Available online at www.sciencedirect.com

www.elsevier.com/locate/gca

Geochimica et Cosmochimica Acta 75 (2011) 2393–2408

Fluvial and hydrothermal input of manganese into theArctic Ocean

R. Middag a,b,⇑, H.J.W. de Baar a,c, P. Laan a, M.B. Klunder a

a Royal Netherlands Institute for Sea Research (Royal NIOZ), P.O. Box 59, 1790 AB Den Burg, Texel, The Netherlandsb Department of Ocean Sciences & Institute of Marine Sciences, University of California Santa Cruz, CA 95064, USA

c Department of Ocean Ecosystems, University of Groningen, Groningen, The Netherlands

Received 5 August 2010; accepted in revised form 1 February 2011; available online 13 February 2011

Abstract

A total of 773 samples were analysed for dissolved manganese (Mn) in the Arctic Ocean aboard R.V. Polarstern duringexpedition ARK XXII/2 from 28 July until 07 October 2007 from Tromsø (Norway) to Bremerhaven. Concentrations ofMn were elevated in the surface layer with concentrations of up to 6 nM over the deep Basins and over 20 nM in the LaptevSea. The general distribution of Mn through the water column is consistent with previous studies, but there are differences inthe absolute concentrations that are most likely related to differences in sample area, sampling and filtration.

The elevated concentrations of Mn in the surface layer are related to fresh water input. This was visible in the strong neg-ative correlations observed between dissolved Mn and salinity. The correlation between Mn and salinity and the correlationbetween Mn and the quasi conservative trace water mass tracer PO4

*, showed fluvial and melt water input and the Pacific andAtlantic origin of the surface waters. A large portion of the Mn delivered by the Arctic rivers is removed in the shelf seas anddoes not pass into the central basins. Most likely a benthic flux is at the origin of the elevated concentrations of Mn near thesediments in the Barents and Kara Seas. These elevated concentrations of Mn apparently affected the deep basins as well, asmaxima in the concentrations of Mn were observed that corresponded with lowered transmission over the continental slope.

A maximum in the concentration of Mn in the deep basin corresponded with anomalies in light transmission, potentialtemperature and dissolved iron, confirming the hydrothermal origin. The hydrothermal plume was observed throughoutthe Nansen Basin and over the deep Gakkel Ridge around 2500 m depth and a smaller plume was observed around3200 m. The concentration of Mn at the Mn maximum around 2500 m depth decreased exponentially, consistent with a firstorder scavenging model. The concentrations of Mn were extremely low in the deep Makarov Basin (�0.05 nM) and slightlyhigher in the Eurasian Basin (�0.1 nM) outside the influence of the hydrothermal activity.� 2011 Elsevier Ltd. All rights reserved.

1. INTRODUCTION

In natural oxic seawater dissolved Mn exists at nanomo-lar concentrations as Mn2+ ions (Sunda and Huntsman,1994). However, the Mn ions are thermodynamically unsta-ble in oxic seawater and will precipitate as insoluble Mn

0016-7037/$ - see front matter � 2011 Elsevier Ltd. All rights reserved.

doi:10.1016/j.gca.2011.02.011

⇑ Corresponding author at: Royal Netherlands Institute for SeaResearch (Royal NIOZ), P.O. Box 59, 1790 AB Den Burg, Texel,The Netherlands. Tel.: +31 222 369 464; fax: +31 222 319 674.

E-mail address: rob.middag@nioz.nl (R. Middag).

oxides, which are eventually lost from the water columnby particulate scavenging and sinking. The inorganic oxida-tion kinetics are relatively slow, but Mn oxidation is alsomicrobially mediated (Sunda and Huntsman, 1988 and ref-erences therein; Tebo et al., 2007 and references therein).Within the photic zone, Mn oxides can be reduced byphoto-reduction to soluble Mn ions (Sunda et al., 1983;Sunda and Huntsman, 1988, 1994). However, photo-oxidation of Mn ions to Mn oxides in the presence of humicsubstances has been suggested as well in surface waters(Nico et al., 2002). Dissolved Mn is an essential trace nutri-ent, most notably needed in Photosystem II for the splitting

2394 R. Middag et al. / Geochimica et Cosmochimica Acta 75 (2011) 2393–2408

of water by photoautotrophs to supply electrons to thereaction centre (e.g. Sunda et al., 1983 and references there-in; Raven, 1990). In marine diatoms Mn is also essential inthe enzyme superoxide dismutase, as recently shown (Peersand Price, 2004; Wolfe-Simon et al., 2006).

The balance between (photo-)reduction and (photo-)oxi-dation, atmospheric input (e.g. Landing and Bruland, 1980;Baker et al., 2006) and fluvial input (e.g. Elderfield, 1976;Aguilar-Islas and Bruland, 2006) to the surface layer causethe typical vertical open ocean profile of dissolved Mn withelevated concentrations at the surface. Below the surfacelayer in the absence of these sources for Mn, oxidationand loss from the water column leads to low backgroundconcentrations in the range of 0.1–0.2 nM Mn in the deepocean (Landing and Bruland, 1980, 1987; Statham et al.,1998). Further sources of Mn to the ocean are reductive dis-solution from (nearly) anoxic sediments (e.g. Froelich et al.,1979; Heggie et al., 1987; Johnson et al., 1992; Pakhomovaet al., 2007) and hydrothermal input (Klinkhammer et al.,1977; Klinkhammer and Bender, 1980). Sediments thatare (nearly) anoxic close to the sediment–water interfaceare usually not observed in the deep open ocean, with someexceptions in semi-enclosed deep basins like the Black Seaor Cariaco Trench. In these specific ocean regions alsothe (deep) water column is anoxic, leading to much higherconcentrations of dissolved Mn (e.g. Jacobs et al., 1987;Haraldsson and Westerlund, 1988; Lewis and Landing,1991; Yemenicioglu et al., 2006; Percy et al., 2008). In theproductive shelf seas export of particulate matter to thesea floor and subsequent microbial breakdown can alsolead to anoxic sediments (e.g. Heggie et al., 1987; Slompet al., 1997; Laes et al., 2007). Under these circumstancesMn gets mobilised from the sediments into the pore waterand can subsequently flux out into the overlying watercolumn.

This paper aims to assess the thus far largely unknowndistribution of Mn in the Arctic Ocean. Sources of dis-solved Mn to the Arctic Ocean could be its extensive shelfseas, like on the Eurasian side, the Barents Sea, Kara Seaand Laptev Sea. Furthermore the outflows of the Lena,Ob and Yenisey rivers enter the Laptev Sea (Guay andFalkner, 1997). Moreover, hydrothermal activity has beenobserved over the Gakkel Ridge in the Arctic Ocean(Edmonds et al., 2003). Photo-reduction and atmosphericinput on the other hand, might be less significant due tothe sea ice cover and the relatively small atmospheric dustflux to this remote ocean. Nevertheless, in the Arctic Oceanthe distribution of Mn will most likely be controlled by acombination of all these processes.

2. METHODS

2.1. Cleaning, sampling and filtration

Samples were taken aboard Polarstern during expeditionARK XXII/2 from 28 July until 07 October 2007 fromTromsø (Norway) to Bremerhaven. The Arctic Basin wassampled in five transects where, for dissolved Mn, 44 tracemetal stations (of which 27 deeper than 2000 m) were sam-pled (Fig. 1) with a total of 773 samples. Of the 773 samples

analysed for Mn, 11 samples (1.4%) were suspected outliersand therefore not further used in the data analyses and fig-ures. The dataset will become available through the interna-tional GEOTRACES data centre (http://www.bodc.ac.uk/geotraces) and is also available as an Electronic annex withthis paper.

The Low Density Polyethylene bottles (LDPE, Nalgene)used for the storage of reagents and samples were cleanedas described by Middag et al. (2009). Samples were takenusing 24 internally Teflon-coated PVC 12 litre GO-FLOSamplers (General Oceanics Inc.) mounted on an all-titanium frame, which was connected to a Kevlar wire andcontrolled from onboard (De Baar et al., 2008). In a class100 clean room environment the water was filtered directlyfrom the GO-FLO sampler over a 0.2 lm filter capsule(Sartobran-300, Sartorius) under nitrogen pressure(1.5 atm). Therefore all data reported in this paper is dis-solved Mn. The filtered seawater samples were taken incleaned LDPE sample bottles (125 ml) from each GO-FLObottle. All sample bottles were rinsed five times with about20% of the sample bottle volume with the sample sea water.

Samples for nutrients were collected unfiltered in HighDensity Polyethylene (HDPE) sample bottles, which wererinsed three times with sample water. Samples were storedin the dark at 4 �C prior to analysis (see text Section 2.3).

