contrasting glacial/interglacial regimes in the...

Contrasting glacial/interglacial regimes in the western ArcticOcean as exempli¢ed by a sedimentary record from

the Mendeleev Ridge

Leonid Polyak a;�, William B. Curry b, Dennis A. Darby c, Jens Bischof c,Thomas M. Cronin d

a Byrd Polar Research Center, Ohio State University, Columbus, OH 43210, USAb Woods Hole Oceanographic Institution, Woods Hole, MA 02543, USA

c Old Dominion University, Norfolk, VA 23529, USAd US Geological Survey, Reston, VA 20191, USA

Received 1 August 2002; received in revised form 14 August 2003; accepted 19 September 2003

Abstract

Distinct cyclicity in lithology and microfaunal distribution in sediment cores from the Mendeleev Ridge in thewestern Arctic Ocean (water depths ca. 1.5 km) reflects contrasting glacial/interglacial sedimentary patterns. Weconclude that during major glaciations extremely thick pack ice or ice shelves covered the western Arctic Ocean andits circulation was restricted in comparison with interglacial, modern-type conditions. Glacier collapse events aremarked in sediment cores by increased contents of ice-rafted debris, notably by spikes of detrital carbonates and ironoxide grains from the Canadian Arctic Archipelago. Composition of foraminiferal calcite N

18O and N13C also shows

strong cyclicity indicating changes in freshwater balance and/or ventilation rates of the Arctic Ocean. Light stableisotopic spikes characterize deglacial events such as the last deglaciation at ca. 12 14C kyr BP. The prolonged periodwith low N

18O and N13C values and elevated contents of iron oxide grains from the Canadian Archipelago in the lower

part of the Mendeleev Ridge record is interpreted to signify the pooling of freshwater in the Amerasia Basin, possiblyin relation to an extended glaciation in arctic North America. Unique benthic foraminiferal events provide a meansfor an independent stratigraphic correlation of sedimentary records from the Mendeleev Ridge and other mid-depthlocations throughout the Arctic Ocean such as the Northwind and Lomonosov Ridges. This correlation demonstratesthe disparity of existing age models and underscores the need to establish a definitive chronostratigraphy for ArcticOcean sediments.; 2003 Elsevier B.V. All rights reserved.

Keywords: Arctic Ocean; Quaternary; stratigraphy; paleoceanography; glaciation; ice rafting; Foraminifera; Ostracoda;stable isotopes; provenance

1. Introduction

The Arctic Ocean plays a major role in the lateCenozoic evolution of the Earth’s environmental

0031-0182 / 03 / $ ^ see front matter ; 2003 Elsevier B.V. All rights reserved.doi:10.1016/S0031-0182(03)00661-8

* Corresponding author. Tel. : +1-614-292-2602;Fax: +1-614-292-4697.

E-mail address: [email protected] (L. Polyak).

PALAEO 3228 5-1-04

Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 73^93

www.elsevier.com/locate/palaeo

mailto:[email protected]

http://www.elsevier.com/locate/palaeo

system, including the modern climatic change(e.g., Aagard et al., 1999). The most importantaspects of the interaction between the ArcticOcean and the global climate are (1) vast ice coverthat increases albedo and controls a signi¢cantpart of the global energy balance, and (2) thecontrol on the oceanic thermohaline circulationthrough the water exchange with the Atlanticand Paci¢c oceans (Delworth et al., 1997; Smithet al., 2002). The Quaternary history of the ArcticOcean features multiple dramatic changes associ-ated with glaciations and deglaciations of the arc-tic periphery and sea level £uctuations. Duringglacial periods and low sea levels, extensive areasof shallow arctic shelves were exposed and par-tially covered by ice sheets, which drastically re-duced the circulation within the Arctic Ocean andthe inter-oceanic water exchange. This, togetherwith more severe ice conditions in the ocean in-terior, had a dramatic e¡ect on sedimentary envi-ronments. Despite the importance of the Arctic

Ocean sedimentary archives for understandingthe Earth’s Quaternary climatic evolution, rela-tively few sedimentary records have become avail-able for detailed investigation because of opera-tional constraints in the ice-bound Arctic Oceaninterior. In addition to logistical problems of ob-taining sediment cores, it is di⁄cult to develop anunequivocal stratigraphy of late Cenozoic sedi-ments because of the meagerness of biotic re-mains, low sedimentation rates of mostly below1 cm/kyr, and the unique history of the ArcticOcean. In recent years, several late Cenozoic sedi-mentary records with a moderate temporal reso-lution have been investigated from the ArcticOcean interior, mainly from intermediate waterdepths, such as on the Northwind and Lomono-sov Ridges (Fig. 1; Poore et al., 1993, 1994; Phil-lips and Grantz, 1997; Spielhagen et al., 1997;Jakobsson et al., 2000, 2001). However, di¡erentstratigraphic approaches employed in these stud-ies result in di¡ering age models and contrasting

Fig. 1. (A) Map of the Arctic Ocean with bathymetry in 1-km contour lines. Inset area (B) is shaded. Black circle shows the lo-cation of NP26 cores; un¢lled circles show locations of cores from Northwind and Lomonosov Ridges used for stratigraphic cor-relation (Fig. 10). Arrows show major current systems: Beaufort Gyre and Transpolar Drift. (B) Map of the Mendeleev Ridgeand Chukchi Borderland with bathymetry in 500-m contour lines (from Perry and Fleming, 1986). Black circles show location ofNP26 cores 5 and 32; un¢lled circles indicate other cores referred to in the paper; triangles indicate hydrographic pro¢les shownin Fig. 2.

PALAEO 3228 5-1-04

L. Polyak et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 73^9374

paleoceanographic interpretations. This under-scores the necessity to investigate more sedimen-tary records from the Arctic Ocean aiming at mul-ti-proxy studies of sea£oor areas with relativelyhigh sedimentation rates. In this paper, we presentresults of a multi-proxy investigation of a sedi-ment core record from intermediate water depthsat the Mendeleev Ridge in the western part of theArctic Ocean (Amerasia Basin).

2. Study area

Mendeleev Ridge is an elevated portion of theArctic Ocean £oor extending from the East Sibe-rian shelf towards the North Pole (Fig. 1). On theshallow southern part of the Mendeleev Ridgeand adjacent Chukchi Borderland (Chukchi Pla-teau and Northwind Ridge), sedimentation ratesare higher than in the interior of the AmerasiaBasin (Poore et al., 1993; Phillips and Grantz,1997), probably because of sediment transportfrom the adjacent shelves and more frequentsummer ice melt. Modern sedimentation rateswere estimated to be as low as 0.2^0.3 mm/kyrthroughout the Amerasia Basin by 230Th andpore water chemistry methods (Cranston, 1997;Huh et al., 1997); however, 14C data from theChukchi Borderland show up to two orders ofmagnitude higher sedimentation rates during thedeglaciation and the Holocene (Darby et al.,1997).

The strati¢ed water mass structure of the ArcticOcean includes the low-salinity surface water, sev-eral intermediate water layers, and the bottomwaters (Aagard et al., 1985). Cold and saline bot-tom waters are formed by the exchange with theNorwegian-Greenland Sea via the Fram Straitand by cascading of dense water ejected by iceformation on the shelves. The Amerasia and Eur-asia Basins of the Arctic Ocean are separated bythe Lomonosov Ridge with a sill depth of ca.1500 m; as a result, their deep waters have some-what di¡ering thermohaline properties and resi-dence times. The deep water of the Amerasia Ba-sin is relatively warm (30.5‡C) and may have anotably old age of almost 1000 years (Macdonaldand Carmack, 1991). The major intermediate water

layer with a core at depths of ca. 200^500 moriginates from the North Atlantic Drift ; thiswater retains fairly high heat content resulting intemperatures s 0‡C throughout the Arctic Ocean(Fig. 2). The water column above this Atlantic-derived layer contains the Paci¢c water that in-£ows via the Bering Strait and is characterizedby high nutrient contents and relatively low salin-ity of 6 33 psu. The balance of Atlantic and Pa-ci¢c waters £owing into the Arctic Ocean varieswith time (Carmack et al., 1997; Steele and Boyd,1998). Currently, the front between Atlantic andPaci¢c components in subsurface waters is locatedalong the Mendeleev Ridge (Jones et al., 1998),which underscores the signi¢cance of this area forstudying the long-term changes in the Arctic cir-culation. The surface water of the Arctic Oceanreceives the voluminous river runo¡ that keepsthe salinity low and thus maintains the sea icecover. Ice and the sur¢cial waters are involvedin a wind-driven circulation which forms the anti-cyclonic Beaufort Gyre over the Amerasia Basin(Fig. 1). The major gateway for out£owing low-salinity surface water is the Fram Strait ; addition-

Fig. 2. Hydrographic pro¢les from the Mendeleev Ridge area(Fig. 1; data from Ekwurzel et al., 2001). Shown are poten-tial temperature, salinity, and equilibrium calcite N

18O (vs.PDB) estimated from temperature and water N

18O using theequation of Shackleton (1974). Halocline and thermoclinelevels are highlighted. Right panel shows details of the upperportion of the water column. Gradient in surface salinity andN

18O re£ects the contribution of runo¡ related to the dis-tance from the shelf.

PALAEO 3228 5-1-04

L. Polyak et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 203 (2004) 73^93 75

ally, substantial volumes leave via the straits ofthe Canadian Archipelago and the Barents Sea.

Due to perennial sea ice cover, the productivityin the central Arctic Ocean is low, although recentstudies estimate the amount of primary produc-tion as high as 15 g C/m2/yr, much higher thanpreviously believed (Gosselin et al., 1997). Benthicpopulation densities appear to be higher at shal-lower depths, possibly due to lateral advection offood from shelves (Clough et al., 1997). Biologicalproduction in the Arctic has undergone large tem-poral variations, being greatly reduced during gla-cial periods, but possibly somewhat enhancedduring the episodes of climatic amelioration,such as the early Holocene (Cronin et al., 1995).

