Jochen Knies ( ) ) Christoph Vogt ) Ruediger SteinAlfred-Wegener-Institute for Polar- and Marine Research,Columbusstrasse, D-27568 Bremerhaven, Germany

Geo-Marine Letters (1999) 18 : 195—202 ( Springer-Verlag 1999

J. Knies · C. Vogt · R. Stein

Late Quaternary growth and decay of the Svalbard/Barents Sea ice sheetand paleoceanographic evolution in the adjacent Arctic Ocean

Received: 25 February 1997 / Revision received: 4 March 1998

Abstract The paleoceanography in the Nordic seas wascharacterized by apparently repeated switching on and offof Atlantic water advection. In contrast, a continous influxof Atlantic waters probably occurred along the northernBarents Sea margin during the last 150 ka. Temporaryice-free conditions enhanced by subsurface Atlantic wateradvection and coastal polynyas accelerated the final icesheet build-up during glacial times. The virtually completedissolution of biogenic calcite during interglacial intervalswas controlled mainly by CO

2-rich bottom waters and

oxidation of higher levels of marine organic carbon andindicates intensive Atlantic water inflow and a stable icemargin.

Introduction

During the Late Quaternary, regional and perhaps globalpaleoclimate and paleoceanography were apparently con-trolled by the switching on and off of Atlantic wateradvection to the European sector of the Arctic Ocean (e.g.,Broecker et al. 1990). Recent investigations show thatopen-water conditions existed at least seasonally not onlyduring the interglacials (e.g., Kellogg 1976) but also duringthe Weichselian glaciation (Hebbeln et al. 1994, Dokkenand Hald 1996). It is assumed that the advection of warmAtlantic water to the Norwegian—Greenland seas (NGS)triggered the growth of the Svalbard/Barents Sea ice sheet(SBIS) (Hebbeln et al. 1994) and greatly influenced theatmospheric circulation through the last glacial—inter-glacial transition (Charles et al. 1994).

Up to now, there have been major gaps in the under-standing of preceding glacial—interglacial cycles in the

NGS because of limited stratigraphic resolution of marineand terrestrial records. In particular, detailed informationabout major advances of the SBIS and the onset of de-glaciation, corresponding to Atlantic water advectionduring oxygen isotope stage 6 (Saalian glaciation), arestill under discussion (Lloyd et al. 1996; Hebbeln andWefer 1997).

Here, we present stratigraphical, geochemical, andsedimentological data from a sediment core (PS2138-1;81°32.1@ N, 30°35.6@ E; 995 m water depth) recovered dur-ing the R/» Polarstern cruise ARK-VIII/2 (Rachor 1992)(Fig. 1) that elucidates the glacial—interglacial changesfrom the northern Barents Sea margin during the LateQuaternary. The aims of this study are to reconstruct thetiming and mode of Atlantic water advection to the north-ern Barents Sea margin and to correlate varying inputs ofice-rafted detritus (IRD) with deglaciation patterns of theSBIS during the last two glacial—interglacial cycles.

Methods

Stable oxygen and carbon isotope measurements of 10—12Neogloboquadrina pachyderma sin. specimens from the'63-lm fraction were performed using a Finnigan MAT251 mass spectrometer (AWI, Bremerhaven). Results areexpressed in the d-notation (& vs. PDB) and are calib-rated against the National Institute of Standards andTechnology (NIST) 19 standards.

Total carbon (TC), nitrogen, and total organic carbon(TOC) were determined by means of a Heraeus CHN-O-RAPID elemental analyzer. The carbonate content wascalculated as CaCO

3(%)"(TC—TOC)*8.333, where TC

(%)"total carbon and TOC (%)"total organic carbon.C/N weight ratios characterizing the composition of the

organic matter were calculated as total organic car-bon/total nitrogen ratios. In general, terrigenous organicmatter (TOM) shows C/N-ratios '15, and marine or-ganic matter (MOM) shows C/N-ratios(10 (Scheffer and

Table 1 AMS 14C dates forPS2138-1 Core Depth in Corr. ages Calibrated Lab.

core (cmbsf ) Material 14Cyr ages ref. no.

