Non-Glacial Paleoenvironments and the Extent of Weichselian Ice Sheets

ARTICLE IN PRESS

Quaternary Science Reviews 22 (2003) 2267–2283

Non-glacial paleoenvironments andthe extent of Weichselian ice sheets on Severnaya Zemlya, Russian High Arctic Alexandra Raaba,*, Martin Mellesb, Glenn W. Bergerc, Birgit Hagedornd, Hans-Wolfgang Hubbertene a Institute of Geography, University of Regensburg, D-93040 Regensburg, Germany b Institute for Geophysics and Geology, University of Leipzig, TalstraX e 35, D-04103 Leipzig, Germany c Desert Research Institute, Earth and Ecosystem Sciences, 2215 Reggio Parkway, Reno, NV 89512-1095, USA d Quaternary Research Center, University of Washington, Johnson Hall 19, Box 351360, Seattle, WA 98195-1360, USA e Alfred Wegener Institute for Polar and Marine Research, Research Unit Potsdam, Telegrafenberg A43, D-14473 Potsdam, Germany Received15 September 2002; accepted19 April 2003

Abstract

The extent of the Barents-Kara Sea ice sheet (northern Europe andRussia) duringthe Last Glacial Maximum (LGM), in Marine Isotope Stage (MIS) 2 is controversial, especially along the southern andnortheastern (Russian High Arctic) margins. We conducted a multi-disciplinary study of various organic and mineral fractions, obtaining chronologies with 14C andluminescence dating methods on a 10.5 m long core from Changeable Lake (4 km from the Vavilov Ice Cap) on Severnaya Zemlya. The numeric ages indicate that the last glaciation at this site occurred during or prior to MIS 5d-4 (Early Middle Weichselian). Deglaciation was followed by a marine transgression which affected the Changeable Lake basin. After the regression the basin dried up. In late Middle Weichselian time (ca 25–40 ka), reworked marine sediments were deposited in a saline water body. During the Late Weichselian (MIS 2), the basin was not affectedby glaciation, andlacustrine sedimentswere formedwhich reﬂect coldandaridclimate conditions.During the termination of the Pleistocene andinto the Holocene, warmer andwetter climate conditionsthan before led to a higher sediment input. Thus, our chronology demonstrates that the northeastern margin of the LGM Barents-Kara Sea ice sheet did not reach the Changeable Lake basin. This result supports a modest model of the LGM ice sheet in northern Europe determined from numeric ice sheet modelling and geological investigations. r 2003 Elsevier Ltd. All rights reserved.

1. Introduction intermediate estimates selected through numerical ice sheet andsolidearth modelling(e.g., Peltier, 1994; Knowledge of the geographic extent and thickness of Siegert andMarsiat, 2001 ; Siegert et al., 2001; Charbit continental ice sheets during the Last Glacial Maximum et al., 2002). (LGM) is important for modelling of past climates and In northern Europe andnorthwestern Russia, the sea level (e.g., Peltier, 1994). However, the Russian High extent of the Barents-Kara Sea ice sheet during the Arctic remains the last continental area, where the LGM LGM has remaineduncertain. This is largely because of is uncertain (Clark andMix, 2002 ). For the Russian disagreement over the age of recognized ice-margin High Arctic, both maximalist (Denton andHughes, deposits. Mangerudet al. (2002) have reviewedthe 1981; Grosswald, 1998; GrosswaldandHughes, 2002 ) glacial geology in northwestern Russia andtheir andminimalist (e.g., Velichko et al., 1997; Svendsen interpretation strongly support a minimalist estimate et al., 1999; Gaultieri et al., 2000; Mangerudet al., 2002 ) for the LGM extent of the Barents-Kara Sea ice sheet. ice-extent estimates have been argued, with some Much of the critical evidence comes from luminescence ages for sand-sized quartz in discontinuous, subaerial deposits of non-glacial sediments from the Pechora *Corresponding author. Tel.: +49-941-943-1572; fax: +49-941-943- 5032. Lowland(e.g., former beaches, exposedﬂuvial terraces). E-mail address: [email protected] These age estimates range stratigraphically from ca 100 (A. Raab). to 20 ka. Finite andlower-limit (‘‘inﬁnite’’) radiocarbon

(14C) ages from relatedmaterial support these quartz only if the precipitation across the Kara Sea was luminescence age estimates. However, goodrecordsof suppressedby a polar desertenvironment. continuous sedimentation related to this controversy have not been reportedbefore. The Taymyr Peninsula (Fig. 1) lies at the northeastern 3. Study site boundary of the minimalist estimates of the Barents- Kara Sea ice sheet (e.g., Svendsen et al., 1999; The Severnaya Zemlya Archipelago is a key area for Alexanderson et al., 2002; Mangerudet al., 2002 ). In the understanding of the Late Quaternary palaeoenvir- this paper we present geological, paleoenvironmental onmental history in the Russian High Arctic. This andgeochronological evidence from a 10.5 m long region is extremely sensitive to environmental changes, sediment core from Changeable Lake (Severnaya due to its geographical position both in the high Zemlya Archipelago, north of the Taymyr Peninsula) latitudes and in the transition zone from West Siberian that provides the first well-dated support from a marine to East Siberian continental climate (Ebel et al., continuous sediment record for ice free conditions at 1999; Hahne andMelles, 1999 ; Harwart et al., 1999). the LGM. We employeda multi-disciplinary approach The archipelago is situatedin northern Siberia (78– to obtain information not only on the glacial history, 81N, 96–106E), between the Kara Sea to the west and but also on past climate andsea-level changes that may the Laptev Sea to the east (Fig. 1). The archipelago have affectedice expansion anddecay. consists of more than 30 islands, with the largest being October Revolution Island, and is the easternmost area affectedby modernglaciation. For instance, the Vavilov Ice Cap near to Changeable Lake covers 1820 km2 and 2. Regional context rises up to 728 m a.s.l. The climatic conditions on the Severnaya Zemlya In order to use a uniform stratigraphical nomencla- Archipelago are extremely severe due to the combina- ture, we follow that proposedby Mangerud(1989) : tion of low air temperature (annual mean: 13 to Early Weichselian (MIS 5d-5a; 117–74 ka); Middle 14C) andstrong winds.The ablation season has Weichselian (MIS 4-3, 74–25 ka); Late Weichselian especially changeable weather conditions. Summer (MIS 2, 25–10 ka), as suggestedby the QUEEN temperatures can rise up to 10–15C, andsnow storms scientific community (Svendsen et al., 1999). can also occur in this season, being similar in magnitude As summarizedby Mangerudet al. (2002) andothers to those in winter (Vaikmae. et al., 1988). The annual (Thiede et al., 2001), the results of recent fieldstudies precipitation on October Revolution Islandvaries preclude the existence of a large Panarctic Ice Sheet between 240 and400 mm, with about 70% falling as during the Late Weichselian. Instead, Asian Arctic snow. The amount of precipitation depends on the glaciation during the Late Weichselian (MIS 2) is distance from both the sea and the Vavilov Ice Dome, consideredto have been restrictedto Europe and which functions as an orographical barrier for air Western Siberia, with alpine-valley glaciation in north- masses from southwest, causing precipitation to be eastern Siberia (Gaultieri et al., 2000). highest on the ice cap (Andreev et al., 1997). In Europe andWestern Siberia, glaciation was The vegetation on the archipelago is typical of polar somewhat larger during Early/Middle Weichselian times desert/high arctic tundra: mosses and lichen dominate, (MIS 5d-4), than during the LGM when it expanded whereas flowering plants are rare. A larger diversity in onto most of the Taymyr Peninsula (Moller. et al., 1999; plant species occurs on river terraces in the central parts Svendsen et al., 1999; Mangerudet al., 2001 ; Andreev of the islands and on climatically favoured sites with et al., 2002; Alexanderson et al., 2002)(Fig. 2). peat growth (Alexandrova, 1988; Andreev et al., 1997). Geomorphological investigations andabsolute age The Severnaya Zemlya Archipelago belongs to the determination (14C, quartz luminescence) of the North Taymyr-Severnaya Zemlya foldarea. October Revolu- Taymyr ice-marginal zone (NTZ) by Alexanderson et al. tion Islandconsists of Paleozoic highly fractured (2001, 2002) indicate that at most a relatively thin Late sedimentary carbonaceous rocks with Ordovician and Weichselian Kara Ice Sheet hadadvancedontoonly the lower Silurian rocks overlain by Quaternary marine northwestern part of the Taymyr Peninsula between 20 sediments (Bolshiyanov, 1985). The geomorphology of and12 ka. This makes a restrictedglaciation on the October Revolution Islandis characterizedby U-shaped Severnaya Zemlya Archipelago, as proposedby Ma- valleys, coldbasedglaciers andoutwash plains ( Ma- keyev andBolshiyanov (1986) and Bolshiyanov and keyev andBolshiyanov, 1986 ). Periglacial processes Makeyev (1995), questionable. According to numerical have produced patterned ground, arctic soils and limited ice sheet modelling by Siegert, M.J. et al. (1999a, 2001), chemical weathering (Pfeiffer et al., 1996). and Siegert andMarsiat (2001) a restrictedglaciation of Changeable Lake (79070N, 95070E) is locatedon the Severnaya Zemlya Archipelago couldhave occurred southwestern October Revolution Island, 4 km to the ARTICLE IN PRESS A. Raab et al. / Quaternary Science Reviews 22 (2003) 2267–2283 2269