3. Materials and methods

Sediment cores used in this study were collectedwith a gravity corer in 1983 from a Russian drift-ing ice camp ‘Severnyj Polyus (North Pole) 26’(further referred to as NP26). All cores were splitand described at VNII Okeangeologia, Russia(Yashin, 1985). Two closely located cores, 5 and32, from intermediate water depths of around 1.5km (Table 1) were continuously sampled by Poly-ak at 1.5^2-cm intervals for a detailed investiga-tion. Samples were taken so as to minimize thecontamination from adjacent intervals by biotur-bation. The upper part of the stratigraphicallylonger core 5 has been damaged; therefore, weuse core 32 with a similar stratigraphy to charac-terize the corresponding portion of the cored sedi-mentary record (Fig. 3). The correlation of thesetwo cores, originally based on the lithologicaloverview, was veri¢ed and detailed using all lith-ological and microfaunal characteristics understudy. Initial results of the investigation of theseand adjacent NP26 cores have been reported ear-

Table 1Geographic position of sediment cores under study

Core No. Latitude Longitude Water depth(‡N) (‡W) (m)

NP26-5 78‡58.7P 178‡09P 1435NP26-32 79‡19.4P 178‡04P 1610

Fig. 3. Stratigraphic characterization of NP26 cores 5 and 32; core 5 is shifted 14 cm down for plotting. Indices of lithologicunits (beds) used in the text are shown to the right of lithologic columns. Brown (interglacial) beds are shaded; a criss-cross pat-tern or darker shading shows pink-white detrital carbonate layers. Arrows next to the planktonic foraminiferal curve show posi-tion of 14C ages. Manganese content was determined by XRF analysis (Yashin, 1985).

PALAEO 3228 5-1-04


lier (Yashin, 1985; Polyak, 1986; Belyaeva andKhusid, 1990; Danilov et al., 1991). We have re-investigated the material (washed residues) fromcores 5 and 32 with an emphasis on foraminiferal,ostracode, stable isotope, and clast provenanceanalyses. Age control for the youngest part ofthe record is provided by six 14C ages (Table 2).

4. Results

4.1. Core stratigraphy

Sediments recovered from the Mendeleev Ridgeare characterized by a distinct alternation ofbrown and yellowish-gray beds of 5^40 cm thick-ness (Fig. 3). Brown beds, which include the sur-¢cial interval, contain moderate amounts of sandand elevated concentrations of organic matter,faunal remnants, and chemical species indicativeof an oxidizing environment, such as Mn andFe3þ (Table 3; Fig. 3). The upper boundary of

brown beds is typically distinct, whereas theirlower boundary is disturbed by bioturbation.Gray beds are almost unfossiliferous and largely¢ne-grained, but some contain prominent sandyintervals near the top and/or bottom of grayunits. We will further use a simpli¢ed lithostratig-raphy, counting brown (B) and gray (G) bedsfrom top to bottom. In addition to this cycliclithostratigraphy, there are several thin intervalsof distinctly pinkish and/or whitish colorationdue to enrichment of sediment with white or pink-ish-yellow carbonate clasts. A similar stratigraphywith alternating brown and gray beds and pink-white intercalations has been described for sedi-ments from the Northwind Ridge, some 800 kmsoutheast of the study area (Fig. 1; Phillips et al.,1992; Poore et al., 1993, 1994; Phillips andGrantz, 1997). Moreover, general patterns of thestratigraphic sequence from the Mendeleev andNorthwind Ridges can be recognized throughoutthe entire central Arctic Ocean (e.g., Belov andLapina, 1961; Jakobsson et al., 2000). Correlationis more tenuous with areas having extremely lowsedimentation rates, such as the Alpha Ridge(Clark et al., 1980; Scott et al., 1989); still itcan be maintained by means of stratigraphicmarkers, including pink-white carbonate layersand sandy intervals.

4.2. Microfauna

We have investigated microfauna in the sizefraction s 150 Wm from 60 samples and 63^100and 100^150 Wm from 19 samples (Figs. 3^5).Brown beds typically contain abundant plankton-

Table 2AMS 14C ages

Lab number Depth in core Reported age(cm) (yr BP)

AA20485 6.0 5 325 S 105AA35047 9.0 9 740 S 95AA20486 12.5 22 030 S 580AA35048 16.0 37 700 S 1200AA35049 21.0 34 000 S 1400AA35050 25.0 38 900 S 1100

All age determinations were performed at the Arizona AMSFacility on tests of Neogloboquadrina pachyderma.

Table 3Characteristic lithological and geochemical features of brown and gray beds in NP26 cores from Mendeleev Ridge (from Yashin,1985)

Feature Brown beds Gray beds

s 100-Wm sand content (%) 4.5 (0.5^14.0) [75] 3.8 (0.4^18.8) [62]CaCO3 (%) 5.3 (0.6^9.2) [26] 2.6 (0^13.2) [48]C org. (%) 0.97 (0.56^2.10) [26] 0.69 (0.13^1.82) [48]Soluble Fe3þ (%) 4.40 (3.06^5.72) [20] 3.99 (2.75^4.69) [22]Soluble Fe2þ (%) 0.36 (0.14^1.07) [20] 0.84 (0.25^1.13) [22]Mn (%) 0.39 (0.09^0.66) [16] 0.16 (0.05^0.59) [23]

Pinkish intercalations rich in carbonate clasts are not included. Notation shows mean values with minimal and maximal values inparentheses; number of analyzed samples is shown in brackets.

PALAEO 3228 5-1-04


ic and benthic foraminifers and ostracodes, aswell as occasional other faunal remains composedof calcite (echinoid spines and pteropodes) or sili-ca (sponge spicules). In contrast, gray beds con-tain only scarce faunal remains that can be par-tially attributed to bioturbation. Planktonicforaminifers throughout the core record are rep-resented predominantly by the only speciesadapted to thrive in polar latitudes, Neogloboqua-drina pachyderma sinistral (left-coiling). In con-trast, benthic faunal assemblages show signi¢cantdowncore changes, superimposed on the cyclic re-currence of faunal-rich intervals. These changesshow a consistent pattern in cores from variousparts of the Arctic Ocean, such as the Mendeleev,Northwind, Alpha, and Lomonosov Ridges(O’Neill, 1981; Polyak, 1986; Scott et al., 1989;Belyaeva and Khusid, 1990; Poore et al., 1994;Ishman et al., 1996; Jakobsson et al., 2001),which enables long-distance correlations by meansof foraminiferal and ostracode stratigraphies.

The oldest fauna recovered on the Northwind,Alpha, and Lomonosov Ridges, but not reachedby NP26 cores, contains almost exclusively arena-ceous foraminifers, which indicates a strong dis-solution of calcareous tests (O’Neill, 1981; Scottet al., 1989; Ishman et al., 1996; Evans and Ka-minski, 1998; Jakobsson et al., 2001). In contrast,calcareous foraminifers predominate in youngersediments. A marked change in calcareous benthicforaminiferal composition occurs throughout theAmerasia Basin, including NP26 cores, at thestratigraphic level of gray bed G5 (Fig. 4; Scottet al., 1989; Poore et al., 1994; Ishman et al.,1996; Jakobsson et al., 2001). Assemblages belowthis level are characterized by high numbers ofBolivina arctica in the small size fraction (63^100Wm), and relatively high content of Cassidulinateretis in coarser fractions. Another characteristicfeature of this assemblage zone is a common oc-currence of foraminifers typical for shelf areas,mostly Elphidium spp. Above, the assemblagesare predominated by Stetsonia horvathi in thesmaller size class and Oridorsalis tener in coarsefractions in cores from water depths s 1 km(Poore et al., 1994; Ishman et al., 1996; Jakobssonet al., 2001); at shallower depths, the prevalent spe-cies isC. teretis, which thrives in the Arctic Ocean

within the depth interval a¡ected by Atlanticwaters (Lagoe, 1979; Polyak, 1990; Ishman andFoley, 1996). The lower part of the O. tener abun-dance zone in cores from water depths between1 and 2 km is characterized by the common oc-currence of Bulimina aculeata that reaches s 50%in unit B4 in the size class s 100 Wm (Fig. 4;Polyak, 1986; Poore et al., 1994; Ishman et al.,1996; Jakobsson et al., 2001). This species is prac-tically absent from the Arctic Ocean today, whichenables the use of its abundance spike as a dis-tinct biostratigraphic marker.

The occurrences of some other species, notablyEpistominella exigua, Nuttalides umboniferus, Ala-baminella weddellensis, and Gyroidinoides spp., arealso con¢ned to certain narrow stratigraphic lev-els (Fig. 4) and indicate conditions unlike those in

Fig. 4. Benthic foraminiferal distribution in NP26 cores (seeFig. 2 for lithology). Upper panel shows abundances pergram (curves, upper scale bars) and percentages (bars, lowerscale bars) in the s 150-Wm size class. Lower panel showstotal abundances per gram and percentages of two species inthe s 63-Wm size class, as well as occurrence ranges of somerare species. Percentages are not shown for levels with verylow abundances.

PALAEO 3228 5-1-04


the modern Arctic Ocean. Some of these species,such as E. exigua, presently inhabit the Norwe-gian-Greenland Sea and a limited adjacent area ofthe Arctic Ocean, but not the Amerasia Basin;other species, such as N. umboniferus, occur atpresent only south of the Norwegian-GreenlandSea.

