Ocean inflow from the North Atlantic and North 247  

Pacific account for an additional 53 Mg a-1 (24%), and sediment diffusion makes up the 248  

remaining 5% of total inputs (11 Mg a-1). Substantial inputs of total Hg from rivers and 249  

coastal erosion to the Arctic Ocean and declining sea ice cover drive high evasion (100 Mg a-250  

1) [Chen et al., 2015; Zhang et al., 2015]. The difference between evasion and inputs from 251  

meltwater and direct atmospheric deposition of total Hg in our budget (-23 Mg a-1) suggests 252  

the Arctic Ocean is a net source to the atmosphere, which is consistent with previous work 253  

[Fisher et al., 2012; Sonke and Heimburger, 2012; Zhang et al., 2015]. 254  

  12  

Figure 4 shows Arctic seawater is enriched in total Hg relative to inflowing waters 255  

from the North Atlantic and North Pacific Oceans at all depths resulting in a 26 Mg a-1 net 256  

loss from the Arctic. We infer from our mass budget that the observed enrichment in total Hg 257  

in Arctic seawater relative to mid-latitude basins is caused by higher relative freshwater 258  

inputs and ice-cover that inhibits losses through evasion. Future increases in mobilization of 259  

Hg from thawing permafrost [Leitch, 2006; Rydberg et al., 2010] could increase Hg inputs to 260  

the Arctic Ocean, offsetting the current declining trend due to enhanced losses from Hg0 261  

evasion with sea ice retreat suggested by Chen et al. [Chen et al., 2015]. 262  

Methylated Hg Species Budget 263  

The stability of MeHg and Me2Hg is enhanced under the dark, cold conditions with 264  

low biological activity found in the deep ocean that contains almost 80% of the methylated 265  

species reservoir (450 Mg) (Figure 2). However, the PML and subsurface ocean are the most 266  

important reservoirs for MeHg accumulation at the base of the food web due to the 267  

concentration of algal production and biological foraging in these regions. Summer peak 268  

concentrations of MeHg in phytoplankton and zooplankton indicate a reservoir of only 0.7 269  

Mg or about 2% of the MeHg reservoir in the upper ocean (0-200 m). 270  

Unlike total Hg, only a small fraction of methylated Hg species are found in benthic 271  

sediment (2.2%) (Figure 2), and we do not calculate a substantial diffusive flux from benthic 272  

sediments to the overlying water column (<1 Mg a-1). Data on Arctic-wide solids 273  

resuspension are needed to further elucidate on the importance of sediments as a MeHg 274  

source to the shelf water column. 275  

Similar to total Hg, the Arctic Ocean is enriched in methylated Hg species at all 276  

depths relative to inflowing seawater from the North Atlantic and North Pacific Oceans 277  

(Figure 4). For methylated Hg species, peak concentrations in excess of 200 fM are found in 278  

subsurface seawater (20-200 m). Subsurface peaks of methylated Hg species in seawater are 279  

  13  

observable across the global oceans [Bowman et al., 2014; Cossa et al., 2011; Sunderland et 280  

al., 2009] but occur at much shallower depths in the Arctic [Heimburger et al., 2015; 281  

Lehnherr et al., 2011]. Due to rapid Me2Hg evasion, a smaller fraction of methylated Hg is 282  

found as Me2Hg in the PML than subsurface water while concentrations of bioavailable 283  

MeHg are similar in the two layers (Table 1). This indicates that the subsurface peak is at 284  

least partly driven by the evasion of Me2Hg in the PML. The subsurface peak of methylated 285  

Hg species may also be enhanced by the pronounced halocline structure in the Arctic 286  

[Heimburger et al., 2015; Schartup et al., 2015b]. Schartup et al. [Schartup et al., 2015b] 287  

showed that salinity driven stratification can concentrate terrestrial dissolved organic matter 288  

(DOM) in a relatively restricted vertical zone, which in turn may stimulate the activity of 289  

methylating bacteria and increase methylated Hg production. Climate driven increases in 290  

stratification of Arctic seawater and declining sea ice cover may thus have a large impact on 291  

future levels of methylated Hg production and loss. 292  

Figure 5 shows net production of MeHg from HgII only occurs in the water column of 293  

the subsurface (20-200 m depth) ocean (20 Mg a-1). A large fraction of MeHg produced in the 294  

subsurface ocean is converted to Me2Hg (17 Mg a-1), much of which diffuses to the PML 295  

(~12 Mg a-1) and is evaded to the atmosphere (15 Mg a-1). Our budget suggests the major net 296  

source of MeHg to the PML is atmospheric redeposition of this evaded Me2Hg (~8.5 Mg a-1) 297  

originally produced in the subsurface ocean. 298  

In addition to deposition, MeHg sources to the PML include rivers (2.5 Mg a-1), 299  

diffusion from the subsurface ocean (1.6 Mg a-1), biotic degradation of Me2Hg in the water 300  

column (0.5 Mg a-1) and inflow from the North Atlantic and Pacific (0.3 Mg a-1). The 301  

reservoir of MeHg in the PML is small (3.8 Mg) and turnover is relatively fast (Figure 5). 302  

Thus, we assume on an annual basis the MeHg budget must balance and infer based on this 303  

assumption that a small net biotic production of MeHg in the range of other inputs occurs in 304  

  14  

the PML water column (5 Mg a-1). Lehnherr et al. [Lehnherr et al., 2011] observed significant 305  

methylation at the base of the PML in the Canadian Arctic Archipelago and Schartup et al. 306  

[Schartup et al., 2015b] measured net dark methylation in oxic estuarine seawater from the 307  

sub-Arctic. Both studies indicate that net biotic methylation in the PML is plausible. 308  

Response to global change and anthropogenic emissions 309  

Biogeochemical cycling of Hg in the Arctic Ocean is highly sensitive to ongoing 310  

climate variability [Fisher et al., 2013; Krabbenhoft and Sunderland, 2013; Stern et al., 311  

2012]. Mean air temperature is increasing rapidly driving decreases in the extent and 312  

thickness of sea ice cover [Bintanja et al., 2011; Kwok and Rothrock, 2009]. In addition, 313  

global shifts in emissions of anthropogenic Hg will affect atmospheric inputs of Hg to the 314  

Arctic ocean, either directly through atmospheric inputs or indirectly through terrestrial runoff 315  

[Stern et al., 2012]. 316  

Figure 6 shows modeled changes in total Hg in Arctic Ocean seawater forced by 317  

differences in atmospheric inputs and changes in sea ice cover. Atmospheric inputs since 318  

1850 are estimated by scaling current inputs using historical deposition scenarios from 319  

Horowitz et al. (2014). A 20% decrease of sea ice since 1975 is based on data from AMAP 320  

[AMAP, 2011b]. Modeled external inputs to the PML increased by 50% between 1850 and 321  

2010 and has an almost linear effect on total Hg concentrations in the PML (44%) and 322  

subsurface water (33%) (Figure 6) reflecting the short turnover time of the total Hg reservoir 323  

in the upper ocean (~3 year for 0-200 meter; Figure 5). Outridge et al. [Outridge et al., 2008] 324  

hypothesized that MeHg concentrations in the Arctic Ocean are unlikely to be affected by 325  

changes in external Hg inputs due to the large inorganic Hg reservoir, which they suggested 326  

provides inertia against changes in loading over time. Our simulations illustrate that for the 327  

most biologically relevant parts of the water column (0-200 meters), regulatory actions that 328  

  15  

decrease Hg emission and associated atmospheric deposition have the capacity to rapidly 329  

affect aquatic Hg concentrations. 330  

Concentrations of methylated Hg can be expected to rapidly follow shifts in inorganic 331  

Hg given rapid exchange between species shown in Figure 5. The impacts of changes in Hg 332  

loading may be obscured or enhanced by other biogeochemical shifts in the ecosystem 333  

affecting the lifetime of Hg species through evasion, methylation and degradation. Separating 334  

these processes allows us to diagnose the impacts of Hg emissions reductions. For example, 335  

Figure 6 shows the rapid decline in ocean Hg concentrations in recent years is for a large part 336  

attributable to increased evasion as sea ice cover disappears [Chen et al., 2015]. 337  

We are unable to simulate the counteracting influence of changes in terrestrial 338  

discharges due to lack of data, which represents a critical uncertainty in scenarios for MeHg 339  

accumulation in the changing Arctic. The relative importance of external MeHg sources to the 340  

Arctic Ocean depends in large part on the stability of MeHg complexes in seawater. Ongoing 341  

changes in the terrestrial landscape such as permafrost melt and increases in freshwater 342  

discharges of DOM are likely to increase MeHg inputs from Arctic rivers to the ocean in the 343  

future [Krabbenhoft and Sunderland, 2013; Rydberg et al., 2010]. Recent work by Jonsson et 344  

al. [Jonsson et al., 2014] suggests riverine MeHg bound to terrestrial DOM is both resistant to 345  

degradation and biologically available. Thus, the overall lifetime of MeHg in the ocean is 346  

likely to be enhanced in the future with increasing terrestrial DOM discharges, which would 347  

lead to increases in the bioavailable MeHg reservoir. 348  

Summary and Conclusion 349  

We have developed a five-box biogeochemical model for the Arctic Ocean to gain 350  

insight into processes and timescales driving changes in MeHg concentrations in the Arctic 351  