Fig. 1. Location of the Severnaya Zemlya Archipelago in the Russian High Arctic (Inlet), andmap of the archipelago with the Changeable Lake situatedon southwestern October Revolution Island. southwest of the Vavilov glacier edge (Fig. 1). The lake outwash plain. The outlet is at the south via an o40 m lies in an oblong, SSW to NNE trending depression deep canyon, which leads to the Kara Sea (Bolshiyanov, (Fig. 3a), which penetrates below the Vavilov Ice Dome 1985). The lake has an ice cover most of the year. andis believedto be an oldkarst form, developedin calcareous andgypsiferous bedrock( Bolshiyanov, 1985). Changeable Lake consists of several basins, 4. Methods divided by sills (Fig. 3b). The lake is 10 km2 large and at its maximum 18 m deep. The modern lake level is 4.1. Coring situatedabout 6 m a.s.l. This hydrologically open, hardwater lake is predominantly fed by glacial melt- Sediment coring in Changeable Lake was conducted water (July–September) from the north across an in two different basins (Fig. 3b). The cores where ARTICLE IN PRESS 2270 A. Raab et al. / Quaternary Science Reviews 22 (2003) 2267–2283

70û N

50û N 50û 80 60û N 60û

N Arctic Ocean

80û N sles British I

rian Sev Sibe 10û W Svalbard erna ew Norwegian N Islands Franz Joseph Land ya Sea Zem lya 140û E ? Laptev Barents Sea North Scandinavia Sea Sea Novaya Zemlya Kara Lena 0û Sea ? Mts. 130û E 70ûN Baltic Byrranga Sea Ya m a l Lowland Peninsula Taymyr

Timan Ri Timan Yenissei ? Arctic Circle Putorana 120û E Plateau

10û E dge Central

a Ural Mounta Ural

e O West Moscow b

Siberian Uplands60û N

driatic S Siberian A

40û N ins

Plain 20û E Black Sea 110û E

50û N

Mediterr C

aspian Sea aspian Late Weichselian glacial maximum anean Sea - according to various sources Aral - according to Svendsen et al. (1999) Sea Early / Middle Weichselian glacial maximum - according to Svendsen et al. (1999) Quaternary glaciation limit (asynchronous) - according to various sources

30û E 40û E 50û E 60û E 70û E 80û E 90û E 100û E Fig. 2. Quaternary glaciation limit, Early/Middle Weichselian and Late Weichselian glacial maximum ice sheet limits for northern Eurasia, derived from geological data after Svendsen et al. (1999). recoveredthrough holes in the ice cover using a continuously in 2 cm segments. The subsamples were percussion piston corer that consists of steel tubes and freeze-dried and the water content per wet bulk sediment inner PVC liners of 6 cm diameter (Melles et al., 1994). was calculated from the difference of the wet and the dry The sediment record was obtained by coring of sample. Further sedimentological, mineralogical, geo- succeeding and partly overlapping up to 3 m long chemical, biogeochemical, geochronological andmicro- sections. At sites PG1238 andPG1239 ( Fig. 3b), coring fossil analyses were conducted only on core PG1238 yielded complete sediment successions of 10.5 and which, according to the analyses described above, shows 12.7 m lengths, respectively (Fig. 4). Their ﬁnal depths a comparable sediment succession to that of core were determined by correlation of the overlapping core PG1239 (Fig. 4). sections. Grain-size analyses were carriedout on a laser particle analyser (Galai-CIS) after sample dispersion with 4.2. Laboratory methods ammonia. The total carbon (TC wt%), total nitrogen (TN wt%) andtotal sulphur (TS wt%) contents were Prior to the opening of the cores, measurements of the determined with an automatic CNS 932 Mikro analyser magnetic susceptibility, p-wave velocity (Vp) and (Leco corp.). The total organic-origin carbon (TOC gamma-ray density (GRD) were conducted at intervals wt%) content was measuredon corresponding samples of 1 cm (Multi-Sensor Core Logger MSCL 14, Geotek after HCl (10%) aciddigestion to remove the carbonate corp.) as described by Weber et al. (1997). on a CS-Analyser Metalyt (Eltra corp.). The total The cores were split along their axis into two halves. inorganic-origin carbon (TIC wt%) was calculatedfrom Following a photodocumentation and sediment descrip- the difference of TC (wt%) and TOC (wt%). tion, smear slides were taken for a ﬁrst microscopical The bulk mineralogy was analysedby X-ray diffrac- investigation of sample characteristics anddiatom tion (Philips PW3020) on powder samples with cor- content. Subsequently, one core half was subsampled undum as standard (5:1) for semi-quantitative analysis ARTICLE IN PRESS A. Raab et al. / Quaternary Science Reviews 22 (2003) 2267–2283 2271

Fig. 3. (A) Sketch map of the Changeable Lake surrounding showing the Changeable Lake depression that penetrates below the Vavilov Ice Cap, the location of marine sediments in the northeastern catchment and the bed dip inclination of basement rocks (crossed arrows) (Bolshiyanov and Makeyev, 1995). (B) Bathymetrical map of Changeable Lake showing several basins divided by sills (Bolshiyanov andMakeyev, 1995 ) andthe coring sites PG1238 andPG1239 in two differentbasins. as described by Vogt (1997). The evaluation of the X-ray AAA-method. The silica portion was removed by heavy 3 spectra was made with the MacDiff 3.3.1 PPC-software liquid(ZnCl 2, 2.0 g/cm ) separation andtreatment with (rby R. Petschik). The results are presentedin peak HF acid, and 14C dating was carried out on the intensity ratios: Imineral /Icorundum(0 3 0). pretreatedandsieved20–100 mm pollen fraction. Un- The bulk geochemistry was determined by X-ray determined insect fragments, undetermined organic fluorescence analysis (SRS 3000, Siemens) on fusedglass matter andmixedbenthic foraminifera were extracted beads of lithium tetraborate (6:1). The major element by wet-sieving (>63 mm), followedby hand-picking andtrace element concentrations were normalizedto from the dry sample under a stereo microscope. The 14C titanum (element/titanum ratio). dating was carried out at the University of Erlangen- Water-soluble cations andanions were extractedby a Nurnberg. (Erl) andat the Leibniz Laboratory in Kiel methodmodifiedfrom Gerasimov andGlazovskaya (Kia). All 14C ages are uncalibrated, and therefore (1965). One gram of sample was mixedwith 25 ml reportedhere as ‘‘BP’’. Calibration wouldnot affect our distilled, deionized water, shaken for 3 min, and finally conclusions. filteredthrough a 45 mm mesh sieve. Prior to the Infrared-stimulated luminescence (IRSL), multi-ali- measurement of element concentrations, electrical con- quot dating (e.g., Aitken, 1998; Berger andDoran, 2001 ) ductivity (EC) and pH of the water extracts were was conducted on the polymineral fine-silt fraction, and determined. Cations were analysed with an Inductively blue-photon-stimulated-luminescence (B-PSL) dating of CoupledPlasma Optical Emission Spectrometer sand-size (90–180 mm) quartz was carriedout by the (ICP-OES, Perkins Elmer) andanions with an Ion single-aliquot-regenerative-dose (SAR) method (e.g., Chromatograph (IC2001, Eppendorf, Biotronik). For Murray andWintle, 2000 ). Details of these methods all chemical analyses international standards (NBS, and results (dose-rate and luminescence data) as well as SRM) were usedto check the analytical precision. full age-interpretation of the results will be published Analytical errors are below 5% for major ions and elsewhere. In order to obtain the required sample size of anions, andbelow 10% for trace elements. 2 g quartz for the SAR method, we isolated the sand 14C dating was conducted by accelerator mass from 9.5 and10.5 cm thick segments by wet sieving. spectrometry (AMS) on different organic fractions. Further treatment anddatingwas carriedout at the Humic acids were extracted by the standard AAA- Nordic Laboratory for Luminescence Dating, Ris^ method (acid-alkali-acid). One pollen sample was National Laboratory, in Roskilde (Denmark). Eight extractedby a modifiedmethodaccordingto Brown IRSL ages were obtainedfrom feldspar grains in the et al. (1989, 1992). This sample was pretreatedwith the fine-silt fraction (4–11 mm) from 2.5 cm thick segments. ARTICLE IN PRESS 2272 A. Raab et al. / Quaternary Science Reviews 22 (2003) 2267–2283

