Astakhov V. 2004. Middle Pleistocene Glaciations of the Russian North

ARTICLE IN PRESS

Quaternary Science Reviews 23 (2004) 1285–1311

Middle Pleistocene glaciations ofthe Russian North Valery Astakhov* Geological Faculty, St. Petersburg University, Universitetskaya 7/9, St. Petersburg 199034, Russia

Abstract

Geological data on the pre-Eemian glaciations ofnorthern Russia, including the latest results by the Russian–Norwegian PECHORA project, are synthesized in order to present evidence for comparison with other early glaciations around the Arctic. The bulk ofevidence indicates that Arctic and Subarctic regions ofEuropean Russia, ofwestern and central Siberia during the Middle Pleistocene were at least 4 times covered by large ice sheets, which advanced mainly from the shelf ice domes, partly from Fennoscandia and the Putorana Plateau. Ice accumulations in the Ural Mountains were insignificant and did not form any noticeable ice dispersal centres. Unlike the classical glaciated areas, ice sheets ofnorthern Russia acted mainly on a soft,perennially frozen substrate, which was heavily glacitectonised. The Middle Pleistocene ice sheets were much larger than the Weichselian ones. The Fennoscandian ice dispersal centre was most active in northern European Russia during the penultimate glaciation (OIS 6) when shelf-centred ice domes were relatively weaker. Larger continental ice sheets were formed in preceding ice ages, when Kara Sea ice dispersal centre dominated. The lowland ice sheets reached their maximum extent at different stages, from Cromerian Don glaciation in European Russia to OIS 8 in West Siberia. Therefore, the maximum ice limit is time-transgressive in northern Russia. r 2003 Elsevier Ltd. All rights reserved.

1. Introduction surveys, led to graphical generalisations in the form of synthetic Quaternary maps for the entire Soviet Union The volume and extent ofMiddle Pleistocene ice (Ganeshin, 1973) and separately for each of the super- sheets exceeded those ofthe Late Pleistocene by far, regions ofnorthern Eurasia such as the Russian Plain, especially on the eastern flank ofglaciated Eurasia (e.g. the Urals, West and Central Siberia. These maps are Ganeshin, 1973). Accordingly, their impact on the principal sources ofhard data about the size offormer geological structure and environments ofthe Arctic ice sheets obtained by generations ofmapping geolo- and Subarctic regions was more profound. However, gists. Lately they were used to compile the digital map of geological research within the QUEEN framework Pleistocene ice limits as part ofthe INQUA project has largely been focused on the Late Pleistocene (Astakhov, 2003). history ofthe Russian North. Only a fewsections of After Sachs and Yakovlev numerous articles and Middle Pleistocene drift have been studied by regional monographs were dealing with the Middle QUEEN members on the Russian mainland. Hence Pleistocene glacial deposits ofthe North in terms of the bulk ofdata discussed below is derived fromRussian stratigraphy (Arkhipov and Matveyeva, 1964; Lazukov, literature. 1970; Zubakov, 1972; Yakhimovich et al., 1973; Over the last half-century various attempts have been Kaplyanskaya and Tarnogradsky, 1974; Arkhipov made to synthesize data on Middle Pleistocene glacia- et al., 1986, 1994; Velichko and Shick, 2001), lithology tions collected by hundreds ofRussian researchers. The (Kuznetsova, 1971; Zemtsov, 1973a, b; Sukhorukova only work in which all Quaternary ofthe Russian Arctic et al., 1987; Andreicheva, 1992; Andreicheva et al., and Subarctic is discussed, is the monumental volume by 1997), palaeogeography and geomorphology (Isayeva, Sachs (1953). Another outstanding contribution is a 1963; Troitsky, 1975; Arkhipov et al., 1976). Structural stratigraphic monograph with a Quaternary map of geological and sedimentological works are noticeably European Russia by Yakovlev (1956). In the 1960–1970s scarcer (e.g. Zakharov, 1968; Kaplyanskaya and Tarno- the huge influx ofdata, especially fromgeological gradsky, 1975; Astakhov et al., 1996). Rare attempts to reconstruct dimensions and flow patterns ofthe ice sheets were mostly based on till lithologies (Zubakov, *Fax: +7-812-328-39-16. 1972; Sukhorukova et al., 1987) and occasionally on E-mail address: [email protected] (V. Astakhov). glaciological considerations (Voronov, 1964).

