The neodymium isotopic composition of waters masses in the eastern Pacific sector of the Southern Ocean

Available online at www.sciencedirect.com

Geochimica et Cosmochimica Acta 79 (2012) 41–59 www.elsevier.com/locate/gca

The neodymium isotopic composition of waters masses in the eastern Pacific sector of the Southern Ocean P. Carter a,b, D. Vance a,⇑, C.D. Hillenbrand c, J.A. Smith c, D.R. Shoosmith c a

Bristol Isotope Group, Department of Earth Sciences, University of Bristol, Wills Memorial Building, Bristol BS8 1RJ, UK b RWE npower Renewables, Auckland House, Lydiard Fields, Swindon SN5 8ZT, UK c British Antarctic Survey, High Cross, Madingley Road, Cambridge CB3 0ET, UK Received 18 July 2011; accepted in revised form 21 November 2011; available online 1 December 2011

Abstract The Antarctic Circumpolar Current is one of the key components of ocean circulation, and a knowledge of its isotopic composition is essential to the use of neodymium (Nd) isotopes to trace circulation now and in the past. Here we present 57 new analyses of the Nd isotopic composition of the water column in the eastern Pacific sector of the Southern Ocean, documenting both the variation in three dimensions as well as the controls on that variability. Nd isotopic data for the middle of the water column demonstrate the homogeneity of Circumpolar Deep Water (CDW) at an eNd value of 8.7 ± 0.1. This homogeneity reflects the large Nd inventory in the ACC flow, and the degree to which this large inventory buffers the Nd characteristics of the ACC against extra-oceanic inputs from the continents, either via dust from the atmosphere or through dissolved and particulate material from the adjacent continents. CDW upwells onto the Amundsen Sea shelf and, even here, its Nd isotopic properties are close to conserved in the middle of the water column (eNd = 8.0 ± 0.2 at 600 m). At the top and bottom of the shelf water column, however, the Nd isotopic and concentration characteristics are strongly modified (to eNd as high as 4.5). All the shelf water column data obtained here are consistent with net addition of Nd to bottom and surface waters with a contrasting isotopic composition that is matched by local sediment (eNd = 1 to 2), followed by conservative mixing of that water into intermediate levels. Mixing with this shelf composition also leads to significant modification of open ocean surface water (Nd isotopic shift around 1 epsilon unit) in the ACC as it flows eastwards. Modification of open ocean bottom waters by interaction with sediment is more subtle, but there is marked non-conservative removal of Nd accompanied by significant changes in isotopic composition in waters within 10 m of the seabed. The new data demonstrate the conservativity of Nd in the middle water column, especially for such large volume flows as the ACC. Though boundary exchange-type processes are clearly important in this region, and their imprint on both the shelf and open-ocean surface water is significant, there is no observable impact on the main core of CDW. This finding augurs well for the use of Nd isotopes as a conservative water-mass tracer now and in the past. For example, these data suggest that the only likely control on the temporal variability of southern-component deep water exported northwards into the Atlantic through the last glacial cycle is variations in the input of North Atlantic Deep Water in the Atlantic sector. On the other hand, the data also suggest caution in the use of sedimentary archives of bottom water Nd as records of deep water Nd isotopic characteristics, given the dramatic modification of bottom waters on the shelf and the more subtly non-conservative behaviour in the open ocean bottom water that are both suggestive of modification of the Nd characteristics of bottom waters by sediment. Ó 2011 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Tel.: +44 117 954 5418; fax: +44 117 935 3385.

E-mail address: [email protected] (D. Vance). 0016-7037/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.gca.2011.11.034

42

P. Carter et al. / Geochimica et Cosmochimica Acta 79 (2012) 41–59

1. INTRODUCTION The main source of neodymium (Nd) to the oceans is from the continents (see Frank, 2002 for a review). The radiogenic isotope composition of this input (the 143 Nd/144Nd ratio, commonly expressed as eNd, the parts per 10,000 deviation of any sample from the Chondritic Uniform Reservoir; Jacobsen and Wasserburg, 1980) is variable because of heterogeneities in the age and Sm/Nd ratio of the continents themselves. Since neodymium is efficiently scavenged from the seawater solution, it has a residence time in the ocean that is relatively short (Tachikawa et al., 1999), and comparable to the timescale on which the oceans mix. Thus, Nd isotopes are increasingly used to track convective circulation, both in the modern and past oceans (e.g. Burton et al., 1997; Rutberg et al., 2000; Lacan and Jeandel, 2001; Thomas, 2004; Piotrowski et al., 2005; Foster et al., 2007; Klevenz et al., 2008; Scher and Martin, 2008; Roberts et al., 2010). Such an approach must rest on both good geographic coverage of Nd isotopic measurements in the modern ocean, as well as a sound understanding of the processes that determine the size and isotopic composition of the input and its redistribution by ocean circulation. With reference to the use of Nd isotopes as a water mass tracer in the ancient oceans, the most prominent successes have been obtained in applications that seek to investigate the Atlantic Meridional Overturning Circulation (AMOC), and its importance for oceanic heat transport on tectonic, glacial–interglacial and millennial timescales (e.g. Burton et al., 1997; Rutberg et al., 2000; Piotrowski et al., 2005; Roberts et al., 2010; Crocket et al., 2011). Early work (Piepgras and Wasserburg, 1980, 1987) on Nd isotopes in the ocean established that deep water exported from the North Atlantic in North Atlantic Deep Water (NADW) has an eNd of around 13. In the Southern Ocean, NADW mixes with water of Pacific origin, carrying an eNd around 2 to 4 (Piepgras and Wasserburg, 1980; Piepgras and Jacobsen, 1988; Vance et al., 2004; Amakawa et al., 2009), to produce the isotopic signature of the Antarctic Circumpolar Current (ACC) of around 8 (Piepgras and Wasserburg, 1982; Jeandel, 1993; Rickli et al., 2009). Nd extracted from the hydrogenous component of Cape Basin (South Atlantic) deep-sea sediments shows variations in eNd between 6 and 9 through the last glacial cycle, interpreted as changes in the strength of the unradiogenic NADW contribution to the ACC water that is exported to the deep Cape Basin. These variations have been used to quantify North Atlantic Deep Water strength, and therefore heat transport in the AMOC, through both glacial– interglacial and millennial climate change (e.g. Rutberg et al., 2000; Piotrowski et al., 2005). The success of this approach involves a number of requirements, aside from the basic one that a sedimentary carrier faithfully records deep water Nd isotopic composition. Firstly, the isotopic composition of NADW must be well known and must be stable through time. We now have relatively detailed coverage for the modern North Atlantic (e.g. Piepgras and Wasserburg, 1980, 1987; Lacan and Jeandel, 2004a,b,c, 2005a,b), while a couple of studies

suggest that the eNd of deep water exported from the deep Atlantic southwards has remained constant at least through the last climatic cycle (Van de Flierdt et al., 2006a; Foster et al., 2007). Secondly, it must be established that there are no major processes, other than the NADW contribution, that can modify the Nd isotopic composition of the ACC itself. In this respect, the comparative lack of data for the Southern Ocean in general, either for the modern-day or the past, is a significant concern. The published Southern Ocean dataset is restricted to one value for surface water and eleven for intermediate and deep water (Piepgras and Wasserburg, 1982; Rickli et al., 2009). Of these only two are from the Pacific sector, and for the Pacific in general the majority of the published data come from north of the Equator (Piepgras and Jacobsen, 1988; Vance et al., 2004; Amakawa et al., 2009). The Pacific sector is important, however, because it is water flowing out of the Pacific and into the Atlantic in the ACC that is modified by the addition of unradiogenic deep water exported from the North Atlantic. Thirdly, all studies that seek to use Nd isotopes as a tracer of ocean circulation involve the inherent assumption that Nd behaves quasi-conservatively in the oceans. Clearly, the prominent role of scavenging to, or isotope exchange with, particulate material in the marine geochemical cycle of Nd (e.g. Tachikawa et al., 1997, 1999) implies nonconservative behaviour at least in the surface ocean. But, importantly, recent studies have presented a particular challenge to the assumption of conservativity in the deep ocean, with a set of processes collectively called “Boundary Exchange” (BE) having been identified that clearly document interaction between deep ocean water and sediment (Lacan and Jeandel, 2001, 2004a,b,c, 2005a,b; Arsouze et al., 2007; Jeandel et al., 2007). These processes modify the Nd isotopic composition in a non-conservative fashion, and have been particularly important in understanding the controls on the Nd isotopic composition of end-member northern-sourced water in the Atlantic system (Lacan and Jeandel, 2004a,b,c, 2005a,b). The emerging consensus is that boundary exchange is restricted to ocean margins, and is important in determining the end-member Nd isotopic compositions of deep water, but is less significant once deep water is advected into the open ocean away from the marginal source regions. However, detailed studies that could identify the importance of such boundary exchange in different settings are still scarce, and the specific process by which boundary exchange operates is still poorly understood. In this paper, we present 57 new measurements of water column Nd isotopic composition and Nd concentration from the Pacific sector of the Southern Ocean. Significantly, these new measurements represent a 10% increase in the data for Nd isotopes in the global ocean (see compilation in Jones et al., 2008), and fill a significant gap in the geographic coverage. More importantly, with reference to the above discussion, the new data shed light on the processes that control Nd isotopes in the Southern Ocean. We show that Nd behaves very conservatively in the ACC, largely due to the inertia inherent in the huge flux of water (and Nd) in this water mass. We also show, however, that nonconservative ocean margin processes are as important for the Southern Ocean as they have proved to be in the North