The samples were preparedanddatedatthe Desert of the microfossils, which argues against their redeposi- Research Institute, Reno (USA) by Berger. tion, andon the foraminifera assemblage that represents a typical arctic shelf fauna indicating full marine 4.3. Statistical analysis conditions. Therefore, the lack of stratification of the sediment is probably caused by bioturbation due to a An analysis of variance (ANOVA) was carriedout on rich marine endobenthic fauna. Further support for a the data of selected water-soluble elements, electrical marine origin of the sediment is provided by the conductivity, pH andon selectedmajor andtrace maximum in pyrite concentrations (Fig. 6), occurring elements. The F-test clearly showedthat the differences as framboids and inside foraminifera shells. Marine in the element composition between individual facies conditions are indicated also by the composition of the (see below) are significant. The multiple comparison of water-soluble elements (Table 1), particularly the con- 2 the averages using the LSD-test statistics showed tents in Cl andSO 4 which reflect the ion concentra- significant differences also between the facies at a tion of the interstitial water andeasy soluble salt confidence level a ¼ 0:05 (Tables 1 and2 ). precipitates. Diatoms are absent in the entire core.

5.3. Drying-up facies 5. Facies classiﬁcation and interpretation The drying-up facies (9.90–9.88 m) is representedby a Six facies can be distinguished in the Changeable 2 cm thick layer of calciferous, sulphur-containing Lake sediment cores: (1) a glacigenic facies, (2) an in situ sediment. It is characterized by distinct maxima in Ba/ marine facies, (3) a drying-up facies, (4) a reworked Ti, Sr/Ti, Co/Ti, Mo/Ti andFe/Ti element ratios ( Table marine facies, (5) a saline facies and(6) a freshwater 1), indicating the presence of gypsum. The geochemical facies (Fig. 4). Since both cores show the same facies composition of the facies hints at a periodwhen the lake succession, local sedimentation effects in the lake are basin was isolatedfrom the sea anddriedup. Missing neglegible. Thus, the results from core PG1238, pre- relict sediments and sediment sorting suggest that the sented and discussed in this paper, are considered gypsum is unlikely to have been reworkedinto the lake representative of the evolution of the entire Changeable from the catchment area. Lake andits surroundings. 5.4. Reworked marine facies 5.1. Glacigenic facies The reworked marine facies (9.88–8.50 m) is in The glacigenic facies at the base of the sediment core contrast to the in situ marine facies well laminated. (10.43–10.13 m) is characterizedby a greyish, consoli- The laminae (thickness 0.1–1 cm) are deﬁned by sedi- dated and massive diamicton that has a larger grain-size ment colour (olive-green) andgrain-size (silt/clay) median and a higher gamma-ray density (GRD) than variations. They suggest variable sediment supply from the overlaying sediments (Figs. 4 and5 ). The diamicton the catchment area andthe absence of bioturbating is interpretedas till. It has a sharp upper boundary and marine fauna. Pyrite is abundant (Fig. 6). The water- consists of debris from red-coloured Devonian sand- soluble element composition shows high concentrations + 2 stone. The occurrence of low amounts of TOC andTN in Na ,Cl andSO 4 (Table 1), andthe electrical indicate that older organic material is incorporated. conductivity of the water extracts, being twice as high as Single Quaternary foraminifera andhigh TIC contents in the in situ marine facies suggests formation during the hint at the incorporation of carbonaceous rocks and establishment of a saline water body in the formerly Quaternary marine sediments. dried-up lake basin.

5.2. In situ marine facies 5.5. Saline facies

The in situ marine facies (10.13–9.90 m) is built up by The saline facies (8.50–7.85 m) is a black clay/silt an olive-green coloured, fine grained (clay/silt) and fraction which shows a distinct and variable lamination massive sediment (Fig. 5). Benthic foraminifera from (1–5 mm thicknesses) (Fig. 5). These suggest sediment several taxa (e.g., Amonia, Astrononion, Cassidulina, formation in anoxic bottom water, possibly due to a Cibicides, Elphidiella, Fursenkoina, Islandiella, Lagena, density stratification of the water column in conse- Lobulatulus, Melonis, Nonionella, Triloculina; A. Mack- quence of a high salt concentration and/or a perennial ensen, pers. comm., 1998) are abundant, and single lake-ice cover, both of which preclude water mixing. A ostracodes and fragments of mussel shells are present. high salt concentration is indicated by a high electrical The interpretation of this facies as being of in situ conductivity reflecting extraordinary high concentra- marine origin is basedmainly on the goodpreservation tions of the cations Ca2+,Mg2+,K+,Na+,Sr2+,Ba2 ARTICLE IN PRESS A. Raab et al. / Quaternary Science Reviews 22 (2003) 2267–2283 2273 2 4 (ppm) SO Cl (ppm) P(aq) (ppm) 2+ Ba (ppb) ent core PG1238 2+ Sr (ppb) Si(aq) (ppm) 4+ (ppm) Mn Fe(aq) (ppm) 3+ Al (ppm) + (ppm) Na + (ppm) K 2+ Mg (ppm) 2+ (ppm) pH Ca S/cm) m ( Stdn. error 68.17Stdn. error 0.12Stdn. error 8.49 5.98 0.00 0.10 5.38 0.49 0.00 0.64Stdn. error 0.43 0.00 4.27 0.00 1.90 0.71Stdn. error 0.04 0.00 1.05 0.52 1.34 0.02 0.81Stdn. 0.24 error 0.00 0.10 0.93 0.33 0.00 0.01 1.75Stdn. 0.87 error 0.00 0.08 0.00 0.44 0.68 0.00 0.00 58.79 0.00 0.32Stdn. 0.37 2.95 error 0.45 2.91 0.00 0.69 0.02 0.00 0.00 6.28 0.02 0.00 0.16 0.03 0.47 1.83 0.00 0.00 0.13 0.00 1.67 0.00 0.06 0.00 0.00 0.01 0.27 0.00 36.61 0.05 0.13 0.00 2.05 0.00 0.00 0.00 0.11 0.00 0.00 6.11 1.90 0.00 0.00 0.06 1.08 0.00 0.00 0.00 19.48 0.00 0.06 1.81 2.00 0.00 0.00 0.92 32.87 4.57 0.00 1.06 0.00 4.45 1.92 0.00 2.42 0.00 6.78 0.00 1.29 0.00 0.00 13.42 0.00 0.00 0.00 0.00 0.00 0.00 Stdn. error 20.96 0.12 0.89 0.26 0.45 3.80 0.31 0.17 0.01 0.61 4.42 7.97 0.03 2.50 4.02 3) Average4) Average 797.671) Average 233.75 6.97 55.00 124.00 8.27 36.95 3.2315) 8.50 11.59 Average 3.45 3.13 62.574) 121.66 0.11 2.72 8.84 Average 32.7 45.78 0.06 7.873) 11.69 117.76 5.02 33.86 Average 0.54 31.97 0.15 1.95 3.601) 11.21 127.31 0.55 11.00 Average 1.36 0.02 33.96 519.86 0.121) 0.34 22.37 12.31 135.47 12.26 0.01 Average 8.21 0.51 29.61 0.15 40.07 0.77 7.20 5.90 0.20 16.65 131.41 11.80 28.60 9.82 37.34 1.07 28.83 0.15 0.29 303.90 0.35 0.34 8.54 5.08 9.66 10.27 30.46 0.50 0.83 0.15 0.27 0.25 0.30 21.43 530.15 5.86 19.70 12.10 22.37 1.13 0.25 0.36 13.80 103.79 5.12 0.64 550.9 41.44 0.67 0.36 22.47 0.25 79.79 535.6 5.26 105.12 0.25 44.37 23.87 0.26 738.95 111.25 11.87 89.88 0.26 42.18 130.97 488.92 124.55 68.05 9.00 455.83 86.05 35.2 122.56 9) Average 119.88 7.48 6.26 2.27 3.48 15.80 2.12 1.12 0.02 4.34 37.77 37.13 0.15 10.82 12.41 ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ n n n n n n n n n Table 1 Analysis of variance (ANOVA) on the data of water-soluble elements, electrical conductivity and pH as well as on the data of bulk geochemistry of sedim Water-soluble elements, electrical conductivityFacies and pH EC Reworkedmarine f. ( In situ marine f.Bulk geochemistry ( FaciesFreshwater f.Saline f. ( Reworkedmarine f. ( ( Drying-up f.In situ marine f. ( Si/Ti ( Al/Ti Fe/Ti Mn/Ti Mg/Ti Ca/Ti Na/Ti K/Ti P/Ti S/Ti Ba/Ti Co/Ti Cr/Ti Ni/Ti Sr/Ti Saline f. ( Freshwater f. ( ARTICLE IN PRESS 2274 A. Raab et al. / Quaternary Science Reviews 22 (2003) 2267–2283