0277-3791/$ - see front matter r 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.quascirev.2003.12.011 ARTICLE IN PRESS 1286 V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311

Fig. 1. Location map ofglaciated northern Russia. Solid line is Early Weichselian ice limit. Broken line is the glacial driftlimit in Siberia and Moscow glaciation limit west ofthe Urals, dash-and-dot line is the driftlimit in European Russia (Don glaciation?). Black arcs are largest ice-push ed features. Lines A and B are geological profiles in Fig. 2,CinFig. 7. Black circles are key sections ofMiddle Pleistocene interglacial formations sandwiched between tills: 1—Lake Chusovskoye (Stepanov, 1974); 2—Rodionovo (Loseva and Duryagina, 1973); 3—Kipiyevo (Guslitser and Isaychev, 1983); 4—Seyda (Russian–Norwegian project PECHORA, 1998); 5—Semeika (Kaplyanskaya and Tarnogradsky, 1974); 6—Belogorye Upland (Arkhipov et al., 1978); 7—Khakhalevsky Yar (Levina, 1964; Zubakov, 1972); 8—Bakhtinsky Yar (Arkhipov and Matveyeva, 1964; Zubakov, 1972); 9—Pupkovo (Zubakov, 1972); 10—Novorybnoye (Kind and Leonov, 1982). Open circles indicate lowland tills in the mountains: 11—clayey tills with clasts ofCarboniferous limestones at 500–600 asl ( Astakhov, 1974a, b; see also Fig. 4A); 12—Weichselian end moraine of northern origin at 560 m (Astakhov et al., 1999); 13—three tills with lowland erratics on flat summit 600 m asl (Fainer et al., 1976). Deepest buried valleys filled with tills: 14—borehole at Lebed, 342 m bsl (Arkhipov and Matveyeva, 1964; Zubakov, 1972); 15—borehole at Kosa Kamennaya, 367 bsl (Arkhipov et al., 1994). Arrows are ice flow directions inferred from ice-pushed ridges and clast indicators.

The volume ofexisting data makes a comprehensive The questions ofimmediate concern will be (i) size and overview impossible for any journal paper. Hundreds of flow pattern offormer ice sheets, (ii) possible correla- papers discussing various aspects ofthe pre-Weichselian tions ofice advances within the QUEEN area. glacial history from different standpoints cannot be reviewed here. My task is limited to briefly describing the most reliable geological results in pre-Weichselian 2. General geological setting glacial geology ofnorthern Russia beyond the realm of Fennoscandian glaciations (Fig. 1) in order to present Northern Russia as an area ofinland glaciation is data for comparison and to highlight weak points that very different from the classical glaciated regions of might be of interest for further research. A selection of northwestern Europe and North America. It is largely key evidence in a diverse area ca 4.5 million km2, soft-rock flatland extending offshore as sedimentary although it might seem arbitrary, is actually scale- basins ofthe Barents and Kara seas ( Fig. 1). The weak dependent. It is dictated by the author’s mapping substrate is responsible for the large thickness (up to experience, according to which superregional conclu- 300–400 m) and predominantly fine-grained composi- sions are derived mostly from examination of large tion ofthe Quaternary cover. Only in the east looms the geological features, whereas small details of geological Putorana Plateau, up to 1600 asl, a major source of structure often yield only results of local significance. hard-rock clasts, built ofhorizontally layered dolerites ARTICLE IN PRESS V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311 1287 and lavas. Other smaller sources oferratics are narrow Since the 1950s many authors, based on the low Palaeozoic ranges ofthe Ural and Byrranga Mountains content ofpebbles, local occurrence ofmarine fossils and low ridges ofTiman, Novaya Zemlya and Pai-Hoi and the old idea ofmontane ice dispersal centres, (Fig. 1). The rest ofthe area is underlain by Mezosoic interpreted thick diamict sheets, partly or entirely, as and Cenozoic sand, clay, silt, opoka, diatomite and glacimarine formations (Sachs, 1953; Lazukov, 1970; therefore is apt to produce few visually recognisable Zubakov, 1972). However, sedimentological analysis of clasts. Consequently, the traditional method ofrecon- fine-grained Middle Pleistocene diamictons in their structing former glaciers by mapping boulder trains stratotype sections proved beyond any reasonable doubt would inevitably point to the Central Siberian uplands that they were deposited by huge ice sheets that and the smaller salients offoldedPalaeozoic rocks as advanced southwards (Guslitser, 1973; Kaplyanskaya the only sources ofmoving ice. The seaward sloping and Tarnogradsky, 1974, 1975). The pattern ofice sedimentary basins ofthe Pechora, West Siberian pushed ridges (Fig. 1) and dispersal oflowland clasts and North Siberian lowlands have generally been also indicate ice streams diverging from lowlands to the viewed as an arena ofmarine incursions fromthe Arctic adjacent highlands (Astakhov, 1974a, 1977), contrary to Ocean. ice flows from the mountains predicted by the glacimar- The stratigraphic methods employed and conclusions ine hypothesis. The distribution and thickness ofglacial obtained are different in two natural zones: (a) the deposits, as well as the pattern oftheir imbricate Arctic with marine and glacial formations alternating, accretions (Figs. 1 and 2), is in accordance with N–S and (b) the Subarctic, where only terrestrial stratified directed ice advances from the shelf but is hardly sediments occur sandwiched