The downcore distribution of ostracodes on theMendeleev Ridge also exhibits signi¢cant changesin their composition (Fig. 5). The lower part ofthe sedimentary sequence with calcareous fauna(B5 to G6) contains a relatively poor, low-diver-sity ostracode assemblage represented by Cythero-pteron spp. and some Krithe glacialis and Poly-cope. More diverse and abundant ostracodesoccur upcore, in brown beds B1^B4. Among oth-er species, this fauna consistently contains Aceta-bulastoma sp., whose life cycle is connected withsea ice (Cronin et al., 1994). A distinct assemblagedominated by Henryhowella sp. characterizes thetransition from G2 to B2, whereas the onset ofthe Holocene is marked by a rise in Cytheropteronand Krithe species at the expense of Polycope.Although ostracodes are more scarce in the Lo-monosov Ridge sediments, the general downcorechange in ostracode abundance and composition

there shows a pattern similar to that in NP26cores (Jones et al., 1999; Jakobsson et al.,2001). Furthermore, the transition from the gla-cial/deglacial to Holocene ostracode assemblagesin the Mendeleev Ridge record is similar to thosedescribed from intermediate water depthsthroughout the Arctic Ocean (Cronin et al., 1995).

4.3. Stable isotopes

Simultaneous measurements of N18O and N

13Cwere run on tests of planktonic foraminifer Neo-globoquadrina pachyderma (sin.) and two benthicspecies, Cibicidoides wuellerstor¢ and Oridorsalistener, believed to represent epifaunal and shallowinfaunal habitats, respectively. Analyzed testswere picked from the 150^250-Wm size class tominimize size-dependent e¡ects ; only some C.wuellerstor¢ specimens were larger than 250 Wm.Most N. pachyderma (sin.) tests had a secondarycalcitic crust. Nearly all N. pachyderma (sin.) sam-ples were measured twice; even more replicateanalyses were performed on C. wuellerstor¢ sam-ples. All measurements were performed on a Fin-nigan MAT 252 with a ‘Kiel’ automated deviceusing the standard WHOI methods (Ostermann

Fig. 5. Ostracode distribution in NP26 cores (see Fig. 2 for lithology). Curve on the left shows total abundances per gram in thes 150 Wm-size class; bars show species percentages. Crosses show presence of species in samples with very low abundances.

PALAEO 3228 5-1-04


and Curry, 2000; www.whoi.edu/paleo/mass-spec).

Records of both planktonic N18O and N

13C dis-play pronounced cyclic variations with amplitudesoften exceeding those of the global open oceanrecord (Fig. 6). The pattern of these variationsin the youngest beds (B1^B3) is consistent withbox core data from the more northern area ofthe Mendeleev Ridge (Fig. 1; Poore et al.,1999). Brown units are generally characterizedby intermediate or high N

18O values. Four distinctspikes in excess of 3x occur at the bottom partsof B2, B3, and B5 and immediately above B3. Thelow N

18O spikes reaching 30.5x in averaged val-ues and below 31.5x in individual measure-ments occur mostly within gray beds or at theirboundaries. Below B5, N18O values are more uni-form, being relatively low in both gray and brownsediments. Benthic N

18O variations, especially inCibicidoides wuellerstor¢, have a pattern similarto the planktonic N

18O record, but with a smalleramplitude. However, some levels within G2 andat the bottom of B7 have outstanding light C.wuellerstor¢ N

18O excursions of s 1x.

The planktonic N13C record shows even more

regular, cyclic variations than N18O (Fig. 6). High-

est values reaching 1.5x characterize intergla-cials, whereas gray beds contain light minima ofnear 0x in averaged values and as low as under31x in individual measurements. Similar to theN

18O record, N13C values are relatively stable,

moderately light in the oldest sediments (B6^G6); however, it is in G6 where extremely lightexcursions of N

13C occur. Downcore changes inbenthic N

13C have a general pattern similar tothe planktonic record, but vary in amplitudesand absolute values. Cibicidoides wuellerstor¢has relatively low-magnitude N

13C variations,mostly below 1x, whereas those in Oridorsalistener show an ampli¢ed magnitude of 3x andvalues up to 4x lighter.

4.4. Provenance of ice-rafted debris (IRD)

The provenance of terrigenous sediments wasstudied using the petrographic composition oflithic clasts s 250 Wm and the chemical composi-tion of detrital iron oxide grains (45^250 Wm) de-

Fig. 6. Distribution of N18O and N

13C in planktonic and benthic foraminifers in NP26 cores (see Fig. 2 for lithology). Symbolsshow individual measurements, curves show averages. N18O values in Oridorsalis tener are shifted +0.5x for plotting.

PALAEO 3228 5-1-04


www.whoi.edu/paleo/mass-spec

www.whoi.edu/paleo/mass-spec

termined by electron microprobe analysis (for de-tails of methods see Darby and Bischof, 1996;Darby, 2003). These techniques have been suc-cessfully applied to the Arctic Ocean modernand Quaternary sediments by employing a largedata base (300+ samples) of circum-arctic sourcearea characterizations of shelf and coastal sedi-ments for lithic grain type and detrital Fe oxidegrain ¢ngerprints (Bischof et al., 1996; Bischofand Darby, 1997; Darby et al., 2002). The distri-bution of both coarse clasts and iron oxides in theNP26 core record shows large variations indica-tive of changing dominant sources of sedimentthat is delivered to the Arctic Ocean interior bysea ice or glacier ice (Figs. 3, 7 and 8).

The unevenly distributed abundance of lithicclasts is mostly consistent with the overall sandcontent and features several spikes that indicateepisodes of intensi¢ed sea ice rafting and/or ice-berg discharge (Fig. 3). Lithic fragments are dom-inated by quartz grains ranging from 40% to morethan 80% (Fig. 7). The major source of quartzclose to the Mendeleev Ridge is the Laptev Sea;however, considerable amounts of quartz grainsmay also originate from portions of the CanadianArchipelago, upstream of the Beaufort Gyre (Bis-chof et al., 1996). Next in abundance in NP26cores are carbonate rock fragments, which peak

Fig. 7. Types of lithic clasts s 250 Wm in NP26 cores (seeFig. 2 for lithology and total clast abundance). Black curvesand upper scale bars show grain number per gram (note dif-ferent scales), gray curves and lower scale bars show percentof quartz in total grains and percent of carbonates in rockfragments.

Fig. 8. Iron oxide grain types (weighted %) in NP26 cores (see Fig. 2 for lithology). Fe oxides are matched to source areas afterDarby and Bischof (1996).

PALAEO 3228 5-1-04


at the pink-white layers and indicate the Lauren-tide part of the Canadian Archipelago (Banks andVictoria Islands) as the source area. Remainingclasts are mostly composed of crystalline and clas-tic sedimentary rocks. The latter fragments, whichmay be characteristic of the Barents and Kara Seasource or the Mackenzie area of the CanadianArctic (Bischof et al., 1996), markedly increasein abundance in the upper beds, G2 to B1.

As indicated by the composition of iron oxides,the lower part of the NP26 record is characterizedby elevated inputs from the Canadian Archipela-go, mostly from its Innuitian part (Queen Eliza-beth Islands) (Fig. 8). A distinct spike of Fe ox-ides from the Laurentide source marks the pink-white carbonate layer between B6 and B7. Ele-vated concentrations of Laurentide Fe oxides con-sistently characterize the two other pink-whitelayers further upcore. Another major source ofFe oxides in the NP26 record is the shelf of theLaptev and East Siberian Seas. Some intervals inthe upper part of the record have elevated con-tents of Fe oxides from the Kara and BarentsSeas, consistent with heightened abundances ofsedimentary rock fragments.

5. Discussion

5.1. Age model

Several authors suggest that the apparently cy-clic sequence of gray and brown beds in the ArcticOcean sediments, including cores from the Men-deleev Ridge, re£ects a succession of glacial andinterglacial/interstadial periods (Belov and Lapi-na, 1961; Polyak, 1986; Poore et al., 1994; Phil-lips and Grantz, 1997; Jakobsson et al., 2000).This implies that faunal-rich beds indicate rela-tively highly productive interglacial environments,signi¢cant in£uence of Atlantic water, and pro-duction of well oxygenated bottom waters. Incontrast, gray, faunal-poor layers correspond toglacial periods associated with thicker ice cover,suppressed biological production, weaker advec-tion of Atlantic water, and reduced bottom waterformation. This interpretation is supported by 14Cages from the upper part of the stratigraphic se-

quence from several sites across the Arctic Ocean(Fig. 9; Darby et al., 1997; Poore et al., 1999).Although some of these 14C dates, especially infaunal-poor beds, can be biased due to bioturba-tion and low sediment £uxes, the consistent down-core succession of ages indicates that the twouppermost brown beds (B1 and B2) representthe Holocene and Marine Isotope Stage (MIS)3, respectively. The intervening gray bed (G1)contains an interval of drastically declined sedi-mentation rates, possibly including a hiatus, be-tween ca. 13 ka and at least 19 ka. The prevalentage^depth pattern indicates that this episode pos-sibly started as early as 24^25 ka, assuming thatthe two 14C ages of 19 and 20 ka, inconsistentwith the general pattern, might result from con-tamination by younger material. The non-deposi-tion has been suggested to result from winnowing(Poore et al., 1999). However, the correspondinginterval is characterized by very ¢ne, rather thansandy, sediment (Fig. 3; cf. Darby et al., 1997),which points to an exceptionally heavy ice coverduring the Last Glacial Maximum (LGM), ratherthan winnowing, as a cause of non-deposition.

Fig. 9. Age^depth distribution in sediments from the Mende-leev Ridge constructed using data from this study (NP26-32)and USGS 1994 box cores (Fig. 1; Darby et al., 1997; Pooreet al., 1999). Reservoir correction applied is 3440 yr (Man-gerud and Gulliksen, 1975); however, actual reservoir timecould be larger in the Arctic Ocean, especially during glacialperiods. Shading highlights a possible hiatus between 13 and19 ka. Note that the age of V20 ka in 94B16 is an inver-sion.

PALAEO 3228 5-1-04


This interpretation is consistent with extremelylow abundance or complete absence of biogenicremains during the low-sedimentation intervaland immediately above it.