Table 2 Similarity of the different facies according to the data of water-soluble elements, electrical conductivity (EC) and pH of PG1238 (+ denotes a statistically signiﬁcant difference at a ¼ 0:05)

Freshwater facies Freshwater facies Freshwater facies Reworked marine Saline facies vs. Saline facies vs. in vs. saline facies vs. reworked vs. in situ marine facies vs. in situ reworked marine situ marine facies marine facies facies marine facies facies

EC + + + + pH+++ ++ Ca2+ + ++ Mg2+ + ++ K+ +++ ++ Na+ ++ ++ Al3+ ++ ++ Fe(aq) + + + + Mn4+ + ++ Si(aq) + + Sr2+ + ++ Ba2+ + P(aq) + + + Cl ++ 2 SO4 + ++

2 andanions SO 4 andCl in the water-soluble element stable climatic conditions, also shown by marked assemblage (Table 1). These elements probably derive ﬂuctuations of TOC, TN andTIC ( Fig. 5). from the input of reworkedmarine sediments. Due to Generally, low organic content indicate that biogenic higher evaporation than the sum of precipitation and production was low in the lake and its catchment area, glacial meltwater supply, a concentratedbrine devel- or that the concentration of organic compounds in the opedin the water bodythat ledto an enrichment in sediments was restricted by their dissolution in the oxic some elements andto the precipitation of easily soluble water column or dilution by a high clastic-sediment mixedsalts. Observations on lakes in Antarctica have supply. The TOC/TN-ratio of o10 evidences that the shown that evaporites can be formedeven in lakes with organic material derives mostly from non-vascular an ice cover of up to 4.5 m, still allowing the exchange of plants (e.g. algal production). An exception to this is a gases, liquids and solids (Andersen et al., 1993). sandy layer at 25 cm depth, whose TOC/TN-ratio of >10 points to a single inﬂow event, when more coarse- grainedsediments andmore terrestrial organic material 5.6. Freshwater facies was suppliedto the lake.

The freshwater facies (7.85–0 m) is a reddish-brown silt/clay derived from the red-coloured Devonian 6. Chronology sandstones in the catchment or from the cover loams derived from these sandstones (Bolshiyanov et al., From the in situ marine facies two infinite 14C ages of 1990). The water-soluble element assemblage differs >47.5 and>48.4 ka BP were measuredon benthic distinctly from that in the underlaying saline facies foraminifera (Table 3, Fig. 7). IRSL from fine-silt + 2 (Tables 1 and2 ) with low Na ,Cl andSO 4 values, feldspars gave an age of 5373 ka, but quartz-sand andan absence of marine organisms, andindicating SAR dating gave inconsistent ages of 3473 and deposition in freshwater. Today, Changeable Lake has a 8476 ka within the same relatively thin horizon. We non-stratified, freshwater body (U. Wand, pers. comm., presently do not consider the quartz SAR ages to be as 2002). accurate as the IRSL ages. Little is known of the In the lower part of the facies (7.85–2.56 m) black effectiveness of zeroing of the quartz SAR clock in layers andirregular but well expressedlaminae, defined marine sediments, in contrast to its demonstrated by colour andgrain-size variations, occur. These layers zeroing in eolian sediments (Murray andWintle, indicate that periods with anoxic bottom water condi- 2000). Thus the quartz SAR estimate of 86 ka could tions, characteristic for the underlaying saline facies, still readily be an over-estimate of the deposition age. On the occurredfrom time to time. The upper part ( o2.56 m) other hand, the younger quartz SAR age estimate of of the facies, in contrast, has a reddish-brown colour 34 ka is inconsistent with the limiting ages from 14C andis poorly stratifiedto massive. This may reflect less dating. The 14C dated foraminifera have an abundance, ARTICLE IN PRESS A. Raab et al. / Quaternary Science Reviews 22 (2003) 2267–2283 2275

PG1238 PG1239

Facies Magnetic GRD Facies Magnetic GRD susceptibility (g/cm3) susceptibility (g/cm3) (10-6SI) (10-6SI)

0 200 1.4 1.8 2.2 0 200 1.4 1.8 2.2 0 0

1 1

2 2

3 3

4 4

5 5

Sediment depth (m) 6 6

7 7

8 8

9 9

10 10

Legend freshwater facies drying-up facies

foraminifera saline facies in situ marine facies pyrite reworked marine facies glacigenic facies

Fig. 4. Facies succession vs. the physical properties magnetic susceptibility and gamma-ray density of the sediment cores PG1238 and PG1239.

assemblage, andgoodpreservation that strongly sug- expect the 14C ages to lie in the ﬁnite range of 30–35 ka gests that they are autochthonous, rather than redepos- BP. Possibly this sample of quartz was accidentally ited. If this SAR age estimate was accurate, one would exposedto light duringsampling or sample preparation, ARTICLE IN PRESS 2276 A. Raab et al. / Quaternary Science Reviews 22 (2003) 2267–2283

Fig. 5. Grain-size median, contents of total inorganic carbon (TIC), total organic carbon (TOC), total nitrogen (TN), total sulphur (TS) and TOC/ TN-ratio in sediment core PG1238 (for legend see Fig. 4).

( Fig. 6. Semi-quantitative bulk mineralogy of the sediment core PG1238. Presented are the peak intensity ratios Imineral/Icorundum(0 3 0). 1.4 nm=14 A- minerals, 1.0 nm=10 A-minerals,( Q=quartz (1 0 0), Or=orthoclas (0 0 2), Al=albite (0 0 2), Cc=calcite (1 0 4), Dol=dolomite (1 0 4), Pyrite=pyrite (3 1 1) (for legendsee Fig. 4). ARTICLE IN PRESS A. Raab et al. / Quaternary Science Reviews 22 (2003) 2267–2283 2277

Table 3 Core PG1238: 14C (AMS) ages, IRSL (infrared-stimulated luminescence), and quartz-SAR (single-aliquot-regenerative-dose) luminescence ages

Lab. no. Method Material Sediment depth (cm) Sediment facies Age (yr BP) d13C(%)

Erl-1158 14C Humic acids 11–17 Freshwater f. 109227150 — Pot99-24 IRSL Feldspar fine-silt fraction 20.8 Freshwater f. 44007380 — Pot99-25 IRSL Feldspar fine-silt fraction 200.8 Freshwater f. 40307420 — Erl-1151 14C Humic acids 258–264 Freshwater f. 81897350 — Pot99-26 IRSL Feldspar fine-silt fraction 307.8 Freshwater f. 51207680 — Erl-1230 14C Humic acids 360–362 Freshwater f. 9253771 — KIA-6086 14C Insect remains (undet.) 610–616 Freshwater f. 60207100 35.5170.17 Pot99-28 IRSL Feldspar fine-silt fraction 714.8 Freshwater f. 12 1107680 — Erl-1078 14C Pollen grains 770–776 Freshwater f. 11 377785 — Erl-1078 14C Humic acids 770–776 Freshwater f. 18 4347120 — Pot99-29 IRSL Feldspar fine-silt fraction 811.8 Saline f. 19 20071300 — Pot99-30 IRSL Feldspar fine-silt fraction 841.8 Saline f. 20 5007910 — KIA-6087 14C Insects andorganic remains (undet.) 862–864 Reworkedmarine f. 24 170 7160 13.4470.08 Erl-1079 14C Humic acids 882–888 Reworked marine f. 22 9537161 — KIA-6088 14C Insects andorganic remains (undet.) 924–928 Reworkedmarine f. 27 400 7220 15.5770.27 KIA-6089 14C Insects andorganic remains (undet.) 962–968 Reworkedmarine f. 25 570 7220 19.7970.24 Pot99-32 IRSL Feldspar fine-silt fraction 962.8 Reworked marine f. 33 40072100 — Ris^995404 B-PSL Quartz sandfraction 990.0–999.5 In situ marine f. 34 000 73000 — Pot99-33 IRSL Feldspar fine-silt fraction 996.8 In situ marine f. 53 00073000 — KIA-6090 14C Benthic foraminifera 997–1000 In situ marine f. >47 560 0.2770.11 KIA-6091 14C Benthic foraminifera 1000–1006 In situ marine f. >48 380 0.0270.14 Ris^995403 B-PSL Quartz sandfraction 1000.5–1011.0 In situ marine f. 84 000 76000 —