PS2138-1 80 cm Bivalves 12,600#140/!130 14,796 KlA363PS2138-1 130 cm mixed forams 15,410#130/!130 18,325 KlA1283PS2138-1 160 cm N. pachyderma sin. 16,230#210/!210 19,111 KlA364PS2138-1 200 cm N. pachyderma sin. 16,880#130/!130 20,573 KlA2745PS2138-1 300 cm N. pachyderma sin. 20,040#330/!320 24,007 KlA365PS2138-1 331 cm N. pachyderma sin. 23,100#240/!240 27,185 KlA2744PS2138-1 380 cm mixed forams 34,900#1570/!1310 34,900 KlA1284

Fig. 1 Surface currents and average summer ice conditions in theEuropean sector of the Arctic Ocean. Location of core PS2138-1 isindicated by the solid circle. Abbrevations are TD: Transpolar Drift;BG: Beaufort Gyre; EGC: East Greenland Current; WSC: WestSpitsbergen Current; WSCs: West Spitsbergen Current [submerg-ing]; ESC: East Spitsbergen Current; RAC: Return Atlantic Current;JMPC: Jan Mayen Polar Current. A detailed description of theoceanographic setting in the study area is given in Hebbeln andWefer (1997)

Schachtschabel 1984). The hydrogen index [HI in milli-grams of hydrocarbon (HC) per gram of TOC] was deter-mined by means of Rock-Eval pyrolysis. In immaturesediments, hydrogen indices of (100 mgHC/g TOC aretypical of TOM, whereas organic matter with significantamounts of MOM have HI-values of 200—400 mgHC/gTOC (Tissot and Welte 1984).

Stable carbon isotope ratios of the organic fractionwere determined on decarbonated samples using a Finn-igan MAT Delta-S mass spectrometer (AWI, Potsdam).Accuracy was checked by parallel analysis of international

standard reference material (IAEA-CH-7). Results areexpressed vs. Vienna PDB. d13C

03'values between

!24 and !27& are widely regarded as an indicator ofTOM supply, whereas heavier values indicate significantamounts of MOM in Arctic Ocean sediments (Ruttenbergand Gon8 i 1997).

The dolomite content was determined by means ofa Phillips PW 3020 diffractometer equipped with a cobaltka radiation. Measurements were performed from 2 to100° theta with a 0.02° theta step/s. Detailed description ofpreparation and analytical processing are outlined inVogt (1997).

To estimate the amount of IRD, which is assumed to bedelivered by icebergs and sea ice and which is generallyaccepted to be a useful tool for reconstructing the activityof glaciers on land, the fraction '2 mm was counteddowncore on X-ray radiographs of each centimeter, ap-plying the method of Grobe (1987). The coarse fraction'63 lm was determined by sieving 5-cm3 sediment overa 63-lm mesh. Percentages were calculated after exclud-ing biogenic particles and detrital carbonate.

Mass accumulation rates grams per square centimeterper thousand years of bulk sediment (AR

"6-,) and of indi-

vidual components like total organic carbon (ARTOC

),carbonate (AR

#!3") and the coarse fraction (AR

;63 lm) arecalculated following Van Andel et al. (1975).

Several samples were chosen for accelerator mass spec-trometry (AMS) 14C dating (Leibniz Lab of Kiel Univer-sity) (Table 1). The 14C dates are d13C-normalized andcorrected for reservoir effects equal to 440 yr (Mangerudand Gulliksen 1975). The radiocarbon age was calibratedto a calendar age using the program Calib 3.0 (Stuiver andReimer 1993) for ages(18 14C ka and an extended sec-ond-order fit for the period'18 14C ka (E. Bard, personalcommunication 1994).