Ocean. The model includes compartments representing three ocean and two sediment 352  

  16  

reservoirs: the PML, subsurface and deep ocean water, and active sediment layers on the shelf 353  

and in the Central Basin. 354  

Results from the total Hg mass budget calculations suggest that exchange with both 355  

the atmosphere (direct deposition and snowmelt) and lower latitude ocean results in net loss 356  

of total Hg from the Arctic Ocean. The terrestrial landscape is the only net Hg source. We use 357  

the box-model simulation to demonstrate the importance of the sea ice cover in controlling 358  

surface ocean evasion loss. From this we infer that the observed enrichment in total Hg in 359  

Arctic seawater relative to mid-latitude basins is caused by higher relative freshwater Hg 360  

inputs and ice cover that inhibits losses through evasion. 361  

The methylated Hg budget suggests that most net MeHg production (20 Mg a-1) 362  

occurs in the subsurface ocean (20-200 m). There, it is subsequently converted to Me2Hg (17 363  

Mg a-1), which readily diffuses to the PML and is evaded to the atmosphere (15 Mg a-1). 364  

Me2Hg has an atmospheric lifetime of only a few hours and rapidly photochemically degrades 365  

back to MeHg. We postulate that most evaded Me2Hg is re-deposited to the ocean and 366  

surrounding landscape as MeHg and that atmospheric MeHg deposition is the largest net 367  

source (8.5 Mg a-1) to the biologically productive surface ocean (0-20 m). We find that other 368  

important MeHg sources to the PML are river input (2.5 Mg a-1) and postulated net biotic 369  

methylation (5 Mg a-1). Rivers are likely to be disproportionately important for biological 370  

uptake in marine food webs since binding to terrestrial DOM extends the MeHg lifetime in 371  

the water column. We suggest that the overall lifetime of MeHg in the Arctic ocean is likely 372  

to be enhanced in the future with increasing terrestrial DOM discharges. 373  

We forced our box model of the Arctic Ocean with changes in sea ice and 374  

anthropogenic Hg deposition from 1850-2010. Our results show that because of the short 375  

lifetime (3 years) of total Hg in the most biologically relevant parts of the water column (0-376  

  17  

200 meters), regulatory actions that decrease Hg emission and associated atmospheric 377  

deposition have the capacity to rapidly affect aquatic Hg concentrations. 378  

  18  

Acknowledgements 379  

We acknowledge financial support from the U.S. National Science Foundation (OCE 380  

1130549, PLR 1023213). ALS acknowledges financial support from the Carlsberg 381  

Foundation (2012_01_0860). JAF acknowledges financial support from a University of 382  

Wollongong Vice Chancellor’s Postdoctoral Fellowship. JES acknowledges financial support 383  

from the European Research Council (ERC-2010-StG_20091028). IL and VSL acknowledge 384  

financial support from the Northern Contaminants Program and ArcticNet. 385  

386  

  19  

Table 1. Measured concentrations of Hg species in Arctic Ocean seawater. Cruise tracks for observations are 387  shown in Figure 1. Averages used in our budget are based on raw data from the cruises presented here. 388  Region Total Hg (pM) Methylated Hg

(fM) Fraction Methylated (%)

Me2Hg (fM) Hg0 (fM)

Polar Mixed Layer (<20 m)

1.56 ± 0.32a 2.92 ± 2.88c

2.09 ± 1.85b 0.55–1.40e

0.70 - 1.20j

2.29 ± 2.00f 1.91 ± 9.24g 1.82 ± 0.45h

64 ± 43a 141±79c

474±70j

284, 454, 364k

102f

105±61g

250h

122l

146m

4.5±3.4a 5c

45.2j

30-40k

5.5f

4.4g

13.7h

60±70c

55±20j

37f

50g

168h

34±30l

17±10m

126±51c, 220±110d

643±179j

30f

89g

127h

Best Estimate

2.0±1.8 130±110 90±70 (MeHg)

7% 50±40 180±90

Subsurface Ocean (20-shelf bottom or 200m)

1.10 ± 0.20a 1.43±1.11c 2.02±0.52f 1.92±0.32g 1.57±0.31h 1.1±0.45i

216±95a 445±180c 407f 627g 517h 200±150i

248l

212m

19.3±7.3a 31c 20f, 33g 33h 18±14i

362±184c

277f

331g

427h

138l

65m

171±149c 82f

48g

103h

Best Estimate 1.29±0.53 320±220 110±70 (MeHg)

25% 220±150 110±70

Deep Central Basin and Baffin Bay (>200 m)

1.00±0.25a 201±86a 20.6±9.5a

Best Estimate 1.00±0.25 200±90 20% n/a = not available. Best Estimate is used in mass budget calculations. The Hg0 reservoir in the PML is estimated 389  by subtracting Me2Hg concentrations from measured dissolved gaseous mercury (DGM). The best estimate of 390  MeHg concentrations in the upper ocean is based on the ratio of MeHg:Me2Hg (PML: 2:1 and subsurface: 0.7:1) 391  [Lehnherr et al., 2011]. 392  a[Heimbürger et al., 2015] Aug-Sep 2011 Cruise in the Central Basin. See Figure 1. 393  b[Zdanowicz et al., 2013], Baffin Bay July 2008 394  c[Kirk et al., 2008], Canadian Arctic Archipelago/border to Baffin Bay August-October 2005 (station 1-11). 395  d [Andersson et al., 2008] July-September measurements in Baffin Bay, Shelf and Central Arctic Basin reported 396  as dissolved gaseous Hg (Hg0+Me2Hg). 397  e[Chaulk et al., 2011], Canadian Arctic Archipelago, March-May 2008 398  f[Lehnherr et al., 2011], Canadian Arctic Archipelago, October 2007 399  g[Lehnherr, 2015], Canadian Arctic Archipelago, September 2006. 400  h[Lehnherr, 2015], Canadian Arctic Archipelago, August 2010. 401  i[Wang et al., 2012], Beaufort Sea, March – July 2008 402  j[St Louis et al., 2007], Canadian Arctic Archipelago May 2004 and May 2005k 403  l[Baya et al. 2014], Canadian Arctic Archipelago July-August 2010 and 2011m 404  

405  

  20  

Table 2. Measured concentrations of Hg species in ice, suspended particles in the water column, benthic 406  sediment and air from the Arctic Ocean. 407   408  Medium Total Hg MeHg

Fraction MeHg (%)

Ice (pM) 4.0, 5.6, 12,7 (0.65-61)a

0.5-4.0, max 20b

3.9±1.0c

0.2, 1.35 (<.10-2.64)a

3.5-10a

Best Estimate 7.4 (2.8-10.0) 0.75 (0.15-1.3) 10% Suspended solids (ng g-1 dry weight)

33±26 (4-88)d

36±27f

74± 27 (45-98)g

1.4±1.1 (0.15-3.51)d

1.1e

<0.15f

3.7±1.0 (2.6-4.4)g

<4d

<0.5f

Best Estimate 50 (30-70) 2 (0.15-3.51) 4% Shelf sediment (ng g-1 dry weight)

31±10 (5-55)h

50-100i

23 (8-40)k

41±29l

54±23m

0.15±0.07 (0.03-0.29)h

0.05-0.37i

0.37±0.36 (<0.01-1.41)j

0.43±0.17h

~0.05-0.20i

0.70±0.40j

Best Estimate 47±18 0.19±0.12 0.4% Central Basin/Baffin Bay sediment (ng g-1 dry weight)

<50-70n <0.05-0.10n ~0.05n

Best Estimate 47±18 0.05 (0.01-0.10) 0.1% Marine Boundary Layer 1.7±0.4 (ng Hg0 m-3)o 3.8±3.1 (pg Me2Hg m-3)p Best Estimate 1.7 3.8 Best estimate for each medium is used in mass budget calculations. 409  a[Beattie et al., 2014], Multitear sea ice from three cores in the Beaufort Sea in May, August and September 410  b[Chaulk et al., 2011], Multiyear sea ice from the Canadian Arctic Archipelago in May 411  c[Poulain et al., 2007], multiyear sea ice from the Canadian Arctic Archipelago in June 412  d[Burt, 2012], total suspended solids from the Canadian Arctic Archipelago. 413  e[Lehnherr, 2015], total suspended solids from the Canadian Arctic Archipelago 414  f[Pucko et al., 2014], particulate organic matter from the Canadian Arctic Archipelago 415  g[Graydon et al., 2009], total suspended solids in the Beaufort Sea 416  h[Fox et al., 2013], benthic sediment Chukchi Sea 417  i[Kading and Andersson, 2011], benthic sediment Chukchi Sea and Yermak Plateau 418  j[Fox et al., 2013], benthic sediment Beaufort Sea 419  k[ICES], Bering Sea upper 1-2 cm benthic sediment from 2004, 2006, 2009 420  l[Trefry et al., 2003], upper 2 cm Beaufort Sea benthic sediment 421  m[Canario et al., 2013], Canadian Arctic Archipelago >65N benthic surface sediment 422  n[Kading and Andersson, 2011], Central Basin surface sediment 423  o[Sommar et al., 2010] 424  p[Baya et al., 2014] 425   426  