accounting for an age underestimation. In comparison, ages of 2171 ka from the lower part andof 19 71ka the stratigraphic consistency of the IRSL ages suggest from the upper part of the saline facies (Table 3, Fig. 7) that they are more reliable, including the result of 53 ka delimit its formation during Late Weichselian time. at ca 10 m. Furthermore, fine-silt feldspar luminescence In the freshwater facies, 14C ages were obtainedfrom dating of polar-region lake core sediments has been insect remains (6.0 ka BP), a pollen extract (11.4 ka BP) shown elsewhere (Berger andAnderson, 2000 ; Berger and4 humic acidsamples (bottom to top: 18.4, 9.3, 8.2 andDoran, 2001 ) to be capable of accuracy over the age and10.9 ka BP) ( Table 3, Fig. 7). Two age inversions in range studied here. In any case, all reasonable age this data set (360–362 cm and 11–17 cm depths), along estimates here give a minimum age of Middle Weichse- with large age differences between two samples from the lian (MIS 3–4) for the in situ marine facies as well as the same depth (770–776 cm), as well as an unrealistic age glacigenic facies. close to the sediment surface (11–17 cm), make the 14C In the reworkedmarine facies 14C ages of 25.6, 27.4, age estimates highly questionable. At best, these 14Cage 23.0 and24.2 ka BP were measuredon mixtures of insect estimates hint at a formation of the freshwater facies remains andplant debris, which hadto be combinedto after the Late Weichselian. In contrast, the IRSL age provide sufficient sample sizes, and on humic acids estimates (top to bottom: 4.4, 4.0, 5.1, 12.1 ka) define a (Table 3, Fig. 7). A more detailed determination of the clearer age-depth trend, one that is highly consistent organic remains was excluded due to their poor with the IRSL age-depth trend deeper in the core (Table preservation. Keeping in mindthat AMS 14C age 3, Fig. 7). This was expectedbecause laminatedlake estimates can be influencedby a variety of natural sediments, such as these, are ideal for luminescence contamination andreservoir effects, by sample size and dating of the fine-silt feldspars (Berger, 1990; Berger and sample processing (Bjorck. et al., 1991; Snyder et al., Easterbrook, 1993). Possibly the IRSL age estimate of 1994; Wolfarth et al., 1998; Turney et al., 2000; Kilian 4.470.4 ka for the top-most sample exceeds the deposi- et al., 2002), resulting in 14C ages of up to several tion age, but given the analytical uncertainties (all at a thousand years different from time of deposition, our 67% confidence level) and the uncertainties of sediment– measured 14C ages indicate a formation of the reworked water interface core recovery with the type of coring marine facies during Middle/Late Weichselian time. This device employed here, this IRSL age estimate is reason- is supportedby an IRSL age (silt feldspar)of 33 72ka able. Certainly it is more accurate than the comparable from the lower part of the facies. 14C result (10.9 ka BP), anddoes not affect our main In the saline facies no 14C dating could be carried out conclusions about the paleoenvironmental history of due to insufficient organic content. However, two IRSL this sediment record. ARTICLE IN PRESS 2278 A. Raab et al. / Quaternary Science Reviews 22 (2003) 2267–2283

Fig. 7. Radiocarbon ages of different organic fractions and luminescence (IRSL and SAR) age estimates of different mineral fractions vs. sediment depth in core PG1238. The 8476 ka result from quartz-SAR luminescence dating in the 10-m horizon (Table 3) is not plottedhere. The dashedline links the IRSL results (for legendsee Fig. 4).

7. Climatic and environmental history >44 ka BP. Based on marine sediment records Knies et al. (2001) also propose a grounded ice sheet along the 7.1. Early and Middle Weichselian outer shelf of northern Severnaya Zemlya in at least 340 m water depth during the Middle Weichselian (MIS The glacigenic facies at the base of the sediment 4) glaciation. Relicts of an Early/Middle Weichselian record, consisting of a till, is interpreted as a relict from glaciation on the Taymyr Peninsula, to the south of the the last glaciation of the Changeable Lake area (Fig. 8a). archipelago, are buriedglacier ice bodies, foundat the The till predates the in situ marine facies that from our ‘‘Ice Hill’’ site at the Jenissej River (Astakhov and data is evidently Middle Weichselian in age or older. Isayeva, 1988) andat the Labaz Lake in the Taymyr The glacigenic facies thus couldhave been formed Lowland( Siegert et al., 1999a). Additionally, glacial during the Early Weichselian or early Middle Weichse- erosion features identified in high-resolution seismic lian. This is in agreement with the Early/Middle data from Levinson-Lessing lake are interpreted to be Weichselian glacial limits presentedby Mangerudet al. older than LGM (Niessen et al., 1999). (2001) and Svendsen et al. (1999) (Fig. 2), which indicate The in situ marine facies that overlays the glacigenic a complete ice coverage of Severnaya Zemlya, andwith facies reflects the deglaciation and a marine transgres- tills exposedon October Revolution Islandwhich are sion which ledto full marine conditions in the Change- classified as Early/Middle Weichselian (Alekseev, 1997). able Lake region, probably during Middle Weichselian In addition, Knies et al. (2001) describe a well-defined time (Fig. 8b). This corresponds with ESR ages of 50– moraine ridge, occurring in 385 m water depth west of 60 ka determined on foraminifera and mollusk shells Komsomolets Island, that yielded an ‘‘infinite’’ age of from Quaternary marine sediments occurring near the ARTICLE IN PRESS A. Raab et al. / Quaternary Science Reviews 22 (2003) 2267–2283 2279

Fig. 8. Reconstructedenvironmental history of the Changeable Lake region since its last glaciation duringpresumably Early Weichselian time and resulting facies succession (for legendsee Fig. 4).