Neodymium isotopes in the Southern Ocean

Atlantic and elsewhere (Lacan and Jeandel, 2001, 2004a,b,c, 2005a,b; Arsouze et al., 2007; Jeandel et al., 2007). 2. SAMPLES AND HYDROGRAPHY Samples for the measurement of Nd concentration and isotopic composition were obtained on cruise JR179 aboard RRS James Clark Ross (JCR) in the austral summer of 2008, from six full depth profiles in the Bellingshausen and Amundsen Seas in the eastern Pacific sector of the Southern Ocean (Fig. 1). One of these stations (022) is north of the Polar Front, and one (011) is from the continental shelf in Pine Island Bay (Fig. 1). Of the remaining four, three are within the ACC and one is from just south of its southern boundary (SBACC, Fig. 1). In addition, a series of eight samples were obtained on cruise JR141 with RRS JCR during austral summer 2005/ 6 at six locations and various water depths in the western Amundsen Sea Embayment. Finally, a further two samples were obtained on this cruise from a single station in the southern Bellingshausen Sea (station 029). Though these latter samples are not from a single CTD cast (Fig. 1c), they form a coherent dataset in terms of temperature (T), salinity (S) and Nd and will be treated here as a single depth profile, labelled “SHELF”, in some subsequent figures. The Southern Ocean is dominated by the Antarctic Circumpolar Current (ACC), the largest current in the world ocean in terms of total water transport (e.g. Orsi et al., 1995). Flowing clockwise around the Antarctic continent, the ACC has the unique characteristic of mixing waters from all three major ocean basins. The structure of the regional water masses is dictated by the major frontal systems of the ACC, which usually affect the whole water column down to the seafloor, controlling not only surface temperatures but also the depth of intermediate water masses. Across the ACC, Circumpolar Deep Water (CDW) is known to shoal dramatically towards the continental landmass, allowing deep waters from the north to enter the Southern Ocean and mix with surface waters (Orsi et al., 1995). 2.1. Surface water All open ocean stations studied here, with the exception of station 022 north of the Polar Front (PF), and all shelf stations, have surface waters corresponding to Antarctic Surface Water (AASW, Figs. 2 and 3). AASW has a temperature range from 1.8 °C to 1.0 °C and salinities between 33.0 and 33.7 (Smith et al., 1999). These waters are freshened and warmed by summer ice melting and heating, creating the highly spatially and temporally variable temperatures and salinities seen in Fig. 2. Winter Water (WW) sits below AASW and is identified by a temperature minimum at around 1.5 °C (Fig. 3), with salinities between 33.8 and 34.0. This is the end member for AASW (Sievers and Nowlin, 1984), i.e. WW is formed by winter cooling, and represents the portion of AASW that, even during the summer, retains the h–S signature from the previous winter (Smith et al., 1999). The h–S characteristics of

43

WW are eroded during the austral summer as heating of the upper water column causes instabilities and mixing with AASW and deeper water masses. All open ocean and shelf stations, again with the exception of station 022, have a subsurface temperature minimum related to WW, though by the time of sampling in the late austral summer, not always WW sensu stricto. In the open ocean, the temperature minimum is found at around 60 m water depth, and on the shelf at around 150 m. Surface waters at station 022 (Figs. 1 and 2) within the Antarctic Polar Frontal Zone (APFZ) are characterised by significantly higher temperatures than the waters south of the PF, the position of the PF being defined by the steep temperature gradient between the APFZ and the Antarctic Zone (AZ) to the south. The water mass is loosely defined as Sub-Antarctic Surface Water (SASW), and is the precursor to the pool of Antarctic Intermediate Water (AAIW) formed in the southeast Pacific (Georgi, 1979; Talley, 1996), which is formed primarily by mixing with upwelling CDW. 2.2. Deep and bottom waters Below WW lies Circumpolar Deep Water (CDW), the most voluminous water mass within the ACC (Sievers and Nowlin, 1984). CDW encompasses a wide range of temperatures and salinities, and is found in one or more of its variants at all stations in this study. North of the SBACC, CDW can be split into two varieties, Upper CDW (UCDW) and Lower CDW (LCDW). These variants differ in h –S space due to their different source regions. UCDW, between 200 and 400 m, is characterised by a subsurface temperature maximum, easily identified by its relative difference from the overlying minimum of WW. LCDW has a core between 800 and 1000 m and is characterised by a salinity maximum (Fig. 3) derived from the input of saline NADW, but in the eastern Pacific sector this characteristic is more eroded than in any other region of the Southern Ocean as this is the furthest point from the NADW input to the ACC (Whitworth and Nowlin, 1987). Fig. 2B shows UCDW and LCDW to be present at all oceanic sites north the SBACC. Below the surface waters at station 022 north of the PF, and overlying LCDW, is a warm, low salinity, water mass. The site sits at the southern edge of the region of AAIW formation, and is therefore influenced by the overlying SASW, upwelling CDW and the more southerly WW. The exact location of the Polar Front will determine how these water masses mix at a given time to form the water seen here in the vicinity of the front (Siedler et al., 2001). In the Pacific sector of the Southern Ocean, CDW is able to migrate far onto the West Antarctic shelf and is generally present there in a modified form, being slightly cooler and fresher due to mixing with overlying waters (e.g. Jacobs et al., 1996; Smith et al., 1999; Klinck et al., 2004; Jenkins and Jacobs, 2008). CDW is found at station 011 on the inner shelf in Pine Island Bay, and at shelf sites ASE1, ASE23 and BSS29 on the outer shelves of the Amundsen Sea Embayment and the southern Bellingshausen Sea (Fig. 2B and C). The latter stations are all located in bathymetric troughs that were eroded into the shelf by ice streams

44

P. Carter et al. / Geochimica et Cosmochimica Acta 79 (2012) 41–59

Fig. 1. Maps of seawater sampling sites (black circles) in the eastern Pacific sector of the Southern Ocean. Top Map: Antarctica, featuring major ocean currents and fronts. Middle Map: eastern sector of the Pacific Southern Ocean – all sites sampled are south of the Sub-Antarctic Front (SAF) but station 022 is north of the Polar Front (PF). The Southern Boundary of the Antarctic Circumpolar Current (SBACC) is located close to the coast in this section of the Southern Ocean. Bottom Map: the Amundsen Sea Embayment with CTD sites on the shelf.

Neodymium isotopes in the Southern Ocean

45

Fig. 2. A, B: Salinity versus potential temperature for all open ocean stations and shelf station 011. Circles represent the CTD parameters of water samples collected for Nd analysis. Water mass identification of Antarctic Surface Water (AASW), Winter Water (WW), Circumpolar Deep Water (CDW), split into upper and lower (UCDW and LCDW), and South Pacific Deep Water (SPDW) is based on the water mass definitions of Smith et al. (1999). C: Salinity versus potential temperature (h) for shelf samples from the Amundsen Sea Embayment and the southern Belingshausen Sea (SHELF) and station 011 from Pine Island Bay.

during past glacial periods (Nitsche et al., 2007; Graham et al., 2010; Hillenbrand et al., 2010) and that facilitate the southward protrusion of CDW (Klinck et al., 2004; Walker et al., 2007; Thoma et al., 2008; Holland et al., 2010). CDW at the stations studied here is formed by a very limited amount of mixing with the overlying cold surface

waters, and has temperatures reaching 1.4 °C. At the inner-shelf sites ASE2, ASE7, ASE8 and ASE15, the bottom waters are a highly modified version of CDW, with temperatures reaching only 0.5 °C, even though these stations are also located in glacial troughs connected to the open ocean (Larter et al., 2009).

46

P. Carter et al. / Geochimica et Cosmochimica Acta 79 (2012) 41–59

Fig. 3. Temperature and salinity sections across the Pacific sector of the Southern Ocean (section line given in map). The temperature and salinity values in the trough at around 60°S are probably an artefact of the plotting procedure. Water-mass features discussed in the text are indicated, including the Polar Front (PF), Antarctic Surface Water (AASW), Sub-Antarctic Surface Water (SASW), Winter Water (WW), Circumpolar Deep Water (CDW) and South Pacific Deep Water (SPDW). Sections produced using Ocean Data View (Schlitzer, 2011).

The bottom waters of the eastern Pacific sector of the Southern Ocean are dominated by South Pacific Deep Water (SPDW) and modified versions of LCDW, Ross Sea Bottom Water (RSBW) and Weddell Sea Deep Water (WSDW). Modified RSBW may form the bottom-water mass on the upper continental rise in the Amundsen Sea (Hellmer et al., 1998), while the bottom-water regime on the upper continental rise in the Bellingshausen Sea is characterised by a generally south-westward-flowing countercurrent formed by SPDW, modified LCDW or modified WSDW derived from the Weddell Sea (Camerlenghi et al., 1997; Smith et al., 1999; Giorgetti et al., 2003; Herna´ndez-Molina et al., 2006; Hillenbrand et al., 2008), with modified WSDW possibly extending no more than 60 m above the seafloor (Camerlenghi et al., 1997; Giorgetti et al., 2003). Our station 008 is located close to the crest of a sediment drift, which was the target for a comprehensive oceanographic study by Giorgetti et al. (2003). The location of site 008 was chosen deliberately to try and identify the Nd isotopic characteristics of modified WSDW. The h–S plot of station 008 (Fig. 2A and B) shows temperature and salinity values for the bottom water that are comparable to those previously reported by Camerlenghi et al. (1997) and Giorgetti et al. (2003), but also to those of the bottom waters of all the deeper oceanic stations, including station 022. The bottom-water mass that is present at the deep-water sites is characterised by salinities between 34.70 and 34.71 pointing to SPDW, which originates in the deep southeast Pacific Basin and flows eastward to the Drake Passage (Sievers and Nowlin, 1984), and/or modified LCDW. We infer that SPDW and modified LCDW form the bottom waters on the middle to lower continental

rise and the abyssal plain of our study area, while modified RSBW and modified WSDW may affect the upper continental rise in the Amundsen Sea and the Bellingshausen Sea, respectively. As expected, no bottom-water masses are found at the shelf stations. 3. METHODS FOR ISOTOPIC ANALYSIS Water-column samples were collected in 12 L Niskin bottles attached to a CTD rosette and were filtered (0.2 lm Whatman Nucleopore) on board into pre-cleaned (2 weeks in 20% Analar HCl followed by three rinses with 1018 X water) 4 L Nalgene LDPE bottles. All samples were acidified to pH <2 using distilled HCl on return to the shorebased laboratory. REE were extracted from seawater samples by iron oxide co-precipitation, following techniques described in Vance et al. (2004). Briefly, bottles containing samples were weighed and spiked with a mixed 149 Sm–150Nd tracer. A solution containing Fe (Alfa Aesar Puratronic 22 mesh 99.998% iron powder in HCl) was added such that the Fe added amounted to 8 mg per litre of water. Samples were shaken by hand multiple times through a 24 h period to fully mix the spike and dissolved iron with the seawater. The pH of the samples was raised to 8 using distilled NH4OH to precipitate the iron and co-precipitate the REE. Samples were shaken vigorously and allowed to settle over a 2-day period, after which 3 L of water was slowly poured away without any loss of precipitate. Samples were then transferred to successively smaller Teflon jars to facilitate more controlled pouring off of supernatant water and retention of the precipitate, before being centrifuged and the remaining water discarded. The precipitate was re-dissolved