between thick diamictons. explainable by ice dispersal from the highlands. These zones, being crossed by the Uralian Range, give four stratoregions to be discussed separately: two in European Russia and two in Siberia. In all regions 3 to 5 3. Size and flow pattern of ice sheets diamict sheets up to 60 m thick each are found in borehole profiles (Arkhipov, 1971; Zubakov, 1972; 3.1. European Russia Lavrushin et al., 1989). Tills ofthe penultimate glaciation can be locally seen in exposures ofthe Arctic 3.1.1. Ice sheets inferred from erratics beneath Upper Pleistocene strata, whereas diamictic The ice limits in central European Russia since the formations of preceding ice advances normally lie below XIX century have been attributed to activity of sea level. The most continuous till sheet, observed in the the Fennoscandian ice dispersal center, as suggested by Subarctic zone as the second from the surface, is the distinct boulder trains and configuration ofthe believed to form the drift limit along 60NinWest marginal formations. For the last half-century three Siberia, deviating northwards in Central Siberia and major Middle Pleistocene ice advances called Oka, southwards in European Russia (Fig. 1). Older tills of Dnieper (the maximum glaciation) and Moscow have more restricted occurrence are known just from bore- been correlated with the Elster, Saale and Warthe holes. glaciations ofCentral Europe ( Yakovlev, 1956; Gor- The Middle Pleistocene till sheets gradually plunge etsky et al., 1982). In northern European Russia, northwards and get progressively more eroded in the however, influx ofglacial ice fromthe northeast was high Arctic, especially on the sea floor (Fig. 2). The acknowledged quite early (Ramsay, 1904). The con- occurence ofpre-Weichselian tills is mantle-like: they tributions ofindividual ice domes were thoroughly can be found both in buried valleys at 300–350 m bsl discussed by Yakovlev (1956). Based on provenance of (locations 14 and 15 in Fig. 1) and on plateaus 500– erratics in the European North, Yakovlev distinguished 600 m asl (locations 11 and 13 in Fig. 1). The thickest three major areas ofice sheet growth: (1) Fennoscandia, accretions ofMiddle Pleistocene sediments, up to 200– (2) Novaya Zemlya with the adjacent Barents Sea shelf 300 m, are known from the areas of rougher sub- and (3) the Urals. Quaternary reliefalong the Urals and Central Siberian The dark-grey Novaya Zemlya till is a thick, fine- Plateau, whereas the flat Timan Ridge and central West grained diamicton with strong NE–SW or N–S fabrics. Siberian lowlands often bear a discontinuous cover of It is devoid ofFennoscandian erratics but contains some glacial deposits less than 50 m thick. Above the Arctic fragments of Novaya Zemlya pink and black limestones Circle the Middle Pleistocene sediments are either totally (Andreicheva, 1992). Large limestone blocks trans- eroded or fill in deep depressions. The latter are ported from the Kara Sea coast towards SW across normally obscured by the Weichselian glacial complex the Pai-Hoi Range have been known for a long time which is often glaciotectonically stacked to attain 100 m (Voronov, 1951). Many authors also noted numerous in thickness (Fig. 2). The described regional structure boulders offoreign rocks scattered over flat-topped reflects a lowland position ofprincipal ice dispersal mountains ofthe Polar Urals up to 1000 m asl ( Yakov- centers. lev, 1956). Fragments ofwestern Uralian rocks occur ARTICLE IN PRESS 1288 V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311 (m5 is probably a glaciotectonic repetition of for location. Diamictons: 1—deep buried (pre-Holsteinian); 2—close to sea Fig. 1 . , simplified). See Lavrushin et al. (1989) Lavrushin et al., 1989 .); lIIlh is a lacustrine formation with Likhvin-type pollen spectra according to V.A level (Saalian); 3—surficial (Weichselian); 4—stratifiedm4- silt and clay; 5—sand; 6—borehole. Till sheets are g1 to g5 and marine formations are m1 to m5 Fig. 2. Profiles of Quaternary formations on the Barents Sea coast (from ARTICLE IN PRESS V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311 1289 east ofthe main watershed, limestone clasts fromthe Plain during the maximum Pleistocene glaciation low western piedmont being often found atop glacially (Fig. 3A), west–east direction for the till of the smoothed hills 500–600 m high (Savelyev, 1966). Ac- penultimate glaciation west ofthe Pechora and cording to Yakovlev (1956) ice streams directed from again N–S ice flow east ofthis river ( Fig. 3B). The Novaya Zemlya to the SSW reached the Volga maps by Lavrov et al. (1986) offer no signs of a Uralian catchment area, where they coalesced with SE flowing influence except at several locations with transverse Fennoscandian ice ofthe Dnieper glaciation to formthe fabric (Fig. 3B) reported by Kuznetsova (1971). The largest ice sheet ofEuropean Russia. This ice sheet was pebble content pattern and the main vector oftill thought to have penetrated into the southern steppes by fabric points directly towards the NE across Pai-Hoi two huge ice streams forming the Dnieper Lobe in the and Novaya Zemlya into the Kara Sea, without Ukraine and the Don Lobe in southeastern Russia. In disturbance by ice flows from the Urals or from the this model ice from northeastern sources covered Barents Sea. practically all sedimentary basins ofthe Russian Plain