2.2. Analysis of dissolved Mn

Analyses of dissolved Mn were performed on board withthe method developed by Doi et al. (2004), modified to buf-fer the samples in-line (Middag et al., in press). The systemwas calibrated using standard additions from a 5000 nMMn stock solution (Fluka) to filtered acidified seawaterwith low concentrations of Mn that was collected in the re-gion. A five-points calibration line (0, 0.36, 0.73, 1.46 and2.18 nM standard additions) and blank determination weremade daily. The 3 lowest points (0, 0.36 and 0.73 nM) of thecalibration line were measured in triplicate and the 2 high-est points in duplicate to add more weight to the lower partof the calibration line. The blank was determined by mea-suring column-cleaned seawater. The latter was obtainedby passing seawater over a Toyopearl AF-Chelate 650 Mand an 8-hydroxyquinoline column (Landing et al., 1986),which retained the dissolved Mn. The average blank valuewas 0.02 nM. The limit of detection defined as three timesthe standard deviation of the blank was <0.01 nM. Theflow injection system was rinsed every day with a 0.5 MHCl solution.

A standard was measured in triplicate every day. Thisstandard was a sub-sample of a 25 L volume of filtered sea-water that was taken at the beginning of the cruise in theBarents Sea and stored at room temperature in the dark.The relative standard deviation of the replicate analysis sea-water sample that was analysed 22 times in triplicate on dif-ferent days was 2.19%. The relative standard deviation ofthe triplicate measurements on the separate days was onaverage 1.21%. The average concentration of Mn of thisstandard was 1.0 nM and the deviation from this averagefor a given measuring day was used as a correction factor.To verify whether this correction was decreasing the

Fig. 1. Stations for trace metal sampling during ARK XXII/2. Transects are indicated by numbered lines. Transect 1 goes from the BarentsSea into the Nansen Basin. Transect 2 also goes from the continental shelf into the Nansen Basin, just west of the St Anna Trough. Transect 3goes from the continental shelf, east of the St Anna Trough, into the Nansen and Amundsen Basin, over the Lomonosov Ridge into theMakarov Basin and onto the Alpha Ridge. Transect 4 goes from the Mendeleyev Ridge into the Makarov Basin and extends until just in theAmundsen Basin. Transect 5 follows the deep Gakkel Ridge and extends into the Laptev Sea. The eastuaries of the major rivers are indicatedas well. The dotted black arrows indicate the Atlantic water flowing in through Fram Strait and the counterclockwise boundary current, thegrey arrow indicates Pacific water flowing in through Bering Strait. AB: Amundsen Basin, AR: Alpha Ridge, BS: Bering Strait, CB: CanadaBasin, FS: Fram Strait, GR: Gakkel Ridge, LR: Lomonosov Ridge, MB: Makarov Basin, MR: Mendeleyev Ridge, NB: Nansen Basin, NZ:Novaya Zemlya, SAT: St Anna Trough, SZ: Severnaya Zemlya.

Dissolved Mn in the Arctic Ocean 2395

inter-daily variability in the dataset, every day a sample thatwas collected and measured the previous measuring day,was analysed once again. The deviation between the con-centrations measured on the different days decreased from3.28% to 2.68%, indicating the data correction procedureis beneficial.

To check the accuracy of the method, the samples col-lected on the SAFe cruise (Johnson et al., 2007), were ana-lysed for Mn. The concentrations of Mn found for SAFe

Surface (S1) and SAFe deep (D2) from 1000 m were0.73 ± 0.01 and 0.31 ± 0.01 nM (n = 3), respectively. Thisis slightly lower than the ‘consensus values’ (http://www.geotraces.org/Intercalibration) of 11 participatinglaboratories, which are 0.83 ± 0.08 and 0.37 ± 0.07 for S1and D2 respectively. However, the standard deviations forthe D2 overlap and the relatively large standard deviationis caused by 4 labs that reported values that were consider-ably higher (average 0.46 nM) than the remaining labs.

2396 R. Middag et al. / Geochimica et Cosmochimica Acta 75 (2011) 2393–2408

When excluding these high values, the resulting averageconcentration of 0.32 ± 0.03 nM is much closer to the aver-age concentration we observed. When excluding the samelabs also for the S1 SAFe standard (average 0.92 nM) theaverage concentration decreases to 0.77 ± 0.04 nM, whichis in range with the here reported concentration. The con-centrations of Mn for the SAFe standards are within rangeof what other labs have reported, but apparently still someissues have to be resolved before a true ‘consensus value’with a smaller standard deviation will be established.

2.3. Additional analyses

The salinity (conductivity), temperature and depth(pressure) were measured with two different CTD’s of thesame type (Seabird SBE 911+), one from the NetherlandsInstitute for Sea Research (NIOZ) and one from the AlfredWegener Institute (AWI). Both had been calibrated beforeand after the expedition by the company (Seabird). More-over, the conductivity sensors were calibrated during thecruise against salinity samples measured onboard (Schauer,2008). A transmissometer was in operation on the AWICTD. The transmissometer measures the fraction of light,from a light source, reaching the light detector and is indic-ative of suspended particles in the water column. For thefifth and final transect a spare AWI transmissometer wasinstalled on the NIOZ all-titanium sampling system. TheAWI data is available through their Pangaea databasehttp://www.pangaea.de.

The inorganic major nutrients nitrate (NO3), nitrite(NO2), phosphate (PO4) and silicic acid (Si(OH)4, hereafterabbreviated as Si) were determined following proceduresafter (Grasshoff et al., 1983). The samples for nutrients werestored in a refrigerator and analysed usually within 10hours and always within 16 hours on a Technicon TrAAcs800 autoanalyser. The reproducibility of an internal labora-tory standard mixture of Si, PO4 and NO3 was measureddaily and typically within 0.7% of its average value. More-over the deepest sample analysed of a station of 24 samples,was kept and re-analysed within the next run of the nextstation of 24 samples as another check and the agreementwas excellent (Middag et al., 2009).

The concentrations of iron (Fe) were determined on thesame cruise following procedures described by Klunderet al. (in press). These results are written up separatelybut one vertical profile is shown here (see Section 4.3).

3. HYDROGRAPHY

The hydrographical features of the Eurasian and Maka-rov basins have been described in detail by Anderson et al.(1994) and Rudels et al. (2000) and briefly summarised byMiddag et al. (2009). The latter summary is presented againhere.

Atlantic water flows into the Eurasian Arctic basins viaFram Strait and the Barents Sea along the continental slopewith the counterclockwise boundary current (Fig. 1), influ-encing the surface and intermediate waters. The boundarycurrent is modified by shelf processes and near the conti-nent a branch crosses over the Lomonosov Ridge that sep-

arates the Eurasian Basin from the Makarov Basin whileanother branch flows along this ridge. The branch of theboundary current that crosses the Lomonosov Ridge influ-ences the water column of the Makarov Basin. The forma-tion of dense water on the shelves and subsequent slopeconvection influences the deepest parts of the water columnin the Eurasian Basin.

Pacific water is flowing in through the Bering Strait overthe Siberian shelves and mixes with the freshwater input ofthe Siberian rivers, creating strong stratification in the cen-tral Arctic Ocean. The Pacific inflow influences the watercolumn of the Makarov Basin and this influence can extendinto the surface layer of the Amundsen Basin. The SurfaceMixed Layer (SML) is shallow, typically extending to 25 mdepth in the central Arctic Ocean. Below the SML, there isan upper halocline in the Makarov Basin that sometimesextends into the Amundsen Basin. The upper halocline ismarked by a nutrient maximum and is of Pacific origin.In the Nansen and Amundsen Basins the lower haloclineis more profound, marked by a minimum in nitrate andnitrite.

Below the haloclines, the Atlantic Layer Water (ALW)is found, generally extending to about 600 m depth. Thewarmer and more saline ALW does not cross over theLomonosov Ridge into the Makarov Basin, which showsin the strong front in temperature and salinity over theLomonosov Ridge. However, following the boundary cur-rent the Atlantic water flows into the Canadian Basin viathe shelf edge. This results in a relatively colder less salineALW in the Canadian Basin compared to the Eurasian Ba-sin. Below about 800 m the Makarov Basin becomes war-mer and more saline than water at a similar depth in theEurasian Basin. This difference remains the case throughoutthe intermediate depth and deep basin waters. In the deepbasins below the ALW, several water masses have been dis-tinguished such as the Intermediate Depth Water (IDW),which is a mixture of overlying ALW and underlying Eur-asian Basin Deep Water. Finally below the latter EurasianBasin Deep Water one may find Eurasian Basin BottomWater.