Neodymium isotopes in the Southern Ocean

in distilled 6 M HCl, transferred to a Teflon vial, and dried down. Samples were dissolved again in 6 M HCl, dried down and finally dissolved in 1.5 ml of 1 M HCl ready for loading onto cation exchange columns. REE were separated from major elements by chromotographic exchange using cation exchange resin. Nd and Sm were separated using Ln spec resin (Stoll et al., 2007). Nd and Sm (for Sm concentrations via isotope dilution) isotope compositions were measured on a Thermo-Finnigan Neptune Multi Collector ICPMS at the University of Bristol, following procedures described previously (Vance and Thirlwall, 2002; Stoll et al., 2007). 143Nd/144Nd ratios were normalised to 146Nd/144Nd = 0.7219 to correct for instrument-induced mass discrimination. A secondary correction was applied using correlations between 146Nd/144Nd-normalised 142Nd/144Nd and 143Nd/144Nd (Vance and Thirlwall, 2002) and normalisation to 142Nd/144Nd = 1.141876. Repeated (n = 10–15) measurements of a 50 ppb La Jolla Nd standard (17 ng total Nd measured) within one session gave between 0.511849 ± 0.000008 and 0.511858 ± 0.000005 (2 sigma). For two sessions, 10 ppb standards were analysed (n = 10 for both, 3 ng Nd measured per analysis) and gave 0.511856 ± 23 and 0.511868 ± 27. Sample sizes analysed here varied between 6 and 20 ng. Uncertainties listed in the tables are 2SE of the fifty 4 s integrations in each analysis (i.e. internal errors) and are generally intermediate between the reproducibilities of the high and low concentration standards. The precision of Nd and Sm concentrations was approximately 5%. 147Sm/144Nd measurements were precise to 0.5% (Vance et al., 2004). The total procedural blank for these seawater analyses is dominated by the Fe solution used for co-precipitation. This solution has been well characterised through multiple analyses of large quantities, and the amount of Fe added to each sample corresponds to 275 ± 2 pg Nd with an eNd = 12.2 ± 0.3. The maximum amount of Nd added from the Fe solution is thus 5% of the total Nd analysed. A correction was applied to the data presented here, amounting to a maximum correction of 0.2 epsilon units for the smallest sample measured. The uncertainty on these corrections is negligible given the uncertainty in the amount of Nd in the Fe solution and its isotope composition. A small number of sediment analyses are also presented here (Table 2) in order to facilitate the discussion of the possibility of modification of bottom water on the shelves by interaction with sediment. The samples are from box core BC437B (collected at station 437, 71.600°S, 113.301°W, 619 m – a location identical to water column station ASE1, see Fig. 1). The Fe–Mn oxyhydroxide fraction was extracted, following the removal of carbonate, by leaching with hydroxylamine hydrochloride, using the approach of Gutjahr et al. (2007). The residual detrital fraction was obtained by digestion in hydrofluoric acid, following the removal of any residual Fe–Mn oxides using a second long leach (see Gutjahr et al., 2007). 4. RESULTS Sm–Nd isotopic results, as well as basic hydrographic data for the water-column samples collected for Nd isotopic

47

analysis, are presented in Table 1 and illustrated as depth profiles in Figs. 4 and 5. Near surface samples, at depths ranging from 5 to 32 m, were taken at all stations. With the exception of station 022, north of the PF, these waters have been identified as AASW. In the four samples analysed from the open ocean, AASW eNd values show relatively small but significant variations, from 8.28 ± 0.12 at station 021 to 7.03 ± 0.17 at station 024, whilst concentrations vary little, from 13.2 to 14.4 pmol/kg (Fig. 4). In the two samples studied from the continental shelf, AASW shows substantial variation in Nd isotopic composition, from 7.25 ± 0.22 at station 011, which is similar to open ocean values, to a more radiogenic eNd value of 5.62 ± 0.15 at the Amundsen Sea Embayment station ASE7. Shelf concentrations are also variable, though both are substantially higher than open ocean values (Fig. 4): 21.5 pmol/kg at station 011, and 30.5 at ASE7. No sensu stricto WW samples were collected, though some are close in h–S space to WW (Fig. 2). These samples are indistinguishable in terms of both Nd isotopic composition and Nd concentration from the overlying AASW samples at the same sites, due to the role of WW in the formation of AASW. Therefore, for the purpose of the Nd data in this study, both AASW and WW are considered together, simply as surface waters. In the open ocean the least radiogenic values at all stations, with the exception of station 022, are found at intermediate depths, corresponding to the core of CDW (Fig. 4). South of the Polar Front seven samples have been taken from four open ocean stations, between 400 and 1250 m (Table 1), corresponding to UCDW, LCDW, or a mixture of the two (Fig. 2B). Only one sample is from UCDW, as identified by the h–S characteristics, two are from a mixture of upper and lower CDW and four from LCDW. All are within two standard errors of each other, showing a remarkable consistency in the isotopic composition of CDW. Based on this small sample set, there is no difference between the isotopic composition of UCDW and LCDW, with values measured for both ranging from 8.59 ± 0.19 to 8.72 ± 0.18. Nd concentrations are higher than surface values at these sites, and increase with depth from 400 to 1250 m, with values between 13.8 and 18.7 pmol/kg for CDW. North of the Polar Front at station 022, CDW is found significantly deeper, with the core of LCDW located at 2000 m (Fig. 2). The 2003 m sample is located within LCDW, and the 1504 m sample on a mixing curve between UCDW and LCDW. The corresponding eNd values are barely distinguishable, at 8.44 ± 0.16 and 8.82 ± 0.21 respectively, and also within error of the isotopic composition of CDW south of the PF. Concentrations increase with depth, with the 2003 m sample having a concentration of 21.2 pmol/kg. On the shelf, six samples can be classified as modified CDW (ASE 23, 430 m; ASE 1, 450 m; BSS29, 350 m and 600 m; station 011, 602 m and 1301 m). As with the open ocean sites, CDW is the least radiogenic water-mass on the shelf and, despite showing more variation in isotopic composition than in the open ocean, is still relatively homogeneous, with values ranging from 7.37 ± 0.13 to 8.21 ± 0.24. CDW has the lowest Nd concentration of

48

P. Carter et al. / Geochimica et Cosmochimica Acta 79 (2012) 41–59

Table 1 Hydrographic and Sm–Nd isotopic and concentration data for the Pacific sector of the Southern Ocean, including the shelves of the Amundsen Sea Embayment (ASE) and the southern Bellingshausen Sea (BSS). Station name

Hydrographic data

Isotopic composition Nd/144Nd

rh

143

0.97 0.89 1.82 1.32 0.63 0.36 0.24 0.24

26.8567 27.4689 27.7584 27.8093 27.8355 27.8473 27.8532 27.8532

0.512253 0.512242 0.512190 0.512187 0.512220 0.512245 0.512259 0.512247

33.21 34.73 34.71 34.70

0.15 1.42 0.47 0.19

26.6714 27.7993 27.8429 27.8562

32 602 1301

33.56 34.64 34.72

1.36 1.12 1.19

Station 021 69o13S, 106o40W 4218 m

18 508 1006 1509 2008 2509 3009 3510 4107 4208

33.35 34.71 34.73 34.72 34.71 34.71 34.70 34.70 34.70 34.70

Station 022 60o32S, 108o18W 5190 m

16 503 1003 1504 2003 2503 3004 3504 4006 4507 5082 5182 15 504 1005 1505 2004 2505 3005 3508 3914 4012

Concentration (pmol/kg) Sm/144Nd

2rm

147

Nd

Sm

7.37 7.58 8.64 8.72 8.07 7.60 7.32 7.54

0.19 0.20 0.29 0.18 0.18 0.14 0.16 0.18

0.0980 0.0983 0.0990 0.0992 0.1010 0.0888 0.1064 0.1023

13.68 13.54 13.77 18.06 22.42 26.17 26.81 24.17

2.29 2.27 2.33 3.04 3.82 3.91 4.79 4.17

0.512216 0.512188 0.512217 0.512253

8.10 8.70 8.15 7.43

0.17 0.24 0.15 0.14

0.1029 0.1038 0.1055 0.1072

13.16 17.64 24.70 26.97

2.32 3.11 4.39 4.86

26.9989 27.7482 27.8038

0.512261 0.512233 0.512222

7.25 7.80 8.04

0.22 0.18 0.20

0.1082 0.1068 0.1072

21.50 21.02 20.74

3.84 3.74 3.89

0.45 1.86 1.45 1.07 0.76 0.51 0.29 0.17 0.04 0.03

26.7930 27.7468 27.7986 27.8187 27.8307 27.8412 27.8514 27.8574 27.8627 27.8627

0.512208 0.512188 0.512192 0.512200 0.512210 0.512216 0.512231 0.512242 0.512243 0.512235

8.28 8.68 8.61 8.47 8.27 8.15 7.87 7.67 7.64 7.74

0.12 0.19 0.14 0.15 0.28 0.14 0.14 0.16 0.18 0.25

33.91 34.33 34.61 34.71 34.73 34.72 34.71 34.71 34.70 34.70 34.70 34.70

3.76 2.44 2.24 1.91 1.48 1.10 0.80 0.53 0.27 0.10 0.04 0.04

26.9399 27.3987 27.6377 27.7495 27.7952 27.8162 27.8296 27.8396 27.8516 27.8588 27.8617 27.8615

0.512235 0.512233 0.512218 0.512181 0.512202 0.512201 0.512206 0.512216 0.512233 0.512246 0.512258 0.512195

7.66 7.76 8.07 8.82 8.44 8.45 8.35 8.16 7.83 7.58 7.33 8.57

0.29 0.24 0.20 0.21 0.16 0.27 0.15 0.13 0.17 0.17 0.14 0.17

33.58 34.69 34.73 34.73 34.72 34.71 34.70 34.70 34.70 34.70

0.39 1.94 1.56 1.17 0.86 0.59 0.34 0.18 0.09 0.08

26.9367 27.7264 27.7896 27.8139 27.8268 27.8374 27.8492 27.8568 27.8607 27.8610

0.512270 0.512192 0.512192 0.512218 0.512208 0.512224 0.512229 0.512256 0.512266 0.512272