north of58 N. 3.1.2. Other evidence of ice dispersal The subsequent penultimate glaciation left the reddish The signals from pebble composition and mineralogy brown till widely observable on the surface west of the oftills are oftenmixed. On the western carbonaceous Pechora. This till is unanimously correlated with the piedmont ofthe Northern Urals (63–61 N) Varsanofie- Moscow glacial complex ofCentral Russia. It contains va (1933) found that the clayey till of western numerous western erratics, including the characteristic provenance was covered in places by a clast-supported nepheline syenite from the Kola Peninsula, and there- diamicton full of central Uralian erratics at 20 km west fore must have been deposited by a Fennoscandian ice ofthe mountain front.Based on these data she sheet. East ofthe Pechora the latter coalesced with ice suggested two glaciations: one in the form of a regional streams originating from Novaya Zemlya and the Urals ice sheet which advanced along the Urals and a (Yakovlev, 1956; Potapenko, 1974; Lavrov et al., 1986; subsequent glaciation in the form of restricted piedmont Andreicheva, 1992). ice flows from old alpine troughs. The situation is Individual ice domes are not easily identified by the mirrored on the Siberian slope ofthe Peri-Polar and erratic dispersal alone on which many authors rely Northern Urals between 65N and 62N(Sirin, 1947; heavily. A straightforward interpretation of statistics on Ber, 1948), where a till ofeastern (Siberian) provenance pebble composition and orientation may be controver- mantling the hills up to 500 m asl is superposed by a sial. For example, south of64 N tills ofwestern cobbly diamicton oflocal (central Uralian) provenance. provenance have long been known on the Uralian Sirin (1947) concurred with Varsanofieva to suggest two piedmont. Heaps ofboulders ofPalaeozoic sedimentary glaciations on the eastern slope: a maximum ice sheet rocks from the west occur even at the very foot of the which advanced from the West Siberia onto the Urals highest range ofthe Northern Urals. However, farther and a local glaciation centred in the Uralian axial zone to the west clasts ofsedimentary rocks are mixed with ofmetamorphic rocks. central Uralian crystalline rocks, i.e. the percentage of Thick sequences of lowland drift later found on both Uralian clasts in the clayey basal till apparently slopes ofthe mountain range amply confirm ice increases westwards. Varsanofieva (1933) explained this advances onto the Northern Urals from both NW and paradox by ice ofnorthern origin streaming along the NE but reject local ice caps. The thickest clayey till at Urals and fanning off westwards into the lowlands and 500–600 m asl in the axial zone (Fig. 4A) contains eastwards into the mountains. North of64 N, just west fragments of black Carboniferous limestones picked ofthe highest massifofthe Peri-Polar Urals, clayey tills up in the low piedmont some 20 km to the west are even more enriched in Uralian erratics, which led to (Astakhov, 1974a). The clayey till is capped by the the idea ofthick ice which presumably flowed westwards 50 m high kame-like hummocks built ofwell-washed from ice domes positioned over the Urals (Yakovlev, sand with small and roundish cobbles ofcentral Uralian 1956; Chernov, 1974). Pebble orientation and miner- origin but obviously deposited by the same inland ice alogical composition ofthe till matrix were also used to sheet, most probably ofOIS 6. In contrast, younger suggest a separate Uralian ice dome (Kuznetsova, 1971; local moraines filling alpine troughs higher than 600 m Andreicheva, 1992). (Fig. 4A) consist ofonly clast-supported tills derived However, the question is not that simple. The most from metamorphic rocks. The Siberian slope of the comprehensive work was done by Lavrov et al. (1986) Northern Urals, where no alpine moraines have been who mapped the glacial topography and measured mapped, is in places covered by thick matrix-supported pebble composition in hundreds oftill samples a quarter tills ofeastern provenance ( Fig. 4B), containing frag- ofa cubic meter each. The generalised results unam- ments ofsoftMesozoic rocks ofthe West Siberian biguously show the persistent N–S and NE–SW basin and the West Siberian mineralogical assemblage dispersal ofglacial clasts over the northeastern Russian (Ryzhov, 1974). ARTICLE IN PRESS 1290 V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311

Fig. 3. Pebble content and ice flow indicators in Middle Pleistocene tills ofnortheastern European Russia: A—maximum glaciation; B—penultimate glaciation. By Lavrov et al. (1986) with minor additions from Matveyeva (1967) and Gornostay (1990) for the Timan Ridge and from Astakhov (1974b), Ryzhov (1974), Astakhov et al. (1999) for the Urals. Isolines show volumetric content of pebble fraction 1–5 cm in till samples about a quarter ofcubic meter each. Black wedges are long axes ofpebbles. Arrows are other ice flow indicators: striae, boulder pavements, eskers, transport paths oferratics. Hatched are Palaeozoic salients ofthe Timan Ridge and Urals. Note the pebble content decreasing downglacier and increasing locally in the lee ofthe Timan Ridge at southwestern corner of Fig. 3A. Lowland erratics in basal tills (Fig. 4A) and esker orientation in Fig. 3B are at odds with fabrics of surficial diamicton measured by Kuznetsova (1971) in the upper Pechora area.