  21  

Table 3. Physical characteristics of Arctic Ocean basin used to develop mass budgets for Hg species 427  Parameter Value Reference Volume of Arctic Ocean (m3) 136×1014 [Jakobsson, 2002] Volume of polar mixed layer (0-20 m) (m3) 2.22×1014 [Jakobsson, 2002] Volume of mid-depth water (shelf: 20-bottom; Central Basin and Baffin Bay: 20-200 m) (m3)

16.1×1014 [Jakobsson, 2002]

Volume of water below 200 meters (m3) 118×1014 [Jakobsson, 2002] Volume of sea ice (m3) 0.1×1014 [Serreze et al., 2006] Volume of active shelf sediment (m3) (0.02 m depth) 18×1010 Volume of active deep sediment (m3) (0.01 m depth) 5×1010 Surface area of shelf regions (m2) 608×1010 [Jakobsson, 2002] Surface area of the Central Basin and Baffin Bay (m2)

500×1010 [Jakobsson, 2002]

Total Arctic surface area (m2) 1109×1010 [Jakobsson, 2002] Bacterial biomass (g C m-2) 0.34 [Kirchman et al., 2009] Phytoplankton shelf biomass (g C m-2) 1.75 (1.00-

2.50) [Kirchman et al., 2009]

Phytoplankton Central Basin biomass (g C m-2) 0.50 [Kirchman et al., 2009] Zooplankton shelf biomass (<82oN) (g dw m-2) 6.9±4.1 [Kosobokova and Hirche, 2009] Zooplankton Central Basin biomass (>82oN) (g dw m-2)

2.5±0.5 [Kosobokova and Hirche, 2009]

Zooplankton [MeHg], best estimate (ng g-1 dw) 11 (5-35) Table S2 Carbon to dry weight plankton 2 [Hedges et al., 2002; Redfield et al.,

1963] Average precipitation (mm) 340 [Serreze et al., 2006] Fraction of summer precipitation 0.6 [Serreze and Barry, 2005] Fraction of ice-free water in summer (Apr-Sept., average)

0.65 [Macdonald et al., 2005; NSIDC, 2013]

Fraction of ice-free water in winter (Oct.-March, average)

0.10 [Macdonald et al., 2005; NSIDC, 2013]

Shelf sediment burial rate (m a-1) 30×10-5 [Kuzyk et al., 2013; Pirtle-Levy et al., 2009; Polyak et al., 2009]

MeHg (HgII) deposition velocity 1 cm s-1 [Zhang et al., 2009] Deep sediment burial rate (m a-1) 3×10-5 [Polyak et al., 2009]

Average concentration of Chl a in mixed layer (mg/L)

0.6×10-3 [Galand et al., 2008; Jin et al., 2012]

Average summer shortwave radiation (W m-2) 160 [Hatzianastassiou et al., 2005] Concentration of DOC in water column (mg L-1) 0.8 [Dittmar and Kattner, 2003; Hansell et

al., 2012] Net primary production 0.28 [Arrigo and van Dijken, 2011; Hill et

al., 2013] Average wind speed at 10 m above sea surface 5.6 (4-6) [Nummelin et al., 2015; Serreze and

Barry, 2005] 428   429  

  22  

Table 4. External inputs of Hg species to the Arctic Ocean. SI Figure S2 and S3 report uncertainty ranges on 430  reservoirs and fluxes not given here. 431  Medium Total Hg

MeHg

Fraction MeHg (%)

Rain and fresh snow (pM) 2.6±0.75a

0.13±0.05b

0.80±0.35a

0.50c

30a

Best Estimate for wet deposition (pM)

0.50 (0.15-0.80)

Surface snow (pM) 15.8±15.5d

35.7e 3.5±2.8f 10.2±3.5g

0.25±0.95d

0.1e

0.35±0.15f

1.6d

0.2e

10f

Snowpack (pM) 39i 2.6-8.8j 6.7-19.4k

10-250l

170m

0.37±0.10h

0.35-0.43j

<0.08-0.10k

21-31j

Runoff (pM) 3.6±0.8n

5.8±4.8o 10.1±3.2o

0.77±0.41n

0.31±0.19p 19.9n

3.2p

Best Estimate for melt water (pM)

15 (2.5-30) 0.3 (0.1-0.7) 5 (2-20)

Rivers and rivulets (pM) 45.7±27.5q

61.3±8.2r

35.8s

33.5-67.7t

73v

31, 39, 24, 18, 14x

~0.35q

0.35±0.05r

0.10±0.10u

1.1 (5.2 shelf)q

0.2-1.5r

1.3u

Best Estimate for rivers (pM)

5 (1-10)

Erosion (ng g-1) 114, 61, 67y Best Estimate for erosion (ng g-1)

81 (61-114) 0

a[Zdanowicz et al., 2013], fresh snow from Baffin Bay collected in April. 432  b[Hammerschmidt et al., 2006], rain collected from the Canadian Arctic Archipelago collected in June 433  c[Lehnherr et al., 2012], rain collected from the Canadian Arctic Archipelago collected in July 434  d[St Louis et al., 2005], surface snow from the Canadian Arctic Archipelago collected in April and May 435  e[St Louis et al., 2007], surface snow from the Canadian Arctic Archipelago collected In May (median) 436  f[Zdanowicz et al., 2013], surface snow from Baffin Bay collected in April. 437  g[Poulain et al., 2007], surface snow from the Canadian Arctic Archipelago collected in June 438  h[St Louis et al., 2005], fall-winter snowpack from the Canadian Arctic Archipelago 439  i[Lu et al., 2001], fall-winter snowpack from the Central Basin collected in November-January 440  j[St Louis et al., 2005], spring snowpack from the Canadian Arctic Archipelago in April-May (means), first & second year 441  snow 442  k[St Louis et al., 2007], spring snowpack from the Canadian Arctic Archipelago in May (medians), first and second year 443  snow 444  l[Poulain et al., 2007], spring snowpack from the Canadian Arctic Archipelago in June 445  m[Lu et al., 2001], spring snowpack from the Central Basin collected in February to May 446  n[St Louis et al., 2005], supraglacial runoff from the Canadian Arctic Archipelago collected in June and July 447  o[Zdanowicz et al., 2013], supraglacial runoff from Baffin Bay collected in June 448  p[St Louis et al., 2005], subglacial runoff from the Canadian Arctic Archipelago collected in June and July 449  q[Graydon et al., 2009], Mackenzie River Canada in June to August 450  r[Emmerton et al., 2013], Mackenzie River Canada ice free season 451  s[Leitch et al., 2007], Mackenzie River Canada 452  t[Wang et al., 2012], Mackenzie River and Horton River, Canada in open water (means) 453  u[Zdanowicz et al., 2013], Baffin Bay, Canada in June 454  v[Riget et al., 2011a], Zackenberg, Greenland in May to October 455  x[Amos et al., 2014], Mackenzie, Lena, Ob, Yenisei, and Kolyma rivers 456  y[Leitch, 2006], coastal erosion along the Beaufort Sea coast 457  

458  

  23  

Figure captions 459  

Figure 1. Recent cruises in the Arctic Ocean that collected samples for Hg species 460  

measurements and Arctic Ocean boundaries as specified in this work. Light brown to dark 461  

blue toning represents increasing depth.    462  

463  

Figure 2. Relative distribution of Hg species reservoirs (Mg) in different components of the 464  

Arctic Ocean. 465  

466  

Figure 3. Inputs and losses (Mg a-1) including uncertainty ranges for total Hg. Each ocean 467  

depth is indicated by a different color on the plot. 468  

469  

Figure 4. Total Hg and methylated Hg concentrations in seawater flowing into (red) and out 470  

of (blue) the Arctic Ocean. Atlantic Ocean concentrations for each depth interval are based on 471  

2013 measurements in the North Atlantic (50°N - 62°N) (Text S1). Pacific Ocean 472  

concentrations are from the upper 50 m of the water column close to the shallow sill at the 473  

Bering Strait [Sunderland et al., 2009]. We use Hg species concentrations in seawater and ice 474  

leaving the Arctic through the Barents Sea, Fram Strait, and Davis Strait (Table 1). 475  