7.03 8.59 8.62 8.11 8.31 8.10 7.89 7.38 7.19 7.07

0.17 0.18 0.19 0.13 0.19 0.23 0.19 0.15 0.17 0.12

33.98 34.71

1.84 1.22

0.512367 0.512221

5.06 7.93

34.04 34.57

1.44 0.47

0.512343 0.512274

5.54 6.91

Depth (m)

Salinity

10 79 505 1003 2005 2505 3145 3245

33.52 34.16 34.72 34.73 34.71 34.70 34.70 34.70

Station 010 70o08S, 91o05W 3542 m

10 1003 2506 3528

Station 011 74o21S, 104o43W 1401 m

h (°C)

eNd

Location Bottom depth (m) Full depth profiles Station 008 68o17S, 76o08W 3261 m

Station 024 67o46S, 83o15W 4020 m

Non-depth profile shelf samples ASE1 100 71o36S, 113o18W 450 619 m ASE2 73o23S, 115o09W 993 m

180 900

13.79 15.13 17.78 21.65 22.99 23.62 26.26 29.39 27.26 15.48 9.66 12.16 14.10 15.75 21.23 22.96 22.90 24.22 24.60 26.32 27.87 22.73

4.31 4.08

0.1092 0.1089 0.1097 0.1156 0.1044 0.1148 0.1047 0.1067 0.1102 0.1103

14.42 15.43 18.74 22.62 20.66 20.86 23.99 26.28 28.54 30.07

2.68 2.86 3.48 4.41 3.64 4.05 4.23 4.72 5.28 5.56

0.14 0.18

0.1080 0.1035

31.52 21.97

5.63 3.76

0.86 0.14

0.1050 0.1058

27.83 26.54

4.83 4.65

0.1058 0.1051 0.1115 0.1056 0.1052 0.1058

2.20

4.36 4.68

Neodymium isotopes in the Southern Ocean

49

Table 1 (continued) Station name

Hydrographic data

Isotopic composition

Depth (m)

143

Salinity

h (°C)

144

Nd/

rh

Nd

eNd

Concentration (pmol/kg) 2rm

147

144

Sm/

Nd

Nd

Sm

Location Bottom depth (m) ASE7 73o59S, 116o05W 1019 m

5

33.63

0.56

0.512339

5.62

0.15

0.1093

30.51

5.52

ASE8 74o02S, 115o56W 1226 m

1003

34.54

0.26

0.512355

5.27

0.17

0.1051

28.36

4.93

ASE15 74o14S, 112o32W 1200 m

1190

34.55

0.37

0.512363

4.53

0.23

0.1052

38.56

6.71

ASE23 71o38S, 113o33W 611 m

430

34.71

1.18

0.512260

7.37

0.13

0.1116

24.82

4.58

BSS029 70o37S, 86o15W 691 m

350 600

34.70 34.73

1.44 1.20

0.512235 0.512217

7.72 8.21

0.16 0.24

0.1125 0.1105

25.35 25.84

4.72 4.72

Table 2 Sm–Nd isotopic and concentration data for sediments from the Amundsen Sea Embayment. Depth (cmbsf)

143

Fe–Mn leaches BC437B-1 BC437B-2 BC437B-4 BC437B-7 BC437B-9 BC437B-11 BC437B-13 BC437B-15

1 2 4 7 9 11 13 15

0.512581 0.512585 0.512574 0.512589 0.512589 0.512594 0.512597 0.512590

Detrital fraction BC437B-1 BC437B-4 BC437B-7 BC437B-10 BC437B-13 BC437B-16

1 4 7 10 13 16

0.512503 0.512516 0.512501 0.512500 0.512528 0.512524

Nd/

144

Nd

any water mass on the shelf (21.0–25.8 pmol/kg), being significantly lower than both the surface waters and bottom waters of the Amundsen Sea Embayment stations, though slightly higher than open ocean CDW. At all open ocean sites the Nd isotopic composition becomes more radiogenic below CDW, reflecting mixing between un-radiogenic CDW and more radiogenic water below. The eNd values of the deep waters show little variation between sites south of the PF, with an isotopic composition ranging from 7.07 ± 0.12 to 7.74 ± 0.25. A similar value of 7.33 ± 0.14 is found at 5082 m water depth at station 022, i.e. in SPDW north of the PF. A bottom water sample taken at 5182 m water depth at station 022 has an isotopic composition of 8.57 ± 0.17, similar to a bottom water sample taken north of the PF in the Bellingshausen Sea by Piepgras and Wasserburg (1982), with the latter having an isotopic composition of 8.2 ± 0.6.

2r

147

1.10 1.02 1.26 0.96 0.95 0.85 0.81 0.94

0.08 0.10 0.08 0.07 0.07 0.09 0.06 0.07

0.1210 0.1204 0.1141 0.1179 0.1179 0.1199 0.1183 0.1168

4.2 4.4 5.0 4.4 4.2 3.5 5.5 4.1

2.66 2.52 2.68 2.65 2.33 2.32

0.06 0.09 0.06 0.06 0.06 0.05

0.1164 0.1163 0.1169 0.1172 0.1172 0.1170

25.0 28.8 26.1 26.4 27.9 28.7

eNd

Sm/144Nd

Nd (ppm)

Nd concentrations generally increase with depth at all open ocean sites, ranging from 22.9 to 26.8 pmol/kg at around 3000 m, and 24.6 to 28.5 pmol/kg at around 4000 m. The highest concentration at all sites is the bottom water sample of station 024 with 30.1 pmol/kg. Below 4000 m, station 022 reaches a maximum of 27.9 pmol/kg at 5082 m, though this station has the lowest concentration of all open ocean sites at a given depth below 3000 m. Nd concentrations do, however, decrease in the bottom water samples at three of the four stations, where bottom water was sampled (7–16 m above the seafloor). When compared with the samples taken 100 m above the seafloor, the largest shift in concentration is observed at station 021, where concentrations decrease from 27.3 to 15.5 pmol/kg, whilst both station 008 and 022 show smaller, but significant, shifts. Only at station 024 does the increase of Nd concentration with depth continue to the bottom sample. Whilst

50

P. Carter et al. / Geochimica et Cosmochimica Acta 79 (2012) 41–59

Fig. 4. Nd isotopic composition and concentration for all open ocean and shelf water samples collected in the eastern Pacific sector of the Southern Ocean. All profiles, with the exception of those labelled SHELF, are from a single geographic location. The SHELF profile represents an amalgamation of sites from the shelf of the western Amundsen Sea Embayment and southern Belingshausen Sea.

sites 008, 021 and 022 also show an accompanying shift to less radiogenic values in the bottom samples, only the shift from -7.33 ± 0.14 at 5082 m water depth to 8.57 ± 0.17 at 5182 m water depth at station 022 is beyond analytical uncertainty. At station 008, similar to the h–S data, neither Nd isotopic composition nor Nd concentration give conclusive evidence for modified WSDW forming the bottomwater mass on the continental rise west of the Antarctic Peninsula, but the eNd values are consistently slightly more radiogenic below ca. 1500 m water depth when compared to the eNd values from similar water depths at the other openocean stations located to the south of the PF (Fig. 4a). Data for Amundsen Sea Embayment sediments are presented in Table 2. Bulk sediment neodymium isotopes (eNd = 2.5 ± 0.2) at this site are more radiogenic than the water above (site ASE1), by about 2.5 epsilon units. Leaches of these sediments designed to extract the Fe–Mn oxyhydroxide fraction are more radiogenic still, at eNd = 0.95 ± 0.15. The isotopic mismatch between the leached Nd and the water column has been reported previously for marginal settings (e.g. Bayon et al., 2004), and attributed to the input of continent-derived particulates with extra-oceanic Fe–Mn oxyhydroxide coatings. 5. DISCUSSION The most prominent feature of the dataset presented above is the significant difference between the Nd isotopic composition of shelf waters in the Amundsen and Bellings-

hausen Seas, particularly for surface and bottom waters, relative to those for the open ocean. We will discuss this feature first before moving onto consider the more subtle variability in open-ocean surface waters, the remarkable homogeneity in CDW, and the processes controlling the Nd isotopic composition of bottom water at the openocean sites. 5.1. Shelf Nd: upwelling of CDW and local continental inputs The depth profiles in Figs. 4 and 5 clearly show similarities for both Nd concentration and isotopic composition between the shelf sites at intermediate depths and openocean Circumpolar Deep Water. Above and below intermediate depths, however, there are pronounced deviations in eNd in the shelf sites that are accompanied by increases in dissolved Nd concentrations. This suggests that the key processes that control dissolved Nd on the shelf are: (a) incursion/upwelling of CDW onto the shelf; (b) addition of radiogenic Nd from a local source in both surface and bottom waters. CDW is known to upwell south of the PF before migrating towards the continent and onto the West Antarctic shelf, mainly promoted by the location of the Amundsen Sea low pressure system (Thoma et al., 2008). The cross shelf troughs, carved over successive glacial cycles (Nitsche et al., 2007; Larter et al., 2009; Graham et al., 2009), allow significant encroachment of CDW far onto the shelf (Jenkins et al., 1997; Walker et al., 2007; Thoma et al.,

Neodymium isotopes in the Southern Ocean

51

Fig. 5. Representative depth profiles for selected stations 022 (open ocean north of PF), 024 (open ocean south of PF) and a composite of the shelf stations showing salinity, potential temperature, Nd isotopic composition and Nd concentration. The SHELF panel represents an amalgamation of locations on the shelves of the western Amundsen Sea Embayment and southern Belingshausen Sea. Hence, salinity and potential temperature are not plotted as a continuous profile but as a series of discrete points.