Time control for the Arctic Ocean stratigraphicsequence beyond the range of 14C ages is mostlybased on magnetostratigraphy. A distinct decreasein inclination occurring at a same stratigraphiclevel throughout the Arctic Ocean is largely inter-preted as the Brunhes^Matuyama boundary (e.g.,Clark et al., 1980; Poore et al., 1993). The succes-

sion of brown and gray beds overlying thisboundary has been suggested to match the majorglacial^interglacial events re£ected by the globalmarine N

18O scale, MIS 1 to 19 (Fig. 10, centralpanel; Poore et al., 1993; Phillips and Grantz,1997; Spielhagen et al., 1997). However, an alter-native interpretation of paleomagnetic data fromthe Lomonosov Ridge suggests that the abovedecline in inclination is an excursion rather thaninversion of magnetic ¢eld, which implies a muchyounger age of sediment and, accordingly, higher

Fig. 10. Correlation of NP26 record with cores from the Northwind Ridge (NWR) (P1-88AR-P5, also referred to as NWR 5;Poore et al., 1993, 1994) and Lomonosov Ridge (LR) (96/12-1pc; Jakobsson et al., 2000, 2001). Interglacial brown mud units areshaded and indexed next to NP26 cores; criss-cross pattern shows pink-white detrital carbonate layers. MIS are shown next toNWR 5 and 96/12-1pc records, as interpreted in respective papers. Curves show numbers of planktonic (U103) and benthic(U102, shaded) s 150-Wm foraminifers per gram sediment. Arrows left to the NWR and LR cores show a drop in magnetic in-clination interpreted as the Brunhes^Matuyama boundary in NWR 5, but not in 96/12-1pc. Also shown are selected biostrati-graphic markers, a boundary between agglutinated and calcareous foraminiferal faunas, and a position of glacial erosion on LR.

PALAEO 3228 5-1-04


sedimentation rates (Jakobsson et al., 2000, 2001).This new paleomagnetic approach is coupled withan alternative correlation of glacial^interglacialcycles with the global N18O scale, which takes sub-stages into account (Fig. 10, right panel). Inde-pendent support for this ‘young’ age model isprovided by the coccolith stratigraphy and bythe ¢rst application of optically stimulated lumi-nescence dating to the Arctic Ocean sediments(Jakobsson et al., 2001, 2003). We believe thatthe solution of this stratigraphic dilemma requiresre¢nement of paleomagnetic studies as well as in-dependent chronostratigraphic techniques. In thispaper, we show both age models as applied toNP26 cores by correlation with the Northwind(‘old’) and Lomonosov Ridge (‘young’) stratigra-phies (Fig. 10).

5.2. Stable isotopes

The water N18O composition in the Arctic

Ocean is primarily controlled by the mixing ofisotopically light runo¡ and marine, Atlantic-de-rived water with N

18O composition of near 0x(Schlosser et al., 2000 and references therein). Ac-cordingly, both the lateral and vertical distribu-tions of N

18O in the Arctic Ocean strongly corre-late with salinity (Fig. 2). This distribution maybe locally modi¢ed by the processes of sea icefreezing and melting, which cause the rejectionof isotopically light brines during freezing andthe release of N18O-enriched water when ice melts(e.g., Strain and Tan, 1993). The equilibrium cal-cite (e.c.) N

18O is additionally a¡ected by temper-ature that is linked in the Arctic Ocean with theadvection of Atlantic-derived water. This resultsin a lowering of N

18Oe:c: in the Atlantic layer byalmost 1x in the Eurasia Basin (Bauch et al.,1997; Lubinski et al., 2001), but in the WesternArctic, where the Atlantic in£uence is weaker, thislowering is much smaller (Fig. 2).

Except for the area near the Fram Strait that isa¡ected by warm Atlantic water, N

18O values inNeogloboquadrina pachyderma (sin.) from sur¢cialbottom sediments throughout the Arctic Oceangenerally get lighter towards the Western Arctic(Spielhagen and Erlenkeuser, 1994; Alderman,1994). This trend is consistent with the increase

in runo¡ component in the surface water, but canalso be explained by a change in calci¢cationdepth. In the southwestern area of the EurasiaBasin, the bulk of N. pachyderma calci¢es in theupper part of the intermediate, Atlantic-derivedwater, whereas in the interior of the Arctic Oceancalci¢cation presumably occurs in the surface/halocline water (Bauch et al., 1997; Volkmannand Mensch, 2001). Core top planktonic N

18O val-ues from NP26 and adjacent box cores are con-sistent with this picture: given a N. pachyderma(sin.) N18O disequilibrium of 31^1.5x (Bauch etal., 199; Volkmann and Mensch, 20017), core topvalues match the local halocline N

18Oe:c: composi-tion of 2^3x (Fig. 2). In contrast to planktonicN

18O, not much lateral variation is expected fordeep-sea benthic foraminiferal N18O in the ArcticOcean because of generally homogeneous bottomwater properties below the Atlantic layer. Ourunpublished data con¢rm this assumption byshowing Cibicidoides wuellerstor¢ N

18O values inthe sur¢cial sediments of mostly 3.5^4x withoutany distinct geographic or bathymetric trend be-low 1 km water depth.

From the modern distribution of planktonic fo-raminiferal calcite N

18O in the Arctic Ocean, weexpect that its downcore variations will primarilyre£ect the history of freshwater budget ratherthan temperature changes. Indeed, N

18O £uctua-tions in the Mendeleev Ridge sediments are gen-erally characterized by heavier values in intergla-cial/interstadial intervals and light peaks in glacialbeds (Fig. 6), in contrast to the global open oceanN

18O stratigraphy that extends into the GreenlandSea and adjacent area of the Eurasia Basin (e.g.,NWrgaard-Pedersen et al., 1998, 2003). Theyoungest planktonic N

18O minimum in our recordcorrelates to that in earlier investigated box corerecords from the Mendeleev Ridge (Poore et al.,1999). This spike is consistently centered at ca. 1214C ka and likely represents a major meltwaterevent corresponding to the last deglaciation ofthe Arctic margins (termination I). We infer thatolder N

18O minima in NP26 cores re£ect similardeglacial events. An extended interval of lightN

18O values below B5 presumably indicates a par-ticularly strong £ux and storage of meltwater inthe Amerasia Basin.

PALAEO 3228 5-1-04


The four maxima (heavy spikes) of planktonicN

18O reaching s 3.5x (Fig. 6) are ca. 2x heav-ier than modern values in the Mendeleev Ridgearea and are as high as the heaviest modern N

18Ovalues in the Arctic Ocean occurring in fully sa-line and cold waters near the Fram Strait (Spiel-hagen and Erlenkeuser, 1994). Because tempera-tures in surface waters over the Mendeleev Ridgetoday are just slightly above the freezing point, weinfer that heavy N

18O spikes re£ect anomalousreductions in the amount of freshwater in theAmerasia Basin, which caused the shallowing ofthe halocline and/or the increase in N

18O compo-sition in surface water. These N

18O maxima areassociated with interglacials and/or interstadials,when signi¢cant decrease in riverine £uxes to theArctic was unlikely. Consequently, we proposethat a large portion of the N

18O increase wascaused by a cessation of low-salinity Paci¢c sur-face waters, with a N

18O value ca. 31.5x lowerthan Atlantic waters, when the Bering Strait wasclosed. The present sill depth is 40^50 m, but itmay have been even shallower in the past. Brownbeds with lower N

18O values, notably B1, B4, andB6^7, would then correspond to the highest sealevel stands, such as during the Holocene andMIS 5.1 and 5.5 (e.g., Lambeck and Chappell,2001).

The general consistency of changes in plankton-ic and benthic N

18O downcore records, althoughwith di¡erent amplitudes, suggests the existence ofa mechanism of transmitting the N

18O signal fromsurface to bottom waters in the Arctic Ocean. Inthe modern circulation system, such downwellingis provided by the descent of brines released in theprocess of ice freezing over the Arctic shelves.Another potential mechanism during glaciationscould be by supercooling of surface water, includ-ing glacial meltwater, beneath an ice shelf (Fold-vik and GammelsrWd, 1988). Light excursions of1x in Cibicidoides wuellerstor¢ N

18O co-occurwith light planktonic N

18O values within G2 andat the bottom of B7; another light benthic N

18Oevent is revealed in Oridorsalis tener in the upperpart of G3. Similar excursions in glacial and de-glacial sediments from the Norwegian-GreenlandSea have been interpreted as an intense down-mixing of isotopically light sur¢cial water by ice

freezing or supercooling processes (Vogelsang,1990; Vidal et al., 1998; Dokken and Jansen,1999; Bauch and Bauch, 2001).

The values of N13C in Neogloboquadrina pachy-

derma (sin.) in sur¢cial sediments increase fromthe Fram Strait towards the Amerasia Basin de-spite an increase in ice coverage (Spielhagen andErlenkeuser, 1994). This pattern may be explainedby a shift in habitats of N. pachyderma to surface/halocline waters, because the shelf-born haloclinewater is ventilated better than the Atlantic-derivedwater (e.g., Anderson et al., 1999). Organic pro-ductivity does not appear to a¡ect signi¢cantlythe planktonic N

13C in the Arctic Ocean (Bauchet al., 2000). It is reasonable to assume that highinterglacial planktonic N

13C values (Fig. 6) re£ectthe import of surface and halocline water fromwell-ventilated shelves, such as occurs today (cf.Bauch et al., 2000). In addition cold temperaturesand enhanced air^sea exchange of CO2 wouldhave caused elevated N

13C in surface watergCO2 (Charles et al., 1993). In contrast, duringglacial periods with low sea levels and ice sheetgrowth, £uxes of water from the shelves to theArctic Ocean interior were greatly reduced, whichresulted in low N

13C values. This lowering waspossibly enhanced by a more solid ice cover andreduced air^sea exchange of CO2.