476  

Figure 5. Reservoirs (Mg) and mass flows (Mg a-1) in the Arctic Polar Mixed Layer and 477  

subsurface ocean. Red arrows indicate external sources, blue arrow losses out of the system, 478  

black arrows internal fluxes, and white arrows indicate net fluxes. Figure 3 shows fluxes for 479  

the deep ocean, which cannot be resolved on species specific basis due to lack of data. 480  

481  

Figure 6. Simulated impact of atmospheric mercury loading and sea-ice cover on total Hg 482  

concentrations in Arctic Ocean seawater. Atmospheric inputs are calculated by scaling current 483  

  24  

Arctic atmospheric inputs by the historical trajectory simulated by Horowitz et al. [2014] 484  

based on shifts in global emissions. The trend in sea-ice cover is based on AMAP [2011]. No 485  

data are presently available to simultaneously consider changing inputs from the terrestrial 486  

landscape and the Pacific and Atlantic Oceans. 487  

488  

  25  

Figure 1 489  

490  491  

Marine MeHg obs.Kirk et al. 2008Lehnherr et al. 2011 &Lehnherr et al. 2015Wang et al. 2012Heimburger et al. 2015St. Louis et al. 2007Zdanowicz et al. 2013

12345

65˚N

75˚N

85˚N

180 W

90

W

0˚90

EO

ce

an

Da

ta V

iew

Barents Sea

Laptev Sea

Kara Sea

East Siberian Sea

Chukchi Sea

Beaufort Sea

Central Basin

Baffin Bay

White Sea

Lincoln Sea

Canadian Archipelago(CAA)

1

2

3

4

5

River locationsOb RiverYenisei RiverLena RiverKolyma RiverMackenzie River

Arctic Ocean boundary

  26  

Figure 2 492   493  

494  495  

Deep sediment

1150

2750

2360

420

6%

475

104

PML (87)Ice (15) Plankton (<1)

PML (6)

Ice (<2)Deep sediment (<2)

Shelf sediment (11)

Shelf sedimentDeep ocean

Sub-surface ocean

Deep ocean

Sub-surface ocean

Total Hg distribution (Mg) Methylated Hg distribution (Mg)

  27  

Figure 3 496   497  

498  499  

  28  

Figure 4 500   501  

502  503  

  29  

Figure 5 504   505  

506  507  

snow

direct deposition

evasion

reservoirs (Mg)

photo12

evasion

3.8Me2Hg

HgII

85

MeHgHg0

rivers

coastal erosion

30

50 30

ice meltwater

Polar Mixed Layer (0-20 m depth)

5net biotic (?)

15

2.1 8.0 738390

8300

0.5

2.5

8.5

MeHg: 0.2

North AtlanticHg: 2.5

Hg: 10

North Atlantic

Hg: 1.9

MeHg: 0.1

4.0

0.5

1.6

gnilttes

1.5

5.0

20

gnilttes

37

11.5

37

MeHg: 0.8

35Me2Hg

HgII

MeHgHg0

Subsurface Ocean (20-200 m depth)

950

930 70 34 280203

200

North Atlantic

Hg: 1825

8

gnilttes

0.1+

1.3*

3+

-5*

gnilttes

6+

30*

3.0+

0.9*

MeHg: 5.0MeHg: 0.3

North Atlantic

Hg: 12MeHg: 0.8

noitalucri c

3.8+

noitalucri c

7.0+

Hg: 3.0

net 90

net 3

+Settling/diffusion/circulation associated with the deep ocean*Settling/diffusion associated with shelf sediment

  30  

Figure 6 508   509  

510  

1.0

1.1

1.2

1.3

1.4

PML

External

inputs

rela

tive c

hange

1.5

1.0

1.1

1.2

1.3

year

1850 1890 1930 1970 2010

Subsurface

water

Deep

waterrela

tive c

hange

Atmospheric deposition

Atmospheric deposition

& sea-ice cover

Historic trends in:

1.6

  31  

References 511  

Alfimov,  V.,  G.  Possnert,  and  A.  Aldahan  (2006),  Anthropogenic  iodine-­‐129  in  the  Arctic  512  Ocean  and  Nordic  Seas:  Numerical  modeling  and  prognoses,  Mar  Pollut  Bull,  52(4),  380-­‐513  385  514    515  AMAP  (2011a),  AMAP  Assessment  2011:  Mercury  in  the  Arctic.  Arctic  Monitoring  and  516  Assessment  Programme  (AMAP).  Rep.,  xiv  +  193  pp,  Oslo,  Norway.  517    518  AMAP  (2011b),  Snow,  Water,  Ice  and  Permafrost  in  the  Arctic  (SWIPA):  Climate  Change  519  and  the  Cryosphere.  Arctic  Monitoring  and  Assessment  Programme  (AMAP)Rep.,  538  pp  520  pp,  Oslo,  Norway.  521    522  Amos,  H.  M.,  D.  J.  Jacob,  D.  Kochman,  H.  M.  Horowitz,  Y.  Zhang,  S.  Dutkiewicz,  M.  Horvat,  523  E.  S.  Corbitt,  and  E.  M.  Sunderland  (2014),  Global  biogeochemical  implications  of  524  mercury  discharges  from  tivers  and  sediment  burial,  Environ.  Sci.  Technol.,  48(16),  9514-­‐525  9522,  doi:10.1021/es502134t.  526    527  Andersson,  M.  E.,  J.  Sommar,  K.  Gardfeldt,  and  O.  Lindqvist  (2008),  Enhanced  528  concentrations  of  dissolved  gaseous  mercury  in  the  surface  waters  of  the  Arctic  Ocean,  529  Mar.  Chem.,  110(3-­‐4),  190-­‐194  530    531  Arrigo,  K.  R.,  and  G.  L.  van  Dijken  (2011),  Secular  trends  in  Arctic  Ocean  net  primary  532  production,  J.  Geosphys.  Res.  Oceans,  116  533    534  Arrigo,  K.  R.,  P.  A.  Matrai,  and  G.  L.  van  Dijken  (2011),  Primary  productivity  in  the  Arctic  535  Ocean:  Impacts  of  complex  optical  properties  and  subsurface  chlorophyll  maxima  on  536  large-­‐scale  estimates,  J.  Geosphys.  Res.  Oceans,  116  537    538  Baya,  P.  A.,  M.  Gosselin,  I.  Lehnherr,  V.  L.  St  Louis,  and  H.  Hintelmann  (2015),  539  Determination  of  Monomethylmercury  and  Dimethylmercury  in  the  Arctic  Marine  540  Boundary  Layer,  Environ.  Sci.  Technol.,  49(1),  223-­‐232,  doi:10.1021/es502601z.  541    542  Beattie,  S.  A.,  D.  Armstrong,  A.  Chaulk,  J.  Comte,  M.  Gosselin,  and  F.  Y.  Wang  (2014),  Total  543  and  Methylated  Mercury  in  Arctic  Multiyear  Sea  Ice,  Environ.  Sci.  Technol.,  48(10),  5575-­‐544  5582  545    546  Beszczynska-­‐Moller,  A.,  E.  Fahrbach,  U.  Schauer,  and  E.  Hansen  (2012),  Variability  in  547  Atlantic  water  temperature  and  transport  at  the  entrance  to  the  Arctic  Ocean,  1997-­‐548  2010,  Ices  Journal  of  Marine  Science,  69(5),  852-­‐863  549    550  Bintanja,  R.,  R.  G.  Graversen,  and  W.  Hazeleger  (2011),  Arctic  winter  warming  amplified  551  by  the  thermal  inversion  and  consequent  low  infrared  cooling  to  space,  Nature  552  Geoscience,  4(11),  758-­‐761,  doi:10.1038/ngeo1285.  553    554  Bittrich,  D.  R.,  A.  P.  Rutter,  B.  D.  Hall,  and  J.  J.  Schauer  (2011),  Photodecomposition  of  555  Methylmercury  in  Atmospheric  Waters,  Aerosol  and  Air  Quality  Research,  11(3),  290-­‐299  556    557  