The water column properties encountered are describedin more detail by Middag et al. (2009) and briefly summa-rised here. The relatively cold and fresh surface layer wasdefined as the upper part of the water column with a poten-tial temperature <0 �C and a salinity <34.6 (Fig. 2). Thesurface layer is underlain by the ALW with its characteristicpotential temperature maximum and the IDW that is char-acterised by decreasing potential temperature with increas-ing depth. In the Deep Eurasian Basin Water (DEBW,comprising Eurasian Basin Deep Water and Eurasian BasinBottom Water) the potential temperature remained rela-tively constant while the salinity increased (Fig. 2). In theEurasian basin the IDW generally extended until about2000 m depth. Deeper than about 2000 m depth the DEBWwas observed.

In the Barents Sea and in the Nansen basin the surfacelayer extended to about 75 m depth. Below this layer anAtlantic influenced warmer and more saline bottom layerexisted in the Barents Sea. In the Eurasian basin the surfacelayer extended until about 50 m depth over the shelf and to

Salinity28 29 30 31 32 33 34 35 36

Pote

ntia

l Tem

pera

ture

(°C

)

-3

-2

-1

0

1

2

3

4

SLWALW and IDWDEBW and DMBW

Salinity

34.86 34.88 34.90 34.92 34.94 34.96 34.98 35.00

Pote

ntia

l Tem

pera

ture

(°C

)

-1.5

-1.0

-0.5

0.0

0.5

1.0

ALW and IDWDEBWDMBW

Fig. 2. Plots of potential temperature (�C) versus salinity for the Nansen, Amundsen and Makarov Basins of all transects in the surface layerwater (SLW) (closed circles), Atlantic layer water (ALW) and intermediate depth water (IDW) (open circles) and deep Eurasian Basin water(DEBW) and deep Makarov Basin water (DMBW) (closed triangles). At expanded scale to the right is the potential temperature (�C) versussalinity for the deep Nansen, Amundsen and Makarov basins of all transects. Different colours are used for the different transects and basins;black for transect 1, red for transect 2, blue for the Nansen Basin transect 3, green for the Amundsen Basin transect 3 and 4, grey for theMakarov Basin transect 3 and 4 and brown for transect 5. This figure has been adapted from Middag et al. (2009).

Dissolved Mn in the Arctic Ocean 2397

about 100 m in the Nansen and Amundsen basins. TheALW underneath the surface layer was relatively warmclose to the Eurasian shelf and cools further into the basinsaway from the shelf (Middag et al., 2009). In the Makarovbasin near the Lomonosov Ridge the surface layer extendedto similar depths as in the Eurasian basin, but to about150 m towards the Alpha Ridge. Furthermore, the surfacelayer in the Makarov basin was relatively fresh comparedto the surface layer in the Eurasian basin. The surface layerwas generally more saline furthest away from the continen-tal shelf and became freshest in the Laptev Sea. In the Lap-tev Sea the water also became warmer, with the upper partsof this shallow water column (55–35 m) being the warmestand least saline (salinity = 29.1).

The Gakkel ridge that separates the Nansen andAmundsen Basin is quite deep with depths between 4000and 5200 m and a maximum depth of 5200 m was observedin a deep trench just before the onset of the continentalslope. The water masses overlying the Gakkel Ridge weresimilar to what was observed in the Nansen and Amundsenbasins.

4. RESULTS AND DISCUSSION

4.1. Mn in the Shelf Seas and over the continental slopes

The concentrations of dissolved Mn for the entire watercolumn of transect 1 and 2 are shown in Figs. 3a and 3b. Inthe Barents Sea the concentrations were highest furthestaway from the Nansen basin. In the upper 30 m concentra-tions of Mn were up to 4 nM in the Barents Sea and as lowas around 2 nM in surface layer of the deep Nansen basin.In the Barents Sea concentrations of Mn decreased sharplyin the surface layer to a mid depth minimum, followed byan increase towards the sediment (Fig. 3a). Over the conti-nental slope into the Nansen basin the concentrations ofMn decreased even more sharply from the maximum con-

centrations in the uppermost surface layer to around0.35 nM around 200 m over the continental slope. How-ever, concentrations of Mn started to increase again just be-low 400 m to a maximum between 500 and 750 m depth atall slope stations of transects 1 and 2 (Fig. 4). This maxi-mum was most profound at the stations closest to the con-tinental shelf with a maximum value of over 0.6 nM Mnand coincided with lower light transmission (Fig. 4). Thisindicates the higher concentrations of Mn were associatedwith water containing particles and possible Mn-releasefrom these particles. Deeper than this Mn maximum overthe continental slope, concentrations of Mn were relativelystable around 0.2 nM before they increased towards thesediments.

Over the continental slope of transect 3 (Fig. 5), also aMn maximum was observed over the continental slope, lo-cated around 1000 m depth at the deeper station halfwaydown the slope. The concentration of Mn at this 1000 mmaximum was 0.25 nM, only slightly higher than the con-centrations of �0.2 nM above and below. Unfortunatelythere was no transmissometer on the trace metal samplingsystem at the time and a comparison with transmissioncan only be made when a concurrent regular hydrocastwas done, which was not the case at this station. However,the Mn maximum coincided with lower transmission, takenfrom observations at the regular hydrocast about 8 miles tothe south (Fig. 5), whereas 8 miles further north no trans-mission minimum was observed anymore. The station overthe shallow part of the slope was only 680 m deep and atthis station a mid-depth minimum of about 0.4 nM at300 m depth was observed, followed by an increase towardsthe sediment to a concentration of around 1 nM.

Over the continental shelf of transect 5 (Laptev Sea), theconcentrations near the sediments were around 5 nM nearthe shelf edge and increased to values over 20 nM (and out-side the range of the method used) in the Laptev Sea(Fig. 6). The relatively low concentrations of Mn at the

Fig. 3a. Concentrations of dissolved Mn (nM) over the entire water column in transect 1. The concentrations of Mn in the Barents Sea areshown in the inset at the bottom left and station 239 in the Barents Sea is also shown as vertical profile on the right. From this figure threesuspected outliers were omitted. Station numbers indicated are used in Fig. 4.

Fig. 3b. Concentrations of dissolved Mn (nM) over the entire water column in transect 2. The concentrations of Mn in the deepest part of theNansen Basin at station 261 (not shown in colour plot) are shown as a vertical profile versus depth on the right to avoid excessiveinterpolation. From the colour plot two suspected outliers were omitted. Station numbers indicated are used in Fig. 4, station 267 indicates theposition of the regular hydrocast.

2398 R. Middag et al. / Geochimica et Cosmochimica Acta 75 (2011) 2393–2408

station near the shelf edge corresponded with lower ob-served values for potential temperature and dissolved alu-minium (Middag et al., 2009). At the two deep (deeperthan 1500 m) stations over the continental slope, a Mnmaximum of about 0.4 nM was visible around 900 m depth.The Mn maximum at the deepest station coincided with asmall transmission minimum. At the other deep slope sta-tion, halfway the continental slope, the Mn maximum didnot correspond exactly with the transmission minimum,which was observed just below the Mn maximum. Towardsthe sediments the concentrations of Mn increased from�0.1 to around 0.3 nM and this increase coincided with adecreased transmission (Fig. 6). An even higher concentra-tion of Mn of 0.7 nM that coincided with decreased trans-

mission was observed near the sediments at the shalloweststation over the continental slope, just before the shelf edge.

The higher concentrations of Mn in the deeper parts ofthe water column of the Barents and Kara Seas are mostlikely due to a benthic flux. This benthic flux of Mn byresuspension of the sediment and direct fluxes from thepore water have been described for continental shelves else-where (e.g. Heggie et al., 1987; Slomp et al., 1997; Stathamet al., 1998; Laes et al., 2007). Moreover, the current datashows increasing Mn concentrations below the mid depthminimum in the Barents and Kara Seas, clearly indicatinga benthic source. In the Laptev Sea however, high concen-trations of Mn were found, but no clear increasing trendtowards the sediments was observed. These high

Fig. 4. Dissolved Mn (nM) and transmission (%) as a vertical profile versus depth over the continental slope of transect 1 (upper panels,stations 246 and 255) and 2 (lower panels, stations 267/268 and 266). Left panels show the profiles closest to the shelf, right panels show theprofiles closest to the deep Nansen Basin. From the upper and lower left panel (stations 246 and 268) one suspected outlier was omitted.Transmission profile from the lower left panel (transect 2, station 267) is from the regular hydrocast about 9 miles to the north as noconcurrent regular hydrocast was done at the trace metal station. Positions of the station numbers are also indicated in Figs. 3a and 3b. Notethe different scales between the different panels.