Stratigraphy

The stratigraphic framework is based on oxygen andcarbon isotope records of planktic foraminifera N. pachy-derma sin. (Fig. 2). The definition of oxygen isotope stages(OIS) and their conversion into absolute ages follow thetime scale of Martinson et al. (1987). The stratigraphicalcontrol is further modified by several AMS 14C datings(Table 1).

196

Fig. 2 Isotope stratigraphy of core PS2138-1 based on stable oxy-gen and carbon isotope records of the planktic foraminifer N.pachyderma sin. and several AMS 14C datings (arrows). Isotopestages and events are related to Martinson et al. (1987). The abbreva-tion P.b. indicates the appearence of the benthic foraminiferP. bulloides within substage 5.5—5.3 and 5.1 (J. Wollenburg, personalcommunication)

Although the global oxygen isotope signal may be sig-nificantly compromised by local meltwater events, therecord reflects OIS 1 through upper OIS 6 in detail.Oxygen isotopic events 6.3 and 6.2 are indicated by a shiftto lighter d18O values (3.7&) and heaviest d18O values(4.6&), respectively (Martinson et al. 1987). The lightd18O value (2.1&) at 542 cm core depth is attributed tothe influence of light isotopic deglacial water and is inter-preted as representing the OIS 6/5 boundary (Termina-tion II). Substage 5.5 is indicated by light d18O values.Substages 5.1 and 5.5—5.3 (Pullenia bulloides occurs be-tween substages 5.5 and 5.3 (J. Wollenburg, personalcommunication)) are indicated by the occurrence of thebenthic foraminifera Pullenia bulloides (Haake andPflaumann 1989; Lloyd et al. 1996). The OIS 5/4 bound-ary is indicated by a marked shift in the oxygen isotoperecord (0.7&) to heavier values. The OIS 4/3 boundary iscoincident with an enhanced input of IRD (Fig. 3), datedin the Fram Strait to approximately 50—38 ka (Hebbelnand Wefer 1997) and the Mid-Weichselian deglaciation on

Svalbard (Mangerud and Svendsen 1992) (Fig. 3). TheOIS 3/2 boundary is indicated by a shift to heavier d18Ovalues. The beginning of the last deglaciation (Termina-tion I) is dated at 15.4 14 C ka and is well defined by thetransition to low d18O values and prominent d13C min-ima. The OIS 1/2 boundary is identified by means of themeasured AMS14 C age at 12.8 ka. A general d18O de-crease to Holocene levels is observed at 60 cm. The agecontrol points and oxygen isotope events according toMartinson et al. (1987) were converted to calendar yearsand then linearly interpolated between these points todetermine the numerical age for each sample, assuminguniform sedimentation rates.

Analytical data

The TOC contents range between 0.3 and 1.5% but are ingeneral(1% (Fig. 3). Based on C/N ratios ('10), Rock-Eval data (HI(100 mg HC/g TOC), and d13C

03'values

((!24&), the organic matter is mainly of terrigenousorigin. Lower C/N ratios are caused by high amounts ofinorganic nitrogen in illite-enriched sediments (Knies un-published data). Slightly higher amounts of MOM indicatedby HI'100 mg HC/g TOC and d13C

03'values '!23&

were preserved during substage 6.3, OIS 5 and 4, early OIS 3,and OIS 1. The carbonate contents range between 0 and20%. The highest calcite (without detrital dolomite) contentsare observed during OIS 6, late OIS 3 and OIS 2, whereasdetrital dolomite (up to 4%) are predominant during OIS5 and 1. Variations in calcite concentrations are causedmainly by periodic changes in the levels of planktic andbenthic foraminifera. Increased coarse fraction percentages'63 lm are slightly correlated to higher IRD input('2 mm) and occur significantly during deglacial periodsand to a lesser extent during glacials and interglacials.