Fig. 4. Lowland tills on western (A) and eastern (B) slopes ofthe Northern Urals. A—river Telpos catchment, N 63 500/E 59 (from Astakhov, 1974a); B—between rivers Manya and Mazapatya, N 62150/E 59500 (from Ryzhov, 1974, simplified). 1—laminated fine sand and silt; 2—coarse sand with pebbles; 3—clast-supported diamicton oflocal provenance; 4—matrix-supported diamicton with lowland clasts; 5—borehole. Note kames oflowland glaciation undisturbed by Uralian glaciers in Fig. 4A. ARTICLE IN PRESS V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311 1291

In the lowlands no landforms testifying to ice flow striations across the Palaeozoic structures ofthe Timan from the Urals have ever been mapped (Fig. 1). Ridge (Fig. 3A). Moreover, even on the western piedmont ofthe North- The most reliable indicators ofthick upslope flowing ern Urals, 200–300 m asl, eskers and striae are oriented inland ice are large glaciotectonic disturbances which in either N–S or even NW–SE (Astakhov, 1974b). This is places form arcuate ridges up to 250 km long (Astakhov quite compatible with ice flow directed towards the et al., 1999). In the Pechora Basin such features are mountains and not with a Uralian ice dispersal centre. oriented in accordance with ice flow directions from NE The 90-m thick sequence offine-grained Middle and NW (Fig. 1). Huge rafts of soft Mesozoic rocks, Pleistocene glacial sediments ofa kame field found in including rock salt, have been incorporated into the till the central metamorphic zone (Fig. 4) has evidently ofnortheastern provenance. The till ofthe penultimate never been affected by any glaciers of montane origin. glaciation contains large basalt blocks from Timan Similar kames built offine sand are also known fromthe (Guslitser, 1973). In the peri-Timan area ice flow from east ofthe central range ( Ber, 1948). According to this the west is also clearly indicated by surficial ice-pushed author’s observations, in several places ofthe Northern features. For example, along the Arctic slope of the Urals well-preserved kames are mantled by thin clast- Timan Ridge a large allochtonous stack ofwestward supported diamicton full of fragments of Uralian dipping, imbricated slices ofMesozoic and Quaternary quartzites and schists. This sediment, which is better rocks, testifies to eastward glaciotectonic transport over be interpreted as ablation till, may well belong to the a distance ofsome 40 km ( Gornostay, 1990). Orienta- same ice advance. Thus, no second ice advance is needed tion ofsmall glaciotectonic structures is more scattered. to account for pebbles of Uralian provenance in the In the Arctic zone ice flow features directed from surficial diamicts ofthe piedmonts. The crystalline lowlands into the mountains are abundant as well. A pebbles mostly originate from mountains higher than sand pit on Hanmei river near Labytnanghi on the 500 m and therefore are rare in the basal till but eastern piedmont ofthe Polar Urals, examined during abundant in the ablation till, which was enriched in the PECHORA project expedition in 2000, displays clasts from nunataks during the lowering of the ice Middle Pleistocene glaciofluvial gravels lying atop a very surface. This agrees with the fact that the percentage of clayey till (Fig. 9). Surprisingly, these sediments contain Uralian crystalline rocks in diamictons ofthe Pechora mostly fragments of soft Mesozoic rocks from West Basin is minimal in the oldest tills ofthicker ice sheets, Siberia but no clasts ofresistant ultramafic rocks from but clearly increases upwards in the till succession the nearby Rai-Iz massif. Gravelly cross-beds all dip (Andreicheva, 1992). This can be explained by the towards the Urals. Farther northwards the Late growing influence ofmontane sources with progressive Pleistocene matrix-supported Sopkay moraines strike thinning ofeach subsequent ice sheet, but should not W–E, i.e. transverse to the Urals, reflecting an ice flow be taken as a proofofa separate ice dome over the from the shelf (Astakhov, 1979). Even the youngest Urals. bouldery moraines on the northwestern tip ofthe Urals The above facts indicate that diamictons with west– at 560 m asl show only ice push from the north, i.e. from east fabric and predominant central Uralian clasts may the Kara Sea (Astakhov et al., 1999). The SE-facing not represent separate glaciations but belong to the arcuate ridges on the Siberian slope (Fig. 1), outlining a ablation subcomplex ofthe same ice age. Stagnating ice piedmont morainic apron south ofthe Arctic Circle sheets, while thinning, might acquire reverse surface (Astakhov, 1997), can hardly be ofa Uralian origin, gradients forcing supraglacial boulders from Uralian because they (i) occur only downglacier ofold wide nunataks to slide westwards. This could happen in the troughs crossing the narrow mountain range and (ii) do end ofevery ice age, thus adding to the volume of not have any symmetrical counterparts on the western, crystalline clasts already delivered to the lowlands by more humid slope. Therefore, the morainic apron of the small alpine glaciers which probably acted prior to the eastern piedmont most likely originated from outlet main ice advances from the shelf. If the thin diamictons glaciers that drained a Middle Pleistocene ice sheet of ofUralian provenance had been formed subglacially by European Russia via Uralian through valleys to the SE. westward ice flow, as Varsanofieva (1933), Kuznetsova This suggestion agrees with the boulder trains of (1971), Andreicheva (1992) and others suggested, there northwestern provenance traced across the Urals should have been abundant features of glacial erosion (Yakovlev, 1956). and deposition transverse to the longitudinal topogra- For