Vogt et al 2001
Transcription
Vogt et al 2001
Global and Planetary Change 31 Ž2001. 23–44 www.elsevier.comrlocatergloplacha Detailed mineralogical evidence for two nearly identical glacialrdeglacial cycles and Atlantic water advection to the Arctic Ocean during the last 90,000 years Christoph Vogt a,) , Jochen Knies b,1, Robert F. Spielhagen c,2 , Ruediger Stein b,1 a FB Geowissenschaftenr Geo Sciences, UniÕersitat ¨ Bremen, Post Box 330440, 28334 Bremen, Germany Alfred Wegener Institute for Polar and Marine Research, Columbusstr., D-27568 BremerhaÕen, Germany GEOMAR Research Center for Marine Geosciences, Kiel UniÕersity, Wischhofstr. 1-3, D-24148 Kiel, Germany b c Received 5 December 1999; accepted 23 May 2001 Abstract Three cores recovered off the northwest of Svalbard were studied with respect to glacialrinterglacial changes of clay and bulk mineralogy, lithology and organic geochemistry. The cores cover the Late Quaternary Marine Isotope Stages ŽMIS. 6–1 Žca. 170,000 years. and are located in the vicinity of the Polar Front which separates the warm Atlantic water of the Westspitsbergen Current and the cold Polar Water of the Transpolar Drift. Globally driven changes in the paleoenvironment like the variable advection of warm Atlantic water into the Arctic Ocean can be distinguished from regional events by means of source mineral signatures and organic geochemistry data. In particular, a combination of high organic carbon and low carbonate contents, high CrN-ratios, a particular lithology and a distinct bulk and clay mineral assemblage can be related to Svalbard ice sheet developments between 23,000 and 19,500 14C years. This complex sediment pattern has been traced to the northwest of Spitsbergen as far north as 828N. Additionally, the same signature has been recognized in detail in upper MIS 5 sediments. The striking similarity of the history of the SvalbardrBarents Sea Ice Sheet during the late and earlyrmiddle Weichselian is elaborated. Both sediment horizons are intercalated between biogenic calcite rich core sequences which contain the so-called AHigh Productivity ZonesB or ANordway EventsB related to the increased advection of warm Atlantic water to the Arctic Ocean. This study provides further evidence that the meridional circulation pattern has been present during most of the Weichselian and that the ice cover was often reduced in the northeastern Fram Strait and above the Yermak Plateau. Our findings contradict the widely used reconstructions in modelling of the last glaciation cycle and reveal a much more dynamic system in the Fram Strait and southwestern Eurasian Basin of the Arctic Ocean. q 2001 Elsevier Science B.V. All rights reserved. Keywords: Quaternary; Fram Strait; Yermak Plateau; Atlantic water advection; lithology; mineralogy; ice sheet developments 1. Introduction ) Corresponding author. Tel.: q49-421-218-9007; fax: q49421-218-7123. E-mail addresses: [email protected] ŽC. Vogt., [email protected] ŽR.F. Spielhagen., [email protected] ŽR. Stein.. 1 Fax: q49-471-4831-1580. 2 Fax.: q49-431-600-2941. The Arctic Ocean is very sensitive to climatic change and might even drive global oceanographic and climatic changes ŽAlley, 1995.. It is adjacent to the northern boundaries of the northern hemisphere ice sheets. Thus, sediments of the Arctic Ocean have 0921-8181r01r$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 1 - 8 1 8 1 Ž 0 1 . 0 0 1 1 1 - 4 24 C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 great potential to archivate the history of northern hemisphere ice sheets, the fate of the produced icebergs, and changes in the sea-ice cover. In the NW of Svalbard, relative warm Atlantic Water of the northward flowing Westspitsbergen Current ŽWSC. submerges at the Polar Front beneath cold Polar Water ŽFig. 1.. The positions of the Polar Front and the sea-ice edge depends on the strength of the WSC and the outflow of the Polar Water, which is strongly connected to the fresh water influx into the Arctic Ocean Že.g. Anderson et al., 1994; Aagaard and Carmack, 1994.. Previous studies on sediment cores in the Fram Strait and NE of Svalbard revealed rapid changes in the Atlantic Water influx and its influence on the built-up and decay of the SvalbardrBarents Sea Ice Sheet ŽSBIS. ŽHebbeln et al., 1994; Dokken and Hald, 1996; Lubinski et al., 1996; Knies et al., 1999.. Here, we present a detailed lithological, mineralogical, and geochemical dataset of undisturbed late Quaternary sediments NW of Svalbard to elucidate paleoceanographic changes during the Weichselian and to enlarge the information on rapid oceanographical and climatological changes in the Polar Front region NW of Svalbard. The sediment cores are well positioned to the north and the south of the Polar Front to record changes in its development through time ŽFig. 1.. 2. Materials and methods Gravity cores PS2122-1 and PS2123-2 and the Kastenlot core PS2212-3 were recovered from the NW continental margin of Svalbard and the Yermak Plateau during expeditions ARK-VIIIr2 and 3 in summer 1991 with RV Polarstern ŽFutterer, 1992; ¨ Rachor, 1992.. Short sediment cores containing the Fig. 1. Surface currents in the European sector of the Arctic Ocean and locations of investigated cores PS2122-1, PS2123-2, PS2212-3 Žpositions and water depth are listed in Fig. 2. and cores PS1533-3, PS2138-1, and NP90-39 for comparison ŽTPD: Transpolar Drift; TPD Sib : Siberian Branch; BG: Beaufort Gyre; EGC: East Greenland Current; WSC: West Spitsbergen Current; WSC s : West Spitsbergen Current Žsubmerging.; ESC: East Spitsbergen Current; RAC: Return Atlantic Current; JMPC: Jan Mayen Polar Current, compiled from Manley et al., 1992; Hebbeln et al., 1994; Nowaczyk et al., 1994.. Position of Spitsbergenbanken with Jurassic Shale Source rock and proposed transport path of fine fraction material Žafter Andersen et al., 1996.. C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 uppermost undisturbed 30–45 cm were taken from the same sites ŽFig. 1.. We routinely sampled every 5 cm Ž1-cm-thick sediment slice.. Each sample was split into two parts. One part was dried and pulverized, and then divided into subsamples. One of these subsamples Žapproximately 30 mg. was examined for carbonate, organic carbon, and nitrogen using a Heraeus CHNelemental-analyser. Carbonate contents were calculated as CaCO 3 Ž%. s Žtotal carbony total organic carbon ŽTOC.. = 8.333. The carbon and nitrogen measurements have a standard deviation of 0.06% and 0.02%, respectively. The CrN weight ratios were calculated, and hydrogen-index ŽHI. and Tmaxtemperature were determined by Rock–Eval pyrolysis ŽKnies and Stein, 1998; reproducibility "8%.. Another subsample Žabout 3 g. of the dried bulk sample was used for the evaluation of bulk mineralogy by means of XRD-measurements with a Philips PW 3020 diffractometer equipped with cobalt k aradiation, automatic divergence slit, graphite monochromator, and automatic sample changer ŽTable 1.. Individual bulk mineral contents were expressed as percentages of bulk sediment. The quartz content was determined by using the QUAX software package ŽEmmermann and Lauterjung, 1990.. The dolo- 25 mite content was inferred from the peak intensity ratio of dolomite and calcite multiplied with the carbonate content from the elemental analysis and, additionally, controlled using QUAX Žstandard deviation "2% for quartz and "1% for the carbonates; cf. Vogt, 1997.. Peak intensity ratios were used for the evaluation of the quartz and feldspar mineralogy. The second part of the original sample Ž2–3 g. was treated with 3–10% H 2 O 2 to oxidize the organic matter, disaggregated, and finally wet sieved through a 63-mm mesh. The coarse fraction Ž) 63 mm. was studied under the light microscope to gain an overview of the terrigenous components. Five to ten specimens of Neogloboquadrina pachyderma sinistral Žsize: 125–250 mm. were picked for stable isotope measurements. The - 63-mm fraction was separated into silt Ž2–63 mm. and clay Ž- 2 mm. by the Atterberg settling tubes method Žaccording to Stoke’s law; Muller, 1967.. The amount of ice-rafted ¨ debris ŽIRD. was estimated by counting terriginous particles ) 2 mm in each centimeter of an X-ray radiography according to the method of Grobe Ž1987.. The clay mineral assemblage was determined by standard preparation and analysis techniques as outlined by Petschick et al. Ž1996, Table 1.. The peak areas of the clay mineral groups ŽTable 1. were Table 1 Ža. Running conditions of XRD measurements for bulk and clay mineral analysis Žrange in 82 u , stepsize in 82 ur1 s, slow scan 82 ur2 s.. Co k alpha-radiation Measurement Sample preparation Range Stepsize Bulk sample powder unoriented pressed pellets ŽPS2212. 