52

P. Carter et al. / Geochimica et Cosmochimica Acta 79 (2012) 41–59

2008). Fig. 4 shows that the Nd isotopic characteristics at depths of ca. 600–700 m on the shelf are identical to those in the open ocean at all depths above about 3000 m. Nd concentrations (Fig. 4) at 600–700 m on the shelf are comparable to those at around 2000–3000 m depth in the open ocean. All of this implies that Nd in the intermediate waters of the shelf could derive from upwelling of open-ocean CDW from 2500 m, without any further modification of its Nd isotopic systematics. Clearly, however, the Nd isotopic characteristics of the shelf stations are significantly modified with respect to the open ocean values, both above and below these intermediate depths. Marginal sources of Nd have been shown to contribute significantly to the dissolved Nd inventory of seawater (see Jeandel et al., 2007 for a discussion), and to alter seawater Nd isotopic compositions by both addition and removal of Nd. The process has been shown to occur in multiple locations along both basaltic and granitic margins (Lacan and Jeandel, 2001, 2004a,b,c, 2005a,b) and is thought to occur in dynamic regions where sediment/particulate loads, and thus the potential for sediments or particulates to act as both sources and sinks of Nd, are high. In order to investigate the potential importance of such a source, we measured the Nd isotopic characteristics of bottom sediment obtained from the outer shelf in the Amundsen Sea Embayment (Table 2). The sedimentary data are compared to those for the water column in Fig. 6. It is clear from Fig. 6a that much of the shelf water column data could be explained by mixing between Nd that is similar to the core of CDW (i.e. open water stations) with Nd similar to that seen in the sediments. A relevant and interesting question here is the extent to which the modification of the Nd isotope signature in both surface and deep waters on the shelf is achieved by net addition of Nd with a different isotope signature, versus exchange of Nd isotopes between particulates and the dissolved phase. In previous work concerned

with modification of water column signatures by interaction with sediments (Lacan and Jeandel, 2001, 2004a,b,c, 2005a,b; Arsouze et al., 2007; Jeandel et al., 2007), the emphasis has been placed on exchange as the dominant mechanism, since the modification of the Nd isotope signature is not generally accompanied by significant increases in concentration. If the interpretation of simple mixing between CDW and a radiogenic sedimentary source is correct, then the modification of the Nd isotope signature in this particular case must be achieved by the net addition of Nd (Fig. 6a). Though the array for the water column data is far from tight, and thus a role for exchange of Nd isotopes between particulates and the dissolved load is certainly permitted, the dominant process must be the net addition of Nd with a radiogenic isotope signature. In this simple view, the water column data in Fig. 5a suggests such modification can result in an approximate tripling of Nd concentrations relative to CDW. The plot of salinity versus eNd in Fig. 6b separates the surface- and deep-water data for the shelf in a way that is not achieved in Fig. 6a. The dataset is small, and it is plausible that the arrays seen in Fig. 6b (dashed lines) have no significance beyond co-incidence, but if they do they suggest that the shelf surface waters represent a modification of open-ocean surface waters towards higher salinities and more radiogenic Nd, while bottom waters look more likely to be a modification of CDW water towards slightly lower salinities at more radiogenic Nd. We do not seek to suggest that the Nd isotope additions at the surface and bottom of the shelf stations occur in even a remotely conservative fashion, but that they are perhaps indicative of the overall process. It is, however, remarkable that for the middle of the water column, the addition of radiogenic Nd appears to be pseudo-conservative with salinity. We note that the intersection of the array containing the bottom water data on Fig. 6b intersects the sediment eNd band at salinities around 34.5, close to those measured

Fig. 6. Nd isotopic data for shelf water column samples versus reciprocal Nd concentration (a) and salinity (b), compared to Nd isotopic data for Amundsen Sea Embayment sediments (green band, the range encompassing both data for both detrital material and leached Fe–Mn oxides). The shelf data are plotted as red squares, with the open symbols used to distinguish station 011, from which a full profile is available. Open-ocean CDW data are indicated as orange circles. In Fig. 5b, the black triangles represent data for open ocean surface water south of the PF. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Neodymium isotopes in the Southern Ocean

for bottom water on the shelf (Table 1, Fig. 5). The most likely explanation for the Nd isotope variation with salinity for the intermediate water shelf samples is the addition of Nd to the bottom water samples through interaction with seabed sediments, and then the mixing upwards of this signal into waters that are influenced by the upwelling/incursion of CDW onto the shelf. Indeed, it is possible, given the fact that the two arrays on Fig. 6b appear to intersect at a salinity close to 34.5 (again, close to shelf bottom water: Table 1, Fig. 5) and within the band representing the isotopic composition of Amundsen Sea Embayment sediments, that this water is also the cause of the modification of surface water, perhaps by lateral spreading away from the continental margin. 5.2. Open ocean surface water: modification across the ASE embayment There is small but significant variation in both Nd concentration (9.7–14.4 pmol kg 1) and Nd isotopic composition (eNd = 8.3 to 7.0) in open ocean surface waters of the study area. These variations are explored in the context of both surface waters from the shelves and CDW in Fig. 7. On this diagram, it is apparent that surface water Nd could, at least, be controlled by mixing between three different end-members. As noted in Section 5.1, the shelf data extend towards a component that is both highly enriched in Nd and radiogenic. The Nd data for the westernmost open ocean sites south of the PF (stations 010 and 021) could be explained by mixing of SASW from north of the PF (station 022) with CDW (Fig. 7), consistent with the physical oceanography. However, the easternmost open ocean sites (stations 008 and 024) lie within the mixing triangle, and significantly off the SASW-CDW mixing line (Fig. 7).

Fig. 7. Nd isotopic data for all open ocean samples with potential densities <27, as well as data for shelf surface water, for which potential density data are not available. Symbols are as in Fig. 5, with the addition of the surface water point for station 022, north of the Polar Front (blue triangle). Of the samples from the open ocean south of the Polar Front (black diamonds), the two easternmost samples from the Bellingshausen Sea (stations 008 and 024) lie within the triangle. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

53

This small but significant shift off a potential mixing trend for the eastern open ocean sites 008 and 024 is related to a small change in eNd of surface waters by P1 epsilon unit towards more radiogenic values in an eastward direction. Due to the isolating effect of the PF, it is probable that any surface input of Nd comes from the shelves in the Amundsen Sea and the Bellingshausen Sea, respectively. But there are actually a number of other potential Nd sources, which we will assess first. For example, the bulk and <63 lm fractions of core-top sediments from the West Antarctic continental margin, to the South of the ACC, reveal significantly more radiogenic values close to the northern tip of the Antarctic Peninsula (shelf: eNd = +1.5, Walter et al., 2000; upper continental rise: eNd = +0.9, Roy et al., 2007). However, given both the westward decrease of eNd values in sediments on the western Antarctic Peninsula shelf (Walter et al., 2000) and the strength of the eastward ACC flow in the surface waters north of the SBACC, any significant westward transport of Nd seems unlikely. Another possible source for a radiogenic input from outside the water column is via dust, for example from South America, where loess has eNd values ranging up to 0.8 (Smith et al., 2003). However, given that the dominant winds are westerly in the region, it is doubtful, whether such dust would reach our study area. Moreover, a simple mass balance calculation can rule dust out. Physical oceanographic studies have shown that shear in the ACC is limited to the bottom few hundred metres of the current (http:// oceancurrents.rsmas.miami.edu/southern/antarctic-cp.html) so that an assumption of constant water flow with depth is a reasonable approximation for the present purpose. The mean total flow in the ACC is 100–150 Sv and the current is ca. 2000–4000 m thick (http://oceancurrents.rsmas.miami. edu/southern/antarctic-cp.html). Thus the flow, even in only the top 50 m, would be around 1–4 Sv. This flow (given the concentration of Nd at the westernmost station 021) carries 0.6–2.5  108 g Nd yr 1. If any potential dust input had an eNd as high as 0, then shifting the eNd from 8.3 (station 021) to 7 (station 024) along the surface ACC flow trajectory would require the addition of 1.6– 6.3  107 g Nd yr 1. At a Nd concentration of 30 ppm, this requires 0.5–2  1012 g dust yr 1 in the unlikely event that all of it is dissolved and added to the top 50 m of the water column. Modern aeolian fluxes to the Southern Ocean are 1–10  104 g km 2 yr 1 (Duce and Tindale, 1991). The distance between sites 021 and 024 is of the order of 1000 km (Fig. 1), and the maximum width of the ACC is less than 2000 km (Smith et al., 2008; this value gives the maximum ACC width in the Atlantic sector, in the Pacific sector it is narrower), implying a maximum annual total dust flux to the ACC across the study region of 0.2–2  1011 g yr 1. Thus, the maximum estimate of the actual dust flux falls short of the absolute minimum required to modify the eNd as observed by an order of magnitude. Clearly, the addition of dust-derived Nd to the water column can in no way explain the change in the Nd isotopic composition of the surface ACC waters in our study area. This conclusion is consistent with those of other authors that dust flux cannot explain magnetic susceptibility data in sediments from the Scotia Sea (Walter et al., 2000; Pugh et al., 2009), or clay

54

P. Carter et al. / Geochimica et Cosmochimica Acta 79 (2012) 41–59

mineral assemblages in sediments from the western Antarctic Peninsula rise (Hillenbrand et al., 2008). Given all the above, it is likely that the eastward modification of Nd in the open surface ocean results from interaction with the shelves in the Amundsen and Bellingshausen Seas, respectively. This finding is in congruence with conclusions elsewhere (Lacan and Jeandel, 2001, 2004a,b,c, 2005a,b; Arsouze et al., 2007; Jeandel et al., 2007), that processes at ocean margins are important in modifying oceanic Nd concentrations and isotopic compositions. Though the data in Fig. 7 are consistent with the suggestion that the eastward change in open ocean surface waters arises through conservative mixing of upwelling CDW, surface ocean water from north of the PF and a shelf end-member, it is to be remembered (Section 5.1) that the ultimate control on the shelf end-member itself is non-conservative interaction with sediments. 5.3. Open ocean intermediate water: conservative behaviour in the middle water column In contrast to the shelf data in our study area, all open ocean sites show increasing Nd concentrations with water depth. This trend is observed at many other locations and conventionally attributed to Nd removal in the surface ocean through particulate scavenging and its release at depth. In the study area, Nd concentrations at almost all water depths are higher south of the PF than north of it, presumably reflecting the increased Nd input from the Antarctic continent to waters proximal to it. The increased gradient of Nd concentration with depth, seen by Piepgras and Wasserburg (1982) south of the PF relative to a station to the north, is not observed here, although the greater sampling resolution of our study may account for this. CDW at intermediate water depths north and south of the PF has similar eNd, though CDW lies deeper in the water column at station 022, as CDW generally upwells towards the Antarctic continent. Contrary to the large scale, circum-Antarctic eNd variability of CDW, noted by Van de Flierdt et al. (2006b) from the Fe–Mn crust data of Albarede et al. (1997), CDW has a remarkably consistent eNd in the eastern Pacific sector of the Southern Ocean. The average eNd of CDW of around 8.7 is, however, different from the only other direct seawater measurements of CDW taken south of the PF in the Atlantic sector of the Southern Ocean ( 6.3 to 8.1 for variants of Circumpolar Water (CPW): Jeandel, 1993). In order for Nd to be a useful water mass tracer it needs to be shown that over a significant distance a definable water mass retains its Nd isotope composition in the same way as it conserves its salinity and temperature characteristics. In the Atlantic sector, this property is easily demonstrated (Piepgras and Wasserburg, 1987; Jeandel, 1993), as the vertical structure of the water column is well defined, with NADW separating two southern sourced water masses with significantly different Nd isotopic compositions. In the eastern Pacific sector of the Southern Ocean, the relatively small range in Nd isotope composition found in all water masses makes assessment of the conservation of Nd more difficult.