the interpretation ofthe ice sheet flow patterns it phy and structures ofthe Urals. Nothing ofthe kind has is crucial that the sedimentological evidence oflowland ever been detected by remote sensing or surface glacial advance into the Urals are in accord with the mapping. On the contrary, all large structural and pattern ofice-pushed ridges and terminal moraines in geomorphic features suggest ice flow towards or parallel the lowlands. Large ice-pushed ridges generally strike to the Urals. The NNE–SSW ice flow reflected in the parallel to the Arctic shoreline and nowhere fringe the clast distribution is also supported by large-scale Uralian range except narrow morainic aprons along the ARTICLE IN PRESS 1292 V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311 footofthe mountains south of67 N(Astakhov, 1977, emanated such long strings oferratics and flow features, 1997; Astakhov et al., 1986). Where discrepancies exist obviously were too thick to be obstructed by most ice between the lithological composition ofthe glacial advances from Fennoscandia. deposits and the pattern ofice-pushed ridges, the latter Yakovlev (1956), who saw the difficulty, suggested an should be taken as decisive evidence, because fresh additional accumulation area on the adjacent Barents large-scale geomorphic features can only be produced by Sea shelf. However, no traces of ice flow from the NW the latest ice flows. The above mentioned horseshoe- onto the Russian mainland have been mapped for the shaped end moraines ofthe Kara ice sheet, shoved up- maximum glaciation: anywhere east of43 N all features valley in the Polar Urals (Astakhov et al., 1999), consist are directed NE–SW or N–S. Features ofW–E and ofheaps oflocal boulders reoriented by ice push from NW–SE orientation and clasts ofwestern provenance the north. These and many other similar observations first appear in surficial tills ofthe penultimate (Vycheg- are good evidence that the Urals were generally over- da=Moscow) glaciation, when the influence ofthe riden or bypassed by ice streams that originated north of northeastern ice dispersal centre decreased (Matveyeva, the mountains, most likely on the western Kara shelf. 1967; Andreicheva, 1992; Andreicheva et al., 1997). This The low and narrow mountain range ofthe Urals was underlines the dominant position ofthe Kara Sea shelf not a major ice dispersal center either in the Late as a major source ofinland ice in northern European Pleistocene (Astakhov et al., 1999), or in pre-Eemian Russia throughout the Pleistocene. Only during the times ofgreater continental ice sheets. penultimate glaciation the shelfice domes were over- powered by a Fennoscandian ice sheet which advanced 3.1.3. Dominant ice dispersal centre across the Timan Ridge at right angles to the ice flow Thus, the available data indicate that during the from the Kara Sea. Middle Pleistocene northern European Russia was the realm of a powerful inland ice that flowed from the NE 3.2. Siberia upslope and eroded mostly soft Meso-Cenozoic sedi- ments to deposit unusually thick and clayey tills. Lateral 3.2.1. Signatures of ice motion erosion ofthe Ural Mountains by transit ice streams East ofthe Urals the thickness ofinland ice and its contributed a small amount ofcentral Uralian clasts to flow pattern are more controversial issues, especially for the generally fine-grained drift. Although the main ice the West Siberian Plain. Historically there have always streams were basically directed to the south and been competing hypotheses oflarge Middle Pleistocene southwest, inland ice also flowed laterally across the glaciers versus concepts ofthinner local ice sheets that Urals, overriding their flat-topped mountains when the were assumed to reflect the drier Siberian climates. The ice was over 1 km thick. When the European ice sheet first paradigm is commonly maintained by practicing was thinner, it could reach the Siberian slope ofthe geologists, who observe ubiquitous glacial features in Urals only in the form of outlet valley glaciers. various areas and at various altitudes, whereas the Karpukhin and Lavrov (1974) considered the Yakov- second trend ofthinking is largely motivated by general lev’s Novaya Zemlya ice sheet too large and preferred to palaeogeographical considerations. explain the NE–SW fabric in the Upper Volga area by a One ofthe main glaciation centres was inferred long diverted flow ofFennoscandian ice. According to them ago from the pyroxene abundance in till matrix and the traces ofa Novaya Zemlya glaciation and erratics ofNE mafic rock fragments scattered over the Central Siberian provenance occur mainly east ofriver Severnaya Dvina. Plateau built ofPalaeozoic sedimentary formations, Immediately west of50 E the drift limit turns almost Triassic basalts and dolerites (Obruchev, 1931; Urvant- straight south in accordance with the margin of sev, 1931; Sachs, 1953). Three morainic belts rich in Fennoscandian ice. For comparison, Velichko et al. dolerite boulders concentrically surround the Putorana (1977), based on the peculiar composition ofthe fine- Plateau (Isayeva, 1963). In the east the Putorana ice grained Don till, suggest that even the huge Don Lobe coalesced with a small ice sheet ofthe Anabar Plateau. It on 52N was produced by the northeastern ice dispersal is interesting that during the penultimate glaciation the centre! easternmost Anabar ice sheet, similarly to the European The strong NNE–SSW orientation ofvarious glacial situation, was overpowered by eastward ice flow from features, persistent over thousands of kilometers from Putorana (Andreyeva and Isayeva, 1974). Ice flow from the Pai-Hoi Range on 68N to the Upper Volga on the High Arctic was inferred early basing