2–100 0.02 Clay fraction ) 18 h glycolated slow scan textural oriented textural oriented textural oriented 2–18 2–40 28.5–30.5 0.02 0.02 0.005 ˚ . for the quantitative evaluation of bulk and clay mineralogy Žthe smectite group includes all expandable Žb. Used XRD peaks Ž d-value in A ˚ OLEMs ordered layered expandable minerals ŽReynolds, 1970., illite includes minerals with a peak of the glycolated sample near 17 A, non expandable mixed layers.. Peaks were recognized graphically by using the MacDiff program ŽPetschick et al., 1996. Bulk and clay fraction sample Clay fraction sample quartz Ž3.34, 4.26., feldspar Ž3.24, 3.18., QzrFsp Ž4.26rŽ3.24 and 3.18.., calcite Ž3.035., dolomite Ž2.89., APyroxene indexB Ž2.995–2.92. OLEM Ž30–22 glyc, 11–12.5., smectite Ž17 glyc., illite Ž10, 5, 4.5., kaolinite Ž7, 3.58., chlorite Ž7, 3.54. 26 C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 calculated and transformed into relative clay mineral percentages by means of Biscaye factors based on the assumption that the clay fraction consists only of clay minerals Žcf. Wahsner et al., 1999; reproducibility: "3%.. 3. Stratigraphy The stratigraphic framework of cores PS2122-1 and PS2123-2 was deduced from their oxygen and carbon isotope records ŽFig. 2., which are, in general, fairly correlateable to the global isotope curve ŽSPECMAP stack; Martinson et al., 1987.. An AMS 14 C date in PS2122-1 at 321 cm bsf Žcm below core surface. helps to pinpoint Marine Isotope Stage ŽMIS. 3. The stratigraphic framework is also corroborated by the occurrence of the benthic foraminifera Pullenia bulloides, a stratigraphic marker for the MIS 5r4 boundary in the Nordic Seas Žcf. Haake and Plaumann, 1989.. Additionally, a typical decrease of Fig. 2. Chronostratigraphies and d18 O- and d13 C-records of cores PS2122-1rPS2123-2 and PS2212-3: Marine Isotope Stage ŽMIS. assignments are based on Martinson et al. Ž1987., an AMS-14 C date at 321 cm in PS2122-1 Žbivalve shell, sample KIA367 measured at the Leibniz Laboratory, Kiel University: 36800" 3030 14 C-years; reservoir correction: 440 years; Mangerud and Gulliksen, 1975., and the occurence of the benthic foraminifera P. bulloides at the MIS 5r4 boundary. Additionally, the chronostratigraphy of core PS2212-3 is based on the correlation of lithological, geochemical and mineralogical data between NP90-39, PS2138-1 and PS2212-3 Žcompare Figs. 1 and 3: AMS 14 C-dated Event I., the occurrence of S. rolshauseni ŽWollenburg et al., 2001. and coccolith abundances and paleomagnetic data of Nowaczyk et al. Ž1994. Žincluding a hiatus in MIS 5; NGS—Norwegian–Greenland-Sea-Event.. C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 the carbon isotope values from late MIS 5 to lowest values in early MIS 3 is observed and used as a regional stratigraphic fixpoint Žcf. Dokken and Hald, 1996; Nørgaard-Pedersen et al., 1998.. Hence, the cores PS2122-1 and PS2123-2 most likely represent MIS 1–5. Well-dated melting events during Termination I Že.g. at 14.5 ka. are also recognized Žcf. Hebbeln et al., 1994.. All ages are reported according to the SPECMAP timescale Žka: 1000 years.. A well defined sequence of sedimentological, mineralogical and organic–geochemical parameters occurs at the MIS 3r2 boundary in all sediment cores ŽŽFigs. 3, 4 and 8.; cf. Andersen et al., 1996.. The central part of this layer exhibits lamination and contains high amounts of mature terrestrial organic material, very low carbonate Žbeing mainly dolomite. and a very distinct Žclay. mineralogy with extremely low smectite, but high kaolinite percentages accompanied by the occurrence of ordered layered expandable minerals ŽOLEM; Fig. 3; cf. Andersen et al., 1996.. It is labeled Event I in accordance with Knies and Stein Ž1998. and has been deposited between 22.5 and 19.5 14 C ka as based on several AMS 14 C-ages ŽFig. 3; approximately 26–22 ka calendar 27 years according to Voelker et al., 1998.. Here, we assume it to be a synchronous deposit in all three sediment cores and use it as one indicator of the lower MIS 2. The stable isotope records of core PS2212-3 generally agree with the outlined stratigraphic concept, although several sediment horizons are barren of carbonate Že.g. Termination I, Figs. 2 and 4.. Therefore, correlation of mineralogical and organic–geochemical parameters with adjacent well dated cores help to determine the exact position of the MIS 2r1 boundary at 40 cm bsf Žcf. Pagels, 1991 for the age of carbonate free horizons during Termination I; Stein et al., 1994.. Similarly, we use the Event I Ž19.5–22.5 ka. which occurs in PS2212-3 at about 110 cm bsf as one important stratigraphic marker ŽFigs. 2–5. to position the MIS boundary 3r2 at 112 cm bsf. In addition, the small d18 O-shift to higher values at 95 cm bsf and lower values of d13 C could indicate early MIS 2 ŽFig. 2.. Here and between 120 and 140 cm bsf, Wollenburg et al. Ž2001. found two occurrences of the agglutinated benthic foraminifera Siphotextularia rolshauseni, the lower being indicative of MIS 3.2 in the Nordic Seas ŽNees and Struck, Fig. 3. Sediment characteristics versus depth of Core 90-39 ŽAndersen et al., 1996. and radiocarbon dates. Marine Isotope Stages 1–3 are indicated by bold numbers and grey shade Žinterglacialsrinterstadials.. 28 C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 Fig. 4. Sediment characteristics versus depth of Kastenlot core PS2212-3 ŽYermak Plateau., gravity core PS2122-1 and gravity core PS2123-2 ŽSpitsbergen Coast Domain.. Marine Isotope Stages 1–6 are indicated by bold numbers and grey shade Žinterglacialsrinterstadials.. C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 ˚ peak.rfeldspar Ž3.24 and 3.18 A˚ peaks.-ratio and Fig. 5. Mineralogical data of cores PS2122-1, PS2123-2, and PS2212-3 versus depth. Bulk quartz content and quartz Ž4.26 A ˚ ., and relative weight percents of clay mineral groups in the clay fraction Ž - 2 mm. are shown. k-feldsparrplagioclase ratio Ž3.24r3.18 A 29 30 C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 1994.. As in PS2122-1 and in PS2123-2, the occurrence of P. bulloides and the decrease of the carbon isotope values delineate the MIS 5r4 and MIS 4r3 boundaries. Noticeably, the carbonate free horizon is again situated in a deglaciation event at MIS 4r3 boundary. Based on this new data, our stratigraphy differs in MIS 1–3 from the published framework of Nowaczyk et al. Ž1994. derived from paleomagnetic and coccolith evidence which is discussed in more detail in Mathiessen et al. Ž2001.. Even with our improvements absolute ages might still differ by up to a few thousand years. 4. Results 4.1. Lithology, IRD and grain-size distribution In all cores the dominant lithotypes are silty clays to clayey silts. Brown colours dominate the MIS 1–5 in core PS2212-3 while cores PS2122-1 and PS2123- 2 show regular alternations of brown and grey sediment colours ŽKnies, 1994; Vogt, 1997.. The cores usually contain a few layers of coarser material which, however, are only visible in the X-ray radiographs. The coarse fraction content ranges between 0 and 25 wt.% with one exception in core PS2122-1 Ž46 wt.