Dissolved Mn in the Arctic Ocean 2399

concentrations of Mn in the Laptev Sea were most likelythe result of fluvial input (see Section 4.2 below). Theapparent benthic flux of Mn from the sediments was onlyobserved in the Barents and Kara Seas, but its influencealso appears to affect the deep basins. The mid depth max-imum of Mn that corresponded with lowered transmission(see Figs. 4–6) at all transects crossing the continental slope,indicates Mn and particles from the shelf seas are trans-ported towards the deep basins. Both the minima in trans-mission and maxima in Mn fade away with increasingdistance from the shelf and are no longer distinguishableas pronounced minima and maxima in the deep basin be-yond the continental slope. Even so, the deep concentra-tions of Mn were higher in the Eurasian Basin comparedto the Makarov basin (Fig. 7). This could be caused by

hydrothermal input at the Gakkel Ridge as well as by inputfrom the shelf seas. The concentrations of dissolved alumin-ium and barium were also elevated in the deep Eurasian Ba-sin and it has been argued this was caused by input from theshelf seas via deep slope convection (Middag et al., 2009).Based on these observations, input from the shelf seas intothe deep basins most likely does affect the deep distributionof dissolved Mn but this effect can not be further quantifiedbased on the available data.

4.2. Mn in the Surface layer

High concentrations of Mn were observed in the upper30 m of the surface layer. These concentrations decreasedsteeply to subnanomolar levels within the surface layer to

Fig. 5. Concentrations of dissolved Mn (nM) in the upper 500 m of the water column (upper left panel) and over the remainder of the watercolumn (lower left panel) in transect 3. Two suspected outliers were omitted from this figure. Dissolved Mn (nM) and transmission (%) overthe continental slope (station 291) are shown as a vertical profile versus depth in the upper right panel. The transmission profile is from theregular hydrocast about 8 miles to the south as no concurrent regular hydrocast was done at the trace metal station. Dissolved Mn (nM) as avertical profile versus depth is shown for 5 profiles (stations 295, 299, 301, 302 and 306) in the deep Nansen basin in the lower right panel.

2400 R. Middag et al. / Geochimica et Cosmochimica Acta 75 (2011) 2393–2408

the low and quite uniform concentrations of Mn in theunderlying water masses (Fig. 7). The concentrations ofMn in the upper 30 m were as high as 5 nM around the edgeof the continental shelf and decreased to concentrations be-low 2 nM at transects 1 and 2 (Figs. 3a and 3b). At transect3 concentrations of Mn first decreased into the basin beforeincreasing again to concentrations of around 5 nM in the(geographic) middle of the Makarov basin. Towards theCanada Basin, the concentrations of Mn started to decreaseagain to values of around 3.5 nM (Fig. 5). The concentra-tions of Mn in the upper surface layer of transect 5 in-creased from 2 nM furthest away from the shelf to almost6 nM halfway the continental slope towards the LaptevSea (Fig. 6). The highest concentrations of Mn in the sur-face layer of transects 3, 4 and 5 coincided with lowest salin-ity values (<32) and strong negative correlations betweenMn and salinity were observed (Fig. 8a and b), indicatinga fresh water source for the increased concentrations ofMn. In transects 1 and 2 no low salinities were observedand hence no strong correlation between Mn and salinitywas found.

Strong negative correlations between Mn and salinitywere also observed by Yeats and Westerlund (1991). Lowsalinity can be an indication of river input, ice melt or Pa-cific inflow, which all have a relatively low salinity com-pared to the Atlantic water that flows in via Fram Strait(Ekwurzel et al., 2001). The quasi conservative water masstracer PO4

* (= PO4 + O2/175–1.95 lmol/kg; Broeckeret al., 1985, 1998) that compensates for organic respirationcan be used to partly distinguish between the differentsources of low salinity water to the surface layer as shownby Ekwurzel et al. (2001). River input and sea ice melt bothhave a relatively low PO4

* of �0.1 and 0.4 lM kg�1, respec-tively, Atlantic water has an intermediate PO4

* value of�0.7 lM kg�1 and Pacific water has a high PO4

*-value of�2.4 lM kg�1 (Ekwurzel et al., 2001, their Table 1). Whenplotting Mn versus PO4

* for the entire surface layer of tran-sect 3, no relation is apparent at first glance. However, atcloser inspection when plotting Mn versus PO4

* per basin,relations do appear. When plotting only the Eurasian basin,a negative correlation appears (Fig. 8c). Even though thiscorrelation is quite weak, it is significant ([PO4

*][lM

Fig. 6. Concentrations of dissolved Mn (nM) in the upper 500 m of the water column (upper left panel) and over the remainder of the watercolumn (lower left panel) in transect 5. Dissolved Mn (nM) and transmission (%) over the continental slope at station 389 and 385 are shownas a vertical profile versus depth in the upper and lower right panel, respectively. Concentrations at the shallowest station in the Laptev Seawere out of range of the analytical method. Station numbers indicated are used in Fig. 10.

Dissolved Mn in the Arctic Ocean 2401

kg�1] = �8.4[Mn][nM kg�1] + 5.8 with R2 = 0.4 andP < 0.001). All PO4

* values are below 0.7 lM kg�1, indicat-ing this water is a mixture of Atlantic water and fresh waterinput from either sea ice or rivers.

For the Makarov basin, however, two different correla-tions appear, both with a negative slope but with differenty-axis intercepts (Fig. 8d). The correlation with the highestintercept is found in the upper 25 m of the Makarov Basin.All PO4

* values in this correlation are below 2.4 and almostall (20 out of 22) above 0.7 lM kg�1. This indicates thewater is a mixture of Pacific water and fresh water inputfrom either sea ice or rivers. The correlation with the lowerintercept (Fig. 8d) is found in the deepest part of the surfacelayer, below about 100 m depth. The highest PO4

* value inthis correlation is 0.57 lM kg�1, indicating this water is amixture of Atlantic water and fresh water input from eithersea ice or rivers. The remaining data points in between thetwo correlations are from depths sampled in between theupper surface layer of Pacific origin and lower surface layerof Atlantic origin and most likely reflect mixing betweenthese layers.

Transect five went from the deep basin into the LaptevSea. The Lena River drains into the Laptev Sea, delivering

fresh water and Mn to the Arctic Ocean. When plotting thedissolved Mn versus salinity in the surface layer, a goodcorrelation appears in the deep basin region when excludingthe upper surface (25 m) of the three stations over the con-tinental slope (Fig. 8b). The concentrations of Mn in theupper surface layer over the slope and in the Laptev Seaare all much higher and do not fit the correlation observedover the deep basin. The PO4

* values in the surface layer oftransect 5 were, with one exception, below 0.7 lM kg�1 andcorrelated negatively with Mn ([PO4

*][lM kg�1] =�10.0[Mn][nM kg�1] + 6.5 with R2 = 0.61 and P < 0.001)indicating the water is of Atlantic origin with freshwater in-put from sea ice or rivers.

Overall, the PO4* tracer shows there is mixing between

Pacific and Atlantic water and that there is fresh water in-put from sea ice melt and/or rivers with high concentrationsof Mn along the entire transects 3, 4 and 5. Pacific waterdoes not appear to be an important source of Mn, other-wise a positive relation between PO4

* and Mn would havebeen found.

The influence of Atlantic, Pacific and fresh water fromrivers and sea ice on the surface layer in the Arctic Oceanhave already been assessed in previous reports based on

Mn (nM)

0.0 0.2 0.4 0.6 0.8 1.0 4.0 5.0

Dep

th (m

)

0

1000

2000

3000

4000

Mendeleev Ridge Makarov BasinAmundsen Basin

Fig. 7. Dissolved Mn (nM) as a vertical profile versus depth for thethree stations of transect 4 over the Mendeleev Ridge (circles), inthe Makarov Basin (squares) and in the Amundsen Basin(diamonds).

2402 R. Middag et al. / Geochimica et Cosmochimica Acta 75 (2011) 2393–2408

conservative tracers (e.g. Anderson et al., 1994; Rudelset al., 2000; Ekwurzel et al., 2001). These assessments giveuseful information about the volumes of water from the dif-ferent sources. However, when considering the importanceof these sources for non-conservative constituents like dis-solved Mn, the biogeochemical cycling and initial concen-trations of Mn should be taken into account. This isapparent in transect five that extended from the deep basininto the Laptev Sea. When considering the salinities in theupper surface layer and in the Laptev Sea, the concentra-tions of Mn were higher than would be expected based onthe Mn-salinity relation observed in the basin (Fig. 8aand b). Thus extrapolating the Mn-salinity relations ob-served in the surface layer of the deep basins to zero salinity(24 and 36 nM Mn for transects 3, 4 and 5, respectively)underestimates the Mn delivered to the Arctic Ocean bythe Lena River that drains into the Laptev Sea. This is con-firmed by Holemann et al. (2005) who reported concentra-tions of dissolved Mn between 82 and 1471 nM andconcentrations of particulate Mn between 13.1 and434 nM (From SPM and its metal content) in the riverand 111 ± 70 nM Mn in the Lena River plume (salinity<10) in the Laptev Sea. The latter river plume data was col-lected in the same time of the year as the data here pre-sented (September). The large range in concentrations ofdissolved Mn and Mn in suspended particulate matter is re-lated to variations in the discharge of the river (Holemannet al., 2005) and large variations have been observed inother river systems (Aguilar-Islas and Bruland, 2006 andreference therein).