Within the limitations outlined above, however, the data collected for this study do point to strong conservation of Nd in the Southern Ocean, with water mass mixing accounting for almost all the variability in eNd. This is well demonstrated in Fig. 8, where a strong correlation is observed between eNd and potential density for almost all samples with a potential density (rh) greater than 27.77 (below 1000 m at all sites, except station 022, where it includes all samples below 2000 m). The linear gradient at all open ocean sites reflects mixing between slightly more un-radiogenic CDW at intermediate depths and slightly more radiogenic SPDW. The only significant outliers on this curve are the bottom waters, discussed in detail in the next section. Perhaps the most remarkable feature of the Nd characteristics of CDW is the extreme homogeneity, despite many observations suggesting vertical movement of Nd by biogeochemical processes in the oceans generally (see Goldstein and Hemming (2003) for a review) as well as specific local continental inputs in this region (see previous sections for example) nearer to the coast. In addition, station 022 from north of the PF has lower Nd concentrations throughout the water column, including the core of CDW, despite the fact that the Nd isotope composition of the water column at the depth of CDW is the same north and south of the PF. All of this re-enforces the suggestion that the Nd signal in the CDW is dominated by oceanic processes, particularly the rapid horizontal advection of Nd in CDW. It was noted in Section 5.2 that the maximum possible flux of dust-derived Nd to the Southern Ocean was not enough to significantly modify the isotopic composition of the surface ACC. The same conclusion applies to CDW at intermediate water depths to a much greater degree due to the much higher Nd inventory of the intermediate water column. Thus, it is perhaps not surprising that vertical processes have very little effect on the middle water column in this region. The implication is that all the Nd characteristics of the ACC at intermediate depths, isotopic composition and concentration, are pre-formed outside of the study area. Because of the huge water (and Nd) flux in the ACC, the Nd in the middle water column of the ACC flows through the region decoupled from processes occurring above, and possibly beneath it (see next section). 5.4. Bottom water: a limited but significant role for boundary processes? Fig. 8 highlights that near the bottom some Nd profiles show data that deviate markedly from the generally simple, conservative mixing trend (cf. Figs. 4 and 5). In fact, significant non-conservative changes in Nd concentration occur at the bottom of all open ocean profiles, with as much as a 15 pmol/kg difference between the bottom water sample and the sample 100 m above. The h–S data show no sharp change in hydrological properties in the bottom few hundred metres at any site (Figs. 3 and 5), thus a change in water masses seems to be unlikely. It thus appears that there is significant alteration of the Nd characteristics in bottom waters close to the sediment–water interface, with an obvious potential origin in sediment–water interaction.

Neodymium isotopes in the Southern Ocean

Lacan and Jeandel (2004b) demonstrated the potential for a continental margin to enrich Nd concentrations and alter Nd isotopic compositions in bottom waters, whilst it has also been observed that an exchange of Nd can occur between water and sediment, altering isotopic compositions though not necessarily concentrations (Lacan and Jeandel, 2001). For the Nd profiles presented here, the Nd concentrations appear to be most altered with respect to standard water column properties (such as potential density). In this case, bottom water concentrations are generally lower than in the overlying water, suggesting non-conservative removal of Nd. The eNd values, on the other hand, appear to be generally similar to the overlying water. The clear exception is station 022, where the bottom water is significantly less radiogenic than the overlying waters, and the deviation from conservative behaviour with respect to potential density is clearly visible in Fig. 8a. There is also a hint of the same process at station 021, with the bottom two samples also slightly displaced off the conservative mixing trend towards less radiogenic Nd isotopic compositions. Again, an important issue here is the degree to which the changes in Nd concentration are coupled to changes in eNd, even though both are non-conservative relative to other

55

water column properties, such as potential density. The plot of eNd versus 1/[Nd] in Fig. 8c is instructive in this regard, in that it shows that, in fact, Nd concentrations at least appear to be conservative with respect to isotopic compositions in all samples except the bottom sample at station 021. At this station the bottom sample, relative to those depths just above it, is characterised by significant removal of Nd with only a small change in eNd to slightly less radiogenic values. The apparent addition of unradiogenic Nd to the bottom water at station 022, and to a much smaller extent, perhaps at station 021, is at odds with the available data on the Nd isotope composition of marine sediments in this region. Seabed surface sediments in this sector of the Southern Ocean are generally more radiogenic than seawater, reflecting the young crustal ages of West Antarctica (Roy et al., 2007). Holocene core-top sediment in the northern Bellingshausen Sea (north of the PF) has an eNd value of 4.6, significantly more radiogenic than the advected water mass (Walter et al., 2000). There is the possibility, however, that less radiogenic material from East Antarctica has been transported to site 022 by the ACC. Walter et al. (2000) measured eNd values < 8 in sediments that were deposited

Fig. 8. Nd and potential density data for intermediate and deep-water samples for the open ocean sites. All data are for potential densities greater than 27.77 (P1000 m for stations 008, 010, 021 and 024 and P2000 m for station 022). These plots illustrate the extremely conservative behaviour of Nd at intermediate water depths of 3000–4000 m, with some indications of deviation from such behaviour only in bottom waters.

56

P. Carter et al. / Geochimica et Cosmochimica Acta 79 (2012) 41–59

at two sites in the Scotia Sea and one site in the South Atlantic during the last interglacial period (Marine Isotope Stage 5) highlighting the unradiogenic character of sediment derived from East Antarctic sources. The exact mechanisms of sediment–seawater interaction remain poorly understood (Jeandel et al., 2007). The significant alteration of bottom waters at the open-ocean stations in our study area is surprising in the context of previous boundary exchange observations. Modification is more likely close to continental slopes, and in areas of dynamic flow, such as straits, where high sediment loads and significant re-suspension allows dissolved/particulate exchange (Jeandel et al., 2007). The high current speed of the ACC are perhaps responsible for re-suspending seabed sediments. The most significant bottom-water alteration occurs at station 022, where current speed is likely higher than at the other stations because site 022 lies close to the Polar Front, the axis of maximum current speed (Grose et al., 1995). Whatever the exact causes, the deviations from non-conservative behaviour of Nd in these bottom waters adds to the growing body of evidence for such a phenomenon near ocean margins. 6. CONCLUDING REMARKS We have presented an extensive dataset for Nd concentrations and isotopic compositions in the water-column of the eastern Pacific sector of the Southern Ocean, including the adjacent shelves of the Amundsen Sea Embayment and the southern Bellingshausen Sea. The data suggest a number of types of behaviour for Nd in different parts of the water column. The bulk of the open ocean water column, away from the top or bottom 100 m, and representing more than 4000 m of water, exhibits conservative mixing between Circumpolar Deep Water and South Pacific Deep Water. This phenomenon is reflected in a tight correlation between Nd isotopic compositions, Nd concentrations and hydrographic properties (e.g. potential density). CDW thus has a very clear isotopic composition in the eastern Pacific sector of the Southern Ocean, with eNd in its core of 8.7 ± 0.1. This is identical to previous measurements of CDW from the Drake Passage (eNd = 8.5 to 9; Piepgras and Wasserburg, 1982) and clearly fingerprints CDW flowing into the Atlantic from the Pacific. The only Nd data for CDW in the Atlantic are actually more radiogenic, exhibiting eNd between 6.3 and 8.1 in the South Atlantic (Jeandel, 1993), despite the contribution of unradiogenic NADW to the Atlantic sector. This may reflect modification of CDW as it flows past the radiogenic sediments of the Antarctic Peninsula (e.g. Walter et al., 2000; Roy et al., 2007) or input of radiogenic dust from South America (e.g. Walter et al., 2000; Smith et al., 2003). If dust input is the key process, however, it must be much more important to the mass balance than is apparent in the Pacific sector studied here. The conservative nature of the Nd signal in the middle water column reflects the very strong advection of water, and Nd, across the region. The inventory of Nd in the Antarctic Circumpolar Current represents a very large reservoir

that is strongly buffered against modification by extra-oceanic fluxes of Nd, such as dust from the atmosphere, repackaged particulate material exported to the deep ocean from the surface, or a lithogenic source in the sediment. More remarkably, CDW upwelling onto the shallow shelves of the Amundsen Sea Embayment and the southern Bellingshausen Sea, also remains largely unmodified in its core at around 300–600 m. This is despite the markedly non-conservative processes that clearly modify Nd concentrations and isotopic compositions on the shelf in the surface waters above and the bottom waters beneath the core of CDW. The modification of the Nd systematics on the shelf is very pronounced in surface and bottom waters, where eNd shifts away from open ocean values by up to 4 units, accompanied by a near tripling of Nd concentrations. Unlike observations in some other situations, where significant boundary modification of Nd occurs (e.g. Lacan and Jeandel, 2004b,c), the process here appears to involve the coupled addition of Nd and modification of the Nd isotopic composition. This shelf water then modifies both the Nd concentrations and Nd isotopic compositions of open ocean surface waters, also in a pseudo-conservative fashion (Fig. 7). By contrast, the subtle but significant modifications of open ocean bottom waters cannot be explained by processes involving addition of Nd with a concomitant change in eNd. Rather, these latter modifications involve Nd removal from solution, accompanied by limited modification of isotopic compositions (Fig. 8). Clearly, these and literature data document a range of types of Nd behaviour, suggesting an array of processes, that modify Nd at ocean boundaries, involving Nd addition (Lacan and Jeandel, 2001), isotopic exchange without change in Nd concentration (this study; Lacan and Jeandel, 2004b,c), and Nd removal (this study; Lacan and Jeandel, 2005a). The data presented here also have two significant implications for the use of Nd isotopes as a water mass tracer in palaeo-records – one positive and one negative. The negative finding is that, at least to some degree, the bottom few 10s of metres of the water column may not accurately reflect the dominant deep-water mass in the study area – i.e. the Nd signature of bottom waters may be significantly offset away from the main water mass flowing above the seabed by sediment–water interaction. Though this effect is minor here, it may be more significant at other locations. If sedimentary archives of Nd isotopes dominantly reflect bottom water, then such a suggestion clearly compromises their use as recorders of water column eNd, at least until the mechanisms by which these modifications occur are more fully understood. More positively, the conservativity of Nd in the bulk of the ACC, as it flows past a region where significant modification of its Nd isotopic signature by boundary processes is possible, removes one key uncertainty in the interpretation of palaeo-records. Much attention has focused recently (e.g. Rutberg et al., 2000; Piotrowski et al., 2005) on the potential of Nd isotopes in sedimentary archives to monitor the past export of deep water from the North Atlantic as part of the Atlantic Meridional Overturning Circulation (AMOC). The rationale is that the Nd isotopic composition