% in MIS 1; Figs. 4 and 6.. In the nearshore sediment cores, PS2122-1 and PS2123-2, the coarse fraction is dominated by inorganic terrigenous components. In PS2212-3, sediment layers with lower coarse fraction content are mainly dominated by foraminifera shells Žsizes from 63–500 mm.. In the Western Eurasian Basin, high numbers of calcareous foraminifers mainly occur in intervals with low sand contents ŽFig. 4. where they might constitute up to 50% of the total sand fraction ŽNørgaard-Pedersen et al., 1998.. Dominance of foraminifera in a particular sediment layer can be traced by high carbonater calcite contents of the bulk and silt fraction ŽFig. 4, MIS 3r2.. In general, the pattern of terrigenous IRD and coarse fraction content seems to be very similar Fig. 6. Relation of gravel and sand content of bulk sediment Ž%. to oxygen isotope stratigraphy age ŽMartinson et al., 1987. in cores PS2122-1, PS2123-2 and PS2212-3. C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 Fig. 7. Compilation of the glacial history of Svalbard Žadvance and retreat. along the Isfjord basin for the Weichselian period Žfrom Mangerud et al., 1998., abundance of ice-rafted material ŽIRD. of the cores PS2122-1, PS2123-2 and PS2212-3 and oxygen isotope stratigraphy ŽPS2122-1.. Gray shaded areas show major IRD events. Cross hatched area is Event I. 31 32 Table 2 Brief characterization of main sedimentary environments by means of sediment data used in this paper Bulk mineralogy Clay mineralogy Lithology IRD, coarse fraction Sedimentation rate Oxygen and carbon isotopes Carbonate Organic carbon Open water Quartz- 25%, yFsp, yyQzrFsp, yyKfsrPlg Ill and Chl ) 70% mainly fine grain sizes, brown colors very low low global ocean values qq, qqcalcite, forams, coccoliths yyTOC, qHIvalues Ž )100., yCrN-ratio, low Tma x , qmarine origin Melting at the Marginal Ice Zone qŽclino-. pyroxene qqSmectite, qqKaolinite, qqQzrFsp qqfine fraction, cryokonites, fecal pellets low large some lighter oxygen values medium carbonate, some dissolution, large dilution qTOC, qqHI, qpreservation, mixed maturity Sea-ice cover medium to low Qz and Fsp depending on origin of sea-ice Ill and Chl dominant due to mainly gravitational transport from Svalbard qfine fraction sand contains mainly forams low low stable heavy oxygen values, continously light carbon values qqcalcite due to small dilution by terr. material, good preservation medium TOC, low HI, qCrN, qTma x , terrigenous origin Built-up and deglaciation of adjacent ice sheets rapid changes, built up: qqAmatureB mineralogy, qqQz, yyFsp, yyKfs, deglaciation; qqQz, qqKfs, qqAmf qq kaolinite and OLEM appearance, qqQzrFspratio, rapid changes in clay mineralogy highly variable lamination, also hiatus due to erosional processes possible qqIRD, qqsand fraction largest several meltwater pulse, light oxygen and carbon values ycalcite dues to dilution and strong dissolution, dolomite from Svalbard sources qqTOC, qq CrN, Ahot shaleB organic material, qqpreservation of allochthonous and authochtonous material Relative scale: Žqq. strong inputrincreased content to Žyy. lowest inputrcontent. Qz—quartz; Fsp—feldspar; Plg—plagioclase; Kfs—K-feldspar; Ill—Illite; Chl—Chlorite; Amf—amphibolesrhornblende; OLEM—ordered layered expandable minerals; IRD —ice rafted debris, dropstones; TOC—total organic carbon; CrN—organic carbonrtotal nitrogen ratio; HI-value, Tma x —data from pyrolysis of organic material. C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 Environment C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 in all cores ŽFigs. 6 and 7.. In the deep sea ŽPS2212-3. and lower slope cores ŽPS2122-1., however, the abundance of IRD is lower compared to the near coast core PS2123-2 ŽFigs. 4 and 7.. 4.2. Carbonate and organic carbon contents The carbonate content in cores PS2122-1 and PS2123-2 is mostly less than 5% ŽFig. 4.. Only during MIS 2 values of up to 10% occur. In core PS2123-2 dolomite contents vary between 0% and 4% often comprising most of the carbonate content. Only in upper MIS 5 sediments and at the MIS boundary 3r2 calcite is the dominant carbonate mineral. In core PS2212-3 the carbonate content is highly variable throughout the whole core ŽFig. 4.. At those intervals without N. pachyderma sin. Žsee missing isotope values, Fig. 2. but with significant carbonate content, the carbonate is mainly dolomite with traces of siderite. In all three cores the total organic carbon content ŽTOC. ranges between 0.4% and 1.0% with distinct maxima from 1.5% to 2.7% in MIS 6, lower MIS 5 and at the MIS boundaries 4r3, 3r2 and Termination I ŽFig. 4.. The results of the geochemical investigations reveal a dominant terrigenous input which is consistent with other findings in the Eurasian Basin Že.g. Schubert and Stein, 1996; terrigenous organic carbon: CrN-ratios larger than 15, HI-values below 100 mg HCrg C; Tmax-values above 450 8C.. The values of the CrN-ratio range between 5 and 35. Especially core PS2212-3 displays higher CrNratios above 10, and the HI-values are entirely below 100 mg HCrg C. In all cores the upper MIS 1 sediments show low CrN-ratios and, additionally, in PS2122-1 and PS2123-2 higher HI-values, suggesting a small increase in marine organic material in combination with increased Žbiogenic. calcite contents ŽFig. 4.. 4.3. Mineralogy The mineral assemblage in cores PS2123-2 and PS2212-3 mainly consists of quartz, feldspar, clay minerals, calcite, dolomite, and accessory heavy minerals. Core PS2123-2 reveals quartz contents of 19% to nearly 50% Žaverage 35.5%., core PS2212-3 12% to 45% Žaverage 29.5%; Fig. 4.. The quartzr 33 feldspar-ratios ŽQzrFsp. of the bulk fraction range between 0.19 ŽPS2212-3; average 0.37. and 1.56 ŽPS2123-2; average 0.68. showing a clear difference in the content of feldspar relatively to quartz ŽTable 2; Fig. 5.. Generally, the ratios are closely related to the quartz content and exhibit maxima near the MIS boundary 3r2. In the K-feldsparrplagioclase plot ŽKfsrPlg. PS2123-2 displays several reductions to 0 due to very low or missing K-feldspar, while only a few core horizons in PS2212-3 miss K-feldspar. One of the prominent minima is in Event I sediments ŽFig. 5.. The clay fraction Ži.e. particles smaller than 2 mm. contains at least 90% clay minerals. The illite group is the major constituent of the clay mineral fraction. Chlorite contents are generally constant around 20%. Quartz is the most important non-clay mineral in the clay fraction reaching a maximum of 8% ŽVogt, 1997.. Smaller percentages of plagioclase, carbonates, and heavy minerals were observed. Ordered–layered expandable minerals ŽOLEM; cf. Reynolds, 1970. occur in a few Event I samples. The QzrFsp-ratio of the clay fraction of Event I sediments is also strongly increased ŽFig. 5.. For the paleoceanographic reconstruction, smectite and kaolinite are the most important clay mineral groups. In general, smectite contents range between 0% and 20% and kaolinite between 12% and 30%. Only a few distinct layer show higher values ŽFig. 5.. Overall, the smectite and kaolinite records of cores PS2122-1 and PS2123-2 display a very similar pattern ŽFig. 5.. In core PS2212-3 sediments of Termination I high smectite Ž) 10%. and kaolinite Ž) 20%. values coincide, but the phasing of smectite and kaolinite peaks differs. The most obvious other feature is a low smectiterhigh kaolinite couple during Event I which combines with the only occurrence of OLEM. 5. Discussion Due to the complex sedimentary environment at the Yermak Plateau and off the NW Spitsbergen coast it is difficult to rely on only a few parameter to reconstruct the paleoceanographic history. Various phenomena can influence the sedimentary processes, 34 C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 including surface water and sea-ice transport from the Siberian shelf regions, icebergs, possibly brines from the near Spitsbergen glaciers or from winter sea-ice production, and surface and bottom currents from the Norwegian–Greenland Sea and the central Arctic Ocean. Therefore, a set of distinct indicators of sedimentary environments have been compiled ŽTable 2.. In this paper we will concentrate on two time intervals Ž0–30 and 50–85 ka. during which well defined changes in the influx of Atlantic water were related to the built-up and melt-down of the SBIS Žcf. Hebbeln et al., 1994; Andersen et al., 1996.. The younger time interval from 0 to 30 ka comprises the well investigated upper MIS 3 to 1. We will show very similar developments of the marine sedimentation west Žsee Elverhøi et al., 1995. and northwest of Svalbard Žthis study.. More strikingly, the older time interval Žca. 50–85 ka. exhibits the same sedimentological record. Hence, we assume a similar paleoceanographic and glaciological history for these two Weichselian SBIS cycles. 