Besides river discharge, the melting of sea ice can be asource of fresh water and elevated concentrations of Mn.There is little data on Mn in (sea) ice in the literature.Townsend and Edwards (1998) found up to 0.55 nM (unfil-tered) Mn in Antarctic snow, but these low concentrations

are most likely not representative for the Arctic Ocean thatis in greater proximity to continental sources. In coastal seaice in the Ross Sea (Antarctica) higher concentrations ofdissolved Mn between 0.84 and 10.5 nM and concentra-tions of particulate Mn between 0.13 and 23.2 nM were ob-served. A labile fraction of the particulate Mn (�20%)(Grotti et al., 2005) can partly dissolve when the sea icemelts. Campbell and Yeats (1982) reported trace metal datafor Arctic sea ice from the Eastern Arctic Ocean of 17.3 nMdissolved Mn and 26.8 nM particulate Mn. Apparentlythere can be a wide range of concentrations of Mn in seaice and the concentrations appear to be higher in the ArcticOcean. However, the concentrations of Mn in sea ice aremuch lower than those observed in the Lena River and itsdischarge. Moreover, the extrapolated concentrations ofMn at zero salinity (24 and 36 nM Mn) are higher thanthe reported concentrations of Mn in sea ice, indicating riv-er input must be at least partly responsible for the elevatedconcentrations of Mn in the surface layer. Also, for the Mnpotentially released by melting sea ice, precipitation andsubsequent loss from the surface layer can not be dis-counted. Moreover, the low concentrations of dissolvedAl in the surface layer of the Arctic Ocean indicate little in-put of ice melt (Middag et al., 2009). Thus the elevated con-centrations of Mn in the surface layer appear to be mainlydue to river input.

The work of Ekwurzel et al. (2001) shows that in thecentral Eurasian Basin (their Fig. 7), the fraction of the sur-face layer water that constitutes river run off is twice thefraction that constitutes sea ice. This implies that theextrapolation to zero salinity of the Mn-salinity relation ob-served in surface layer of the deep basins, should render aconcentration of Mn of 2/3 the river Mn concentrationand 1/3 sea ice Mn concentration when assuming, for thesake of argument, Mn is behaving conservatively. Takingthe lower limit of the Lena River water concentration ofMn of 82 nM (Holemann et al., 2005) and the sea ice con-centration of Mn of 17.3 nM (Campbell and Yeats, 1982),results in a concentration of Mn at zero salinity of about60 nM. Even when assuming none of the Mn released frommelting sea ice would remain in solution in the surfacelayer, the resulting concentration of 55 nM Mn exceedsthe extrapolation concentration of Mn for zero salinity(24 and 36 nM Mn) by almost a factor of 2. This indeedconfirms that a considerable fraction of the dissolved Mndelivered towards the Arctic Ocean by river discharge mustbe lost in the estuary and shelf sea. This is consistent withobserved behaviour of Mn in estuaries and salinity gradi-ents and the Mn loss is most likely caused by precipitationand scavenging (Salomons and Forstner, 1984; Noltinget al., 1999 and references therein). In the Columbia Riversystem, however, the concentrations of Mn behaved muchmore conservative over a salinity range of �9–32.5(Aguilar-Islas and Bruland, 2006). Apparently the behav-iour of Mn in estuaries and over salinity gradients dependson multiple factors. These could perhaps be the effects oflight and the balance between (photo-)reduction and(photo-)oxidation (Nico et al., 2002), effects of temperature,differences in organic substances and metal-organic com-plexes that potentially form (Kravtsov et al., 1995), initial

Fig. 8. Dissolved Mn (nM) versus salinity in the surface layer of transect 3 and 4 in the upper left panel and of transect 5 in the upper rightpanel (a). Linear regressions are shown in the graph and R2 = 0.65; n = 103; P < 0.001; Std. Error slope = 0.05 Std. Error intercept = 1.55 fortransect 3 and 4 combined. In transect 5 there is a correlation between Mn and salinity in the surface layer over the deep basin (circles;R2 = 0.94; n = 36; P < 0.001; Std. Error slope = 0.05 Std. Error intercept = 1.61) when excluding the upper surface (25 m) of the three stationsover the continental slope as well as the upper surface (25 m) deepest shelf station (triangles) of transect 5 (upper right panel (b)). Theconcentrations of Mn at the shallowest shelf station were outside the range of the analytical method used and not depicted. Dissolved Mn(nmol/kg) versus PO4

* (lmol/kg) in the surface layer of the Eurasian Basin (transect 3) in the lower left panel (c) (R2 = 0.37; n = 36 andP < 0.001) and in the Makarov Basin (transect 3 and 4) in the lower right panel (d) (Pacific origin R2 = 0.91; n = 15 and P < 0.001; Atlanticorigin R2 = 0.67; n = 9 and P < 0.008). Linear regressions are shown in the graph.

Dissolved Mn in the Arctic Ocean 2403

chemical composition of the source river (Aguilar-Islas andBruland, 2006) and mixing between different river plumeslike in the Arctic Ocean. However the sampling in this studywas mainly focussed on the central Arctic Ocean and fur-ther work is needed on the behaviour of Mn in different riv-er systems to identify the various factors and assess theirrelative importance.

4.3. Mn in the deep Deep Basins

In the deep Nansen basin at transect 1 and 2 concentra-tions of Mn decreased with increasing depth (�0.1 nM) un-

til a Mn maximum (up to 0.95 nM) was observed around2300 m depth at a rh (= (density � 1)�1000 kg dm�3) of28.090 (SD 0.0035). Deeper than this maximum the concen-trations of Mn decreased to concentrations as low as0.1 nM before a second, less profound maximum around3300 m (up to 0.45 nM; rh 28.103; SD 0.0012). The lowestconcentrations of Mn (approximately 0.06 nM) were ob-served at the greatest sampled depths (almost 4000 m).

In the deep basins of transect 3 concentrations of Mn de-creased with depth to values below 0.1 nM in the Amund-sen basin and to even lower values in the Makarov basin(around 0.05 nM). In contrast, in the Nansen basin showed

Transmission (%)90.0 90.5 91.0 91.5 92.0

Dep

th (m

)

0

1000

2000

3000

4000

Mn and Fe (nM)

0 1 2 3 4 5 6

Transmission (%)Mn (nM)Fe (nM)

Potential temperature (°C)-0.95 -0.93 -0.91 -0.89 -0.87 -0.85 -0.83 -0.81

Mn (nM)

0 1 2 3 4 5 62000

2200

2400

2600

2800

3000

3200

3400

3600

3800

Potential temperature (°C)Mn (nM)

Fig. 9. Dissolved Mn (nM) (closed circles), Fe (nM) (open circles) and transmission (%) as a vertical profile versus depth at the station overthe Gakkel Ridge (station 306, see Fig. 5) in the left panel and dissolved Mn (nM) and potential temperature (�C) as a vertical profile versusdepth at the same station between 2000 and 3800 m depth in the right panel.

Transmission (%)95.5 95.6 95.7 95.8 95.9 96.0

Dep

th (m

)

0

1000

2000

3000

4000

Mn (nM)0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Potential temperature (°C)

-1.00 -0.98 -0.96 -0.94 -0.92 -0.90 -0.88 -0.86

Transmission (%) St 372Mn (nM) St 371Mn (nM) St 372Mn (nM) St 373Potential temperature (°C) St 372

Fig. 10. Dissolved Mn (nM), transmission (%) and potentialtemperature (�C) as vertical profiles versus depth. Three profiles areshown for Mn over the deep Gakkel Ridge at transect 5 (stations371, 372 and 373; Fig. 6) and one profile for the transmission andpotential temperature (station 372).