Discussion

Indications of Atlantic water advection during OIS 2 and 6

Two short periods (27—22.5 and 19.5—14.5 14C ka) of rela-tively warm Atlantic water advection have been reportedduring the Late Weichselian in the eastern NGS (e.g.,Hebbeln et al. 1994). As a regional moisture source, theseperiods had a major influence on the SBIS build-up (He-bbeln et al. 1994) and also caused an increased productionof coccoliths and subpolar planktic and benthic foraminif-era (Hebbeln and Wefer 1997). Recurring ice-free condi-tions due to Atlantic water advection to the Fram Straitand NGS are also reported during OIS 6 for at least threetime periods (145, 165 and 180 ka) (e.g., Lloyd et al. 1996;Hebbeln and Wefer 1997).

The predominance of biogenic calcite rather than detri-tal dolomite in deposits along the northern Barents Seamargin during mid and late OIS 3 and OIS 2 may reflect

197

198

bFig. 3 Sediment characteristics of gravity core PS2138-1 (NE-Nordaustlandet). Compilation of total organic carbon (wt. %),carbonate and dolomite content (wt. %), carbon/nitrogen ratios,hydrogen indices (mgHC/gTOC), d13Corg (& , vs. V-PDB), ter-rigenous coarse fraction (without carbonate), ice-rafted detritus('2 mm/10 ccm), and lithological description by X-radiographs ver-sus depth (cmbsf)

the paleoceanographic situation described in the easternNGS with a distinct Atlantic water influx (Fig. 3).

Comparable high concentrations of calcite (up to 15%)during upper OIS 6 resemble the situation during OIS3 and 2 (Fig. 3). Additionally, the abundance of the ben-thic foraminifera Cassidulina teretis during upper OIS6 and OIS 3 and 2 (J. Wollenburg, unpublished data)indicates that at least a subsurface body of Atlantic waterreached the northern Barents Sea margin (e.g., Mackensenet al. 1985). The increased occurrences of planktic andbenthic foraminifera point to seasonally ice-free condi-tions during glacial periods because the production ofzooplankton is largely a function of irradiance and nutri-ent concentrations and, thus, of open water conditions(Smith 1995). However, owing to an extensive sea icecover and only short-term Atlantic water subsurface ad-vection during OIS 2 and 6 compared to OIS 1 and 5 (seelater discussion), we assume that highest accumulationrates of biogenic calcite (up to 6 g/cm2/ky) on the northernBarents Sea margin are probably not related just to Atlan-tic water inflow. Hebbeln and Wefer (1991) and Kohfeldet al. (1996) argued that favorable conditions for biolo-gical productivity and high lithogenic flux resulting fromthe release of IRD by melting (Fig. 4) will also be foundaround polynyas that are caused by the upwelling ofrelatively warm water. For example, investigations byKohfeld et al. (1996) in the North East Water polynya (NEGreenland) confirm that abundances of N. pachydermasin. show maximum fluxes occurring at the same time asmaximum carbon fluxes to the sediments. In contrast,fluxes of N. pachyderma sin. decrease to zero below com-plete ice cover. Therefore, we suggest that an expansion ofa coastal polynya at least to the northern Franz VictoriaTrough, triggered by katabatic winds from the protrudingSBIS and an inflow of subsurface Atlantic water masses,may have supported the seasonally ice-free conditions andthe enhanced flux of biogenic calcite, and provided addi-tional moisture to build-up the SBIS along the northernBarents Sea margin during glacial OIS 3, 2, and 6 (Fig. 5).A continuous terrigenous input along the ice edge or atthe grounding lines of the ice sheet, documented by thehigh lithogenic flux (Fig. 4), and a subsequent intensivesupply of nutrients probably induced by upwelling ofAtlantic water may have fostered the foraminiferal pro-duction in the water column. Indications of higheramounts of MOM, which point to increased surface-waterproductivity, do not occur during OIS 2 and 6 (Fig. 3)because highest accumulation rates of terrigenous organic