5.1. SÕalbardr Barents sea ice sheet adÕances between 27 and 16 ka This distinct Event I signal ŽTable 2; a AmatureB organic–geochemical and mineralogical signature. can be found in all three analyzed cores above the MIS boundary 3r2 ŽFigs. 4–8.. It is intercalated between sediments with high calcite content and increased foraminifera and coccolith abundances ŽFigs. 4 and 7; PS2212-3: cf. Nowacyzk et al., 1994.. The increased abundance of planktonic species Žthe Ahigh productivity zonesB; Dokken and Hald, 1996. suggests seasonally open-water conditions not only at the SW coast of Spitsbergen but also to the north up to the northern Yermak Plateau. The seasonally open water could have supported a moisture supply for the two-step build-up of the northwestern SBIS between 27 and 23 ka and 19 and 16 ka Žcf. Hebbeln et al., 1994; Elverhøi et al., 1995.. Additionally, high numbers of IRD larger than 2 mm ŽFig. 7. suggest an increased iceberg transport before and after Event I. These high numbers of IRD occur together with strongly increased quartz contents and reduced feldspar contents leading to high QzrFspratios ŽFig. 5.. The near-coast core PS2123-2 reveals the highest values ŽFigs. 5–7.. Increased amounts of amphiboles in the bulk fraction have been attributed to the input of crystalline rocks from Fennoscandia and Svalbard ŽAndersen et al., 1996. and were also recognized in our cores. One source region for the quartz-richrfeldspar-depleted material ŽFig. 5. could be the Paleozoic crystalline strata ŽHekla Hoek. and the Devonian clastic wedge rocks which extensively outcrop in northern Spitsbergen ŽWinsnes, 1988.. The Paleozoic material from Spitsbergen and also from Fennoscandia, however, does contain some amount of K-feldspar Žcf. Andersen et al., 1996.. In contrast, the siliciclastic nature of the mature JurassicrCretaceous Shale material, identified as the source material for Event I, could also produce a high quartz content while it is extremely depleted in ŽK-. feldspar ŽAndersen et al., 1996.. The KfsrPlgratio decreases nearly to 0 ŽFig. 5., pointing to a dominant input from these shales during the finefraction sedimentation of Event I. As Event I is developed in all three cores similarly, we regard it as a synchronous regional event representing the advanced SBIS and a time of northward transport of fine fraction material along the western Svalbard continental slope. The reduction of the horizons thickness from S to N Žtens of centimeter W of Spitsbergen to a few at the Yermak Plateau. supports the assumption that currents transported the fine fraction from Ža. reworking by the advanced SBIS at the Spitsbergenbanken to Žb. injection into the intermediate waters of a Paleo-Westspitsbergen Current through dense, suspension-rich bottom-water currents from the Storfjorden Trough Žcf. Hebbeln et al., 1994; Andersen et al., 1996., and Žc. finally reaching the northern Yermak Plateau at 828N and water depth of 2500 m Žcompare Fig. 1.. On Spitsbergenbanken and in the area SE of Spitsbergen only thin Quaternary sediment blankets cover outcrops of Late Triassic to Early Cretaceous sedimentary rocks ŽFig. 1.. An Early Cretaceous Ahot shaleB member of the Mesozoic formations which is rich in mature organic material and possesses a diagenetically mature mineral assemblage ŽFig. 3: OLEM clays, no smectite, low feldspar and no K-feldspar. is of particular importance ŽAndersen et al., 1996.. This set of AmatureB material was found in sediment cores south of our study area ŽFigs. 1 and 3: NP 90-39., additionally manifested C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 Fig. 8. Age versus parameter plot of PS2212-3 from the NE of the Yermak Plateau. Hatched area—TOCrOLEMrkaolinite event; grey shaded areas are MIS 1 and 3. 35 36 C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 by the identification of Mesozoic palynomorphs ŽElverhøi et al., 1995; Andersen et al., 1996.. Event I is also documented in cores to the south and east of our cores combined with high abundances of Mesozoic sediment clasts in the IRD ŽFig. 1; Andersen et al., 1996; Knies and Stein, 1998.. As our cores are rather distal to the source of the Event I material, the thickness of the horizon is strongly reduced Ži.e. more than 1 m in NP9039 ŽFig. 3. to 0.02 m in PS2212-3 ŽFigs. 4 and 5... Therefore, it is difficult to deduce in more detail the developments of the SBIS’s glacier fronts in the SE or NE of Svalbard on the basis of our records. According to Dokken Ž1995. the Event I sediments could also pinpoint a deglazial phase at the southeastern extension of the SBIS while Knies et al. Ž2000. and Kleiber et al. Ž2000. reconstructed the largest extension of the northeastern SBIS in the vicinity of the Franz Victoria Trough at the upper continental slope at approximately 23 ka ŽFig. 1: PS2138-1.. Between 23 and 15.4 ka small undulations of the SBIS front but no larger deglaciations are reported. Partly laminated fine fraction sediments are deposited in the Franz Victoria Trough between approximately 22 and 20.5 ka and interpreted as meltwater plumes in front of the ice sheet ŽKleiber et al. 2000.. Leirdal Ž1997. reconstructed the same environment north of the Hinlopenstrait, northern Spitsbergen. A hold of the ice sheet advance is most probable and the fine fraction material records the meltwater plumes in front of the undulating ice sheet. This fine material would then be redistributed as outlined above. The strong similarities of the cores from the Western Fram Strait and PS2212-3 on the northeastern Yermak Plateau slope leads us to assume a Paleo-WSC very similar to today including a strong intermediate water component. The position on the Yermak Plateau slope excludes the influence of Nansen Basin bottom currents. Sea-ice or icebergs from the Franz Victoria Trough Žcompare Fig. 1. should have transported more coarse fraction and poorly sorted material. 5.2. Early MIS 2 The deposition of dolomite, in particular the input of bulk and clay-fraction dolomite at the northeastern Yermak Plateau site PS2212 was high during early MIS2 ŽFig. 5; cf. Vogt, 1997.. As dolomite derives from the northern Svalbard Paleozoic carbonate rocks, increased iceberg production from the advanced SBIS on the northwestern and northern coast of Spitsbergen can be assumed including glacial rock flour Žclay fraction dolomite.. While IRD and sand sedimentation continued through the entire MIS 2 in the W-Spitsbergen cores, coarse fraction input diminished at about 17 to 15.5 ka in the northern Yermak Plateau core PS2212-3, and fine fraction dominated ŽFig. 8.. Icebergs might be blocked by sea ice from moving to the Yermak Plateau site or meltout of particles might be prevented by cold ŽPolar. water conditions. Core positions closer to the Northern Barents Sea slope yield strongly increased sedimentation of IRD ŽKubisch, 1992; Knies et al., 2000, 2001.. To the west of Spitsbergen, seasonally open water conditions prevailed as indicated by increased calcite contents, low TOC contents, low CrN ratios and slightly increased HI-values as well as coccolith and subpolar planktic foraminifera abundances ŽFig. 4; cf. Hebbeln et al., 1994; Dokken, 1995.. 