2404 R. Middag et al. / Geochimica et Cosmochimica Acta 75 (2011) 2393–2408

a clear maximum around 2500 m over the entire basin(Fig. 5), at a similar rh (28.092; SD 0.001) as for transect1 and 2. An exception was the station at the transition be-tween the continental slope and the deep basin where nodiscernable maximum was observed. The highest Mn con-centration of almost 5 nM was observed at the station overthe Gakkel Ridge and the maximum concentrations of Mnaround 2500 m depth decreased through the Nansen basintowards the Eurasian continental slope (Fig. 5). At the sta-tion over the Gakkel Ridge also the transmission decreased(Fig. 9) and the potential temperature was elevated (Fig. 9)around 2500 m depth. Moreover, the dissolved iron (Fe)concentrations were elevated over the same depth rangeas Mn (Fig. 9). There is no universal relation between Mnand transmission, however, elevated concentrations of Mnin conjunction with decreased transmission indicates an in-creased amount of particles as the Mn source. Particles tendto sink out of the water column whereas the elevated con-centrations of Mn can propagate over much longer dis-tances and thus be found in the absence of decreasedtransmission. Hydrothermal input is usually found by look-ing at light scattering (transmission or turbidity), tempera-ture and concentrations of Mn (e.g. Klinkhammer et al.,2001; Edmonds et al., 2003) and the elevated concentrationsof Mn coinciding with maxima of Fe and potential temper-ature and a transmission minimum observed over the Gak-kel Ridge clearly indicate hydrothermal input.

Over the deep Gakkel Ridge the three stations furthestaway from the continental slope had station numbers 371,372 and 373 (Fig. 6), going into the direction of the slope.At these stations two separate Mn maxima were observed(Fig. 10) around 2500 m (rh 28.091; SD 0.0018) and be-tween 3000 and 3300 m depth (rh 28.102: SD 0.0010), atsimilar rh as for the previous transects. The middle stationof the three (372) had the highest maximum concentrationsof over 0.3 and 1.2 nM at 2500 and 3130 m depth, respec-

tively. Deeper than the observed maxima, the concentra-tions of Mn at the three stations matched closely as theydecreased to values of around 0.15 nM until the greatestsampled depth close to the sediment. The deepest, mostprofound, Mn maximum coincided with a transmissionminimum (Fig. 10). The transmission minimum was, likethe Mn maximum, most profound at the middle station

Distance (km)0 100 200 300 400 500 600

Mn

(nM

)

0

1

2

3

4

5

6

Transect 1 and 2

Transect 3

)*(0058.0e*4.98)nM(Mn distance

)*(0121.0e*4.97)nM(Mn distance

Fig. 11. Concentrations of Mn (nM) at the Mn maximum that wasobserved over the rh range 28.089–28.094 (average 28.091; SD0.003; around 2300–2500 m depth) versus distance (km) fromstation 306 over the Gakkel Ridge (see Fig. 5) for transect 1 and 2(circles) and transect 3 (squares). Exponential equations are shownin the figure.

Dissolved Mn in the Arctic Ocean 2405

(372). Moreover, the Mn maximum coincided with a poten-tial temperature maximum (Fig. 10) at the middle station(372). At the station just before the deep trench the Mnmaxima around 2500 m and between 3000 and 3300 werestill visible in the profile, but they were not profound. Overthe deep trench, however, the Mn maxima were not distin-guishable and concentrations were around 0.1 in the deeptrench.

In transect 4 no hydrothermal input was observed(Fig. 7). Below the surface layer the two profiles on theMendeleev Ridge and in the Makarov basin were similarand concentrations of Mn were lower than in the Amund-sen basin. The lowest concentrations (�0.05 nM) wereagain observed in the deep Makarov basin. In the absenceof hydrothermal input, the concentrations of Mn are rela-tively uniform in the remainder of the water column(Fig. 7). However, the concentrations did tend to decreaseslightly with depth from the concentrations observed belowthe surface layer to the lowest concentrations in the deepestlayers. Within the deep basins, usually no increase towardsthe sediments was observed, indicating no sedimentary fluxin the deep basins.

Hydrothermal input is a known source of dissolved Mnfor the deep ocean (Klinkhammer et al., 1977, 2001;Klinkhammer and Bender, 1980). The Gakkel Ridge is anultraslow-spreading ridge, but has numerous known sitesof hydrothermal activity (Edmonds et al., 2003). Edmondset al. (2003) located nine primary vent sites along the Gak-kel Ridge, of which the most easterly at 85�390N, 84�500Ewas located closest to the observed hydrothermal plumein transects 3. To locate the vent site, Edmonds et al.(2003) used Miniature Autonomous Plume Recorders thatrecorded temperature and light scattering. The latter is, likethe light transmission, indicative of suspended particles. Be-sides the physical parameters light scattering and tempera-ture anomalies, also some samples were taken for totaldissolvable Mn by Edmonds et al. (2003). The depth andthe depth range (between 2000 and 3000 m) at which theplume was detected at 85�390N, 84�500E by Edmondset al. (2003, their Fig. 4) was remarkably similar to whatwas observed for the most profound plume of Mn at85�550N, 91�070E (Fig. 9). The vertical profiles of total dis-solvable Mn had a higher background concentration(�0.1–1 nM) and exceeded 10 nM in the plume. The higherbackground concentrations of Mn confirm the remark ofEdmonds et al. (2003) that some portion of the total dis-solvable Mn signal could result from leaching of Mn oxidecoatings off suspended particles.

The depth range at which the Mn maximum and trans-mission anomaly at transect 5 were observed, is consistentwith the depth range for which an additional hydrothermalplume was observed at the closest and several other loca-tions sampled by Edmonds et al. (2003). Furthermore, amaximum of dissolved Mn was observed throughout theEurasian basin in the 3000–3300 m depth range.

The maxima of Mn of hydrothermal origin that were ob-served in the entire Eurasian Basin could be related to theenclosed nature of this ocean basin. But at the very least,this shows that the influence of hydrothermal vent activityon the distribution of Mn in the deep Arctic Ocean is more

significant than hitherto realised and can indeed extend todistances of over 500 km, as the model of Weiss (1977) pre-dicted. The extent of the hydrothermal influence on the dis-tribution of Mn in other deep ocean basins remains to bestudied in more detail but could be much greater than ex-pected thus far.

The maxima of Mn around 2300–2500 m (rh 28.091; SD0.003) in the Eurasian basin decreased quasi-exponentiallyaway from the station in transect 3 over the Gakkel Ridge(Fig. 11). This was observed for the maxima in transect 3and the maxima in transect 1 and 2 combined. This expo-nential decrease with distance agrees with the model ofWeiss (1977). However, the residence time for Mn thatcan be calculated from the model fit for the different tran-sects (0.4–2 years) is much shorter than the 51 years calcu-lated by Weiss (1977). Therefore the assumption of themodel that eddy diffusion is the only process of influencefor the plume expansion appears incorrect. This suggeststhat also deep water circulation, increased Mn oxidationby bacteria (Sunda and Huntsman, 1988; Tebo et al.,2007) or increased Mn oxidation related to the oxygen con-centrations (Johnson et al., 1996) have to be taken into ac-count. Moreover, it must also be considered that there aremultiple active hydrothermal vent sites (Edmonds et al.,2003). This further complicates the process, which cantherefore not be described solely by a simple turbulent mix-ing and scavenging model. Nevertheless, the observed dis-tribution of Mn around 2300–2500 m (rh 28.091; SD0.003) at various distances from the Gakkel Ridge in a spe-cific direction can be described with an exponential equa-tion without accounting for currents, increased oxidationor multiple sources. This shows that scavenging removalof Mn must play a major role, but nevertheless, the hydro-thermal influence is affecting the Mn distribution on a basinwide scale.

2406 R. Middag et al. / Geochimica et Cosmochimica Acta 75 (2011) 2393–2408

4.4. Comparison with previously reported Mn in the Arctic

Ocean

Previously limited data has been available about the dis-tribution of dissolved Mn in the Arctic Ocean. To the bestof our knowledge the only data published on dissolved Mn(0.4 lm filtered in this case) in the central Arctic Ocean isfrom Yeats (1988). Like the current profiles of Mn, this pro-file also shows elevated concentrations of Mn in the surface.This station extended to about 1350 m depth and the lowestconcentration of about 0.25 nM was found in the AtlanticLayer. This concentration of 0.25 nM in the Atlantic Layeris only slightly higher than the current concentrations andcould be related to spatial and temporal variability or dif-ferences is sampling and filtration. The higher concentra-tions of Mn (around 0.6 nM) in the deeper parts of thewater column reported by Yeats (1988) are most likelythe result of the relatively shallow water column depth.The other profile of Mn reported by Yeats (1988) close toFram Strait was for unfiltered samples and extended toabout 2500 m depth. This profile for unfiltered Mn alsoshows a maximum of Mn at the surface, but the generalconcentrations of the unfiltered samples are much higher.