Neodymium isotopes in the Southern Ocean

of deep South Atlantic water represents a balance between unradiogenic (eNd  13.5; Piepgras and Wasserburg, 1987) North Atlantic Deep Water flowing southward and eventually contributing to the signal in the ACC, and the more radiogenic Nd (because of significant contributions from Pacific Deep Water with eNd around 4: Piepgras and Jacobsen, 1988; Vance et al., 2004) in the ACC flowing northward. But the interpretation of eNd variations in deep South Atlantic water solely in terms of strength of the AMOC requires that either the two end-members do not change back through time or, if they do, that such changes are wellcharacterised. Recently, it has been shown that NADW appears to remain constant at an eNd around 13.5 through the Late Quaternary (Foster et al., 2007), and even during the rapid climate change that characterises the last deglaciation (Van der Flierdt et al., 2006a). Similar constraints on the southern-sourced deep-water mass end-member are absent. The data presented here, and the conclusion that the huge flux of Nd through the ACC is extremely difficult to modify, at least suggest that local continental inputs are highly unlikely to change the eNd of the ACC through time. For example, even the 30 times higher dust fluxes to the Southern Ocean during the last glacial period relative to the Holocene (e.g. Petit et al., 1990) would equate to a total annual Nd flux from the atmosphere that is still two orders of magnitude lower than the annual flux of Nd across the ACC. Recently, Robinson and van de Flierdt (2009) presented deep-sea coral data which suggest that during Heinrich Event 1 (H1) Nd in CDW was shifted towards more radiogenic values by 2–2.5 epsilon units relative to modern values. Their preferred interpretation is that the shift records a shutdown in the AMOC, and a vast reduction in the export of unradiogenic Nd from the North Atlantic to the Southern Ocean. These authors specifically discuss the possibility that the change is due instead to increased Nd input from the Antarctic shelves during H1, but they argue that the shelves are well-developed today and that seawater values close to the coast do not display the radiogenic Nd values characterising the shelves themselves. While the data presented here demonstrate that this is not the case, and that seawater close to the coast is indeed highly modified, it does also appear to be the case that CDW in the ACC flowing past the West Antarctic shelf is unlikely to be modified, even by much increased interaction between the two. These conclusions simplify the use of South Atlantic and ACC palaeo-Nd records by removing a key uncertainty, and suggest that the only major process that can change Nd characteristics in both the ACC and the deep South Atlantic is variability in NADW export. ACKNOWLEDGEMENTS This research was supported by a NERC studentship to P.C. Participation by P.C. in cruise JR179 was supported by the Antarctic Funding Initiative. Ye Zhao helped us make the ODV section in Fig. 3. We thank Associate Editor Sidney Hemming for her efficient handling of the review process, and three anonymous reviewers for comments that improved the paper.

57 REFERENCES

Albarede F., Goldstein S. L. and Dautel D. (1997) The neodymium isotopic composition of manganese nodules from the Southern and Indian oceans, the global oceanic neodymium budget, and their bearing on deep ocean circulation. Geochim. Cosmochim. Acta 61, 1277–1291. Amakawa H., Sasaki K. and Ebihara M. (2009) Nd isotopic composition in the central North Pacific. Geochim. Cosmochim. Acta 73, 4705–4719. Arsouze T., Dutay J.-C., Lacan F. and Jeandel C. (2007) Modeling the neodymium isotopic composition with a global ocean circulation model. Chem. Geol. 239, 165–177. Bayon G., German C. R., Burton K. W., Nesbitt R. W. and Rogers N. (2004) Sedimentary Fe–Mn oxyhydroxides as paleoceanographic archives and the role of aeolian flux in regulating oceanic dissolved REE. Earth Planet. Sci. Lett. 224, 477–492. Burton K. W., Ling H.-F. and O’Nions R. K. (1997) Closure of the Central American Isthmus and its effect on deep water formation in the North Atlantic. Nature 386, 382–385. Camerlenghi A., Crise A., Pudsey C. J., Accerboni E., Laterza R. and Rebesco M. (1997) Ten-month observation of the bottom current regime across a sediment drift of the Pacific margin of the Antarctic Peninsula. Antarct. Sci. 9, 426–433. Crocket K. C., Vance D., Gutjahr M., Foster G. L. and Richards D. A. (2011) Persistent Nordic deep-water overflow to the glacial North Atlantic. Geology 39, 515–518. Duce R. A. and Tindale N. W. (1991) Atmospheric transport of iron and its deposition in the ocean. Limnol. Oceanogr. 36, 1715–1726. Foster G. L., Vance D. and Prytulak J. (2007) No change in the neodymium isotopic composition of deep water exported from the North Atlantic on glacial–interglacial timescales. Geology 35, 37–40. Frank M. (2002) Radiogenic isotopes: tracers of past ocean circulation and erosional input. Rev. Geophys. 40, 1–38. doi:10.1029/2000RG000094. Georgi D. T. (1979) Modal properties of Antarctic Intermediate Water in the southeast Pacific and the south Atlantic. J. Phys.Oceanogr. 9, 456–468. Giorgetti A., Crise A., Laterza R., Perini L., Rebesco M. and Camerlenghi A. (2003) Water masses and bottom boundary layer dynamics above a sediment drift of the Antarctic Peninsula Pacific Margin. Antarct. Sci. 15, 537–546. Goldstein S. L. and Hemming S. R. (2003) Long lived isotopic tracers in oceanography, palaeoceanography and ice-sheet dynamics. In Treatise on Geochemistry, vol. 6 (ed. H. Elderfield). Elsevier, Amsterdam, pp. 453–489. Graham A. G. C., Larter R. D., Gohl K., Hillenbrand C.-D., Smith J. A. and Kuhn G. (2009) Bedform signature of a West Antarctic palaeo-ice stream reveals a multi-temporal record of flow and substrate control. Quat. Sci. Rev. 28, 2774–2793. Graham A. G. C., Larter R. D., Gohl K., Dowdeswell J. A., Hillenbrand C.-D., Smith J. A., Evans J., Kuhn G. and Deen T. (2010) Flow and retreat of the Late Quaternary Pine IslandThwaites palaeo-ice stream, West Antarctica. J. Geophys. Res. 115, F03025. doi:10.1029/2009JF001482. Grose T. J., Johnson J. A. and Bigg G. R. (1995) A comparison between the Fram (Fine Resolution Antarctic Model) results and observations in the Drake Passage. Deep-Sea Res. I 42, 365–388. Gutjahr M., Frank M., Stirling C. H., Klemm V., van de Flierdt T. and Halliday A. N. (2007) Reliable extraction of a deepwater trace metal isotope signal from Fe–Mn oxyhydroxide coatings of marine sediments. Chem. Geol. 242, 351–370.

58

P. Carter et al. / Geochimica et Cosmochimica Acta 79 (2012) 41–59

Hellmer H. H., Jacobs S. S. and Jenkins A. (1998) Oceanic erosion of a floating Antarctic glacier in the Amundsen Sea. In Ocean, Ice, and Atmosphere: Interactions at the Antarctic Continental Margin, vol. 75 (eds. S. Jacobs and R. Weiss). AGU, Washington, DC, pp. 83–99. Herna´ndez-Molina F. J., Larter R. D., Rebesco M. and Maldonado A. (2006) Miocene reversal of bottom water flow along the Pacific Margin of the Antarctic Peninsula: stratigraphic evidence from a contourite sedimentary tail. Mar. Geol. 228, 93–116. Hillenbrand C. D., Camerlenghi A., Cowan E. A., Herna´ndezMolina F. J., Lucchi R. G., Rebesco M. and UenzelmannNeben G. (2008) The present and past bottom-current flow regime around the sediment drifts on the continental rise west of the Antarctic Peninsula. Mar. Geol. 255, 55–63. Hillenbrand C.-D., Larter R. D., Dowdeswell J. A., Ehrmann W., ´ Cofaigh C., Benetti S., Graham A. G. C. and Grobe H. O (2010) The sedimentary legacy of a palaeo-ice stream on the shelf of the southern Bellingshausen Sea: clues to West Antarctic glacial history during the Late Quaternary. Quat. Sci. Rev. 29, 2741–2763. Holland P. R., Jenkins A. and Holland D. M. (2010) Ice and ocean processes in the Bellingshausen Sea, Antarctica. J. Geophys. Res. 115, C05020. doi:10.1029/2008JC005219. Klevenz V., Vance D., Schmidt D. N. and Mezger K. (2008) Neodymium isotopes in benthic foraminifera: core-top systematics and a down-core record from the Neogene south Atlantic. Earth Planet. Sci. Lett. 265, 571–587. Klinck J. M., Hofmann E. E., Beardsley R. C., Salihoglu B. and Howard S. (2004) Water-mass properties and circulation on the west Antarctic Peninsula continental shelf in austral fall and winter 2001. Deep-Sea Res. II 51, 1925–1946. Jacobs S. S., Hellmer H. H. and Jenkins A. (1996) Antarctic ice sheet melting in the Southeast Pacific. Geophys. Res. Lett. 23, 957–960. doi:10.1029/96GL00723. Jacobsen S. J. and Wasserburg G. J. (1980) Sm–Nd isotopic evolution of chondrites. Earth Planet. Sci. Lett. 50, 139–155. Jeandel C. (1993) Concentration and isotopic composition of Nd in the South Atlantic Ocean. Earth Planet. Sci. Lett. 117, 581–591. Jeandel C., Arsouze T., Lacan F., Te´chine´ P. and Dutay J.-C. (2007) Isotopic Nd compositions and concentrations of the lithogenic inputs into the ocean: a compilation, with an emphasis on the margins. Chem. Geol. 239, 156–164. Jenkins A. and Jacobs S. (2008) Circulation and melting beneath George VI Ice Shelf, Antarctica. J. Geophys. Res. 113, C04013. doi:10.1029/2007JC004449. Jenkins A., Vaughan D. G., Jacobs S. S., Hellmer H. H. and Keys J. R. (1997) Glaciological and oceanographic evidence of high melt rates beneath Pine Island Glacier, West Antarctica. J. Glaciol. 43, 114–121. Jones K. M., Khatiwala S. P., Goldstein S. L., Hemming S. R. and vandeFlierdt T. (2008) Modeling the distribution of Nd isotopes in the oceans using an ocean general circulation model. Earth Planet. Sci. Lett. 272, 610–619. Lacan F. and Jeandel C. (2001) Tracing Papua New Guinea imprint on the central Equatorial Pacific Ocean using neodymium isotopic compositions and rare earth element patterns. Earth Planet. Sci. Lett. 186, 497–512. Lacan F. and Jeandel C. (2004a) Denmark Strait water circulation traced by heterogeneity in neodymium isotope compositions. Deep-Sea Res. I 51, 71–82. Lacan F. and Jeandel C. (2004b) Neodymium isotopic composition and rare earth element concentrations in the deep and intermediate Nordic Seas: constraints on the Iceland Scotland Overflow Water signature. Geochem. Geophys. Geosys. 5, Q11006. doi:10.1029/2004GC000742.