5.3. Intermediate water carbonate dissolution before and during early Termination I Core PS2212-3 ŽWD: 2550 m. was affected by complete carbonate dissolution during Termination I as outlined by the absence of carbonate ŽFig. 8; 17–14 ka. and the benthic foraminifera assemblage ŽWollenburg et al., 2001.. During the same time interval, the NW-Spitsbergen cores PS2122-1 and PS2123-2, and the neighboring Fram Strait, Nansen Basin and Gakkel Ridge sediments reveal increased biogenic calcite contents ŽFig. 4; Stein et al., 1994; Andersen et al., 1996.. Nearly the complete sand fraction in Nansen Basin and Gakkel Ridge cores is comprised of foraminifera Ž90–100%; Markussen et al., 1985; Nørgaard-Pedersen et al., 1998.. Hence, the water depth of core PS2212-3 seems to be most affected. Later, only dolomite was preserved delineating reduced dissolution of carbonate at that time. All sediment cores on the deeper slope of the northwestern Barents Sea and the Yermak Plateau show decreased carbonate contents during early Termination I Žaround 15 ka. and are affected by dissolution and in particular dilution of Žbiogenic. carbonate C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 sedimentation Žcf. Pagels, 1991; Elverhøi et al., 1995; Hebbeln and Wefer, 1997; Bauch et al., 1999; Knies et al., 1999.. To explain the differences between PS2212-3 and the adjacent cores we assume the production of carbonate aggressive brines in the vicinity of the ice-sheet andror sea-ice edge north of Spitsbergen following the model of Steinsund and Hald Ž1994.. Since the adjacent Nansen Basin cores show only small dissolution of carbonate during this time span Žcf. Markussen et al., 1985; Pagels, 1991; Vogt, 1997., we regard brine injection as the most probable process leading to complete carbonate dissolution in about 2500 m water depth at the NE Yermak Plateau. Similar processes have been assumed for the Northern Barents Sea Shelf Žcf. Pagels, 1991; Kohler, ¨ 1992; Knies and Stein, 1998.. Brine production could have been caused by processes at the ice-sheet front or during freeze-up of sea-ice in autumn. In slope sediment cores west of Spitsbergen benthic foraminifera register oxygen and carbon isotope values which are indicative of brines ŽLloyd et al., 1996a.. Today, such brines reach water depth of 2000 m ŽQuadfasel et al., 1988; Schauer et al., 1997.. A lowered sea-level and differences in brine andror intermediate water composition might have led to the increased water depths of corrosive shelf waters. 5.4. Surface water enÕironment at N Yermak Plateau during early Termination I If the dissolution of carbonate is due to brine ejection during autumn freeze-up, seasonally open water must have existed near Site PS2212. As biogenic carbonate is absent due to dissolution, other tracers of open water conditions have to be used for reconstruction. Sea-ice and icebergs could have been melted during summer. In this case a sediment of mixed origin would be dominant, indicating strong sea-ice detritus deposition together with IRD from icebergs derived from adjacent Spitsbergen Žcf. Knies et al., 1999, 2000.. Sea-ice sediment input can be deduced from increased smectite contents above 15% combined with increased pyroxene contents ŽFig. 7; Letzig, 1995; Vogt, 1997; Behrends, 1999.. Increased organic carbon contents with a terrigenous ŽCrN-ratio ) 10. and partly marine signature ŽHI- 37 index increase., low IRD counts but increased sand contents and high amounts of the fine fraction Žsilt and clay. are observed ŽFig. 8: ca. 16 and 14 ka.. All this evidence indicates sedimentation from melting sea-ice, entrained on the Siberian shelves and transported to the Yermak Plateau position by the Transpolar Drift ŽNurnberg et al., 1994; Hebbeln and ¨ Wefer, 1997; Knies and Stein, 1998.. The benthic foraminifera assemblage strongly supports the idea of seasonally open water above PS2212-3 ŽWollenburg et al., 2001.. 5.5. The Termination I deglaciation (16–9 ka) In the western Fram Strait high IRD and sand deposition continued ŽFigs. 4, 6, 7–8.. The d18 O plots show light values conspicuous of melting events ŽFigs. 2 and 6.. The mineralogy is typical for Svalbard sources ŽTable 2., which indicates that icebergs from the Svalbard fjords could drift northward in the seasonally open water. Thus, we conclude that seasonally open waters prevailed in western Fram Strait and north of Spitsbergen near the SBIS ice edge around 16–17 ka. During summer the Polar Front reached this region. To the north the Nansen Basin was covered by perennial sea-ice as indicated by low sedimentation rates of mainly fine grained material Žcf. Stein et al., 1994; Vogt, 1997; Nørgaard-Pedersen et al., 1998.. In core PS2212-3 the carbonate free horizon smoothly passes over to sediments with increased HI-indices Žca. 13 ka.. The same development has been observed in other cores from the Fram Strait and the adjacent Nansen Basin ŽAndersen et al., 1996; Vogt, 1997.. This suggests small increases in the input of marine organic carbon ŽFig. 8; Knies et al., 1998.. Higher calcite contents in the NW Spitsbergen cores might also indicate higher productivity during this second phase of early Termination I. High fine fraction, kaolinite and maximum smectite contents in the NE Yermak Plateau core PS2212-3 are probably the result of biologically enhanced Žfecal pellets. sedimentation at the summer ice edge during the ice-melt induced phytoplankton bloom, similar to the recent summer situation in the Fram Strait described by Berner and Wefer Ž1994.. Other productivity proxies do reflect the increased productivity during this time ŽŽe.g. opal, benthic foramini- 38 C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 fera; Nurnberg et al., 1995; Wollenburg et al., 2001.. ¨ The WSC seems to have reached further north, propelling seasonally open water as far north as 828N. This occurred contemporary with an increased transport of sea-ice sediments due to extensive entrainment of Žsmectite rich. shelf sediments during the flooding of western Siberian shelves by the rising sea-level Žcf. Forman et al., 1996; Kassens et al., 1999 and references therein.. Additionally, starting meltdown of the Kara Sea Ice Sheet might have enhanced the remobilization of fine-grained material in the Kara Sea ŽVogt, 1997; Polyak et al., 1997.. A stronger warm water influx with the WSC system andror the sea-level rise triggered glacial retreats in Kara SearSt. Anna Trough area and increased the melting of icebergs above the Yermak Plateau Žcf. Hebbeln and Wefer, 1997; Polyak et al., 1997.. The origin of the PS2212-3 sediment during this time can be clearly related to Siberian shelf sources and the Franz Josef Land region ŽTable 2; i.e. high kaolinite, smectite and pyroxene contents, weak smectite crystallinity, high KfsrPlg-ratio of bulk sample, high sand contents but small gravel contents; cf. Vogt, 1997.. Most of the material has probably been transported by sea-ice and the related surface waters as well as some icebergs from the St. Anna Trough and Franz Josef Land Žhigh kaolinite and K-feldspar contents.. A first IRD peak combined with strongly increased kaolinite contents has been recognized by Knies et al. Ž1999. at the northern Barents Sea slope between 16.8 and 15.4 ka and has been related to the early melting of the northern SBIS between Franz Josef Land and Svalbard. The cores off NW Spitsbergen display highest IRD-counts near the MIS 2r1 stage boundary with material originating mainly from Spitsbergen ŽFig. 7.. Elverhøi et al. Ž1995. and Hebbeln et al. Ž1994. report a distinct melting event of the western SBIS at 14.5 ka in sediment cores which are located south of the study area ŽFig. 1.. This melting event can also be recognized in all cores of this study ŽFigs. 6–8, grey shades. and in cores to the northeast Že.g. PS2138, Knies et al., 1999, 2001. as well as a second melting event around 13 ka ŽNørgaard-Pedersen et al., 1998.. Highest coarse fraction contents during the 13 ka melting occur on the NE Yermak Plateau site ŽFig. 6., which are accompanied by the highest single smectite peak within the whole core ŽFig. 8.. Solely smectite in the fine fraction indicates input from more eastern regions than Franz Josef Land and its adjacent troughs, which would have increased the kaolinite content ŽWahsner et al., 1999.. Our records correspond very well to that of Birks et al. Ž1994., who summarized the climatological development of the Svalbard region and concluded that the deglaciation of the Barents Sea started probably as early as 15 ka, but certainly before 13.3 ka in the central and southern parts. In the north, the earliest deglaciation sediments in the Franz Victoria Trough west of Franz Josef Land are dated to 15.4 ka ŽKleiber et al., 2000.. Between 13 and 12 ka, a rapid retreat of the western SBIS glaciers signalled the fast decrease of the ice sheet ŽMangerud et al., 1998., which is also recorded in the NW Spitsbergen cores by high IRD signals, increased coarse-grained fraction and high quartz contents ŽFigs. 