The general co-variation of downcore changesin planktonic N

13C and N18O indicates the possi-

bility of a common control. This control could bethe restriction of Arctic circulation during glacialperiods and the resulting pooling of surface waterin the western Arctic Ocean (cf. Poore et al.,1999; NWrgaard-Pedersen et al., 2003). Such pool-ing would maintain (1) low N

18O values by abuild-up of runo¡ and/or meltwater and (2) lowN

13C due to a reduced ventilation.Downcore values of benthic N

13C mostly varyconsistently with a planktonic record (Fig. 6), in-dicating that large-scale changes in the ArcticOcean ventilation similarly a¡ected the surfaceand deep waters. However, a large o¡set in Ori-dorsalis tener N13C values and a large amplitude oftheir variations suggest that N

13C signal in sedi-ment is modi¢ed by organic carbon burial rates(cf. Grossman, 1984). The interpretation of thebenthic N

13C signal is complicated by uncertainties

PALAEO 3228 5-1-04


with foraminiferal habitats. It is possible that dur-ing ‘starvation’ periods, infaunal species such asO. tener changed their feeding strategy for surfacescavenging.

5.3. Benthic paleoecology

The succession of benthic foraminiferal and os-tracode assemblages in Mendeleev, Northwind,Alpha, and Lomonosov Ridge sediments indicatessigni¢cant changes in benthic environments in thedeep Arctic Ocean during the Quaternary. Thesechanges could be associated with (1) food supplypatterns, controlled by biological productivity andsedimentation, and (2) properties of bottomwaters. Recent studies emphasize the signi¢canceof productivity/sedimentation conditions for deep-sea benthic meiofauna including foraminifers, es-pecially in oligotrophic areas such as the ArcticOcean (Wollenburg and Mackensen, 1998; Vander Zwaan et al., 1999 and references therein).On the other hand, the distribution of modernbenthic foraminifers and ostracodes in the ArcticOcean displays a generally good relationship withmajor bottom water masses (Lagoe, 1979; Polyak,1990; Ishman and Foley, 1996; Cronin et al.,1994). More studies are needed to understandwhether this relationship is merely coincidentalor re£ects true linkages of the faunal compositionwith water properties, such as temperature, salin-ity, and oxygenation.

The arenaceous foraminiferal fauna indicatesan ocean-wide dissolution of calcareous tests inthe older part of the recovered late Cenozoicstratigraphic sequence (O’Neill, 1981; Scott etal., 1989; Poore et al., 1994; Ishman et al.,1996; Evans and Kaminski, 1998; Jakobsson etal., 2001). The causes for dissolution could in-clude low productivity of calcifying organisms,diminished inputs of high-alkaline surface waters,and enhanced £ux of organic carbon (relative toclastic particle £ux) to the sea£oor (cf. Huber etal., 2000). The latter explanation may be espe-cially important in the context of a ‘young’ agemodel for Arctic Ocean sediments discussedabove (Jakobsson et al., 2000). This model impliesa dramatic increase in sedimentation rates at thetop of the arenaceous foraminiferal assemblage

zone on the Lomonosov Ridge (Jakobsson etal., 2001). This interpretation is consistent witha change in trace fossil assemblages at the samestratigraphic level on the Northwind Ridge (Phil-lips and Grantz, 1997), which indicates a changein sedimentation rates and/or organic carbon£uxes. Alternatively, the enhancement of carbon-ate preservation could be related to the establish-ment of deep-water convection in the Norwegian-Greenland Sea in the Middle Pleistocene (Henrichet al., 2002). The latter change was time-transgres-sive, extending west- and northwards across theNorwegian-Greenland Sea between 1.2 and 0.7Ma, and possibly started a¡ecting the ArcticOcean even later.

The oldest calcareous benthic foraminiferal as-semblages that succeed the arenaceous fauna arecharacterized by elevated contents of Bolivina arc-tica. Foraminifers with elongated shape, such asBolivina, are generally believed to indicate rela-tively highly productive infaunal environments(e.g., Corliss and Chen, 1988). In NP26 core 5,B. arctica peaks at the bottom of B7, togetherwith the occurrence of Epistominella exigua thatreaches almost 30% in s 100 Wm (Fig. 4; cf.Poore et al., 1994). Epistominella exigua indicatesseasonal high £uxes of phytodetritus to the sea-£oor (Gooday, 1993; Thomas and Gooday, 1996)and presently lives in the Arctic Ocean only in theseasonally ice-free area adjacent to the FramStrait. Its presence in B7 implies that the ice mar-gin at that time was located at least seasonally asfar north as 80‡N in the Amerasia Basin ^ amarked contrast with present conditions. This in-terpretation is consistent with a common occur-rence of continental shelf foraminifers, mostly El-phidium spp., in sediments below B5. These testsare presumably transported into the interior ofthe Arctic Ocean by drifting ice (Wollenburg,1995), which requires relatively open-watersummer conditions. Elevated content of Cassidu-lina teretis at several intervals below and through-out B4 is suggested to indicate an enhanced ad-vection of Atlantic water into the Arctic Ocean(Lagoe, 1979; Polyak, 1990; Ishman and Foley,1996), which could be a cause of reduced ice cov-er.

The switch to Oridorsalis tener- and Stetsonia

PALAEO 3228 5-1-04


horvathi-dominated assemblages occurred at thebottom of G5 or at the correlative stratigraphiclevel on the Mendeleev, Northwind, Alpha, andLomonosov Ridges (Fig. 4; Scott et al., 1989;Poore et al., 1994; Ishman et al., 1996; Jakobssonet al., 2001). Oridorsalis tener has been suggestedto be adapted to low-nutrition environments(Mackensen et al., 1985), and the distribution ofS. horvathi appears to be associated with perenni-al ice cover (Wollenburg and Mackensen, 1998).Therefore, we infer that this ocean-wide change inbenthic foraminiferal fauna may signify the de-crease in food supply for benthic organisms dueto the increase in ice cover of the Arctic Ocean.However, signi¢cant changes within the O. tener/S. horvathi abundance zone show that conditionswere not uniformly low-productive. The notablefeature for the lower part of this zone is the pres-ence of Bulimina aculeata, especially abundant inB4. Bulimina, including B. aculeata are commonfor bottom environments with a stable, fairly high£ux of organic carbon to the sea£oor (Lutze andCoulbourn, 1984; Mackensen et al., 1995). A pos-sible cause for such an environment in the ArcticOcean could be a drastic reduction of ice cover,however, this scenario is not supported by stableisotope records in B4 that show values similar tothose in the Holocene (Fig. 6). Another explana-tion for increased biological production may bean enhanced input of nutrient-rich Paci¢c waterto the Amerasia Arctic Ocean. Abundant B. acu-leata may be also associated with oxygen de¢-ciency in interstitial water (e.g., Lutze and Coul-bourn, 1984; Hermelin and Shimmield, 1990). Itis noteworthy that the correlative interval withB. aculeata in the Lomonosov Ridge record hasa speci¢c olive-gray color (Jakobsson et al., 2001),possibly indicating a more reduced geochemicalenvironment than during other interglacials.

Several more stratigraphic intervals have char-acteristic foraminiferal compositions including ac-cessory species that do not live in the modernArctic Ocean, such as Alabaminella weddellensis(G4), Gyroidinoides lamarckianus (B 5^7), Gyroi-dinoides orbicularis (B2^upper G2), and Nuttalidesumboniferus (G2^G3). We believe that occur-rences of these species, although in minor propor-tions, indicate unique hydrography and/or pro-

ductivity regimes in the development of theArctic Ocean. The modern-type benthic forami-niferal fauna characterized by relatively highabundances of Eponides tumidulus horvathi, Quin-queloculina akneriana, and Valvulineria arcticaformed during the deposition of B2, that is, dur-ing MIS 3.

5.4. Paleocirculation and glacial history

The abundance of iron oxide grains from theLaptev and East-Siberian Seas throughout theNP26 record may indicate the importance of seaice for the net transportation of sediment to theMendeleev Ridge (cf. Bischof and Darby, 1997).However, the pulsed nature of the IRD abun-dance and the variability of its compositionmostly re£ect the history of iceberg discharge re-lated to the build-up and disintegration of icesheets on the Arctic Ocean periphery.

The composition of Fe oxide grains in the low-er part of the NP26 record, below G5, has astrong Canadian Archipelago signature (Fig. 8).The interval below or including G5 is also char-acterized by generally light N13C and N

18O values,presumably indicating restricted circulation andthe pooling of freshwater in the Amerasia Basin.Associated benthic assemblages are interpreted toindicate reduced summer ice cover and enhancedin£ow of Atlantic water during interglacials. In amaximal age model, this lower interval (G5 andbelow) is tentatively correlated to MIS 16^12, yetin a ‘young’ model, it corresponds only to MIS 6to 5b or 5d (Fig. 10). Regardless of an absoluteage, the lowermost gray unit G6 correlates to thetop of an interval in Lomonosov Ridge sedimentsassociated with the erosion of the ridge crest to awater depth of 1 km by an ice shelf that extendedfrom the Barents Sea (Jakobsson, 1999; Polyak etal., 2001; Jakobsson et al., 2001). Bedforms onthe Chukchi Borderland extending to 700 m depthalso evidence the grounding of an ice shelf, poten-tially of the same age, that advanced from theAlaskan/Canadian margin (Polyak et al., 2001).This paleogeographic situation suggests a verylarge ice sheet in North America, which shouldhave reduced the water export from the ArcticOcean. This, together with meltwater inputs

PALAEO 3228 5-1-04


from the ice sheet, likely resulted in the pooling oflarge amounts of freshwater in the Amerasia Ba-sin, as re£ected in a prolonged period of low N

13Cand N

18O values. However, the conditions duringthis period were not uniform and included inter-vals of inferred relatively reduced summer ice cov-er and high productivity, especially well pro-nounced in B7.