Lacan F. and Jeandel C. (2004c) Subpolar Mode Water formation traced by neodymium isotopic composition. Geophys. Res. Lett. 31, L14306. doi:10.1029/2004GL019747. Lacan F. and Jeandel C. (2005a) Neodymium isotopes as a new tool for quantifying exchange fluxes at the continent–ocean interface. Earth Planet. Sci. Lett. 232, 245–257. Lacan F. and Jeandel C. (2005b) Acquisition of the neodymium isotopic composition of the North Atlantic Deep Water. Geochem. Geophys. Geosys. 6, Q12008, 10.1029/2005GC00956. Larter R. D., Graham A. G. C., Gohl K., Kuhn G., Hillenbrand C.-D., Smith J. A., Deen T. J., Livermore R. A. and Schenke H.-W. (2009) Subglacial bedforms reveal complex basal regime in a zone of paleo-ice stream convergence, Amundsen Sea embayment, West Antarctica. Geology 37, 411–414. Nitsche F. O., Jacobs S. S., Larter R. D. and Gohl K. (2007) Bathymetry of the Amundsen Sea continental shelf: implications for geology, oceanography, and glaciology. Geochem. Geophys. Geosys. 8, Q10009, 10.1029/2007GC001694. Orsi A. H., Whitworth, III, T. and Nowlin, Jr., W. D. (1995) On the meridional extent and fronts of the Antarctic Circumpolar Current. Deep-Sea Res. I 42, 641–673. Petit J. R., Mourner L., Jouzel J., Korotkevich Y. S., Kotlyakov V. I. and Lorius C. (1990) Palaeoclimatological and chronological Implications of the Vostok core dust record. Nature 343, 56–58. Piepgras D. J. and Jacobsen S. B. (1988) The isotopic composition of neodymium in the North Pacific. Geochim. Cosmochim. Acta 52, 1373–1381. Piepgras D. J. and Wasserburg G. J. (1980) Neodymium isotopic variations in seawater. Earth Planet. Sci. Lett. 50, 128–138. Piepgras D. J. and Wasserburg G. J. (1982) Isotopic composition of of neodymium in waters from the Drake Passage. Science 217, 207–214. Piepgras D. J. and Wasserburg G. J. (1987) Rare earth element transport in the western North Atlantic inferred from Nd isotopic observations. Geochim. Cosmochim. Acta 51, 1257– 1271. Piotrowski A. M., Goldstein S. L., Hemming S. R. and Fairbanks R. G. (2005) Temporal relationships of carbon cycling and ocean circulation at glacial boundaries. Science 307, 1933–1938. Pugh R. S., McCave I. N., Hillenbrand C.-D. and Kuhn G. (2009) Circum-Antarctic age modelling of Quaternary marine cores under the Antarctic Circumpolar Current: ice-core dust–magnetic correlation. Earth Planet. Sci. Lett. 284, 113–123. Rickli J., Frank M. and Halliday A. N. (2009) The hafnium– neodymium isotopic composition of Atlantic seawater. Earth Planet. Sci. Lett. 280, 118–127. Roberts N. L., Piotrowski A. M., McManus J. F. and Keigwin L. D. (2010) Synchronous deglacial overturning and water mass source changes. Science 327, 75–78. Robinson L. F. and van de Flierdt T. (2009) Southern Ocean evidence for reduced export of North Atlantic Deep Water during Heinrich event 1. Geology 37, 195–198. Roy M., van de Flierdt T., Hemming S. R. and Goldstein S. L. (2007) 40 Ar/39 Ar ages of hornblende grains and bulk Sm/Nd isotopes of circum-Antarctic glacio-marine sediments: implications for sediment provenance in the southern ocean. Chem. Geol. 244, 507–519. Rutberg R. L., Hemming S. L. and Goldstein S. L. (2000) Reduced North Atlantic Deep Water flux to the glacial Southern Ocean inferred from neodymium isotope ratios. Nature 405, 935–938. Scher H. D. and Martin E. E. (2008) Oligocene deep water export from the North Atlantic and the development of the Antarctic Circumpolar Current examined with neodymium isotopes. Paleoceanography 23, PA1205. doi:10.1029/2006PA001400. Schlitzer R. (2011) Ocean Data View, http://odv.awi.de.

Neodymium isotopes in the Southern Ocean Siedler G., Gould J. and Church J. (2001) Ocean Circulation and Climate: Obser ving and Modelling the Global Ocean. Academic Press, London, pp. 715. Sievers H. A. and Nowlin W. D. (1984) The stratification and water masses at Drake Passage. J. Geophys. Res. (Oceans) 89, 489– 514. Smith D. A., Hofmann E. E., Klinck J. M. and Lascara C. M. (1999) Hydrography and circulation of the west Antarctic Peninsula continental shelf. Deep-Sea Res. I 46, 925–949. Smith J., Vance D., Kemp R., Archer C., Toms P., King M. and Zarate M. (2003) Isotopic and geochemical constraints on the origin of Argentinian loess: implications for the geographic origin of glacial ice-core dust. Earth Planet. Sci. Lett. 212, 181– 196. Stoll H. M., Vance D. and Arevalos A. (2007) Nd isotope records from Bay of Bengal constrain effect of Northern Hemisphere cooling on ITCZ movement. Earth Planet. Sci. Lett. 255, 213– 228. Tachikawa K., Jeandel C. and Dupre´ B. (1997) Distribution of rare earth elements and neodymium isotopes in settling particulate material of the tropical Atlantic (EUMELI site). Deep-Sea Res. I 44, 1769–1792. Tachikawa K., Jeandel C. and Roy-Barman M. (1999) A new approach to the Nd residence time in the oceans: the role of atmospheric inputs. Earth Planet. Sci. Lett. 170, 433–446. Talley L. D. (1996) Antarctic intermediate water in the South Atlantic. In The South Atlantic: Present and Past Circulation (eds. G. Wefer, W. H. Berger, G. Siedler and D. Webb). Springer-Verlag, Berlin, pp. 219–238. Thoma M., Jenkins A., Holland D. and Jacobs S. (2008) Modelling Circumpolar Deep Water intrusions on the Amundsen Sea continental shelf, Antarctica. Geophys. Res. Lett. 35, L18602, 10.1029/2008GL034939.

59

Thomas D. J. (2004) Evidence for deep-water production in the North Pacific Ocean during the early Cenozoic warm interval. Nature 430, 65–68. Vance D. and Thirlwall M. F. (2002) An assessment of mass discrimination in MC-ICPMS using Nd isotopes. Chem. Geol. 185, 227–240. Vance D., Scrivner A. E., Beney P., Staubwasser M., Henderson G. M. and Slowey N. C. (2004) The use of foraminifera as a record of the past neodymium isotope composition of seawater. Paleoceanography 19, PA2009, 10.1029/2003PA000957. Van de Flierdt T., Robinson L. F., Adkins J. F., Hemming S. R. and Goldstein S. L. (2006a) Temporal stability of the neodymium isotope signature of the Holocene to glacial North Atlantic. Paleoceanography 21, PA4102, 10.1029/2006PA001294. Van de Flierdt T., Hemming S. R., Goldstein S. L. and Abouchami W. (2006b) Radiogenic isotope fingerprint of Wilkes Land – Adelie Coast Bottom Water in the circum-Antarctic Ocean. Geophys. Res. Lett. 33, L12606, 10.1029/2006GL026020. Walker D. P., Brandon M. A., Jenkins A., Allen J. T., Dowdeswell J. A. and Evans J. (2007) Oceanic heat transport onto the Amundsen Sea shelf through a submarine glacial trough. Geophys. Res. Lett. 34, L02602. doi:10.1029/2006GL028154. Walter H. J., Hegner E., Diekmann B., Kuhn G. and Rutgers ven der Loeff M. M. (2000) Provenance and transport of terrigenous sediment in the South Atlantic Ocean and their relations to glacial and interglacial cycles: Nd and Sr isotopic evidence. Geochim. Cosmochim. Acta 64, 3813–3827. Whitworth T. and Nowlin W. D. (1987) Water masses and currents of the Southern-Ocean at the Greenwich Meridian. J. Geophys. Res. (Oceans) 92, 6462–6476. Associate editor: Sidney R. Hemming