material (up to 0.42 g/cm2/ky) reflected by C/N ratios'15, low HI values, and light d13Corg values due tomelting processes probably dilute the marine organic sig-nal (Fig. 4). Visual inspection of the coarse fraction cor-roborates the high input of glacially reworked TOC-richsiltstones, which are indicative for a Svalbard/Barents Seasource, and confirms the high terrigenous organic accu-mulation rates. A similar scenario from the Antarctic icesheet was proposed by Melles (1991). He argued, based onhigher foraminiferal abundances in upper slope sedimentsof the Weddell Sea during OIS 2, that a coastal polynyatriggered by katabatic winds existed in front of the Ant-arctic ice sheet. A seasonally coastal polynya in front ofthe protruding SBIS at least during OIS 2 (and presum-ably during upper OIS 6) (Fig. 5) would explain the dis-tinct shift in sedimentation rates from 5 cm/ky during latesubstage 3.1 up to 38 cm/ky during OIS 2 due to release ofIRD by melting of the nearby ice-sheet (Fig. 5). However,compared to OIS 2, the reduction of sedimentation ratesby a factor of 4 or 5 resulted in lower terrigenous andbiogenic fluxes for the upper OIS 6 (Fig. 4). Although thediminished sedimentation rates might be, at least partly,artificial due to a lower stratigraphic resolution, they mayalso reflect a closer sea ice cover and/or reduced meltingprocesses during upper OIS 6 in relation to OIS 2.

Deglaciation patterns along the northernBarents Sea margin

Deglaciation patterns in the Fram Strait and the NGS areindicated by an intense input of IRD from the surround-ing ice sheets (e.g., Elverh+i et al. 1995; Fronval andJansen 1997; Hebbeln and Wefer 1997). Distinct layers ofsand and gravel and the IRD data indicate recurringperiods of increased glacial activity and enhanced supplyof terrigenous material by glaciomarine processes alongthe northern Barents Sea margin (Fig. 3). In OIS 6, higheramounts of IRD were deposited during the deglacialsubstage 6.3. Short-term IRD events during substage 6.2correlate with layers of gravel and sand, light d13C

03'values, and biogenic calcite peaks, and probably indicatehigher calving rates of the SBIS. Differences in the IRDcomposition during substages 6.3 and 6.2 indicated byd13C

03'values may delineate different source regions out-

lined in Hebbeln and Wefer (1997).The OIS 6/5 boundary is marked by a strong meltwater

signal (d18O: 2.1& ; d13C: !0.56&) and an increasedsupply of IRD. Prior to this event, a second IRD peak wasrecorded in late OIS 6. Despite a lower stratigraphicresolution of OIS 6, this peak could be correlated with theinitial melting of the ice sheets surrounding the NGS inlate OIS 6 (before 130 ka) suggested by Fronval andJansen (1997). A distinctly lower input of coarse-grainedmaterial and IRD during OIS 5 document that most ofthe SBIS had already disappeared at the beginning of theEemian. Moderate IRD input and low levels of coarse-grained material in mid OIS 5 do not reflect a glacier

199

Fig. 4 Sedimentation rate (cm/ky), total mass accumulation rate(ARbulk), accumulation rate of total organic carbon (ARTOC), car-bonate (ARCaCO

3), dolomite (AR

Dolomite) and terrigenous coarse

fraction (AR'63 lm) (all in g/cm2/ky) versus calibrated calendar years

advance as known from the onshore glacial history ofSvalbard (Mangerud and Svendsen 1992).

At the OIS 4/3 boundary and early OIS 3, a distinctlyincreased supply of IRD was caused by enhanced ice-berg drifting produced by the rapid retreat of the SBISduring the Mid-Weichselian deglaciation on Svalbard(Mangerud and Svendsen 1992). However, in contrast tothe LGM and upper OIS 6, OIS 4 is marked by very lowsedimentation rates, carbonate dissolution and heavierd13C values and, thus, does not seem comparable with theSaalian and Late Weichselian glaciation.

The SBIS build-up during the Late Weichselian is in-dicated by a very low input of IRD on the northernBarents Sea margin. Only short-term pulses of IRDaround 20 ka and 16.8 ka may reflect distinct deglaciation

signals in between periods of glacier growth within theLate Weichselian. In contrast, Elverh+i et al. (1995) sug-gested that the IRD pulse around 20 ka documents thefirst significant ice advance beyond the present coastlineon the western Svalbard margin.