A noticeable change occurred during the inter-val between B6 and B5. Stable isotopic recordsabove this transition display regular, high-ampli-tude £uctuations indicative of alternating glacialand interglacial regimes. Benthic foraminiferal as-semblages dominated by Oridorsalis tener andStetsonia horvathi are generally interpreted to in-dicate more severe sea ice cover. This conditioncould be possibly connected with the increasedimport of the Atlantic water into the Arctic Oceanacross the Barents Sea rather than via the FramStrait, which enhances the cooling of this water(e.g., Schauer et al., 1997). Such a change in therouting of Atlantic water would be caused by aprogressive deepening of the Barents Sea shelf byglacial erosion during the Quaternary (Faleide etal., 1996; Rasmussen and Fjeldskaar, 1996). Asigni¢cant e¡ect of changes in the Barents Seabathymetry on the passage of Atlantic water tothe Arctic has been demonstrated by ocean circu-lation modeling (Butt et al., 2002) and exempli¢edby paleoceanographic records from the last degla-ciation and the Holocene (Lubinski et al., 2001).Another possible long-term e¡ect of glaciationson the Arctic circulation may have been the deep-ening of the Canadian Archipelago straits, whichfacilitates the export of surface water from theAmerasia Basin. Reaching a critical level of thisexport would cause a shift towards heavier N

18Ovalues, such as occur on top of G5.

The existence of at least two sets of glacigenicbedforms on the Chukchi Borderland (Polyak etal., 2001) implies several episodes of ice shelf for-mation over the western Arctic Ocean during gla-cial maxima. A likely candidate for an extensiveice shelf cover is the penultimate glacial period(G2) that is characterized by a prolonged deposi-tion of very ¢ne, mostly unfossiliferous sediment,bound by sand/IRD spikes (Fig. 3). The centralportion of G2 is characterized by a distinct

N18O minimum in both planktonic and benthic

foraminifers (Fig. 6) and is followed by a pecu-liar benthic fauna that includes Nuttalides umbo-niferus, potentially a low-productivity indicator(Gooday, 1993; Loubere and Fariduddin, 1999).The most recent ice shelf over the Amerasia Basinmight have existed during the LGM, as re£ectedby extremely low sedimentation rates and almostcomplete absence of biogenic remains during theinterval of 13 to at least 19 ka (Figs. 3 and 9). Apronounced N

13C and N18O minimum at ca. 12 ka

emphasizes the dramatic e¡ect of the last deglaci-ation on the Amerasia Basin (cf. Poore et al.,1999).

The disintegration of ice shelves over the Amer-asia Basin would have caused surges and out-bursts of icebergs into the Arctic Ocean from ad-jacent ice sheets (cf. Darby et al., 2002). Themagnitude of such events is exempli¢ed by a dis-charge of estimated 80 000 km3 from just one icestream at the northern Laurentide margin shortlybefore 10 ka (Clark and Stokes, 2001) and up to106 km3 ice reaching Fram Strait from this icesheet over 1^2 kyr (Darby et al., 2002). We expectsuch discharges to be re£ected in sedimentary re-cords by distinct IRD spikes exempli¢ed by pink-white carbonate layers indicative of material fromthe Laurentide part of the Canadian Archipelago(cf. Bischof et al., 1996; Darby et al., 2002). Threepronounced discharge events from this area arerecorded in NP26 cores by detrital carbonatelayers and Laurentide iron oxide grains (Figs. 3,7 and 8). The V10-ka ice stream event is notclearly expressed here, possibly due to insu⁄cientresolution, but forms a distinct spike in coresfrom the Chukchi Rise (J. Bischof, personal com-munication, 2000). The increase in IRD inputsfrom the Barents and Kara Seas in the youngestpart of the NP26 record probably indicates theexpansion of ice sheets in that sector of the Arcticshelf in the latest Pleistocene.

6. Summary and conclusions

The combined paleontological, stable isotopic,and IRD provenance data from NP26 sedimentcores provide new insights into the Quaternary

PALAEO 3228 5-1-04


paleoceanographic evolution of the Arctic Ocean,despite the age model uncertainty. The most evi-dent feature of the investigated sedimentary rec-ord is the cyclic alternation of glacial and inter-glacial regimes, expressed in lithology andpaleontological and stable isotopic compositions,consistent with results from other areas in thecentral Arctic Ocean with similar sedimentationrates (Phillips et al., 1992; Poore et al., 1993,1994; Phillips and Grantz, 1997; Jakobsson etal., 2000). Fluctuations in N

13C and N18O records

acquired regularity and highest amplitudes afterthe glacial interval G5; the onset of this intervalalso marks noticeable changes in benthic assem-blages and in mineral grain composition. We sug-gest that these changes indicate a transition fromthe environment characterized by a large fresh-water component in surface waters and a some-what reduced summer sea ice cover to more stableperennial ice and periodically lessened freshwateramounts. Although the causes of this transitionare unclear, we hypothesize that they were possi-bly associated with large-scale glaciations at theArctic Ocean periphery. Glacial erosion has deep-ened the Barents Sea shelf, which opened a routefor importing Atlantic water colder than that en-tering via the Fram Strait. In a similar way, thedeepening of the Canadian Archipelago straitscould intensify the drainage of surface watersfrom the Amerasia Basin during interglacials.We note that further back in time the ArcticOcean had experienced another pronouncedchange in hydrographic and/or sedimentation re-gime, as expressed in the establishment of condi-tions favorable for preservation of carbonates(O’Neill, 1981; Scott et al., 1989; Poore et al.,1994; Evans and Kaminski, 1998; Jakobsson etal., 2001).

Pronounced £uctuations in foraminiferal calciteN

18O and N13C in the upper portion of the record

from the Mendeleev Ridge depict profoundchanges in freshwater balance and/or ventilationrates of the Arctic Ocean. Lighter stable isotopicvalues indicate higher freshwater inputs and/orresidence time, and weaker ventilation of the sur-face and halocline waters in the Amerasia Basinduring glacial periods. The lightest stable isotopicspikes characterize deglacial events such as the

last deglaciation at ca. 12 kyr BP (cf. Poore etal., 1999). Intervals with anomalously heavyN

18O values were possibly related to interstadialperiods with the closed Bering Strait, when thePaci¢c surface waters with relatively low salinityand N

18O composition ceased to enter the ArcticOcean.

Extremely low numbers or complete absence ofbiogenic remains in sediments from some glacialintervals support the notion that the western Arc-tic Ocean was possibly covered by very thick packice or extensive ice shelves during Pleistocene gla-ciations (e.g., Grosswald and Hughes, 1999;NWrgaard-Pedersen et al., 2003). The lowermostlithological unit recovered by NP26 cores (G6)correlates to the unit deposited during and/or im-mediately after the inferred ice shelf erosion onLomonosov Ridge, constrained by di¡erent agemodels to some time between MIS 6 and 16 (Ja-kobsson et al., 2001). Younger ice shelves over theAmerasia Basin may have existed during the pen-ultimate and, possibly, the last glaciation. Thelatter would have advanced at 19 ka or somewhatearlier and disintegrated at ca. 13 ka, followed bya pronounced meltwater event indicated by stableisotopic data. We infer that ice shelf collapsescaused dramatic discharges of ice into the ArcticOcean from the North American ice sheets (e.g.,Clark and Stokes, 2001; Darby et al., 2002); suchevents originating from the Laurentide part of theCanadian Archipelago are presumably marked insedimentary records by detrital carbonate layersand iron oxide grains from Banks and VictoriaIslands.

Some stratigraphic intervals of the MendeleevRidge record have unique paleontologic signa-tures that have no analogs in the modern ArcticOcean. The most evident indicator of a non-anal-ogous paleoceanographic situation is a benthic fo-raminiferal assemblage with abundant Buliminaaculeata centered in interglacial unit B4; this as-semblage zone occurs throughout the AmerasiaBasin, including the Lomonosov Ridge (Poore etal., 1994; Ishman et al., 1996; Jakobsson et al.,2001). We infer that the B. aculeata assemblagere£ects enhanced biological productivity, possiblyconnected with elevated in£ow of Paci¢c water,and/or reduced oxygenation of bottom water.

PALAEO 3228 5-1-04


We note that our understanding of the Quater-nary history of the Arctic Ocean is restrained bythe uncertainties with age models beyond therange of 14C dating. By means of bio- and litho-stratigraphy, we correlate the Mendeleev Ridgerecord with those from other areas of the ArcticOcean, such as the Northwind and LomonosovRidges (Fig. 10), but we are cautious to make ade¢nitive judgement regarding an age model forthese sediments. Because age model choicestrongly a¡ects the interpretation of paleoceano-graphic evolution, this uncertainty prompts anurgency in re¢ning and verifying the chronostra-tigraphy for the Arctic Ocean sediments.

Acknowledgements

This work was partially supported by the USANSF awards OPP-9614451, 9817051, and9817054. We are grateful to Ye.V. Telepnev andI.A. Alekseev, who collected the cores, and to allpeople who assisted with sample processing andanalytical e¡orts at various stages of this work.This is Byrd Polar Research Center PublicationNo. C-1281 and Woods Hole Oceanographic In-stitution Contribution No. 11002.

References

Aagard, K., Swift, J.H., Carmack, E.C., 1985. Thermohalinecirculation in the arctic mediterranean seas. J. Geophys.Res. 90, 4833^4846.

Aagard, K., Darby, D., Falkner, K., Flato, G., Grebmeier, J.,Measures, C., Walsh, J., 1999. Marine Science in the Arctic:A Strategy. Arctic Research Consortium of the UnitedStates (ARCUS), Fairbanks, AK, 84 pp.

Alderman, S.E., 1994. Correlation of Coretop N18O and Mod-

ern Salinity Gradients in the Arctic Ocean. M.S. Thesis,MIT-WHOI, Woods Hole, MA.

Anderson, L.G., Jones, E.P., Rudels, B., 1999. Ventilation ofthe Arctic Ocean estimated by a plume entrainment modelconstrained by CFCs. J. Geophys. Res. 104, 13423^13429.

Bauch, D., Carstens, J., Wefer, G., 1997. Oxygen isotope com-position of living Neogloboquadrina pachyderma (sin.) in theArctic Ocean. Eart Planet. Sci. Lett. 146, 47^58.