on character- 58N, is truly amazing, especially compared to the istic granite boulders transported onto the dolerite radially diverging pattern ofFennoscandian ice flow. It plateaus ofPutorana fromthe Nordenskjold Archipe- seems impossible for a small elongated ice dome over the lago and northern shores ofthe Taimyr Peninsula across narrow Novaya Zemlya archipelago to produce the the c. 400 m high Byrranga Mountains. The phenomen- extensive ice streams which deflected the Fennoscandian on was first thought to result from a reverse topographic ice. The lowland ice sheets ofnortheastern origin, which gradient ofthe Ice Age ( Urvantsev, 1931). Today it is ARTICLE IN PRESS V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311 1293 seen as evidence ofan ice dome thicker than 2 km on the continuous permafrost of the Arctic. Detached blocks Kara Sea shelf. This Arctic ice sheet prevented the and drag structures in glacial drift of the region suggest Putorana ice from flowing northwards (Andreyeva, that the cold glaciers proceeded by involving into 1978; Kind and Leonov, 1982). motion a 300–400 m layer ofvery unstable although The difference between glacial features of Central frozen clayey substrate (Astakhov et al., 1996). This type Siberia and West Siberia is striking. In Central Siberia of inland glaciation is dynamically different from other U-shaped valleys, deeply cut into bedrock and in places glaciated regions, which is also evident from the poor barred by boulder-rich morainic arcs, radiate from the preservation ofinterglacial formations. Inter-till palaeo- Putorana ice dispersal centre. They are accompanied by sols, common in North America, have never been smoothed, striated, sometimes fluted and drumlinised observed in West Siberia. surfaces, by occasional eskers and glaciofluvial fans The ice flow pattern is hard to recognise in the West (Isayeva, 1963; Arkhipov et al., 1976). There are such Siberian Plain, because the predominantly soft bedrock, typical signatures ofwet-based sliding as boulder consisting ofCretaceous and Palaeogene sand, silt, clay, pavements on the Yenissei (Troitsky, 1975). opoka and diatomite provide neither streamlined On the contrary, eskers, flutes and marginal ridges are features, nor enough clasts for statistical analysis. totally absent west ofthe Yenissei. They are replaced by Before the onset of modern glacial sedimentology and the englacial glaciotectonic imbrications, kame fields glaciotectonic research in the 1970s the only ice flow and late glacial sandurs. Especially characteristics are indicators used for reconstructions were clasts of thick fine-grained West Siberian tills with ubiquitous Uralian and Central Siberian crystalline rocks. local and far-travelled blocks of soft sediments, includ- Although already Obruchev (1931), basing on the ing Quaternary and Palaeogene loose sands with configuration ofthe fewknown morainic chains, original lamination preserved. In many cases sedimen- suggested a lowland ice dome on the Taz and Gydan tary rafts hundreds of meters long were transported over peninsulas, subsequent mapping failed to find its hundreds ofkilometers ( Shatsky, 1965; Zakharov, 1968; material signatures, and ensuing overviews acknowl- Kaplyanskaya and Tarnogradsky, 1974). The glaciotec- edged only ice dispersal paths from the Urals and tonised structure ofthe glacial deposits, that is common Central Siberian uplands (Sachs, 1953; Zarrina et al., for most sections of northern West Siberia, often 1961; Lazukov, 1970; Zubakov, 1972; Arkhipov et al., precludes visual tracing ofstratigraphic contacts. Such 1977). Nevertheless, Zemtsov (1973a, b), who inter- tills cannot have been deposited by lodgement, but most preted the composition ofseveral thousand boulders likely reflect the structure ofdirty basal parts ofhuge and mineralogical samples oftill matrix, concluded that stagnant ice sheets which slowly melted out in a there was a central zone in the north ofWest Siberia degrading permafrost environment (Astakhov and with no fragments of Uralian rocks or central Siberian Isayeva, 1988; Astakhov et al., 1996). dolerites. Tills ofthe central zone contained mostly The wet-based versus dry-based features at the same material oflocal Mesozoic and Cenozoic formations latitudes in Siberia clearly point out to the difference with admixture ofTaimyr rocks. Zemtsov thought this between the downslope sliding over rigid bedrock in zone to have been influenced by an additional ice Central Siberia and slow glaciotectonic motion ofthe dispersal centre on the Taimyr Peninsula, and he did not entire ice-permafrost couplet upslope in West Siberia. rule out a possibility ofice flow froma hypothetical Especially suggestive are the huge composite imbrica- lowland ice dome by Obruchev (1931). Also Kaplyans- tions and the lack ofany ice retreat features. In West kaya and Tarnogradsky (1975) interpreted marine Siberia vast ice fields seem to have simultaneously lost fossils found in diamictons of the Lower Yenissei area mobility to decay for a very long time, being protected as signatures ofice flow fromthe Kara Sea. from beneath by thick permafrost (Astakhov et al., 1996). Masses ofburied glacial ice are ubiquitous in the 3.2.2. Puzzling erratics Upper Pleistocene tills ofArctic Siberia. Fossil ice may Again, as in European Russia, in West Siberia a have even survived from Middle Pleistocene glaciations discrepancy exists between the strong signatures of conserved in the West Siberian permafrost, as suggested upslope ice flow from lowland ice domes (Astakhov, by very thick massive ice sometimes found in deep 1976, 1977) and the dominant petrographic and miner- boreholes. alogical composition ofthe glacial drift,which