  32  

Black,  F.  J.,  C.  H.  Conaway,  and  A.  R.  Flegal  (2009),  Stability  of  Dimethyl  Mercury  in  558  Seawater  and  Its  Conversion  to  Monomethyl  Mercury,  Environ.  Sci.  Technol.,  43(11),  559  4056-­‐4062  560    561  Bowman,  K.  L.,  C.  R.  Hammerschmidt,  C.  H.  Lamborg,  and  G.  Swarr  (2014),  Mercury  in  562  the  North  Atlantic  Ocean:  The  U.S.  GEOTRACES  zonal  and  meridional  sections,  Deep-­‐Sea  563  Research  II,  doi:DOI:  10.1016/j.dsr2.2014.07.004.  564    565  Braune,  B.,  et  al.  (2015),  Mercury  in  the  marine  environment  of  the  Canadian  Arctic:  566  Review  of  recent  findings,  Sci.  Total.  Environ.,  509,  67-­‐90,  567  doi:10.1016/j.scitotenv.2014.05.133.  568    569  Burt,  A.  E.  (2012),  Mercury  uptake  and  dynamics  in  sea  ice  algae,  phytoplankton  and  570  grazing  copepods  from  a  Beaufort  Sea  Arctic  marine  food  web,  92  pp,  University  of  571  Manitoba,  http://hdl.handle.net/1993/8907.  572    573  Cai,  P.,  M.  R.  van  der  Loeff,  I.  Stimac,  E.  M.  Nothig,  K.  Lepore,  and  S.  B.  Moran  (2010),  Low  574  export  flux  of  particulate  organic  carbon  in  the  central  Arctic  Ocean  as  revealed  by  Th-­‐575  234:U-­‐238  disequilibrium,  J.  Geosphys.  Res.  Oceans,  115  576    577  Canario,  J.,  L.  Poissant,  M.  Pilote,  C.  Blaise,  P.  Constant,  J.-­‐F.  Ferard,  and  F.  Gagne  (2013),  578  Toxicity  survey  of  Canadian  Arctic  marine  sediments,  J.  Soils  Sediments,  14,  196-­‐203,  579  doi:DOI  10.1007/s11368-­‐013-­‐0792-­‐1.  580    581  Chaulk,  A.,  G.  A.  Stern,  D.  Armstrong,  D.  G.  Barber,  and  F.  Y.  Wang  (2011),  Mercury  582  Distribution  and  Transport  Across  the  Ocean-­‐Sea-­‐Ice-­‐Atmosphere  Interface  in  the  Arctic  583  Ocean,  Environ.  Sci.  Technol.,  45(5),  1866-­‐1872  584    585  Chen,  L.,  Y.  Zhang,  D.  J.  Jacob,  A.  Soerensen,  J.  A.  Fisher,  H.  M.  Horowitz,  E.  S.  Corbitt,  and  586  X.  Wang  (2015),  Differences  in  decadal  trends  of  atmospheric  mercury  between  the  587  Arctic  and  northern  mid-­‐latitudes  suggest  a  decline  in  Arctic  Ocean  mercury,  Geophys.  588  Res.  Lett.,  available  online,  doi:10.1002/2015GL064051.  589    590  Clough,  L.  M.,  W.  G.  Ambrose,  J.  K.  Cochran,  C.  Barnes,  P.  E.  Renaud,  and  R.  C.  Aller  (1997),  591  Infaunal  density,  biomass  and  bioturbation  in  the  sediments  of  the  Arctic  Ocean,  Deep-­‐592  Sea  Res.,  Part  li,  44(8),  1683-­‐1704  593    594  Cossa,  D.,  L.  E.  Heimburger,  D.  Lannuzel,  S.  R.  Rintoul,  E.  C.  V.  Butler,  A.  R.  Bowie,  B.  595  Averty,  R.  J.  Watson,  and  T.  Remenyi  (2011),  Mercury  in  the  Southern  Ocean,  Geochim.  596  Cosmochim.  Acta,  75(14),  4037-­‐4052,  doi:DOI  10.1016/j.gca.2011.05.001.  597    598  Curry,  B.,  C.  M.  Lee,  and  B.  Petrie  (2011),  Volume,  Freshwater,  and  Heat  Fluxes  through  599  Davis  Strait,  2004-­‐05,  J.  Phys.  Oceanogr.,  41(3),  429-­‐436  600    601  Dastoor,  A.,  and  D.  Durnford  (2014),  Arctic  Ocean:  Is  it  a  sink  or  a  source  of  atmospheric  602  mercury?,  Environmental  Science  and  Technology,  48(3),  1707-­‐1717,  603  doi:10.1021/es404473e.  604    605  

  33  

Dittmar,  T.,  and  G.  Kattner  (2003),  The  biogeochemistry  of  the  river  and  shelf  ecosystem  606  of  the  Arctic  Ocean:  a  review,  Mar.  Chem.,  83(3-­‐4),  103-­‐120  607    608  Emmerton,  C.  A.,  J.  A.  Graydon,  J.  A.  L.  Gareis,  V.  L.  St  Louis,  L.  F.  W.  Lesack,  J.  K.  A.  Banack,  609  F.  Hicks,  and  J.  Nafziger  (2013),  Mercury  export  to  the  Arctic  Ocean  from  the  Mackenzie  610  River,  Canada,  Environmental  Science  and  Technology,  doi:10.1021/es400715r.  611    612  Evenset,  A.,  G.  N.  Christensen,  T.  Skotvold,  E.  Fjeld,  M.  Schlabach,  E.  Wartena,  and  D.  613  Gregor  (2004),  A  comparison  of  organic  contaminants  in  two  high  Arctic  lake  614  ecosystems,  Bjornoya  (Bear  Island),  Norway,  Sci.  Total.  Environ.,  318(1-­‐3),  125-­‐141  615    616  Fisher,  J.  A.,  D.  J.  Jacob,  A.  L.  Soerensen,  H.  M.  Amos,  A.  Steffen,  and  E.  M.  Sunderland  617  (2012),  Riverine  source  of  Arctic  Ocean  mercury  inferred  from  atmospheric  618  observations,  Nature  Geoscience,  5(7),  499-­‐504  619    620  Fisher,  J.  A.,  D.  J.  Jacob,  A.  L.  Soerensen,  H.  M.  Amos,  E.  S.  Corbitt,  D.  G.  Streets,  Q.  Wang,  R.  621  M.  Yantosca,  and  E.  M.  Sunderland  (2013),  Factors  driving  mercury  variability  in  the  622  Arctic  atmosphere  and  ocean    over  the  past  thirty  years,  Glob.  Biogeochem.  Cycles,  27(1-­‐623  10),  doi:10.1002/2013GB004689.  624    625  Fox,  A.  L.,  E.  A.  Hughes,  R.  P.  Trocine,  J.  H.  Trefry,  S.  V.  Schonberg,  N.  D.  McTigue,  B.  K.  626  Lasorsa,  B.  Konar,  and  L.  W.  Cooper  (2013),  Mercury  in  the  northeastern  Chukchi  Sea:  627  Distribution  patterns  in  seawater  and  sediments  and  biomagnification  in  the  benthic  628  food  web,  Deep-­‐Sea  Res.,  Part  li,  doi:10.1016/j.dsr2.2013.07.012.  629    630  Graydon,  J.  A.,  C.  A.  Emmerton,  L.  F.  W.  Lesack,  and  E.  N.  Kelly  (2009),  Mercury  in  the  631  Mackenzie  River  delta  and  estuary:  Concentrations  and  fluxes  during  open-­‐water  632  conditions,  Sci.  Total.  Environ.,  407(8),  2980-­‐2988  633    634  Hammerschmidt,  C.  R.,  W.  F.  Fitzgerald,  C.  H.  Lamborg,  P.  H.  Balcom,  and  C.  M.  Tseng  635  (2006),  Biogeochemical  cycling  of  methylmercury  in  lakes  and  tundra  watersheds  of  636  Arctic  Alaska,  Environ.  Sci.  Technol.,  40(4),  1204-­‐1211  637    638  Hansell,  D.  A.,  C.  A.  Carlson,  and  R.  Schlitzer  (2012),  Net  removal  of  major  marine  639  dissolved  organic  carbon  fractions  in  the  subsurface  ocean,  Glob.  Biogeochem.  Cycles,  26  640    641  Hatzianastassiou,  N.,  C.  Matsoukas,  A.  Fotiadi,  K.  G.  Pavlakis,  E.  Drakakis,  D.  642  Hatzidimitriou,  and  I.  Vardavas  (2005),  Global  distribution  of  Earth's  surface  shortwave  643  radiation  budget,  Atmos.  Chem.  Phys.,  5,  2847-­‐2867  644    645  Hedges,  J.  I.,  J.  A.  Baldock,  Y.  Gelinas,  C.  Lee,  M.  L.  Peterson,  and  S.  G.  Wakeham  (2002),  646  The  biochemical  and  elemental  compositions  of  marine  plankton:  A  NMR  perspective,  647  Mar.  Chem.,  78(1),  47-­‐63  648    649  Heimburger,  L.  E.,  J.  Sonke,  D.  Cossa,  D.  Point,  C.  Lagane,  L.  Laffont,  B.  Galfond,  M.  650  Nicolaus,  B.  Rabe,  and  M.  M.  Rutgers  van  der  Loeff  (2015),  Shallow  methylmercury  651  production  in  the  marginal  sea  ice  zone  of  the  central  Arctic  Ocean,  Scientific  reports,  5,  652  doi:10.1038/srep10318.  653    654  