The first distinct signal of ice recession on the northernBarents Sea margin, which probably reflects the onset ofSBIS decay, is marked by a small but significant input ofIRD around 16.2 ka. This is consistent with a first distinctdeglacial phase at 16 ka at the southwestern Barents Seamargin proposed by Vorren et al. (1988). The drasticclimatic change and the gradual retreat of the SBIS on thenorthern Barents Sea during the last deglaciation (Termi-nation I) is documented by a maximum pulse of IRDat 15.4 ka. A subsequent second pulse of deglaciationmarked by a distinct meltwater event (d18O: 3.2&; d13C:!0.57&) and high IRD input occurred between 14 and13 ka. This is widely consistent with the deglaciation pat-terns on the East Greenland margin (Nam et al. 1995) andwell correlated with major meltwater anomalies in theArctic Ocean and the Fram Strait (e.g., Jones and Keigwin

200

Fig. 5 Paleoenvironmentalmodel of the SBIS during LastGlacial Maximum (at 19 ka).Extension of ice sheets andAtlantic water advection arebased on Vorren et al. (1988) andHebbeln et al. (1994). Allinvestigated cores (except PS2447and PS2448) show enhancedabundances of planktic andbenthic foraminifera in the coarsefraction ('63 lm) during lateOIS 3 and OIS 2 (Knies,unpublished data) and can becorrelated to PS2138-1. (WSC:Westspitsbergen Current; WSCs:submerging WestspitsbergenCurrent)

1988, Stein et al. 1994). Although the stratigraphic resolu-tion of Termination I is much higher and the IRD eventsare more significant compared to Termination II, the twosmaller IRD peaks at the transition of OIS 6 to OIS 5 mayreflect similar deglaciation patterns. The lower magnitudeof IRD events during Termination II might be explainedby a more extensive sea ice cover, which might haveprevented higher supply and accumulation of IRD at thecontinental slope (Nam et al. 1995).

Low to moderate IRD input during the Holocene iscomparable with OIS 5 and does not indicate majorglacial activity.

Interglacial sedimentary characteristics

According to Hebbeln and Wefer (1997), substages 5.5, 5.1,and the Holocene were the globally warmest periods inthe Fram Strait over the last 180 ka. However, indicationsfor a surface ocean warming influenced by an intensiveAtlantic water inflow and thus higher input of MOM onthe northern Barents Sea margin are not documentedby the data (Fig. 3). Only heavier d13C

03'values (up to

!22.5&) indicate a slightly higher amount of MOM inthe sediments (Fig. 3). Additionally, the carbonate contentused as a surface water productivity indicator in the FramStrait (Hebbeln and Wefer 1997) is also of limited value inthe study area. Particularly, the transitions from Termina-tions II and I to interglacial periods OIS 5 and OIS 1 arecharacterized by an extraordinary drop of biogenic cal-cite. The carbonate content is completely composed ofdetrital dolomite (Fig. 3). According to Steinsund andHald (1994), the dissolution of biogenic calcite in sedi-ments of the eastern and northern Barents Sea has in-creased during deglacial and interglacial periods due to

a combination of Atlantic water influx, the annual forma-tion of sea ice, dense bottom water formation, and surfacewater productivity blooms. During seasonal sea iceformation in the Barents Sea brines are ejected by sea ice.These brines form a highly saline, CO

2- and oxygen-en-

riched water mass that descends in troughs and depress-ions on the eastern and northern Barents Sea margin(Steinsund and Hald 1994). Oxidation of MOM, which ishighly accumulated near the ice edge (Hebbeln and Wefer1991), is enhanced under such conditions and causes thenet increase of metabolized CO

2. When such a dense

water mass overlies the organic-rich sediments, the degra-dation products of MOM will not be recirculated into thesurface water, but will be concentrated in the bottomwaters and thus reinforce dissolution of biogenic calcite.