Slightly more data on dissolved Mn is available for theEastern Arctic Ocean near Greenland and the CanadianArchipelago by Campbell and Yeats (1982) and Yeats andWesterlund (1991) (both 0.4 lm filtered). In the Central Baf-fin Bay near Greenland (Campbell and Yeats, 1982) concen-trations of Mn decreased from a surface maximum of about5.3–1.6 nM at the greatest sampled depth of 2300 m. Thisprofile shape, a surface maximum and lower concentrationsin the deep, is similar to the shape of the current vertical pro-files of Mn. However, the concentrations of Mn in this en-closed bay appear to be much higher. The concentrations ofMn in the shelf sea at the edge of the Canada Basin at81�43.40N, 93�25.00W (Yeats and Westerlund, 1991) wereconsiderably lower. A maximum of Mn of 1.58 nM was ob-served at the shallowest depth of 60 m, followed by a steep de-crease to 0.24 nM at 150 m depth and concentrations of 0.51and 0.26 nM at the greatest sampled depths of 270 and 280 m,respectively. The latter general profile shape and concentra-tions are comparable to the current data for shallow stationsin the shelf seas with the exception of the relatively low con-centration of Mn near the sediments in the previous work.

5. CONCLUSIONS

This study of dissolved Mn in the central Arctic Oceanshows elevated concentrations in the surface layer with con-centrations up to 6 nM. The highest concentrations of Mnwere observed in the shelf seas, notably the Laptev Sea withconcentrations over 20 nM. In the deep basin the concen-trations of Mn decreased steeply from the surface valuesto low background concentrations of just below 0.1 nMin the Eurasian Basin and even lower at about 0.05 nM inthe Makarov Basin. The general profile shape is consistentwith three previous studies of dissolved Mn in the ArticOcean. Observed differences in the absolute concentrationsare most likely related to differences in sample area, sam-pling and filtration.

The elevated concentrations of Mn in the surface layerhave a fresh water source and show a negative relation withsalinity. Both fluvial input and sea ice melt have to be takeninto account as fresh water and Mn source, but fluvial inputappears to be the most important in both volume and con-centration of Mn. By extrapolating the Mn-salinity relationobserved in the surface layer of the deep basin it is apparentthat a large portion of the Mn delivered by the Arctic riversis removed in the shelf seas and does not pass into the cen-tral basins.

In the shelf seas a mid depth minimum was observed,indicating a benthic source. From the shelf seas Mn is ex-ported along the continental slope into the deep basinand this process results in higher concentrations of Mnand lower light transmission over the slope. The effect ofthis Mn input over the slope faded out with distance intothe basin and was not observed beyond the continentalslope.

In the deep basins the lowest concentrations of Mn wereobserved. The concentrations of Mn tended to decreaseslightly with depth below the initial decrease in the first200–300 m, but were overall quite uniform within the deepbasins. The higher concentrations in the Eurasian Basincompared to the Makarov Basin might be related to the ob-served input from the shelf seas, but also the hydrothermalinput plays a role. Distinct hydrothermal activity and Mninput was observed over the deep Gakkel Ridge. Thehydrothermal plume of dissolved Mn could be seen in aclear Mn maximum around 2300–2500 m depth throughoutthe Eurasian Basin and a smaller plume was observedaround 3200 m. The Mn concentration at the 2300–2500 m maximum decreased exponentially with distancefrom the Gakkel Ridge, which is consistent with a first or-der scavenging model. However, also the deep basin circu-lation, oxidation of Mn and the multiple sources of Mnhave to be considered when assessing the hydrothermalMn plume dispersion.

ACKNOWLEDGEMENTS

The authors are most grateful to the master S. Schwarze, offi-cers and crew of R.V. Polarstern for their splendid support duringthe expedition. Special thanks to Sven Ober and Lorendz Boomfor the operation of the all-titanium CTD sampling system fortrace metals, Loes Gerringa for her constructive comments andCharles-Edouard Thuroczy for his help with sampling and generalsupport. Last but not least many thanks to all participants ofcruise ARK XXII/2 for their help and support and for makingthe long cruise enjoyable. This work was funded by The Nether-lands Organisation for Scientific Research, IPY-NL-GEOTRAC-ES sub-project: 851.40.101. Constructive comments of threeanonymous reviewers and the associate editor are appreciatedand have led to significant improvements of the manuscript.

APPENDIX A. SUPPLEMENTARY DATA

Supplementary data associated with this article can befound, in the online version, at doi:10.1016/j.gca.2011.02.011.

Dissolved Mn in the Arctic Ocean 2407

REFERENCES

Aguilar-Islas A. and Bruland K. W. (2006) Dissolved manganeseand silicic acid in the Columbia River plume: a major source tothe California Current and coastal waters off Washington andOregon. Mar. Chem. 101(3–4), 223–247.

Anderson L. G., Bjork G., Holby O., Jones E. P., Kattner G.,Koltermann K. P., Liljeblad B., Lindegren R., Rudels B. andSwift J. H. (1994) Water masses and circulation in the EurasianBasin: results from the Oden 91 expedition. J. Geophys. Res.

99(C2), 3273–3283.

Baker A. R., Jickells T. D., Witt M. and Linge K. L. (2006) Trendsin the solubility of iron, aluminium, manganese and phospho-rus in aerosol collected over the Atlantic Ocean. Mar. Chem.

98(1), 43–58.

Broecker W. S., Takahashi T. and Takahashi T. (1985) Sources andflow patterns of deep-ocean waters as deduced from potentialtemperature, salinity, and initial phosphate concentration. J.

Geophys. Res. 90(NC4), 6925–6939.

Broecker W. S., Peacock S. L., Walker S., Weiss R., Fahrbach E.,Schroeder M., Mikolajewicz U., Heinze C., Key R., Peng T. H.and Rubin S. (1998) How much deep water is formed in theSouthern Ocean?. J. Geophys. Res. 103(C8) 15833–15843.

Campbell J. A. and Yeats P. A. (1982) The distribution ofmanganese, iron, nickel, copper and cadmium in the waters ofBaffin-Bay and the Canadian arctic archipelago. Oceanol. Acta

5(2), 161–168.

De Baar H. J. W., Timmermans K. R., Laan P., de Porto H. H.,Ober S., Blom J. J., Bakke M. C., Schilling J., Sarthou G., SmitM. G. and Klunder M. (2008) Titan: A new facility forultraclean sampling of trace elements and isotopes in the deepoceans in the international Geotraces program. Mar. Chem.

111(1–2), 4–21.

Doi T., Obata H. and Maruo M. (2004) Shipboard analysis ofpicomolar levels of manganese in seawater by chelating resinconcentration and chemiluminescence detection. Anal. Bioanal.

Chem. 378(5), 1288–1293.

Edmonds H. N., Michael P. J., Baker E. T., Connelly D. P., SnowJ. E., Langmuir C. H., Dick H. J. B., Muhe R., German C. R.and Graham D. W. (2003) Discovery of abundant hydrother-mal venting on the ultraslow-spreading Gakkel ridge in theArctic. Nature 421(6920), 252–256.

Ekwurzel B., Schlosser P., Mortlock R. A., Fairbanks R. G. andSwift J. H. (2001) River runoff, sea ice meltwater, and Pacificwater distribution and mean residence times in the ArcticOcean. J. Geophys. Res. 106(C5), 9075–9092.

Elderfield H. (1976) Manganese fluxes to the oceans. Mar. Chem.

4(3), 103–132.

Froelich P. N., Klinkhammer G. P., Bender M. L., Luedtke N. A.,Heath G. R., Cullen D., Dauphin P., Hammond D., HartmanB. and Maynard V. (1979) Early oxidation of organic matter inpelagic sediments of the eastern equatorial Atlantic: suboxicdiagenesis. Geochim. Cosmochim. Acta 43(7), 1075–1090.

Guay C. and Falkner K. (1997) Barium as a tracer of Arctichalocline and river waters. Deep-Sea Res. Pt. II 44(8), 1543–

1569.

Grasshoff K., Erhardt M. and Kremling K. (1983) Methods of

Seawater Analysis. Verlag Chemie, Weinheim, Germany, 419pp..

Grotti M., Soggia F., Ianni C. and Frache R. (2005) Trace metalsdistributions in coastal sea ice of Terra Nova Bay, Ross Sea,Antarctica. Antarc. Sci. 17(2), 289–300.

Haraldsson C. and Westerlund S. (1988) Trace metals in the watercolumns of the Black Sea and Framvaren Fjord. Mar. Chem.

23(3–4), 417–424.

Heggie D., Klinkhammer G. and Cullen D. (1987) Manganese andcopper fluxes from continental-margin sediments. Geochim.

Cosmochim. Acta 51(5), 1059–1070.