  34  

Heimbürger,  L.  E.,  J.  E.  Sonke,  D.  Cossa,  D.  Point,  C.  Lagane,  L.  Laffont,  B.  Galfond,  M.  655  Nicolaus,  B.  Rabe,  and  M.  R.  van  der  Loeff  (2015),  Shallow  methylmercury  production  in  656  the  marginal  sea  ice  zone  of  the  central  Arctic  Ocean  Scientific  Reports  657    658  Heyes,  A.,  R.  P.  Mason,  E.  H.  Kim,  and  E.  Sunderland  (2006),  Mercury  methylation  in  659  estuaries:  Insights  from  using  measuring  rates  using  stable  mercury  isotopes,  Mar.  660  Chem.,  102(1-­‐2),  134-­‐147,  doi:10.1016/J.Marchem.2005.09.018.  661    662  Hill,  V.  J.,  P.  A.  Matrai,  E.  Olson,  S.  Suttles,  M.  Steele,  L.  A.  Codispoti,  and  R.  C.  Zimmerman  663  (2013),  Synthesis  of  integrated  primary  production  in  the  Arctic  Ocean:  II.  In  situ  and  664  remotely  sensed  estimates,  Progress  in  Oceanography,  110,  107-­‐125  665    666  Hollweg,  T.  A.,  C.  C.  Gilmour,  and  R.  P.  Mason  (2010),  Mercury  and  methylmercury  667  cycling  in  sediments  of  the  mid-­‐Atlantic  continental  shelf  and  slope,  Limnology  and  668  Oceanography,  55(6),  2703-­‐2722  669    670  Horvat,  M.,  J.  Kotnik,  M.  Logar,  V.  Fajon,  T.  Zvonaric,  and  N.  Pirrone  (2003),  Speciation  of  671  mercury  in  surface  and  deep-­‐sea  waters  in  the  Mediterranean  Sea,  Atmos.  Environ.,  37,  672  S93-­‐S108  673    674  ICES  ICES  Data  Portal,  http://ecosystemdata.ices.dk/  Extraction  02-­‐12-­‐2014;  675  Contaminants  and  biological  effects  of  contaminants  on  sediments.  ICES,  Copenhagen,  676  edited.  677    678  Jakobsson,  M.  (2002),  Hypsometry  and  volume  of  the  Arctic  Ocean  and  its  constituent  679  seas,  Geochemistry  Geophysics  Geosystems,  3  680    681  Jonsson,  S.,  U.  Skyllberg,  M.  B.  Nilsson,  E.  Lundberg,  A.  Andersson,  and  E.  Bjorn  (2014),  682  Differentiated  availability  of  geochemical  mercury  pools  controls  methylmercury  levels  683  in  estuarine  sediment  and  biota,  Nature  Communication,  5,  doi:10.1038/ncomms5624.  684    685  Joos,  F.,  J.  C.  Orr,  and  U.  Siegenthaler  (1997),  Ocean  carbon  transport  in  a  box-­‐diffusion  686  versus  a  general  circulation  model,  J.  Geosphys.  Res.  Oceans,  102(C6),  12367-­‐12388  687    688  Kading,  T.,  and  M.  Andersson  (2011),  Mercury  and  methylmercury  distribution  in  Arctic  689  Ocean  sediments,  paper  presented  at  International  Conference  on  Mercury  as  a  Global  690  Pollutant,  Halifax,  Canada.  691    692  Kirchman,  D.  L.,  V.  Hill,  M.  T.  Cottrell,  R.  Gradinger,  R.  R.  Malmstrom,  and  A.  Parker  693  (2009),  Standing  stocks,  production,  and  respiration  of  phytoplankton  and  694  heterotrophic  bacteria  in  the  western  Arctic  Ocean,  Deep-­‐Sea  Res.,  Part  li,  56(17),  1237-­‐695  1248  696    697  Kirk,  J.  L.,  V.  L.  St  Louis,  and  M.  J.  Sharp  (2006),  Rapid  reduction  and  reemission  of  698  mercury  deposited  into  snowpacks  during  atmospheric  mercury  depletion  events  at  699  Churchill,  Manitoba,  Canada,  Environ.  Sci.  Technol.,  40(24),  7590-­‐7596,  700  doi:10.1021/es061299.  701    702  

  35  

Kirk,  J.  L.,  V.  L.  S.  Louis,  H.  Hintelmann,  I.  Lehnherr,  B.  Else,  and  L.  Poissant  (2008),  703  Methylated  Mercury  Species  in  Marine  Waters  of  the  Canadian  High  and  Sub  Arctic,  704  Environ.  Sci.  Technol.,  42(22),  8367-­‐8373  705    706  Kirk,  J.  L.,  et  al.  (2012),  Mercury  in  Arctic  marine  ecosystems:  Sources,  pathways  and  707  exposure,  Environ.  Res.,  119,  64-­‐87  708    709  Kosobokova,  K.,  and  H.  J.  Hirche  (2009),  Biomass  of  zooplankton  in  the  eastern  Arctic  710  Ocean  -­‐  A  base  line  study,  Progress  in  Oceanography,  82(4),  265-­‐280  711    712  Krabbenhoft,  D.  P.,  and  E.  M.  Sunderland  (2013),  Global  Change  and  Mercury,  Science,  713  341(6153),  1457-­‐1458  714    715  Kuzyk,  Z.  Z.  A.,  C.  Gobeil,  and  R.  W.  Macdonald  (2013),  210Pb  and  137Cs  in  margin  716  sediments  of  the  Arctic  Ocean:  Controls  on  boundary  scavenging,  Glob.  Biogeochem.  717  Cycles(27),  1-­‐18,  doi:10.1002/gbc.20041.  718    719  Kwok,  R.,  and  D.  A.  Rothrock  (2009),  Decline  in  Arctic  sea  ice  thickness  from  submarine  720  and  ICESat  records:  1958-­‐2008,  Geophys.  Res.  Lett.,  36,  doi:10.1029/2009gl039035.  721    722  Lalonde,  J.  D.,  A.  J.  Poulain,  and  M.  Amyot  (2002),  The  role  of  mercury  redox  reactions  in  723  snow  on  snow-­‐to-­‐air  mercury  transfer,  Environ.  Sci.  Technol.,  36(2),  174-­‐178,  724  doi:10.1021/es010786g.  725    726  Lehnherr,  I.  (2014),  Methylmercury  biogeochemistry:  a  review  with  special  reference  to  727  Arctic  aquatic  ecosystems,  Environmental  Reviews,  22(3),  229-­‐243,  doi:10.1139/er-­‐728  2013-­‐0059.  729    730  Lehnherr,  I.  (2015),  Methylmercury  Cycling  in  the  Arctic:  What  do  we  need  to  know  in  731  the  face  of  a  changing  climate,  in  AGU-­‐GAC-­‐MAC-­‐CGU,  edited,  Montreal,  Canada  (3-­‐7  May  732  2015).  733    734  Lehnherr,  I.,  and  V.  L.  St  Louis  (2009),  Importance  of  Ultraviolet  Radiation  in  the  735  Photodemethylation  of  Methylmercury  in  Freshwater  Ecosystems,  Environ.  Sci.  Technol.,  736  43(15),  5692-­‐5698  737    738  Lehnherr,  I.,  V.  L.  St  Louis,  H.  Hintelmann,  and  J.  L.  Kirk  (2011),  Methylation  of  inorganic  739  mercury  in  polar  marine  waters,  Nature  Geoscience,  4(5),  298-­‐302  740  Lehnherr,  I.,  V.  L.  St.  Louis,  C.  A.  Emmerton,  J.  D.  Barker,  and  J.  L.  Kirk  (2012),  741  Methylmercury  cycling  in  High  Arctic  wetland  ponds:  Sources  and  sinks,  Environ.  Sci.  742  Technol.,  46(19),  10514-­‐10522  743    744  Leitch,  D.  R.  (2006),  Mercury  distribution  in  water  and  permafrost  of  the  lower  745  Mackenzie  Basin,  their  contribution  to  the  mercury  contamination  in  the  Beaufort  Sea  746  marine  ecosystem,  and  potential  effects  of  climate  variation,  130  pp,  University  of  747  Manitoba,  Winnipeg,  Manitoba.  748    749  