Hence, we conclude that dissolution of biogenic calciteand low MOM reflect the strongest advection of Atlanticwater and enhanced surface-water productivity along thenorthern Barents Sea continental margin during OIS5 and 1 and presumably during early OIS 3 and substage6.3. Based on heavier d13C

03'values and virtually com-

plete dissolution of biogenic calcite, we also speculatethat during OIS 4 a continuous influx of Atlantic wateroccurred.

Conclusions

Our data document a continuous influx of Atlantic wateralong the northern Barents Sea margin during the last150 ka. During the Late Weichselian and Saalian gla-ciations, moisture supply enhanced by seasonally variableAtlantic water advection and coastal polynyas, increasedsnowfall, and a minimal loss of ice as indicated by a lower

201

input of IRD accelerated the final SBIS build-up. Thecoastal polynya in front of the SBIS expanded at least tothe northern Franz Victoria Trough. Varying inputs ofIRD indicate major calving events during the last 150 ka.The almost complete dissolution of biogenic calcite duringthe interglacials was controlled mainly by dense, cold,saline, and CO

2-rich bottom water and subsequent oxida-

tion of higher contents of MOM. This indicates the stron-gest advection of Atlantic water and a stable marginal icezone during OIS 5, the Holocene, and presumably duringearly OIS 3.

Acknowledgments We thank the captain and the crew of theR/» Polarstern for their cooperation during the ARKVIII/2-expedi-tion. Thanks to J. Hefter, J. Matthiessen, and C. Schubert for veryhelpful comments on an earlier draft and to two anonymous re-viewers for formal review. For mass spectrometer operations,G. Meyer, G. Traue, and L. Schoenicke are greatly acknowledged.This is contribution No. 1401 of the Alfred Wegener Institute forPolar and Marine Research.

References

Broecker W, Bond G, and Klas M (1990) A salt oscillator in theGlacial Atlantic. Paleoceanography 5 : 469—477

Charles CD, Rind D, Jouzel J, Koster RD, and Fairbanks RG (1994)Glacial—interglacial changes in moisture sources for Greenland:Influences on the ice core record of climate. Science 263 : 508—511

Dokken TM and Hald M (1996) Rapid climatic shifts during isotopestages 2—4 in the Polar North Atlantic. Geology 24(7) : 599—602

Elverh+i A, Andersen ES, Dokken T, Spielhagen R, Svendsen JI,S+rflaten M, R+rnes A, Hald M, and Forsberg CF (1995) Thegrowth and decay of the Late Wechselian ice sheet in westernSvalbard and adjacent areas based on provenance studies ofmarine sediments. Quaternary Research 44 : 303—316

Fronval T and Jansen E (1997) Eamian and early Weichselian(140—60 ka) paleoceanography and paleoclimate in the Nordicseas with comparisons to Holocene conditions. Paleoceanogra-phy, 12 : 443—462

Grobe H (1987) A simple method for determination of ice rafteddebris in sediment cores. Polarforschung 57 : 123—126

Haake FW and Pflaumann U (1989) Late Pleistocene foraminiferalstratigraphy on the Voring Plateau, Norwegian Sea. Boreas18 : 343—356

Hebbeln D and Wefer G (1991) Effects of ice coverage and ice-raftedmaterial on sedimentation in the Fram Strait. Nature350 : 409—411

Hebbeln D and Wefer G (1997) Late Quaternary paleoceanographyin the Fram Strait. Paleoceanography 12 : 65—79

Hebbeln D, Dokken T, Andersen ES, Hald M, and Elverh+i A (1994)Moisture supply for northern ice-sheet growth during the LastGlacial Maximum. Nature 370 : 357—359

Jones GA and Keigwin LD (1988) Evidence from Fram Strait (78°N)for early deglaciation. Nature 336 : 56—59