4–7.. The early Termination I sequence in core PS2212-3 yields high quartz contents but low QzrFsp- and KfsrPlg-ratios in the bulk fraction ŽFig. 5.. This is rather indicative of crystalline rocks like the Hekla Hoek basement of Svalbard. In contrast, PS2123-2 exhibits fairly low quartz contents, the third high carbonatercalcite peak in MIS 2, lowest TOC contents, CrN ratios below 5, and the beginning increase of HI values which are all indicative of Žseasonally. open water ŽTable 2, Figs. 4 and 5.. A drastic increase in sediment accumulation rates is observed Žfrom about 12 to ) 20 grcm2 ka; Knies, 1994.. As the siltrclay-ratio increases ŽFig. 4., an increase in bottom currents can be assumed, which is today related to a strong WSC current activity Žcf. Boulton, 1990.. Finally, high IRD-input and coarse fraction contents mark the complete breakdown of the SBIS after 10 ka. 5.6. SÕalbard ice sheet adÕance during stage 5 similar to the 23 ka eÕent? A comparison of the sedimentological, organic– geochemical and mineralogical signature of Event I in core PS2212-3 ŽFig. 8. with the climatic records from the NW Spitsbergen cores PS2122-1 and PS2123-2 during the middle to upper MIS 5 reveals striking similarities ŽFigs. 4–7 and 9: ca. 85 to 75 C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 Fig. 9. Age versus parameter plot of PS2122-1 NW of Svalbard. Hatched area—TOCrOLEMrkaolinite event; grey shaded areas are MIS 3 and 5. 39 40 C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 ka.. Assuming the same development as during MIS 3 and 2, the occurrence of the Event I-like sediment layer indicates another SBIS advance to the SE in late MIS 5. During this time interval core PS2123-2 shows a peak in the IRD record ŽFig. 7.. Unfortunately, parts of this time interval are missing in core PS2212-3 because of a hiatus. The hiatus might have been caused by increased bottom currents at the NE slope of the Yermak Plateau. Currents might indicate a stronger and deeper WSC or dense bottom water formed at the front of the advanced northern SBIS andror brine formation during the winter sea-ice built-up. Alternatively, a debris flow or turbidity current could have eroded the sediments on the NE slope of the Yermak Plateau, which was also observed with sub-bottom echosounder investigations in that region ŽPARASOUND data; Bergmann, 1996.. At the adjacent site PS1533 Žca. 2000 m water depth, 11 km to the west of PS2212-3. the sediments point to seasonally open-water conditions and winter sea-ice cover with low iceberg occurrences during MIS 5a ŽPagels, 1991; Kubisch, 1992.. Some seasonally open water is also indicated by the increased calcite and foraminifera contents, low CrN-ratios and mainly fine fraction sediment in PS2122-1 and PS2123-2 before and after the event ŽFigs. 4 and 9.. Open water would provide moisture for the build-up of the SBIS as during MIS 2 and 3, and outlined by Hebbeln et al. Ž1994.. The younger IRD-peak in PS2122-1rPS2123-2 and the Event I-like sediment layer point to another SBIS advance in MIS 5a. It is again intercalated between sediments with increased calcite contents, one indicator for seasonally open water conditions ŽFig. 9.. This is in accordance with the latest landbased reconstruction of the SBIS, where early advances of the mid-Weichselian SBIS are envisioned ŽFigs. 7 and 9; Mangerud et al., 1998.. Due to the limited data base for glaciation of Spitsbergen, some advances registered in the marine record during upper MIS 5 might be missing in the onshore record Žcf. Mangerud et al., 1998.. The cores investigated by Elverhøi et al. Ž1995. do not reach back into MIS 5. Hence, this is the first indication of an Event I-like development in an older record. IRD records of Lloyd et al. Ž1996b. and Hebbeln and Wefer Ž1997. also show multiple IRD peaks in MIS 5Žb?.. Investigations of cores N and NE of Spitsbergen gave evidence of earlier advances of the northern rim of the SBIS, which coincide with a strong northward influx of Atlantic water into the Arctic Ocean by the WSC during MIS 6 ŽVogt, 1997; Knies and Stein, 1998.. Fig. 3a shows at least two such events in MIS 6 ŽPS2212-3: increased calcite content.. However, during MIS 5 there is no indication of an intensive advance at the northern rim of the SBIS according to IRD-data ŽKnies et al., 1999, 2001.. This might be due to blocking of icebergs by a more or less closed sea-ice cover. At the isotopic stage 5r4 boundary and during early stage 4, a marked increase of IRD, higher quartz contents, CrN-ratios between 10 and 15, and HI values - 50 mg HCrg C in the near coastal core PS2123-2 indicate dominantly terrigenous sedimentation and the readvance of the Spitsbergen glaciers ŽFigs. 4–8.. Further offshore, only stage 4 sediments show the increase in IRD while stage 5r4 boundary sediments contain several indicators of increased sea-ice cover Žlow sedimentation rate, mainly fine fraction sedimentation, low carbonate content being mainly dolomite, a marked decrease in d13 C-values.. Only the fairly and solely high smectite contents in PS2122-1 and PS2212-3 could point to some melting of sea-ice during summer. This could also point to a continuous sea-ice transport of the Transpolar Drift with probably Laptev Sea sources Žcf. Vogt, 1997.. Latest reconstructions of the Kara Sea Ice Sheet assume a strong middle Weichselian glaciation Žcf. Astakhov et al., 1999.. Hence, sea-ice during MIS 4 could only be built east of the Kara Sea. While in the central Arctic Ocean a continuos ice-cover is assumed ŽDarby et al., 1997; Nørgaard-Pedersen et al., 1998 and references therein., a calcite peak in early MIS 4 records of PS2212-3 ŽFig. 4. and other cores near the Svalbard continental margin delineate some influx of Atlantic Water even to the Northeast of Svalbard ŽDokken and Hald, 1996; Knies and Stein, 1998.. 5.7. Deglaciation at the MIS 4 r 3 boundary Upper MIS 4 sediments of PS2212-3 from the northeastern Yermak Plateau suggest input of very fine-grained material with extremely low IRD and sand fraction contents. No coccoliths were observed C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 in PS2212-3 ŽNowacyzk et al., 1994.. Thus, a perennial sea ice-cover can be assumed during MIS 4 at the Yermak Plateau in accordance with published reconstructions Žcf. Pagels, 1991; Kohler, 1992; Ku¨ bisch, 1992; Hebbeln and Wefer, 1997.. Planktonic foraminifera occurrences are extremely low on the Yermak Plateau, and in the Nansen Basin Žcf. Pagels, 1991. the calcite contents are reduced ŽFig. 4.. Increased dolomite contents in MIS 4 sediments point to Svalbard sources of the sediment ŽFig. 4.. The cores NW of Spitsbergen still receive coarse fraction material being highest in the near coast core PS21232 ŽFigs. 4, 6 and 7.. The isotope data show a few light values indicating several melting events west of Spitsbergen ŽFig. 6.. Hence, at least some seasonally open water conditions were present. Increased quartz contents, and high KfsrPlg- and clay fraction Qzr Fsp-ratios as well as increased smectite and kaolinite contents point to melting of sea-ice with probably Siberian origin. Similar to the way the advance sequence of the SBIS between upper MIS 5 and MIS 4 resembles the last glaciation at the MIS 3r2 boundary, the deglaciation record at the MIS 4r3 boundary resembles Termination I data ŽFigs. 4–9.. On the northeastern Yermak Plateau, a PS2212-3 core interval with full carbonate dissolution is overlain by sediments with increased coarse fraction and IRD contents ŽFig. 4.. The mineralogical data indicates eastern sources. Increased kaolinite and smectite contents and increased KfsrPlg-ratios and QzrFsp-ratios of the clay fraction are indicative of Franz Josef Land and Siberian shelves east of the Archipelago ŽVogt, 1997.. IRD of Nansen Basin sediments during the early MIS 3 deglaciation have been attributed to Siberian origin ŽKubisch, 1992; Vogt, 1997, Nørgaard-Pedersen et al., 1998.. West of Spitsbergen, Svalbard and Fennoscandian sources were dominant Žcf. Hebbeln and Wefer, 1997.. Hence, the Polar Front between warmer Atlantic and cold Polar Waters has been in a similar position as today. 