  36  

Leitch,  D.  R.,  J.  Carrie,  D.  Lean,  R.  W.  Macdonald,  G.  A.  Stern,  and  F.  Y.  Wang  (2007),  The  750  delivery  of  mercury  to  the  Beaufort  Sea  of  the  Arctic  Ocean  by  the  Mackenzie  River,  Sci.  751  Total.  Environ.,  373(1),  178-­‐195  752    753  Lin,  C.-­‐J.,  and  S.  O.  Pehkonen  (1999),  The  chemistry  of  atmospheric  mercury:  a  review,  754  Atmospheric  Environment,  33(13),  2067-­‐2079  755    756  Loseto,  L.  L.,  G.  A.  Stern,  D.  Deibel,  T.  L.  Connelly,  A.  Prokopowicz,  D.  R.  S.  Lean,  L.  Fortier,  757  and  S.  H.  Ferguson  (2008),  Linking  mercury  exposure  to  habitat  and  feeding  behaviour  758  in  Beaufort  Sea  beluga  whales,  J.  Mar.  Sys.,  74(3-­‐4),  1012-­‐1024,  759  doi:10.1016/j.jmarsys.2007.10.004.  760    761  Lu,  J.  Y.,  W.  H.  Schroeder,  L.  A.  Barrie,  A.  Steffen,  H.  E.  Welch,  K.  Martin,  L.  Lockhart,  R.  V.  762  Hunt,  G.  Boila,  and  A.  Richter  (2001),  Magnification  of  atmospheric  mercury  deposition  763  to  polar  regions  in  springtime:  the  link  to  tropospheric  ozone  depletion  chemistry,  764  Geophys.  Res.  Lett.,  28(17),  3219-­‐3222  765    766  Macdonald,  R.  W.,  T.  Harner,  and  J.  Fyfe  (2005),  Recent  climate  change  in  the  Arctic  and  767  its  impact  on  contaminant  pathways  and  interpretation  of  temporal  trend  data,  Sci.  768  Total.  Environ.,  342(1-­‐3),  5-­‐86  769    770  Mason,  R.  P.,  K.  R.  Rolfhus,  and  W.  F.  Fitzgerald  (1995),  Methylated  and  Elemental  771  Mercury  Cycling  in  Surface  and  Deep-­‐Ocean  Waters  of  the  North-­‐Atlantic,  Water,  Air,  Soil  772  Pollut.,  80(1-­‐4),  665-­‐677  773    774  Mason,  R.  P.,  A.  L.  Choi,  W.  F.  Fitzgerald,  C.  R.  Hammerschmidt,  C.  H.  Lamborg,  A.  L.  775  Soerensen,  and  E.  M.  Sunderland  (2012),  Mercury  biogeochemical  cycling  in  the  ocean  776  and  policy  implications,  Environ.  Res.,  119,  101-­‐117,  doi:10.1016/j.envres.2012.03.013.  777    778  Monperrus,  M.,  E.  Tessier,  D.  Amouroux,  A.  Leynaert,  P.  Huonnic,  and  O.  F.  X.  Donard  779  (2007),  Mercury  methylation,  demethylation  and  reduction  rates  in  coastal  and  marine  780  surface  waters  of  the  Mediterranean  Sea,  Mar.  Chem.,  107(1),  49-­‐63  781    782  Moran,  S.  B.,  K.  M.  Ellis,  and  J.  N.  Smith  (1997),  Th-­‐234/U-­‐238  disequilibrium  in  the  783  central  Arctic  Ocean:  implications  for  particulate  organic  carbon  export,  Deep-­‐Sea  Res.,  784  Part  li,  44(8),  1593-­‐1606  785    786  Nagorski,  S.  A.,  D.  R.  Engstrom,  J.  P.  Hudson,  D.  P.  Krabbenhoft,  E.  Hood,  J.  F.  Dewild,  and  787  G.  R.  Aiken  (2014),  Spatial  distribution  of  mercury  in  southeastern  Alaskan  streams  788  influenced  by  glaciers,  wetland,  and  salmon,  Environmental  Pollution,  184,  62-­‐72,  789  doi:10.1016/j.envpol.2013.07.040.  790    791  NSIDC  (2013),  National  Snow  and  Ice  Data  Center:  website:  792  http://nsidc.org/cryosphere/sotc/sea_ice.html,  Access  date:  September  2013,  edited,  793  NSIDC.  794    795  Nummelin,  A.,  C.  Li,  and  L.  H.  Smedsrud  (2015),  Response  of  Arctic  Ocean  stratification  796  to  changing  runoff  in  a  column  model,  Journal  of  Geophysical  Research:  Oceans,  120,  797  doi:10.1002/2014JC010571.  798  

  37  

 799  Outridge,  P.  M.,  R.  W.  Macdonald,  F.  Wang,  G.  A.  Stern,  and  A.  P.  Dastoor  (2008),  A  mass  800  balance  inventory  of  mercury  in  the  Arctic  Ocean,  Environmental  Chemistry,  5(2),  89-­‐111  801    802  Panteleev,  G.  G.,  D.  A.  Nechaev,  and  M.  Ikeda  (2006),  Reconstruction  of  summer  Barents  803  Sea  circulation  from  climatological  data,  Atmosphere-­‐Ocean,  44(2),  111-­‐132  804    805  Pirtle-­‐Levy,  R.,  J.  M.  Grebmeier,  L.  W.  Cooper,  and  I.  L.  Larsen  (2009),  Chlorophyll  a  in  806  Arctic  sediments  implies  long  persistence  of  algal  pigments,  Deep-­‐Sea  Res.,  Part  li,  807  56(17),  1326-­‐1338  808    809  Polyak,  L.,  et  al.  (2009),  Late  Quaternary  stratigraphy  and  sedimentation  patterns  in  the  810  western  Arctic  Ocean,  Global  and  Planetary  Change,  68(1-­‐2),  5-­‐17  811    812  Porter,  D.  F.,  J.  J.  Cassano,  and  M.  C.  Serreze  (2011),  Analysis  of  the  Arctic  atmospheric  813  energy  budget  in  WRF:  A  comparison  with  reanalyses  and  satellite  observations,  Journal  814  of  Geophysical  Research-­‐Atmospheres,  116  815    816  Poulain,  A.  J.,  E.  Garcia,  M.  Amyot,  P.  G.  C.  Campbell,  and  P.  A.  Arlya  (2007),  Mercury  817  distribution,  partitioning  and  speciation  in  coastal  vs.  inland  High  Arctic  snow,  Geochim.  818  Cosmochim.  Acta,  71(14),  3419-­‐3431  819    820  Pucko,  M.,  A.  Burt,  W.  Walkusz,  F.  Wang,  R.  W.  Macdonald,  S.  Rysgaard,  D.  G.  Barber,  J.  E.  821  Tremblay,  and  G.  A.  Stern  (2014),  Transformation  of  Mercury  at  the  Bottom  of  the  Arctic  822  Food  Web:  An  Overlooked  Puzzle  in  the  Mercury  Exposure  Narrative,  Environ.  Sci.  823  Technol.,  48(13),  7280-­‐7288,  doi:10.1021/es404851b.  824    825  Rachold,  V.,  H.  Eicken,  V.  V.  Gordeev,  H.  Grigoriev,  H.-­‐W.  Hubberten,  A.  P.  Lisitzin,  L.  826  Shevchenko,  and  L.  Schirmeister  (2004),  Modern  Terrigeneous  Organic  Carbon  Input  to  827  the  Arctic  Ocean,  in  The  Organic  Carbon  Cycle  in  the  Arctic  Ocean,  edited  by  R.  Stein  and  828  R.  W.  Macdonald,  pp.  33-­‐55,  Springer,  Berlin  Heidelberg.  829    830  Redfield,  A.  C.,  B.  H.  Ketchum,  and  F.  A.  Richards  (1963),  The  influence  of  organisms  on  831  the  composition  of  seawater,  in  The  Sea,  vol.  2.,  edited  by  M.  N.  Hill,  pp.  26-­‐77,  832  Interscience,  New  York.  833    834  Rice,  C.  P.,  and  S.  M.  Chernyak  (1997),  Marine  arctic  fog:  An  accumulator  of  currently  835  used  pesticide,  Chemosphere,  35(4),  867-­‐878  836    837  Riget,  F.,  M.  P.  Tamstorf,  M.  M.  Larsen,  J.  Sondergaard,  G.  Asmund,  J.  M.  Falk,  and  C.  838  Sigsgaard  (2011a),  Mercury  (Hg)  Transport  in  a  High  Arctic  River  in  Northeast  839  Greenland,  Water,  Air,  Soil  Pollut.,  222(1-­‐4),  233-­‐242  840    841  Riget,  F.,  et  al.  (2011b),  Temporal  trends  of  Hg  in  Arctic  biota,  an  update,  Sci.  Total.  842  Environ.,  409(18),  3520-­‐3526  843    844  Rydberg,  J.,  J.  Klaminder,  P.  Rosen,  and  R.  Bindler  (2010),  Climate  driven  release  of  845  carbon  and  mercury  from  permafrost  mires  increases  mercury  loading  to  sub-­‐arctic  846  lakes,  Sci.  Total.  Environ.,  408(20),  4778-­‐4783,  doi:10.1016/j.scitotenv.2010.06.056.  847  