Bauch, D., Carstens, J., Wefer, G., Thiede, J., 2000. The im-print of anthropogenic CO2 in the Arctic Ocean: evidencefrom planktic N

13C data from watercolumn and sedimentsurfaces. Deep-Sea Res. II 47, 1791^1808.

Bauch, D., Bauch, H., 2001. Last glacial benthic foraminiferalN

18O anomalies in the polar North Atlantic: a modern ana-logue evaluation. J. Geophys. Res. 106 (C5), 9135^9143.

Belov, N.A., Lapina, N.N., 1961. Donnye otlozheniya Arkti-cheskogo bassejna (Sea£oor sediments of the Arctic Basin).Morskoj Transport, Leningrad, 150 pp. (in Russian).

Belyaeva, N.V., Khusid, T.A., 1990. Distribution patterns ofcalcareous foraminifers in Arctic Ocean sediments. In: Bleil,U., Thiede, J. (Eds.), Geological History of the PolarOceans: Arctic versus Antarctic. NATO ASI Ser. C 308,311^316.

Bischof, J.F., Darby, D.A., 1997. Mid- to Late Pleistocene icedrift in the Western Arctic Ocean: evidence for a di¡erentcirculation in the past. Science 277, 74^78.

Bischof, J.F., Clark, D.L., Vincent, J.-S., 1996. Pleistocene pa-leoceanography of the central Arctic Ocean: The sources ofice rafted debris and the compressed sedimentary record.Paleoceanography 11, 743^756.

Butt, F.A., Drange, H., ElverhWi, A., Ottera, O.H., Solheim,A., 2002. Modelling Late Cenozoic isostatic elevationchanges in the Barents Sea and their implications for oceanicand climatic regimes: preliminary results. Quat. Sci. Rev. 21,1643^1660.

Carmack, E.C., Aagard, K., Swift, J.H., Macdonald, R.W.,McLaughlin, F.A., Jones, E.P., Perkin, R.G., Smith, J.N.,Ellis, K., Kilius, L., 1997. Changes in temperature and con-taminant distributions within the Arctic Ocean. Deep-SeaRes. II 44, 1487^1502.

Charles, C.D., Wright, J.D., Fairbanks, R.G., 1993. Thermo-dynamic in£uences on the marine carbon isotope record.Paleoceanography 8, 691^697.

Clark, C.D., Stokes, C.R., 2001. Extent and basal character-istics of the M’Clintock Channel ice stream. Quat. Int. 86,81^101.

Clark, D.L., Whitman, R.R., Morgan, K.A., Mackey, S.D.,1980. Stratigraphy and glacial-marine sediments of theAmerasian Basin, central Arctic Ocean. Geol. Soc. Am.Spec. Pap. 181, 57 pp.

Clough, L.M., Ambrose, W.G., Jr., Cochran, K.J., Barnes, C.,Renaud, P.E., Aller, R.V., 1997. Infaunal density, biomassand bioturbation in the sediments of the Arctic Ocean.Deep-Sea Res. II 44, 1683^1704.

Corliss, B.H., Chen, C., 1988. Morphotype patterns of Nor-wegian Sea deep-sea benthic foraminifera and ecological im-plications. Geology 16, 716^719.

Cranston, R.E., 1997. Organic carbon burial rates across theArctic Ocean from the 1994 Arctic Ocean Section expedi-tion. Deep-Sea Res. II 44, 1705^1724.

Cronin, T.M., Holtz, T.R., Stein, R., Spielhagen, R., Fu«tterer,D., Wollenburg, J., 1995. Late Quaternary paleoceanogra-phy of the Eurasian Basin. Arctic Ocean. Paleoceanogr. 10,259^281.

Cronin, T.M., Holtz, T.R., Jr., Whatley, R.C., 1994. Quater-nary paleoceanography of the deep Arctic Ocean based onquantitative analysis of Ostracoda. Mar. Geol. 119, 305^332.

Danilov, I.D., Telepnev, Ye.V., Chugunov, A.B., Belyaeva,N.V., Khusid, T.A., Virina, Ye.I., Polyakova, Ye.I., 1991.

PALAEO 3228 5-1-04


Paleogeographic investigation of bottom sediments of thecentral Arctic Ocean (Mendeleyev Ridge) (translated fromRussian). Oceanology (USSR) 31, 77^82.

Darby, D.A., 2003, Sources of sediment found in sea ice fromthe western Arctic Ocean, new insights into processes ofentrainment and drift patterns. J. Geophys. Res. 108, 13-1to 13-10.

Darby, D.A., Bischof, J.F., 1996. A statistical approach tosource determination of lithic and Fe-oxide grains: an ex-ample from the Alpha Ridge, Arctic Ocean. J. Sediment.Res. 66, 599^607.

Darby, D.A., Bischof, J.F., Jones, G.A., 1997. Radiocarbonchronology of depositional regimes in the western ArcticOcean. Deep-Sea Res. II 44, 1745^1757.

Darby, D.A., Bischof, J.F., Spielhagen, R.F., Marshall, S.A.,Herman, S.W., 2002. Arctic ice export events and their po-tential impact on global climate during the late Pleistocene.Paleoceanogr. 15-1 to 15-17.

Delworth, T.L.S., Manabe, S., Stou¡er, R.J., 1997. Multide-cadal climate variability in the Greenland Sea and surround-ing regions: a coupled simulation. Geophys. Res. Lett. 24,257^260.

Dokken, T., Jansen, E., 1999. Rapid changes in the mechanismof ocean convection during the last glacial period. Nature401, 458^461.

Ekwurzel, B., Schlosser, P., Mortlock, R.A., Fairbanks, R.G.,Swift, J.H., 2001. River runo¡, sea ice meltwater and Paci¢cwater distribution and mean residence times in the ArcticOcean. J. Geophys. Res. 106, 9075^9092.

Evans, J.R., Kaminski, M.A., 1998. Pliocene and Pleistocenechronostratigraphy and paleoenvironment of the centralArctic Ocean, using deep water agglutinated foraminifera.Micropaleontology 44, 109^130.

Faleide, J.I., Solheim, A., Hjelstuen, B.O., Andersen, E.S.,Vanneste, K., 1996. Late Cenozoic evolution of the westernBarents Sea-Svalbard continental margin. Global Planet.Change 12, 53^74.

Foldvik, A., GammelsrWd, T., 1988. Notes on Southern Oceanhydrography, sea-ice and bottom-water formation. Palaeo-geogr. Palaeoclimatol. Palaeoecol. 67, 3^17.

Gooday, A.J., 1993. Deep-sea benthic foraminiferal specieswhich exploit phytodetritus: characteristic features and con-trols on distribution. Mar. Micropaleontol. 22, 129^146.

Gosselin, M., Levasseur, M., Wheeler, P.A., Horner, R.A.,Booth, B.C., 1997. New measurements of phytoplanktonand ice algal production in the Arctic Ocean. Deep-SeaRes. II 44, 1623^1644.

Grossman, E.L., 1984. Stable isotope fractionation in livebenthic foraminifera from the southern California border-land. Palaeogeogr. Palaeoclimatol. Palaeoecol. 47, 301^327.

Grosswald, M.G., Hughes, T.J., 1999. The case for an ice shelfin the Pleistocene Arctic Ocean. Polar Geogr. 23, 23^54.

Henrich, R., Baumann, K.-H., Huber, R., Meggers, H., 2002.Carbonate preservation records of the past 3 Myr in theNorwegian-Greenland Sea and the northern North Atlantic:implications for the history of NADW production. Mar.Geol. 184, 17^39.

Hermelin, J.O.R., Shimmield, B.G., 1990. The importance ofthe oxygen minimum zone and sediment geochemistry in thedistribution of recent benthic foraminifera in the NorthwestIndian Ocean. Mar. Geol. 91, 1^29.

Huber, R., Meggers, H., Baumann, K.-H., Henrich, R., 2000.Recent and Pleistocene carbonate dissolution in sediments ofthe Norwegian-Greenland Sea. Mar. Geol. 165, 123^136.

Huh, C.A., Pisias, N.G., Kelley, J.M., Maiti, T.C., Grantz, A.,1997. Natural radionuclides and plutonium in sedimentsfrom the Western Arctic Ocean: sedimentation rates andpathways of radionuclides. Deep-Sea Res. II 44, 1725^1744.

Ishman, S.E., Foley, K.M., 1996. Modern benthic foraminiferdistribution in the Amerasian Basin, Arctic Ocean. Micro-paleontology 42, 206^220.

Ishman, S.E., Polyak, L., Poore, R.Z., 1996. An expandedrecord of Pleistocene deep Arctic change: Canada Basin,western Arctic Ocean. Geology 24, 139^142.

Jakobsson, M., 1999. First high-resolution chirp sonar pro¢lesfrom the central Arctic Ocean reveal erosion of LomonosovRidge sediments. Mar. Geol. 158, 111^123.

Jakobsson, M., LWvlie, R., Al-Hanbali, H., Arnold, E., Back-man, J., Mo«rth, M., 2000. Manganese/color cycles in ArcticOcean sediments constrain Pleistocene chronology. Geology28, 23^26.

Jakobsson, M., LWvlie, R., Arnold, E., Backman, J., Polyak,L., Knutsen, J.O., Musatov, E., 2001. Pleistocene stratigra-phy and paleoenvironmental variation from LomonosovRidge sediments, central Arctic Ocean. Global Planet.Change 31, 1^21.

Jakobsson, M., Backman, J., Murray, A., LWvlie, R., 2003.Optically stimulated luminescence dating supports centralArctic Ocean cm-scale sedimentation rates. Geochem. Geo-phys. Geosyst. 2.

Jones, E.P., Anderson, L.G., Swift, J.H., 1998. Distribution ofAtlantic and Paci¢c waters in the upper Arctic Ocean: im-plications for circulation. Geophys. Res. Lett. 25, 765^768.

Jones, R.L., Whatley, R.C., Cronin, T.M., Dowsett, H.J.,1999. Reconstructing late Quaternary deep-water masses inthe eastern Arctic Ocean using benthonic Ostracoda. Mar.Micropaleontol. 37, 251^272.