41 clay fraction can be related to Late Quaternary glacialrinterglacial changes. The use of single-source minerals as provenance indicators allows to decipher different transport and sedimentation processes in the Arctic Ocean during glacial and interglacial times. Other geochemical and sedimentological parameters support these changes Ži.e. isotope record, carbonate and organic carbon content and composition, grainsize distribution and IRD content.. A very distinct signal combination including high TOC content and CrN-ratios, high kaolinite content, low carbonate and randomly ordered smectite contents, and the OLEM occurrence is related to the 23–19 ka initial advance of the SvalbardrBarents Sea Ice Sheet. We could add additional organic geochemistry and mineralogical information to this Event I especially with regard to its occurrence on the north eastern Yermak Plateau. Furthermore, we were able to recognize a Event I-like signal during late MIS 5 for the first time. This Svalbard ice sheet advance might not be properly recorded in onshore data. Using a combination of organic geochemical and mineralogical parameters, intensified warm Atlantic water influx as far north as the Yermak Plateau could be stated for late MIS 4rearly MIS 3, late MIS 3rearly MIS 2, early Termination I and the Holocene. Termination I and the MIS boundary 4r3 exhibit a very similar pattern of strong Atlantic water influx into the Arctic Ocean, leading to a northward migration of the summer ice edge. Early MIS 3 can be compared to Termination I. Paleoceanographic and ice sheet developments of the Svalbard region can be recognized by means of mineralogical and grain size parameter being independent from productivity and stable isotope data. This enables us to use sediments from deeper water cores for paleoceanographic reconstructions, although they have been influenced by corrosive bottom water conditions and carbonate dissolution. v v v v Acknowledgements 6. Conclusion Downcore variations of mineral assemblages and distributions within the bulk sediment and the v The authors thank Antje Volker and Trond Dokken ¨ for critical comments and thoughtful reviews which improved and strengthened the manuscript. We thank the captain and the crew of the RV Polarstern for 42 C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 excellent cooperation during the 1991 expeditions. For technical assistance and data discussion we thank M. Wahsner, D. Nurnberg, N. Nørgaard-Pedersen, ¨ C.J. Schubert, R. Frohlking, M. Siebold, and H. ¨ Grobe and M. Seebeck for grain size data and clay sample preparation. B. Diekmann and D.-K. Futterer ¨ critically read the manuscript in an early stage which is much appreciated. The English was improved by Emma Eades. This work was partially funded by DFG ŽDeutsche Forschungsgemeinschaft. under the contracts Fi-443r1,2 and 3. We benefited from the European Science Foundation project APolar North Atlantic Margins ŽPONAM.B including early data discussions on PONAM workshops and the QUEEN program. Data of this paper are stored in the PANGAEA database at http:rrwww.pangaea.de. References Aagaard, K., Carmack, E.C., 1994. The Arctic Ocean and climate: a perspective. In: Johannessen, O.M., Muench, R.D., Overland, J.E. ŽEds.., The Polar Oceans and their Role in Shaping the Global Environment: The Nansen Centennial Volume. Geophysical Monograph, vol. 85. American Geophysical Union, Washington, DC, pp. 5–20. Alley, R.B., 1995. Resolved: the arctic controls global climate change. In: Smith Jr., W.O., Grebmeier, J.M. ŽEds.., Arctic Oceanography: Marginal Ice Zones and Continental shelves. Coastal and Estuarine Studies, vol. 49. American Geophysical Union, Washington, DC, pp. 263–284. Andersen, E.S., Dokken, T.M., Elverhøi, A., Solheim, A., Fossen, I., 1996. Late Quaternary sedimentation and glacial history of the western Svalbard margin. Mar. Geol. 133, 123–156. Anderson, L.G., Bjork, ¨ G., Holby, O., Jones, E.P., Kattner, G., Koltermann, K.P., Liljeblad, B., Lindegren, R., Rudels, B., Swift, J., 1994. Water masses and circulation in the eurasian basin: results from the Oden 91 North Pole expedition. J. Geophys. Res. 99 ŽC2., 3273–3283. Astakhov, V.I., Matiouchkov, A., Svendsen, J.I., Mangerud, J., Maslenikova, O., Tveranger, J., 1999. Marginal formations of the last Kara and Barents ice sheets in northern European Russia. Boreas 28 Ž1., 23–45. Bauch, H.A., Erlenkeuser, H., Fahl, K., Spielhagen, R.A., Weinelt, M.S., Andruleit, H., Henrich, R., 1999. Evidence for steeper Eemian than Holocene sea surface temperature gradient between Arctic and sub-Arctic regions. Palaeogeogr., Palaeoclimatol., Palaeoecol. 145 Ž1–3., 95–117. Behrends, M., 1999. Reconstruction of sea-ice drift and terrigenous sediment supply in the late quaternary: heavy-mineral associations in sediments of the Laptev-Sea continental margin and the central Arctic Ocean. Reports on Polar Research, vol. 310. Alfred Wegener Institut Bremerhaven, Bremerhaven, 167 pp. Bergmann, U., 1996. Interpretation of digital Parasound echosounder records of the eastern Arctic Ocean on the basis of sediment physical properties. Rep. Pol. Res., vol. 183. Alfred Wegener Institute, Bremerhaven, 164 pp. Berner, H., Wefer, G., 1994. Clay–mineral flux in the Fram Strait and Norwegian Sea. Mar. Geol. 116, 327–345. Birks, H.H., Paus, A., Alm, T., Mangerud, J., Landvik, J.Y., 1994. Late Weichselian environmental change in Norway including Svalbard. J. Quat. Sci. 9 Ž2., 133–145. Boulton, G.S., 1990. Sedimentary and sea level changes during glacial cycles and their control on glacimarine facies architecture. In: Dowdeswell, J.A., Scourse, J.D. ŽEds.., Glacimarine Environments: Processes and Sediments. Geol. Soc. Spec. Publ. vol. 53. The Geological Society, London, pp. 15–52. Darby, D.A., Bischof, J.F., Jones, G.A., 1997. Radiocarbon chronology of depositional regimes in the western Arctic Ocean. Deep-Sea Res. II 44 Ž8., 1745–1757. Dokken, T.M., 1995. Poleoceanographic changes during the last Interglacial–Glacial cycle from the Svalbard–Barents Sea margin: Implications for ice sheet growth and decay. Dr. Scient. Thesis Žunpublished., Inst. Biol. Geol., Univ. of Tromsø, 175 pp. Dokken, T.M., Hald, M., 1996. Rapid climatic shifts during isotope stages 2–4 in the Polar North Atlantic. Geology 24 Ž7., 599–602. Elverhøi, A., Andersen, E.S., Dokken, T., Hebbeln, D., Spielhagen, R.F., Svendsen, J.I., Sørflaten, M., Rørnes, A., Hald, M., Forsberg, C.F., 1995. The growth and decay of the Late Wechselian Ice Sheet in western Svalbard and adjacent areas based on provenance studies of marine sediments. Quat. Res. 44, 303–316. Emmermann, R., Lauterjung, J., 1990. Double X-ray analysis of cuttings and rock flour: a powerful tool for rapid and reliable determination of borehole lithostratigraphy. Sci. Drill. 1, 269– 282. Forman, S.L., Lubinski, D., Miller, G.H., Matishov, G., Korsun, S., Snyder, J., Herlihy, F., Weihe, R., Myslivets, V., 1996. Postglacial emergence of Western Franz Josef Land, Russia and retreat of the Barents Sea Ice Sheet. Quat. Sci. Rev. 15, 77–90. Futterer, D.K. ŽEd.., 1992. ARCTIC ’91: The Expedition ARK¨ VIIIr3 of RV APolarsternB in 1991. Rep. Pol. Res., vol. 107. Alfred Wegener Institute, Bremerhaven, 267 pp. Grobe, H., 1987. A simple method for determination of ice rafted debris in sediment cores. Polarforschung 57 Ž3., 123–126. Haake, F.W., Plaumann, U., 1989. Late Pleistocene foraminiferal stratigraphy on the Vøring Plateau, Norwegian Sea. Boreas 18 Ž4., 343–356. Hebbeln, D., Wefer, G., 1997. Late Quaternary paleoceanography in the Fram Strait. Paleoceanography 12 Ž1., 65–78. Hebbeln, D., Dokken, T., Andersen, E.S., Hald, M., Elverhøi, A., 1994. Moisture supply for northern ice-sheet growth during the Last Glacial Maximum. Nature 370, 357–359. Kassens, H., Bauch, H.A., Dmitrenko, I., Eicken, H., Hubberten, H.