Vavilov glacier margin andin the Changeable Lake The subsequent drying up of the marine basin, catchment (Makeyev et al., 1992). A partial marine reflectedby the drying-up facies, was the consequence transgression is also reportedfor the North Siberian of a regression, an aridclimate anda limitedmeltwater Lowland; however, this transgression was estimated to supply (Fig. 8c). The dried basin probably lasted for less have taken place somewhat later, between 50 and26 ka than 20 ka, as inferredfrom IRSL ages in the overlaying BP (KindandLeonov, 1982 ). (3372 ka) andunderlaying (53 73 ka) facies. Erosion, ARTICLE IN PRESS 2280 A. Raab et al. / Quaternary Science Reviews 22 (2003) 2267–2283 as another possible explanation for the time gap, is only in about 4 km distant, the Vavilov Ice Cap did not regarded as unlikely, because neither a relict sediment reach Changeable Lake during the LGM. This supports nor compaction or deformation structures occur in the earlier investigations on the Severnaya Zemlya Archi- underlaying sediment. Interestingly, the interval 30– pelago, where Alekseev (1997), Velichko et al. (1984) 50 ka was a time of development of peat beds on the and Stievenardet! al. (1996) suppose small andrelatively Taymyr Peninsula (Moller. et al., 1999), anddeposition thin ice for the Late Weichselian maximum between 18 of peaty sandin the Taymyr Lowland( Andreev et al., and14 ka BP. Accordingto Makeyev andBolshiyanov 2002). It was also a time of soil development (hence, (1986) the ice sheet coveredthe Changeable Lake warmer climatic intervals) at middle latitudes through- depression, but was still restricted to local ice caps on out the world, based on luminescence and 14C dating of October Revolution Island( Fig. 1). The latter inter- paleosols from Alaska, USA, New ZealandandChina pretation is basedon mammoth findsnext to the (Berger, 2003). modern margin of the Vavilov Ice Cap, dated to 24.9, Sedimentation commenced again some time before 20.0, 19.3 and11.5 ka BP (summarizedby Vasilchuk 3372 ka. During the remaining Middle Weichselian and et al., 1997). However, the lake cores reportedin this parts of the Late Weichselian, until ca 21 ka, the paper demonstrate that a glacier expansion did not take reworked marine facies was formed, reflecting redeposi- place andthat fine-grainedlake sedimentation was tion of marine sediments from the catchment of continuous throughout this period. This suggests that Changeable Lake (Fig. 8d). The input of reworked during the LGM the Eurasian ice sheet did not extend marine sediments with the corresponding elements led to eastwardonto the Severnaya Zemlya Archipelago, and the formation of a saline water body. The fluvial by implication, perhaps also did not extend onto the sediment supply suggests a comparably warmer and Taymyr Peninsula. Further numeric dating of relevant more humidclimate than before. deposits on the Taymyr Peninsula is needed to test this In general, an unstable climate with alternating implication. warmer and cooler phases is described for the Middle A scenario of limitedglaciation during the LGM of Weichselian in the Taymyr region (Andreeva and Kind, the Eurasian Arctic has been modelled by Siegert et al. 1982). Melles et al. (1996), Hahne andMelles (1999) and (1999b) andby Siegert andMarsiat (2001) . They Siegert et al. (1999a) foundevidencefor predominantly propose for the LGM climate of the Eurasian Arctic coldandaridcontinental climate conditions, interrupted that the eastern margin of the LGM ice sheet, andthe by warmer periods with still very cold winters but with central area of the Barents andKara Seas experienced summer temperatures sufficient for the development of very low amounts of ice accumulation (o200 mm yr1), increasedvegetation. andthat the rate of ice accumulation over the Kara Sea was less than 100 mm yr1, making this area similar to 7.2. Late Weichselian the polar desert conditions in central East Antarctica. The regions to the east of Severnaya Zemlya, andthe Parts of the Late Weichselian are representedin the greater part of the Taymyr Peninsula, were certainly not Changeable Lake recordby the saline facies ( Fig. 8e). affectedby glaciation during the LGM ( Fig. 2). This is Formation of this facies startedearlier than 21 ka and demonstrated by lacustrine sediments, massive ground lasted until after 19 ka. It reflects sedimentation under a ice, thick Yedoma (ice-loess) complexes and numerous permanent lake ice cover, with a limitedmeltwater and mammoth finds in these areas (e.g., Isayeva, 1984; precipitation supply that ledto a precipitation of readily Sulerzhitsky, 1995; Vasilchuk et al., 1997; Hahne and soluble salts due to evaporation of the saline water, and Melles, 1999; Siegert et al., 1999a). An exception is the in a stratifiedwater column with anoxic bottom water. northwestern Taymyr Peninsula, where a thin glacier These processes suggest a particularly coldanddry originating from the Kara Ice Sheet inundated the climate. An extreme continental climate is assumedfor present landarea forming the North Taymyr ice- the Late Weichselian of Northwestern Siberia by Kind marginal zone (NTZ) between 20 and12 ka BP andLeonov (1982) and Velichko et al. (1997). The same (Alexanderson et al., 2001, 2002). conclusion was drawn by Siegert et al. (1999a) from permafrost sequences at Labaz Lake, Eastern Taymyr 7.3. Latest Weichselian and Holocene Lowland, which show salt accumulation in a dried lake depression, connected with evaporation and freezing- The global warming at the termination of the out processes in the active layer. Pleistocene ledto an enhancedmeltwater supply and The continuous limnic sedimentation and the lack of to a freshwater body in Changeable Lake earlier than glacial and glacio-lacustrine sediments during the Late 12 ka (Fig. 8f). In the lower part of this freshwater facies, Weichselian in the Changeable Lake depression also during the transition from the Late Weichselian to indicate a small glaciation of October Revolution Island the Holocene, anoxic bottom water conditions develo- at the LGM. Although at present the glacier margin is pedpossibly dueto perennial ice cover of the lake. ARTICLE IN PRESS A. Raab et al. / Quaternary Science Reviews 22 (2003) 2267–2283 2281

Therefore, these represent temporary coldperiods. environmental history of the Severnaya Zemlya Archi- During most of the remaining Holocene the climate pelago since Early Weichselian time. seems to have been warmer andmore stable, resulting in well developed laminations caused by annual meltwater * discharge and by low but constant biomass production. The last ice advance to the Changeable Lake is During the later part of the Holocene the lack of representedby a till. This glaciation took place prior lamination andstronger fluctuations in the biogeochem- to ca 45 ka but most probably just before 53 ka. * ical data suggest that conditions became more instable. Deglaciation of the area commencedsometime after Thereafter, sedimentation conditions pass into modern >53 ka, andwas followedby marine transgression climate conditions, being characterized by a direct into the Changeable Lake basin. * dependence of the sedimentation processes on the air The basin dried up between 53 and 33 ka, suggesting temperature (Bolshiyanov, 1985). a time of an aridclimate anda geomorphological The global ice retreat at the Weichselian/Holocene stability. * transition ledto a sea-level rise that resultedin the Enhanced humidity or temperature during the Mid- separation of the Severnaya Zemlya Archipelago from dle to Late Weichselian transition (ca 33–21 ka) led to the continental mainland( Alekseev, 1997). The low the development of a lake in the basin. The lake water relative landuplift, indicatedby the low marine limit on was saline, in consequence of a fluvial supply of western October Revolution Islandis a further argu- marine sediments and the corresponding ions from ment against a large Late Weichselian ice cover on the the catchment area. * Severnaya Zemlya Archipelago, since glacial loading During the Late Weichselian (between 21 and19 ka), wouldhave ledto a larger isostatic reboundand, a coldandaridclimate ledto a permanent ice cover consequently, a higher marine limit. on the lake, limitedmeltwater andprecipitation More detailed information about the Holocene supply, a density stratification of the water column climatic andglacial history on the Severnaya Zemlya andsedimentation in anoxic bottom water. This Archipelago comes from ice-core data and investiga- limnic sediment formation, and the lack of glacigenic tions of peat sections. According to oxygen isotope sediments, are evidence for a restricted or only local analyses on ice cores from the Vavilov Ice Dome andthe glaciation of the Severnaya Zemlya Archipelago Academy of Science glacier, the Holocene thermal during the time of the LGM. Our dating of a climate optimum on the Severnaya Zemlya Archipelago continuous non-glacial sedimentary record provides occurredbetween 10 and8 ka BP ( Bolshiyanov et al., a significant constraint on the northeastwardlimit of 1990; Vaikmae,. 1991; Stievenardet! al., 1996 ; Andreev the Eurasian ice sheet during the LGM stage. * et al., 1997) andit is presumedthat the glaciers that During the termination of the Pleistocene andthe existedon the archipelago were smaller than at present. entire Holocene (since about 12 ka) sediment forma- This is indicated by 14C ages of lake-swamp samples tion took place under freshwater conditions. A between about 11.5 and8.8 ka BP andof soil horizons warmer andmore humidclimate ledto higher with woodremains between 10.2 and9.0 ka BP, lying sedimentation rates than during the Weichselian. below the recent Vavilov glacier (Velichko et al., 1984). Holocene climate changes are only weakly reflectedin A milder climate than today, during the Early to the Changeable Lake record. Middle Holocene, is also assumed for regions adjacent to Severnaya Zemlya. This is based, for example, on pollen analyses on 14C dated peat sections from different locations in the Russian Arctic (Makeyev et al., 1992; Acknowledgements Kondratjeva et al., 1993; Alekseev, 1997; Serebryanny andMalyasova, 1998 ; Serebryanny et al., 1998; Siegert D. Yu. Bolshiyanov, T. Muller-Lupp. andA. Zielke et al., 1999a). The same conclusions were drawn from are acknowledged for their competent help in collecting palynological, biogeochemical anddiatominvestiga- the sediment cores during the expedition 1996. We tions on lake sediment cores from the Taymyr Peninsula thank F. Niessen for supporting the physical property (Hahne andMelles, 1999 ; Harwart et al., 1999; Kienel, measurements andA. Mackensen for conducting the 1999). foraminifera analysis. Our special thanks are extended to the reviewers M.J. Siegert andJ. Mangerud,andto J. Rose, U. Kienel, T. Kumke andT. Raab for helpful 8. Conclusions comments and discussions. This study was financially supported by the German Federal Ministry for Educa- Basedon our multi-disciplinary investigation of the tion, Science, Research andTechnology (BMBF; Grant Changeable Lake sediment record, the following con- No. 03PL014A,B), andby the US National Science clusions can be drawn concerning the climatic and Foundation (NSF; grant EAR-9909686 to Berger). ARTICLE IN PRESS 2282 A. Raab et al. / Quaternary Science Reviews 22 (2003) 2267–2283