Lagoe, M.B., 1979. Recent benthonic foraminiferal biofacies inthe Arctic Ocean. Micropaleontology 25, 214^224.

Lambeck, K., Chapell, J., 2001. Sea level change through thelast glacial cycle. Science 292, 679^686.

Loubere, P., Fariduddin, M., 1999. Quantitative estimation ofglobal patterns of surface ocean biological productivity andits seasonal variation on timescales from centuries to millen-nia. Global Biogeochem. Cycles 13, 115^133.

Lubinski, D.A., Polyak, L., Forman, S.L., 2001. Decipheringthe Latest Pleistocene and Holocene in£ows of freshwaterand Atlantic water to the deep northern Barents and Karaseas: Foraminifera and stable isotopes. Quat. Sci. Rev. 20,1851^1879.

Lutze, G.F., Coulbourn, W.T., 1984. Recent benthic forami-nifera from the continental margin of northwest Africa:community structure and distribution. Mar. Micropaleontol.8, 361^401.

PALAEO 3228 5-1-04


Macdonald, R.W., Carmack, E.C., 1991. Age of Canada Basindeep waters: a way to estimate primary production for theArctic Ocean. Science 254, 1348^1350.

Mackensen, A., Sejrup, H.P., Jansen, E., 1985. The distribu-tion of living benthic foraminifera on the continental slopeand rise o¡ southeast Norway. Mar. Micropaleontol. 9,275^306.

Mackensen, A., Schmiedl, G., Harlo¡, J., Giese, M., 1995.Deep-sea foraminifera in the South Atlantic Ocean: ecologyand assemblage generation. Micropaleontology 41, 342^358.

Mangerud, J., Gulliksen, S., 1975. Apparent radiocarbon agesof recent marine shells from Norway, Spitsbergen, and Arc-tic Canada. Quat. Res. 5, 263^273.

NWrgaard-Pedersen, N., Spielhagen, R.F., Thiede, J., Kassens,H., 1998. Central Arctic surface ocean environment duringthe past 80,000 years. Paleoceanography 13, 193^204.

NWrgaard-Pedersen, N., Spielhagen, R.F., Erlenkeuser, H.,Grootes, P.M., Heinemeier, J., Knies, J., 2003. The ArcticOcean during the Last Glacial Maximum: Atlantic and Po-lar domains of surface water mass distribution and ice cover.Paleoceanography 18, 8-1 to 8-19.

O’Neill, B.J., 1981. Pliocene and Pleistocene benthic foraminif-era from the central Arctic Ocean. J. Paleontol. 55, 1141^1170.

Ostermann, D.R., Curry, W.B., 2000. Calibration of stableisotopic data: An enriched N

18O standard used for sourcegas mixing detection and correction. Paleoceanography 15,353^360.

Perry, R.K., Fleming H.S., 1986. Bathymetry of the ArticOcean. Geol. Soc. America, Map and Chart Series MC-56.

Phillips, R.L., Grantz, A., 1997. Quaternary history of sea iceand paleoclimate in the Amerasia basin, Arctic Ocean, asrecorded in the cyclical strata of Northwind Ridge. Geol.Soc. Am. Bull. 109, 1101^1115.

Phillips, R.L., Grantz, A., Mullen, M.W., Rieck, H.J.,McLaughlin, M.W., Selkirk, T.L., 1992. Summary of Lith-ostratigraphy and Stratigraphic Correlations in Piston Coresfrom Northwind Ridge, Arctic Ocean, from USCGC PolarStar, 1988. US Geol. Survey Open-¢le Report 92-426, 114pp.

Polyak, L.V., 1986. New data on microfauna and stratigraphyof bottom sediments of the Mendeleev Ridge, Arctic Ocean.In: Andreev, S.I. (Ed.), Sedimentogenes i konkrecioobrazo-vanie v okeane (Sedimentogenesis and nodule-formation inthe Ocean). Sevmorgeologia, Leningrad, pp. 40^50 (in Rus-sian).

Polyak, L.V., 1990. General trends of benthic foraminiferaldistribution in the Arctic Ocean. In: Kotlyakov, V.M., So-kolov, V.E. (Eds.), Arctic Research: Advances and Pros-pects. Part 2. Nauka, Moscow, pp. 211^213.

Polyak, L., Edwards, M.H., Coakley, B.J., Jakobsson, M.,2001. Ice shelves in the Pleistocene Arctic Ocean inferredfrom glaciogenic deep-sea bedforms. Nature 410, 453^457.

Poore, R.Z., Ishman, S.E., Phillips, L., McNeil, D., 1994. Qua-ternary stratigraphy and paleoceanography of the CanadaBasin, western Arctic Ocean. US Geol. Survey Bull. 2080,32 pp.

Poore, R.Z., Phillips, L., Rieck, H.J., 1993. Paleoclimate rec-ord for Northwind Ridge, western Arctic Ocean. Paleocean-ography 8, 149^159.

Poore, R.Z., Osterman, L., Curry, W.B., Phillips, R.L., 1999.Late Pleistocene and Holocene meltwater events in the west-ern Arctic Ocean. Geology 27, 759^762.

Rasmussen, E., Fjeldskaar, W., 1996. Quanti¢cation of thePliocene-Pleistocene erosion of the Barents Sea frompresent-day bathymetry. Global Planet. Change 12, 119^133.

Schauer, U., Muench, R.D., Rudels, B., Timkhov, L., 1997.Impact of eastern Arctic shelf waters on the Nansen Basinintermediate layers. J. Geophys. Res. 102, 3371^3382.

Schlosser, P., Ekwurzel, B., Khatiwala, S., Newton, B., Mas-lowski, W., P¢rman, S., 2000. Tracer studies of the Arcticfreshwater budget. In: Edward, L.L. et al. (Eds.), TheFreshwater Budget of the Arctic Ocean. NATO Sci. Ser. 270, 453^478.

Scott, D.B., Mudie, P.J., Baki, V., MacKinnon, K.D., Cole,F.E., 1989. Biostratigraphy and late Cenozoic paleoceanog-raphy of the Arctic Ocean: foraminiferal, lithostratigraphic,and isotopic evidence. Geol. Soc. Am. Bull. 101, 260^277.

Shackleton, N.J., 1974. Attainment of isotopic equilibrium be-tween ocean water and the benthonic foraminifera genusUvigerina : isotopic changes in the ocean during the last gla-cial. In: Labeyrie, L. (Ed.), Colloque international sur lesme¤thodes quantitatives d’e¤tude des variation du climat ausours du Pleistocene. Coll. Int. C.N.R.S. 219, 203^209.

Smith, L.M., Miller, G.H., Otto-Bliesner, B., Shin, S., 2002.Sensitivity of the Northern Hemisphere climate system toextreme changes in Holocene Arctic sea ice. Quat. Sci.Rev. 22, 645^658.

Spielhagen, R.F., Erlenkeuser, H., 1994. Stable oxygen andcarbon isotopes in planktic foraminifers from Arctic Oceansurface sediments: re£ection of the low salinity surface waterlayer. Mar. Geol. 119, 227^250.

Spielhagen, R.F., Bonani, G., Eisenhauer, A., Frank, M.,Frederichs, T., Kassens, H., Kubik, P.W., Mangini, A., Nor-gaard-Pedersen, N., Nowaczyk, N.R., Schper, S., Stein, R.,Thiede, J., Tiedemann, R., Wahsner, M., 1997. Arctic Oceanevidence for late Quaternary initiation of northern Eurasianice sheets. Geology 25, 783^786.

Steele, M., Boyd, T., 1998. Retreat of the cold halocline layerin the Arctic Ocean. J. Geophys. Res. 103, 10419^10435.

Strain, P.M., Tan, F.C., 1993. Seasonal evolution of oxygenisotope-salinity relationships in high-latitude surface waters.J. Geophys. Res. 98, 14589^14598.

Thomas, E., Gooday, A.J., 1996. Cenozoic deep-sea benthicforaminifera: tracers for changes in oceanic productivity?Geology 24, 355^358.

Van der Zwaan, G.J., Duijnstee, I.A.P., den Dulk, M., Ernst,S.R., Jannink, N.T., Kouwenhoven, T.J., 1999. Benthic for-aminifers: proxies or problems? A review of paleoecologicalconcepts. Earth Sci. Rev. 46, 213^236.

Vidal, L., Labeyrie, L., van Weering, T., 1998. Benthic N18O

records in the North Atlantic over the last glacial period(60^10 kyr): evidence for brine formation. Paleoceanogra-phy 13, 245^251.

PALAEO 3228 5-1-04


Vogelsang, E., 1990. Pala«o-Ozeanographie des Europa«ischenNordmeeres an Hand stabiler Kohlensto¡- und Sauersto¡-isotope. Ber. Sonderforsch. 313, Univ. Kiel, 23, 136 pp.

Volkmann, R., Mensch, M., 2001. Stable isotope composition(N18O, N13C) of living planktic foraminifers in the outer Lap-tev Sea and the Fram Strait. Mar. Micropaleontol. 42, 163^188.

Wollenburg, J.E., 1995. Bentische foraminiferenfaunen alswassermassen-, Produktions- und Eisdriftanzeiger in Arkti-schen Ozean. Ber. Polarforsch. 179, 227 pp.

Wollenburg, J.E., Mackensen, A., 1998. Living benthic fora-minifers from the central Arctic Ocean: faunal composition,standing stock and diversity. Mar. Micropaleontol. 34, 153^185.

Yashin, D.S. (Principal Investigator), 1985. Veschestvennyjsostav i geokhimiya donnyh otlozhenij zapadno-arkticheskihmorej SSSR (Mineral composition and geochemistry of bot-tom sediments, Western Arctic seas of the USSR). TechnicalReport, Sevmorgeologia, Leningrad (in Russian).

PALAEO 3228 5-1-04


contrasting glacial/interglacial regimes in the...

Documents