  38  

 848  Schartup,  A.  T.,  U.  C.  Ndu,  P.  H.  Balcom,  R.  P.  Mason,  and  E.  M.  Sunderland  (2015a),  849  Contrasting  Effects  of  Marine  and  Terrestrially  Derived  Dissolved  Organic  Matter  on  850  Mercury  Speciation  and  Bioavailability  inSeawater,  Envionmental  Science  and  851  Technology,  49(10),  5965-­‐5972,  doi:10.1021/es506274x.  852    853  Schartup,  A.  T.,  R.  S.  D.  Calder,  A.  L.  Soerensen,  P.  H.  Balcom,  R.  P.  Mason,  and  E.  M.  854  Sunderland  (2015b),  Estuarine  Mercury  Dynamics  in  the  Subarctic  Region,  PNAS,  855  doi:10.1073/pnas.1505541112.  856    857  Schroeder,  W.  H.,  K.  G.  Anlauf,  L.  A.  Barrie,  J.  Y.  Lu,  A.  Steffen,  D.  R.  Schneeberger,  and  T.  858  Berg  (1998),  Arctic  springtime  depletion  of  mercury,  Nature,  394(6691),  331-­‐332  859  Serreze,  M.  C.,  and  R.  G.  Barry  (2005),  The  Arctic  Climate  System,  Cambridge  University  860  Press,  United  Kingdom.  861    862  Serreze,  M.  C.,  A.  P.  Barrett,  A.  G.  Slater,  R.  A.  Woodgate,  K.  Aagaard,  R.  B.  Lammers,  M.  863  Steele,  R.  Moritz,  M.  Meredith,  and  C.  M.  Lee  (2006),  The  large-­‐scale  freshwater  cycle  of  864  the  Arctic,  J.  Geosphys.  Res.  Oceans,  111(C11)  865    866  Shaw,  W.  J.,  and  T.  P.  Stanton  (2014),  Vertical  diffusivity  of  the  Western  Arctic  Ocean  867  halocline,  Journal  of  Geophysical  Research:  Oceans,  119(8),  5017-­‐5038  868    869  Smedsrud,  L.  H.,  R.  Ingvaldsen,  J.  E.  O.  Nilsen,  and  O.  Skagseth  (2010),  Heat  in  the  Barents  870  Sea:  transport,  storage,  and  surface  fluxes,  Ocean  Science,  6(1),  219-­‐234  871    872  Soerensen,  A.  L.,  E.  M.  Sunderland,  C.  D.  Holmes,  D.  J.  Jacob,  R.  M.  Yantosca,  H.  Skov,  J.  H.  873  Christensen,  S.  A.  Strode,  and  R.  P.  Mason  (2010),  An  Improved  Global  Model  for  Air-­‐Sea  874  Exchange  of  Mercury:  High  Concentrations  over  the  North  Atlantic,  Environ.  Sci.  Technol.,  875  44(22),  8574-­‐8580,  doi:10.1021/Es102032g.  876    877  Sommar,  J.,  M.  E.  Andersson,  and  H.  W.  Jacobi  (2010),  Circumpolar  measurements  of  878  speciated  mercury,  ozone  and  carbon  monoxide  in  the  boundary  layer  of  the  Arctic  879  Ocean,  Atmos.  Chem.  Phys.,  10(11),  5031-­‐5045  880    881  Sonke,  J.  E.,  and  L.  E.  Heimburger  (2012),  Environmental  science:  Mercury  in  flux,  Nature  882  Geoscience,  5,  447-­‐448,  doi:10.1038/ngeo1508.  883    884  St  Louis,  V.  L.,  H.  Hintelmann,  J.  A.  Graydon,  J.  L.  Kirk,  J.  Barker,  B.  Dimock,  M.  J.  Sharp,  and  885  I.  Lehnherr  (2007),  Methylated  mercury  species  in  Canadian  high  arctic  marine  surface  886  waters  and  snowpacks,  Environ.  Sci.  Technol.,  41(18),  6433-­‐6441  887    888  St  Louis,  V.  L.,  M.  J.  Sharp,  A.  Steffen,  A.  May,  J.  Barker,  J.  L.  Kirk,  D.  J.  A.  Kelly,  S.  E.  Arnott,  889  B.  Keatley,  and  J.  P.  Smol  (2005),  Some  sources  and  sinks  of  monomethyl  and  inorganic  890  mercury  on  Ellesmere  island  in  the  Canadian  high  arctic,  Environ.  Sci.  Technol.,  39(8),  891  2686-­‐2701  892    893  St  Pierre,  K.,  V.  L.  St  Louis,  J.  Kirk,  I.  Lehnherr,  S.  Wang,  and  C.  La  Farge  (2015),  The  894  importance  of  open  marine  waters  to  the  enrichment  of  total  mercury  and  895  

  39  

monomethylmercury  in  lichens  in  the  Canadian  High  Arctic,  Envionmental  Science  and  896  Technology,  49(10),  5930-­‐5938,  doi:10.1021/acs.est.5b00347.  897    898  Steffen,  A.,  I.  Lehnherr,  A.  Cole,  P.  Ariya,  A.  Dastoor,  D.  Durnford,  J.  Kirk,  and  M.  Pilote  899  (2015),  Atmospheric  mercury  in  the  Canadian  Arctic.  Part  I:  A  review  of  recent  field  900  measurements,  Sci.  Total.  Environ.,  509,  3-­‐15  901    902  Stern,  G.  A.,  et  al.  (2012),  How  does  climate  change  influence  arctic  mercury?,  Sci.  Total.  903  Environ.,  414,  22-­‐42  904    905  Strode,  S.  A.,  L.  Jaegle,  and  S.  Emerson  (2010),  Vertical  transport  of  anthropogenic  906  mercury  in  the  ocean,  Glob.  Biogeochem.  Cycles,  24  907    908  Sunderland,  E.  M.,  F.  A.  P.  C.  Gobas,  B.  A.  Branfireun,  and  A.  Heyes  (2006),  Environmental  909  controls  on  the  speciation  and  distribution  of  mercury  in  coastal  sediments,  Mar.  Chem.,  910  102(1-­‐2),  111-­‐123,  doi:10.1016/J.Marchem.2005.09.019.  911    912  Sunderland,  E.  M.,  D.  P.  Krabbenhoft,  J.  W.  Moreau,  S.  A.  Strode,  and  W.  M.  Landing  913  (2009),  Mercury  sources,  distribution,  and  bioavailability  in  the  North  Pacific  Ocean:  914  Insights  from  data  and  models,  Glob.  Biogeochem.  Cycles,  23(Artn  Gb2010),  915  doi:10.1029/2008gb003425.  916    917  Sunderland,  E.  M.,  J.  Dalziel,  A.  Heyes,  B.  A.  Branfireun,  D.  P.  Krabbenhoft,  and  F.  A.  P.  C.    918  Gobas  (2010),  Response  of  a  Macrotidal  Estuary  to  Changes  in  Anthropogenic  Mercury  919  Loading  between  1850  and  2000,  Environ.  Sci.  Technol.,  44(5),  1698-­‐1704,  920  doi:10.1021/Es9032524.  921    922  Trefry,  J.  H.,  R.  D.  Rember,  R.  P.  Trocine,  and  J.  S.  Brown  (2003),  Trace  metals  in  923  sediments  near  offshore  oil  exploration  and  production  sites  in  the  Alaskan  Arctic,  924  Environmental  Geology,  45(2),  149-­‐160  925    926  Trefry,  J.  H.,  R.  P.  Trocine,  L.  W.  Cooper,  and  K.  H.  Dunton  (2014),  Trace  metlas  and  927  organic  carbon  in  sediments  of  the  northeastern  Chukchi  Sea,  Deep-­‐Sea  Res.,  Part  li,  In  928  press,  doi:10.1016/j.dsr2.2013.07.018.  929    930  Wang,  F.  Y.,  R.  W.  Macdonald,  D.  A.  Armstrong,  and  G.  A.  Stern  (2012),  Total  and  931  Methylated  Mercury  in  the  Beaufort  Sea:  The  Role  of  Local  and  Recent  Organic  932  Remineralization,  Environ.  Sci.  Technol.,  46(21),  11821-­‐11828  933    934  Weiss-­‐Penzias,  P.  S.,  C.  Ortiz,  R.  P.  Acosta,  W.  Heim,  J.  P.  Ryan,  D.  Fernandez,  J.  L.  Collett,  935  and  A.  R.  Flegal  (2012),  Total  and  monomethyl  mercury  in  fog  water  from  the  central  936  California  coast,  Geophys.  Res.  Lett.,  39  937    938  Zdanowicz,  C.,  E.  M.  Krummel,  D.  Lean,  A.  J.  Poulain,  E.  Yumvihoze,  J.  B.  Chen,  and  H.  939  Hintelmann  (2013),  Accumulation,  storage  and  release  of  atmospheric  mercury  in  a  940  glaciated  Arctic  catchment,  Baffin  Island,  Canada,  Geochim.  Cosmochim.  Acta,  107,  316-­‐941  335  942    943  

  40  

Zhang,  L.,  L.  P.  Wright,  and  P.  Blanchard  (2009),  A  review  of  current  knowledge  944  concerning  dry  deposition  of  atmospheric  mercury,  Atmospheric  Environment,  43(37),  945  5853-­‐5864  946    947  Zhang,  Y.,  D.  J.  Jacob,  S.  Dutkiewicz,  H.  M.  Amos,  and  C.  J.  Sunderland  (2015),  948  Biogeochemical  drivers  of  the  fate  of  riverine  mercury  disharged  to  the  global  and  Arctic  949  oceans,  Glob.  Biogeochem.  Cycles,  29,  doi:10.1002/2015GB005124  950