References Brown, T.A., Nelson, D.E., Mathews, R.W., Vogel, J.S., Southon, J.R., 1989. Radiocarbon dating of pollen by accelerator mass Aitken, M.J., 1998. An Introduction to Optical Dating. Oxford spectrometry. Quaternary Research 32, 205–212. University Press, Oxford, 262pp. Brown, T.A., Farwell, G.W., Grootes, P.M., Schmidt, F.H., 1992. Alekseev, M.N., 1997. Paleogeography andgeochronology in the Radiocarbon AMS dating of pollen extracted from peat samples. Russian Eastern Arctic during the second half of the Quaternary. Radiocarbon 34, 550–556. Quaternary International 41/42, 11–15. Charbit, S., Ritz, C., Ramstein, G., 2002. Simulations of Northern Alexanderson, H., Hjort, C., Moller,. P., Antonov, O., Pavlov, M., Hemisphere ice-sheet retreat: sensitivity to physical mechanisms 2001. The North Taymyr ice-marginal zone, Arctic Siberia—a involvedduringthe Last Deglaciation. Quaternary Science Reviews preliminary overview anddating.Global andPlanetary Change 31, 21, 243–265. 427–445. Clark, P.U., Mix, A.C., 2002. Ice sheets andsea level of the Last Alexanderson, H., Adrielson, L., Hjort, C., Moller,. P., Antonov, O., Glacial Maximum. Quaternary Science Reviews 21, 1–7. Eriksson, S., Pavlov, M., 2002. Depositional history of the North Denton, G.H., Hughes, T.J., 1981. The Last Great Ice Sheets. Wiley, Taimyr ice-marginal zone, Siberia—a landsystem approach. New York, 484pp. Journal of Quaternary Science 17, 361–382. Ebel, T., Melles, M., Niessen, F., 1999. Laminatedsedimentsfrom Alexandrova, V.D., 1988. Vegetation of the Soviet Polar Deserts. Levinson-Lessing Lake, Northern Central Siberia—a 30,000 year Cambridge University Press, Cambridge, 228pp. recordof environmental history ? In: Kassens, H., Bauch, H.A., Andersen, D.W., Wharton, Jr., R.A., Squires, S.W., 1993. Terrigenous Dmitrenko, I.A., Eicken, H., Hubberten, H.-W., Melles, M., clastic sedimentation in Antarctic Dry Valley Lakes. In: Green, Thiede, J., Timokhov, L.A. (Eds.), Land-Ocean Systems in the W.J. (Ed.), Physical and Biogeochemical Processes in Antarctic Siberian Arctic: Dynamics andHistory. Lecture Notes in Earth Lakes, Vol. 59, Antarctic Research Series, pp. 71–81. American Science. Springer, Berlin, pp. 425–435. Geophysical Union, Washington, D.C. Gaultieri, L., Glushkova, O., Brigham-Grette, J., 2000. Evidence for Andreev, A.A., Nikolaev, V.I., Bolshiyanov, D.Yu., Petrov, V.N., restrictedice extent duringthe last glacial maximum in the Koryak 1997. Pollen andisotope investigations of an ice core from Vavilov Mountains of Chukotka, far eastern Russia. Geological Society of Ice Cap, October Revolution Island, Severnaya Zemlya Archipe- America Bulletin 112, 1106–1118. lago, Russia. Geographie! physique et Quaternaire 51, 379–389. Gerasimov, I.P., Glazovskaya, M.A, 1965. Fundamentals of Soil Andreev, A.A., Siegert, C., Klimanov, V.A., Derevyagin, A.Yu., Science andSoil Geography. Israel Program for Scientific Shilova, G.N., Melles, M., 2002. Late Pleistocene andHolocene Translations. vegetation andclimate on the Taymyr Lowland.Quaternary Grosswald, M.G., 1998. Late-Weichselian ice sheets in Arctic and Research 57, 138–150. Pacific Siberia. Quaternary International 45/46, 3–18. Andreeva, S.M., Kind, N.V., 1982. Karginsk deposits (in Russian). In: Grosswald, M.G., Hughes, T.J., 2002. The Russian component of an Kind, N.V., Leonov, B.N. (Eds.), The Antropogen of Taymyr. Arctic Ice Sheet during the Last Glacial Maximum. Quaternary Nauka, Moscow, pp. 78–114. Science Reviews 21, 121–146. Astakhov, V., Isayeva, L.L., 1988. The Ice Hill—an example of Hahne, J., Melles, M., 1999. Climate andvegetation history of the retarded deglaciation in Siberia. Quaternary Science Reviews 7, Taymyr Peninsula since Middle Weichselian time—palynological 29–40. evidence from lake sediments. In: Kassens, H., Bauch, H.A., Berger, G.W., 1990. Effectiveness of natural zeroing of the thermo- Dmitrenko, I.A., Eicken, H., Hubberten, H.-W., Melles, M., luminescence in sediments. Journal of Geophysical Research 95, Thiede, J., Timkhov, L.A. (Eds.), Land-Ocean Systems in the 12375–12397. Siberian Arctic: Dynamics andHistory. Lecture Notes in Earth Berger, G.W., 2003. Luminescence chronology of late Pleistocene Science. Springer, Berlin, pp. 407–423. loess-paleosol andtephra sequences near Fairbanks, Alaska. Harwart, S., Hagedorn, B., Melles, M., Wand, U., 1999. Lithological Quaternary Research, in press. and biochemical properties in sediments of Lama lake as indicators Berger, G.W., Anderson, P.M., 2000. Extending the geochronometry for the late Pleistocene andHolocene ecosystem development of the of arctic lake cores beyondthe radiocarbonlimit by using southern Taymyr Peninsula, Central Siberia. Boreas 28, 167–180. thermoluminescence. Journal of Geophysical Research 105 (D12), Isayeva, L.L., 1984. Late Pleistocene glaciation of North-Central 15439–15455. Siberia. In: Velichko, A.A. (Ed.), Late Quaternary Environments Berger, G.W., Doran, P.T., 2001. Luminescence-dating zeroing tests in of the Soviet Union. University of Minnesota Press, Minneapolis, Lake Hoare, Taylor Valley, Antarctica. Journal of Paleolimnology pp. 21–30. 25, 519–529. Kienel, U., 1999. Late Weichselian to Holocene diatom succession in a Berger, G.W., Easterbrook, D.J., 1993. Thermoluminescence dating sediment core from Lama Lake, Siberia and presumed ecological tests for freshwater, glaciomarine and floodplain sediments from implications. In: Kassens, H., Bauch, H.A., Dmitrenko, I.A., western Washington andBritish Columbia. CanadianJournal of Eicken, H., Hubberten, H.-W., Melles, M., Thiede, J., Timokhov, Earth Sciences 30, 1815–1828. L.A. (Eds.), Land-Ocean Systems in the Siberian Arctic: Dynamics ! Bjorck,. S., Infgolfsson,! O., Skog, G., 1991. Radiocarbon dates from andHistory. Lecture Notes in Earth Science. Springer, Berlin, pp. the Antarctic Peninsula region—problems andpotential. Quatern- 377–405. ary Proceedings 1, 55–65. Kilian, M.R., van der Pflicht, J., van Geel, B., Goslar, T., 2002. Bolshiyanov, D.Yu., 1985. Sedimentation on the recent subglacial lake Problematic 14C-AMS dates of pollen concentrates from Lake (by the example of the Izmenchivoye Lake, Severnaya Zemlya Gosciaz (Poland). Quaternary International 88, 21–26. Archipelago) (in Russian). Vestnik DGU 7, 43–50. Kind, N.V., Leonov, B.N., 1982. The Anthropogene of Taymyr (in Bolshiyanov, D.Yu., Makeyev, V.M., 1995. Severnaya Zemlya Russian). Nauka, Moscow, 184pp. Archipelago: glaciation andenvironmental history (in Russian). Knies, J., Kleiber, H.-P., Matthiessen, J., Muller,. C., Nowaczyk, N., Gidrometeoizdat, St. Petersburg, 216pp. 2001. Marine ice-rafted debris records constrain maximum extent Bolshiyanov, D.Yu., Klementyev, O.L., Korotkov, I.M., Nikolaev, of Saalian andWeichselian ice-sheets along the northern Eurasian V.I., 1990. Study of the moraine containing ice core of Vavilov margin. Global andPlanetary Change 31, 45–64. glacier on Severnaya Zemlya (in Russian). Academy of Sciences of Kondratjeva, K.A., Khrutzky, S.F., Romanovsky, N.N., 1993. the USSR Soviet Geophysical Committee. Changes in the extent of permafrost during the Late Quaternary ARTICLE IN PRESS A. Raab et al. / Quaternary Science Reviews 22 (2003) 2267–2283 2283