Holemann J. A., Schirmacher M. and Prange A. (2005) Seasonalvariability of trace metals in the Lena River and the southeast-ern Laptev Sea: impact of the spring freshet. Global Planet.

Change 48(1–3), 112–125.

Jacobs L., Emerson S. and Huested S. S. (1987) Trace metalgeochemistry in the Cariaco Trench. Deep-Sea Res. 34(5–6),

965–981.

Johnson K. S., Berelson W. M., Coale K. H., Coley T. L., Elrod V.A., Fairey W. R., Iams H. D., Kilgore T. E. and Nowicki J. L.(1992) Manganese flux from continental margin sediments in atransect through the oxygen minimum. Science 257(5047),

1242–1245.

Johnson K. S., Boyle E., Bruland K., Measures C., Moffett J.,Aguilarislas A., Barbeau K., Cai Y., Chase Z., Cullen J., DoiT., Elrod V., Fitzwater S., Gordon M., King A., Laan P.,Laglera-Baquer L., Landing W., Lohan M., Mendez J., MilneA., Obata H., Ossiander L., Plant J., Sarthou G., Sedwick P.,Smith G. J., Sohst B., Tanner S., Van Den Berg S. and Wu J.(2007) Developing standards for dissolved iron in seawater. Eos

Trans. Am. Geophys. Union 88(11), 131–132.

Johnson K. S., Coale K. H., Berelson W. M. and Gordon R. M.(1996) On the formation of the manganese maximum in theoxygen minimum. Geochim. Cosmochim. Acta 60(8), 1291–1299.

Klinkhammer G., Bender M. and Weiss R. F. (1977) Hydrother-mal manganese in Galapagos Rift. Nature 269(5626), 319–320.

Klinkhammer G. P. and Bender M. L. (1980) The distribution ofmanganese in the Pacific Ocean. Earth Planet. Sci. Lett. 46(3),

361–384.

Klinkhammer G. P., Chin C. S., Keller R. A., Dahlmann A.,Sahling H., Sarthou G., Petersen S., Smith F. and Wilson C.(2001) Discovery of new hydrothermal vent sites in BransfieldStrait, Antarctica. Earth Planet. Sci. Lett. 193(3–4), 395–

407.

Klunder M. B., Laan P., Middag R., De Baar H. J. W. and VanOoijen, J. C. (in press) Dissolved Fe in the Southern Ocean(Atlantic Sector). Deep-Sea Res. Pt. II. doi:10.1016/j.dsr2.2010.10.042.

Kravtsov V. A., Gordeev V. V. and Pashkina V. I. (1995) Labiledissolved forms of heavy metals in the Kara Sea waters.Oceanology 34(5), 611–618 (English translation).

Laes A., Blain S., Laan P., Ussher S. J., Achterberg E. P., TreguerP. and de Baar H. J. W. (2007) Sources and transport ofdissolved iron and manganese along the continental margin ofthe Bay of Biscay. Biogeosciences 4(2), 181–194.

Landing W. M. and Bruland K. W. (1980) Manganese in the NorthPacific. Earth Planet. Sci. Lett. 49(1), 45–56.

Landing W. M. and Bruland K. W. (1987) The contrastingbiogeochemistry of iron and manganese in the Pacific-ocean.Geochim. Cosmochim. Acta 51(1), 29–43.

Landing W. M., Haraldsson C. and Paxeus N. (1986) Vinylpolymer agglomerate based transition metal cation-chelatingion-exchange resin containing the 8-hydroxyquinoline func-tional group. Anal. Chem. 58(14), 3031–3035.

Lewis B. L. and Landing W. M. (1991) The biogeochemistry ofmanganese and iron in the Black Sea. Deep-Sea Res. 38(Suppl.

1.2), 773–803.

Middag R., De Baar H. J. W., Laan P. and Bakker K. (2009)Dissolved aluminium and the silicon cycle in the Arctic Ocean.Mar. Chem. 115(3–4), 176–195.

Middag R., De Baar H. J. W., Laan P. and Van Ooijen J. (in press)Dissolved manganese in the Atlantic section of the SouthernOcean. Deep-Sea Res. Pt. II. doi:10.1016/j.dsr2.2010.10.043.

2408 R. Middag et al. / Geochimica et Cosmochimica Acta 75 (2011) 2393–2408

Nico P. S., Anastasio C. and Zasoski R. J. (2002) Rapid photo-oxidation of Mn(II) mediated by humic substances. Geochim.

Cosmochim. Acta 66(23), 4047–4056.

Nolting R. F., Helder W., De Baar H. J. W. and Gerringa L. J. A.(1999) Contrasting behaviour of trace metals in the Scheldtestuary in 1978 compared to recent years. J. Sea Res. 42(4),

275–290.

Pakhomova S. V., Hall P. O. J., Kononets M. Yu., Rozanov A. G.,Tengberg A. and Vershinin A. V. (2007) Fluxes of iron andmanganese across the sediment–water interface under variousredox conditions. Mar. Chem. 107(3), 319–331.

Peers G. and Price N. M. (2004) A role for manganese insuperoxide dismutases and growth of iron-deficient diatoms.Limnol. Oceanogr. 49(5), 1774–1783.

Percy D., Li X. N., Taylor G. T., Astor Y. and Scranton M. I.(2008) Controls on iron, manganese and intermediate oxidationstate sulfur compounds in the Cariaco Basin. Mar. Chem.

111(1–2), 47–62.

Raven J. A. (1990) Predictions of Mn and Fe use efficiencies ofphototrophic growth as a function of light availability forgrowth and of C assimilation pathway. New Phytol. 116(1),

1–18.

Rudels B., Muench R. D., Gunn J., Schauer U. and Friedrich H. J.(2000) Evolution of the Arctic Ocean boundary current north ofthe Siberian Shelves. J. Mar. Syst. 25, 77–99.

Salomons W. and Forstner U. (1984) Metals in the Hydrocycle, firsted. Springer-Verlag, Berlin, Heidelberg, 349 p.

Schauer U. (Ed.) 2008. The Expedition ARKTIS-XXII/2 of theResearch Vessel “Polarstern” in 2007. Berichte zur Polar- und

Meeresforschung 579, 265 p.Slomp C. P., Malschaert J. F. P., Lohse L. and VanRaaphorst W.

(1997) Iron and manganese cycling in different sedimentaryenvironments on the North Sea continental margin. Cont. Shelf

Res. 17(9), 1083–1117.

Statham P. J., Yeats P. A. and Landing W. M. (1998) Manganesein the eastern Atlantic Ocean: processes influencing deep andsurface water distributions. Mar. Chem. 61(1–2), 55–68.

Sunda W. G., Huntsman S. A. and Harvey G. R. (1983)Photoreduction of manganese oxides in seawater and itsgeochemical and biological implications. Nature 301(5897),

234–236.

Sunda W. G. and Huntsman S. A. (1988) Effect of sunlight onredox cycles of manganese in the southwestern Sargasso Sea.Deep-Sea Res. 35(8), 1297–1317.

Sunda W. G. and Huntsman S. A. (1994) Photoreduction ofmanganese oxides in seawater. Mar. Chem. 46(1–2), 133–152.

Tebo B. M., Clement B. G. and Dick G. J. (2007) Biotransforma-tions of manganese. In Manual of Environmental Microbiology

(eds. C. J. Hurst, R. L. Crawford, J. L. Garland, D. A. Lipson,A. L. Mills and L. D. Stetzenbach), third ed. ASM Press,

Washington, DC, pp. 1223–1238.

Townsend A. T. and Edwards R. (1998) Ultratrace analysis ofAntarctic snow and ice samples using high resolution induc-tively coupled plasma mass spectrometry. J. Anal. Atom.

Spectrom. 13(5), 463–468.

Weiss R. F. (1977) Hydrothermal manganese in deep-sea –scavenging residence time and Mn–He-3 relationships. Earth

Planet. Sci. Lett. 37(2), 257–262.

Wolfe-Simon F., Starovoytov V., Reinfelder J. R., Schofield O.and Falkowski P. G. (2006) Localization and role of manganesesuperoxide dismutase in a marine diatom. Plant Physiol. 142(4),

1701–1709.

Yeats P. A. (1988) Manganese, nickel, zinc and cadmium distri-butions at the Fram-3 and Cesar Ice Camps in the ArcticOcean. Oceanol. Acta 11(4), 383–388.

Yeats P. A. and Westerlund S. (1991) Trace-metal distributions atan Arctic-Ocean Ice Island. Mar. Chem. 33(3), 261–277.

Yemenicioglu S., Erdogan S. and Tagrul S. (2006) Distribution ofdissolved forms of iron and manganese in the Black Sea. Deep-

Sea Res. Pt. II 53(17–19), 1842–1855.

Associate editor: Timothy J. Shaw