Kellogg TB (1976) Late Quaternary climatic changes: Evidence fromdeep-sea cores of Norwegian and Greenland Seas. In: Cline RM

and Hays JD (Eds.), Investigation of Late Quaternary Pale-oceanography and Paleoclimatology. Geolocal Society AmericaMemoir 145 : 77—110

Kohfeld KE, Fairbanks RG, Smith SL, and Walsh ID (1996) Neo-gloquadrina pachyderma (sinistral coiling) as paleoceanographictracers in polar oceans: Evidence from Northeast Water Polynyaplankton tows, sediment traps, and surface sediment samples.Paleoceanography 11 : 679—700

Lloyd JM, Kroon D, Boulton GS, Laban C, and Fallick A (1996) Icerafting history from the Spitsbergen ice cap over the last 200 kyr.Marine Geology 131 : 103—121

Mackensen A, Sejrup HP, and Jansen E (1985) The distribution ofliving benthic foraminifera on the continental slope and rise offsouthwest Norway. Marine Micropaleontology 9 : 275—306

Mangerud J and Gulliksen S (1975) Apparent radiocarbon ages ofrecent marine shells from Norway, Spitsbergen, Arctic Canada.Quaternary Research 5 : 273—296

Mangerud J and Svendsen JI (1992) The last interglacial—glacialperiod on Spitsbergen, Svalbard. Quaternary Science Reviews11 : 633—664

Martinson DG, Pisias NG, Hays JD, Imbrie J, Moore TC, andShakleton NJ (1987) Age dating and the orbital theory ofthe ice ages: Development of a high-resolution 0 to 300 000 yearschronostratigraphy. Quaternary Research 27 : 1—27

Melles M (1991) Late Quaternary paleoglaciology and paleoceano-graphy at the continental margin of the southern Weddell Sea,Antarctica. Reports on Polar Research 81 : 190 pp

Nam S-I, Stein R, Grobe H, and Hubberten H (1995) Late Quarter-nary glacial—interglacial changes in sediment composition at theEast Greenland continental margin and their paleoceanographicimplications. Marine Geology 122 : 243—262

Rachor E (1992) Scientific report of RV ‘‘Polarstern’’ cruise ARK-VIII/2. Reports on Polar Research 115 : 150 pp

Ruttenberg KC and Gon8 i MA (1997) Phosphorus distribution, CNPratios, and d13C

03'in arctic, temperate, and tropical coastal

sediments: Tools for characterizing bulk organic matter. MarineGeology 139 : 123—145

Scheffer F and Schachtschabel P (1984) Lehrbuch der Bodenkunde:Stuttgart: Enke Verlag, 442 pp

Smith WO (1995) Primary productivity and new production in theNortheast Water (Greenland) Polynya during summer 1992.Journal of Geophysical Research 100(C3) : 4341—4356

Stein R, Nam S-I, Schubert C, Vogt C, Futterer D, and HeinemeierJ (1994) The last deglaciation event in the Eastern Central ArcticOcean. Science 264 : 692—696

Steinsund PI and Hald M (1994) Recent calcium carbonate dissolu-tion in the Barents Sea: Paleoceanographic applications. MarineGeology 117 : 303—316

Stuiver M and Reimer PJ (1993) Extended 14C data base and revisedCALIB 3.0 14C age calibration program. Radiocarbon 35 :215—230

Tissot BP and Welte DH (1984) Petroleum Formation and Occur-rence. Berlin: Springer-Verlag, 699 pp

Van Andel TH, Heath GR, and Moore TC (1975) Cenozoic historyand paleoceanography of the central equatorial Pacific.Geolocial Society America Memoir 143 : 134 pp

Vogt C (1997) Regional and temporal variations of mineral assem N.pachyderma sin.blages in Arctic Ocean sediments as climaticindicator during glacial/interglacial changes. Reports on PolarResearch 251 : 335 pp

Vorren TO, Hald M, and Lebesbye E (1988) Late Cenozoic environ-ment in the Barents Sea. Paleoceanography 3 : 601—612

202