periodin the territory of the former Soviet Union. Permafrost and Siegert, M.J, Dowdeswell, J.A., Melles, M., 1999b. Late Weichselian Periglacial Processes 4, 113–119. glaciation of the Russian high arctic. Quaternary Research 52/4, Makeyev, V.M., Bolshiyanov, D.Yu., 1986. Formation of the relief 273–285. and deposits in areas of present-day glaciation on Severnaya Siegert, M.J., Dowdeswell, J.A., Hald, M., Svendsen, J.I., 2001. Zemlya. Polar Geography andGeology 10, 331–338. Modelling the Eurasian ice sheet through a full (Weichselian) Makeyev, V.M., Bolshiyanov, D.Yu., Malakhovskiy, D.B., Barans- glacial cycle. Global andPlanetary Change 31, 367–385. kaya, A.I., Ponomareva, D.P., 1992. Stratigraphy andgeochronol- Snyder, J.A., Miller, G.H., Werner, A., Jull, A.J.T., Stafford Jr., T.W., ogy of the Pleistocene deposits of Severnaya Zemlya (in Russian). 1994. AMS-radiocarbon dating of organic-poor lake sediment, an Geokhronologiya Chetvertichnogo Periots, 132–137. example from Linnevatnet,! Spitsbergen, Svalbard. The Holocene 4, Mangerud, J., 1989. Correlation of the Eemian and Weichselian with 413–421. the deep-sea oxygen isotope stratigraphy. Quaternary International Stievenard,! M., Nikolaev, V., Bolshiyanov, D.Yu., Flehoc,! C., Jouzel, 3/4, 13–19. J., 1996. Pleistocene ice at the bottom of the Vavilov Ice Cap, Mangerud, J., Astakhov, V.I., Murray, A., Svendsen, J.I., 2001. The Severnaya Zemlya, Russian Arctic. Journal of Glaciology 42, chronology of a large ice-dammed lake and the Barents-Kara ice 403–406. sheet advances, Northern Russia. Global and Planetary Change 31, Sulerzhitsky, L.D., 1995. Features of radiocarbon chronology of the 321–336. mammoths (Mammutus primigenus) of Siberia andNortheastern Mangerud, J., Astakhov, V., Svendsen, J.I., 2002. The extent of the Europe (in Russian). In: Baryshnikov, G. (Ed.), Studies of Barents-Kara ice sheet during the Last Glacial Maximum. Pleistocene andRecent Mammals. Proceedingsof the Zoological Quaternary Science Reviews 21, 111–119. Institute. Russian Academy of Sciences, St. Petersburg, Melles, M., Kulbe, T., Overduin, P.P., Verkulich, S., 1994. The pp. 163–183. Expedition Bunger Oasis 1993/1994 of the AWI Research Unit Svendsen, J.I., Astakov, V.I., Bolshiyanov, D.Yu., Demidov, I., Potsdam. In: Melles, M. (Ed.), The Expedition Norilsk/Taymyr Dowdeswell, J.A., Gataullin, V., Hjort, C., Hubberten, H.-W., 1993 andBunger Oasis 1993/1994 of the AWI Research Unit Larsen, E., Mangerud, J., Melles, M., Moller,. P., Saarnisto, M., Potsdam. Reports on Polar Research 148, pp. 3–25. Siegert, M.J., 1999. Maximum extent of the Eurasian ice sheets in Melles, M., Siegert, C., Hahne, J., Hubberten, H.-W., 1996. Klima- the Barents andKara Sea region duringthe Weichselian. Boreas undUmweltgeschichte desn ordlichen. Mittelsibiriens im Spatquart-. 28, 234–242. ar—erste. Ergebnisse. Geowissenschaften 14, 376–380. Thiede, J., Bauch, H.A., Hjort, C., Mangerud, J., 2001. The Late Moller,. P., Bolshiyanov, D.Yu., Bergsten, H., 1999. Weichselian Quaternary stratigraphy andenvironments of northern Eurasia geology andpaleoenvironmental history of the central Taymyr andthe adjacentArctic seas—new contributions for QUEEN. Peninsula, Siberia, indicating no glaciation during the last global Global andPlanetary Change 31, vii–x. glacial maximum. Boreas 28, 92–114. Turney, C.S.M., Coope, G.R., Harkness, D.D., Lowe, J.J., Walker, Murray, A.S., Wintle, A.G., 2000. Luminescence dating of quartz M.J.C., 2000. Implications for the dating of Wisconsinan using an improvedsingle-aliquot regenerative-doseprotocol. (Weichselian) Late-Glacial events of systematic radiocarbon age Radiation Measurements 32, 57–73. differences between terrestrial plant macrofossils from a site in SW Niessen, F., Ebel, T., Kopsch, C., Fedorov, G.B., 1999. High- Ireland. Quaternary Research 53, 114–121. resolution seismic stratigraphy of lake sediments on the Taymyr Vaikmae,. R., 1991. Palaeoenvironmental data from less-investigated Peninsula, Central Siberia. In: Kassens, H., Bauch, H.A., polar regions. In: Well, G., Wilson, C.L., Severin, A.B. (Eds.), Dmitrenko, I.A., Eicken, H., Hubberten, H.-W., Melles, M., International Conference on the Role of the Polar Regions on Thiede, J., Timokhov, L.A. (Eds.), Land-Ocean Systems in the Global Change, pp. 611–616. Siberian Arctic: Dynamics andHistory. Lecture Notes in Earth Vaikmae,. R., Punning, Ya.-M.K., Romanov, V.V., Barkov, N.I., Science. Springer, Berlin, pp. 437–456. 1988. Stratigraphy of the Vavilov Ice Dome in Severnaya Zemlya Peltier, W.R., 1994. Ice age paleotopography. Science 265, 195–201. using isotopic geochemical methods. Data of Glaciological Studies Pfeiffer, E.M., Gundelwein, A., Nothen,. T., Becker, H., Guggenber- 40, 127–135. ger, G., 1996. Characterization of the organic matter in permafrost Vasilchuk, Y., Punning, J.M., Vasilchuk, A., 1997. Radiocarbon ages soils andsediments of the Taymyr Peninsula/Siberia andSevernaya of mammoths in Northern Eurasia: implications for population Zemlya/Arctic Region. In: Bolshiyanov, D.Yu., Hubberten, H.-W. development and Late Quaternary environment. Radiocarbon 39, (Eds.), Russian-German Cooperation: The Expedition Taymyr 1–19. 1995 andthe Expedition Kolyma 1995. Reports on Polar Research Velichko, A.A., Isayeva, L.L., Makeyev, V.M., Matishov, G.G., 211, pp. 46–84. Faustova, M.A., 1984. Late Pleistocene glaciation of the Arctic Serebryanny, L., Malyasova, E., 1998. The Quaternary vegetation and Shelf, andthe reconstruction of Eurasian Ice Sheets. In: landscape evolution of Novaya Zemlya in the light of palynological Velichko, A.A. (Ed.), Late Quaternary Environments of the records. Quaternary International 45/56, 59–70. Soviet Union. University of Minnesota Press, Minneapolis, Serebryanny, L., Andreev, A., Malyasova, E., Tarasov, P., Roma- pp. 35–41. nenko, F., 1998. Lateglacial andearly-Holocene environments of Velichko, A.A., Kononov, Y.M., Faustova, M.A., 1997. The last Novaya Zemlya andthe Kara Sea region of the Russian Arctic. glaciation of the Earth: size andvolume of ice-sheets. Quaternary The Holocene 8, 323–330. International 41/42, 43–51. Siegert, C., Derevyagin, A.Y., Shilova, G.N., Hermichen, W.D., Vogt, C., 1997. Regional andtemporal variations of mineral Hiller, A., 1999a. Paleoclimatic indicators from permafrost assemblages in Arctic Ocean sediments as climatic indicators sequences in the eastern Taymyr Lowland. In: Kassens, H., Bauch, during glacial/interglacial changes (in German). Reports on Polar H.A., Dmitrenko, I.A., Eicken, H., Hubberten, H.-W., Melles, M., Research 251, 309pp. Thiede, J., Timokhov, L.A. (Eds.), Land-Ocean Systems in the Weber, M.E., Niessen, F., Kuhn, G., Wiedicke, M., 1997. Calibration Siberian Arctic: Dynamics andHistory. Lecture Notes in Earth andapplication of marine sedimentaryphysical properties using a Science. Springer, Berlin, pp. 477–499. multi-sensor core logger. Marine Geology 136, 151–172. Siegert, M.J., Marsiat, I., 2001. Numerical reconstructions of LGM Wolfarth, B., Skog, G., Possnert, G., Holmquist, B., 1998. Pitfalls in climate across the Eurasian Arctic. Quaternary Science Reviews 20, the AMS radiocarbon-dating of terrestrial macrofossils. Journal of 1595–1605. Quaternary Science 13, 137–145.