-W., Melles, M., Thiede, J., Timokhov, L. ŽEds.., 1999. C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 Land–Ocean Systems in the Siberian Arctic: Dynamics and History. Springer-Verlag, Berlin, Heidelberg, 711 pp. Kleiber, H.P., Knies, J., Niessen, F., 2000. The late weichselian glaciation of the Franz Victoria Trough, northern Barents sea: ice sheet extent and timing. Mar. Geol. 168 Ž1–4., 25–44. Knies, J., 1994. Spatquatare ¨ ¨ Sedimentation am Kontinentalhang nordwestlich Spitzbergens. Der letzte GlazialrInterglazialZyklus Žunpublished Dipl. thesis. Justus-Liebig-Universitat, ¨ Gießen, 95 pp. Knies, J., Stein, R., 1998. New aspects of organic carbon deposition and its paleoceanographic implications along the northern Barents Sea margin during the last 30,000 years. Paleoceanography 13 Ž4., 384–394. Knies, J., Vogt, C., Stein, R., 1999. Late Quaternary growth and decay of the SvalbardrBarents sea ice sheet and paleoceanographic evolution in the adjacent Arctic ocean. Geo-Mar. Lett. 18, 195–202. Knies, J., Nowaczyk, N., Muller, C., Vogt, C., Stein, R., 2000. A ¨ multiproxy approach to reconstruct the environmental changes along the Eurasian continental margin over the last 150,000 years. Mar. Geol. 163 Ž1–4., 317–344. Knies, J., Kleiber, H.-P., Nowaczyk, N., Mathiessen, J., Muller, ¨ C., Niessen, F., Stein, R., Weiel, D., 2001. Marine ice-rafted debris records constrain maximum extent of Saalian and Weichselian ice-sheets along the northern Eurasian Margin. Global and Planetary Change 31, 45–64. Kohler, S.E.I., 1992. Spatquartare En¨ ¨ ¨ palao-ozeanographische ¨ twicklung des Nordpolarmeeres und Europaischen Nord¨ meeres anhand von Sauerstoff-und Kohlenstoffisotopenverhaltnissen der planktischen Foraminifere Neogloboquadrina ¨ pachyderma Žsin... GEOMAR Rep., vol. 13. GEOMAR, Kiel, 104 pp. Kubisch, M., 1992. Die Eisdrift im Arktischen Ozean wahrend der ¨ letzten 250,000 Jahre. GEOMAR Rep., vol. 16. GEOMAR, Kiel, 100 pp. Leirdal, G., 1997. Senkværtær Utvikling Av Kontinentalmarginen Nord for Svalbard. unpubl. Ms. Thesis, University of Oslo, Oslo, 141 pp. Letzig, T., 1995. Sea ice-transported lithogenic fine fraction of late quaternary deep-sea sediments of the central eastern Arctic Ocean and the Fram Strait. Reports on Polar Research, vol. 162. Alfred Wegener Institute, Bremerhaven, 98 pp. Lloyd, J.M., Kroon, D., Laban, C., Boulton, G.S., 1996a. Deglaciation history and paleoceanography of the western Spitsbergen margin since the last glacial maximum. In: Andrews, J.T., Austin, W.E.N., Bergsten, H., Jennings, A.E. ŽEds.., Late Quaternary Paleoceanography of the North Atlantic Margins, vol. 111. Geol. Soc. Spec. Publ, London, pp. 289–301. Lloyd, J.M., Kroon, D., Boulton, G.S., Laban, C., Fallick, A., 1996b. Ice rafting history from the Spitsbergen ice cap over the last 200 kyr. Mar. Geol. 131, 103–121. Lubinski, D.J., Korsun, S., Polyak, L., Forman, S.L., Lehman, S.J., Herlihy, F.A., Miller, G.H., 1996. The last deglaciation of the Franz Victoria Trough, northern Barents Sea. Boreas 25, 89–100. 43 Mangerud, J., Gulliksen, S., 1975. Apparent radiocarbon ages of recent marine shells from Norway, Spitsbergen, and Arctic Canada. Quat. Res. 5, 273–296. Mangerud, J., Dokken, T., Hebbeln, D., Heggen, B., Ingolfsson, ´ Ó., Landvik, J.I., Mejdahl, V., Svendsen, J.I., Vorren, T.O., 1998. Fluctuations of the Svalbard Barents Sea Ice Sheet during the last 150,000 years. Quat. Sci. Rev. 17 Ž1–3., 11–42. Manley, T.O., Bourke, R.H., Hunkins, K.L., 1992. Near-surface circulation over the Yermak Plateau in northern Fram Strait. J. Mar. Syst. 3, 107–125. Markussen, B., Zahn, R., Thiede, J., 1985. Late Quaternary sedimentation in the eastern Arctic Basin: Stratigraphy and depositional environment. Palaeogeogr., Palaeoclimatol., Palaeoecol. 50, 271–284. Martinson, D.G., Pisias, N.G., Hays, J.D., Imbrie, J., Moore, T.C., Shackleton, N.J., 1987. Age dating and the orbital theory of the ice ages: development of a high-resolution 0 to 300,000 years chronostratigraphy. Quat. Res. 27, 1–27. Mathiessen, J., Knies, J., Nowaczyk, N.R., Stein, R., 2001. Late Quaternary dinoflagellate cyst stratigraphy at the Eurasian continental margin, Arctic Ocean: indications for Atlantic water inflow in the past 150,000 years. Global and Planetary Change 31, 65–86. Muller, G., 1967. Methods in sedimentary petrology. In: von ¨ Engelhardt, W., Fuchtbauer, H., Muller, G. ŽEds.., Sedimen¨ ¨ tary Petrology, vol. 1. Schweizerbart, Stuttgart, 283 pp. Nees, S., Struck, U., 1994. The biostratigraphy and paleoceanographic significance of Siphotextularia rolshauseni Phleger and Parker in Norwegian–Greenland Sea sediments. J. Foraminiferal Res. 24 Ž4., 233–240. Nørgaard-Pedersen, N., Spielhagen, R.F., Thiede, J., Kassens, H., 1998. Central Arctic surface ocean environment during the past 80,000 years. Paleoceanography 13 Ž2., 193–204. Nowaczyk, N.R., Fredrichs, T.W., Eisenhauer, A., Gard, G., 1994. Magnetostratigraphic data from late Quaternary sediments from the Yermak Plateau, Arctic Ocean: evidence for four geomagnetic polarity events within the last 170 Ka of the Brunshes Chron. Geophys. J. Int. 117, 453–471. Nurnberg, D., Wollenburg, I., Dethleff, D., Eicken, H., Kassens, ¨ H., Letzig, T., Reimnitz, E., Thiede, J., 1994. Sediments in Arctic sea ice: implications for entrainment, transport and release. Mar. Geol. 119, 185–214. Nurnberg, D., Schubert, C.J., Stein, R., Vogt, C., 1995. Biogenic ¨ barium and opal in Arctic Ocean sediments - do they reflect paleoproductivity? EOS Trans. AGU 76 Ž16rSpring Meet. Suppl.., S172. Pagels, U., 1991. Sedimentologische Untersuchungen und Bestimmungen der Karbonatlosung in spatquartaren ¨ ¨ ¨ Sedimenten des ostlichen Arktischen Ozeans. GEOMAR Rep., vol. 10. GEO¨ MAR, Kiel, 106 pp. Petschick, R., Kuhn, G., Gingele, F.X., 1996. Clay mineral distribution in surface sediments of the South Atlantic: sources, transport, and relation to oceanography. Mar. Geol. 130, 203– 229. Polyak, L., Forman, S.L., Herlihy, F.A., Ivanov, G., Krinitsky, P., 44 C. Vogt et al.r Global and Planetary Change 31 (2001) 23–44 1997. Late Weichselian deglaciation history of the Svyataya ŽSaint. Anna Trough, northern Kara Sea, Arctic Russia. Mar. Geol. 143, 169–188. Quadfasel, D., Rudels, B., Kurz, K., 1988. Outflow of dense water from a Svalbard fjord into the Fram Strait. Deep-Sea Res. I 35 Ž7., 1143–1150. Rachor, E., 1992. Scientific Report of RV APolarsternB Cruise ARK-VIIIr2. Rep. Pol. Res., vol. 115. Alfred Wegener Institute, Bremerhaven, 150 pp. Reynolds Jr., R.C., 1970. The nature of interlayering in mixedlayer illite–montmorillonites. Clays Clay Mineral. 18, 25–36. Schauer, U., Muench, R.D., Rudels, B., Timokhov, L., 1997. Impact of eastern Arctic shelf waters on the Nansen Basin intermediate layer. J. Geophys. Res. 102 ŽC2., 3371–3382. Schubert, C.J., Stein, R., 1996. Deposition of organic carbon in late quaternary Arctic ocean: terrigenous supply vs. marine productivity. Org. Geochem. 24 Ž4., 421–436. Stein, R., Schubert, C., Vogt, C., Futterer, D., 1994. Stable ¨ isotope stratigraphy, sedimentation rates, and salinity changes in the Latest Pleistocene to Holocene eastern central Arctic ocean. Mar. Geol. 119, 333–355. Steinsund, P.I., Hald, M., 1994. Recent calcium carbonate dissolu- tion in the Barents Sea: paleoceanographic applications. Mar. Geol. 117, 303–316. Voelker, A.H.L., Sarnthein, M., Grootes, P.M., Erlenkeuser, H., Lay, C., Mazaud, A., Nadeau, M.-J., Schleicher, M., 1998. Correlation of marine 14 C ages from the Nordic Seas with the GISP2 isotope record: implications for 14 C calibration beyond 25 ka BP. Radiocarbon 40 Ž1., 517–534. Vogt, C., 1997. Regional and temporal variations of mineral assemblages in Arctic Ocean sediments as climatic indicator during glacialrinterglacial changes. Rep. Pol. Res., vol. 251. Alfred Wegener Institute, Bremerhaven, 309 pp. Wahsner, M., Muller, C., Stein, R., Ivanov, G., Levitan, M., ¨ Shelekhova, E., Tarasov, G., 1999. Clay–mineral distribution in surface sediments of the Eurasian Arctic Ocean and continental margin as indicator for source areas and transport pathways—a synthesis. Boreas 28 Ž1., 215–233. Winsnes, T.S., 1988. Geological map 1:1000000. Bedrock map of Svalbard and Jan Mayen. Nor. Polar. Temakart. 3, 12 pp. Wollenburg, J.E., Kuhnt, W., Mackensen, A., 2001. Changes in Arctic Ocean paleoproductivity and hydrography during the last 145 kyr: the benthic foraminiferal record. Paleoceanography 16 Ž1., 65–77.