see- The described combination ofglacial features indi- mingly is evidence to the contrary. E.g. the most cates large ice sheets that were somewhat sluggish due to comprehensive study oftill composition in West Siberia the cold stored in the lithosphere over the entire (Sukhorukova et al., 1987) reveals three major clastic Pleistocene. This accummulated cold is evident from provinces roughly coinciding with the earlier conclu- numerous boreholes in West Siberia, which find a thick sions by Zemtsov (1973a, b): Central Siberian, Uralian perennially frozen layer at 150–200 m from the surface and West Siberian (Fig. 6). Sukhorukova et al. (1987) even on 59–60N, i.e. far south from the present-day maintain that these provinces reflect the relative ARTICLE IN PRESS 1294 V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311 significance ofthree major ice dispersal paths: fromthe reversed ice surface gradient after the late glacial Putorana Plateau, from the Urals and from the Kara collapse ofthe lowland ice domes. Sea shelf. First, this interpretation contradicts the pattern ofice-pushed features which everywhere only 3.2.3. Ice thickness and major ice dome reflects ice flow directed upslope (Astakhov, 1977). The mapped configuration ofthe glacial driftlimit Second, it is glaciologically impossible: there is no (Fig. 1) and large directional features are clear indica- obstacles in the very low central West Siberian Plain tions ofthick ice upslope moving during glacial that would prevent Uralian and Central Siberian ice culminations, irrespective ofthe predominant mechan- streams from moving farther south beyond the mapped isms ofcrystalline clasts transport to the lowlands. In drift limit (Figs. 1 and 6). Especially strange is the the adjacent mountains ofthe Urals and Central Siberia sudden increase ofUralian clasts in the Ob tills south of there is ample evidence ofupslope ice flow, such as 64N, which made Sukhorukova et al. (1987) suggest an lowland tills found high in the mountains (Fig. 4, loc.11, ice dome in the Northern Urals instead ofthe Polar 12, 13 in Fig. 1). Urals, as would follow from the general N–S ice The crucial indication ofsouthbound flow ofthick direction oflocal boulder trains. In addition, the till Middle Pleistocene ice in West Siberia is the system of fabric along the Yenissei shows only ice movement arcuate ice-shoved imbrications ofintricately disturbed towards the south and not westwards across the valley. soft rocks, often topographically expressed as hill–hole The suggestion by Kaplyanskaya and Tarnogradsky pairs. The system ofthe largest ice pushed ridges runs (1975), that upland ice divides, formed in the beginning roughly parallel to the coast ofthe Arctic Ocean ( Fig. 1). of each ice age, afterwards might have shifted into the The largest zones ofthrusted and tightly folded lowland, only partially helps. sediments are up to 200 km long and 20–25 km wide It must be taken into account that the West Siberian (Zakharov, 1968; Troitsky, 1975; Arkhipov et al., 1976; tills are typical lowland diamictons consisting largely of Astakhov, 1979; Astakhov et al., 1986). The distur- fine-grained material derived from the underlying bances may penetrate down to 400 m below the surface, Mesozoic and Palaeogene sediments. The admixture of which is probably the world record. The tectonic style of hard rock pebbles is very low (less than 1%) and these structures indicates their deformation in a frozen decreases northwards. Another interesting feature is flat- state and is evidence ofdeep crumpling ofsoftsubstrate iron and wedge-shaped forms among the pebbles, during ice movement (Astakhov et al., 1996). These noticeably more frequent than in other glaciated structures, normally consisting ofparaautochtonous regions. The ideal shape ofglacial abrasion is often slices or detached blocks oflocal provenance, as a rule, found in the same pebble sample on very dissimilar occur far upglacier from the ice margin. Along the ice materials, such as hard dolerites and soft opokas, unlike margin they are often replaced by huge rafts of far- the situation in Urals and Central Siberia, where wedge- travelled Mesozoic, Palaeogene and Quaternary sedi- shaped clasts mostly consist ofschists and limestones. ments (the famous Yugan, Samarovo and Semeyka Also, up to 40% ofthe West Siberian glacial pebbles erratics ofsoftJurassic and Paleogene rocks hundreds of show palimpsest aqueous roundness. Clasts in till are meters long). They sometimes have been transported as normally found to be well sorted by their size, which far as 600 km downglacier (Shatsky, 1965; Kaplyans- should be characteristic offluvial rather than glacial kaya and Tarnogradsky, 1974; Arkhipov et al., 1976). transport (Sukhorukova et al., 1987). All these are The above facts imply a thickness of the West Siberian unmistakable signs ofvery long paths ofglacial and ice sheets in the order ofkilometers, not hundreds of fluvial transport that involved multiple redeposition meters as was thought in the 1950–1960s (e.g. Lazukov, which certainly distorted the initial boulder trains. 1970; Zubakov, 1972). Therefore, the mapped ice limits, in agreement with Independent evidence ofvery thick Middle Pleisto- the flow pattern reflected in ice-pushed features, show cene ice in West Siberia is provided by numerous glacial only the direction ofthe final, most powerfulice valleys with irregular bottom profiles that are buried by advances from the north but not any earlier paths of drift, in places 300–400 m thick (Arkhipov and Mat- clast dispersal. veyeva, 1964; Zubakov, 1972; Arkhipov et al., 1976, However, the progressive mixing and integration of 1994). These overdeepened, predominantly N–S striking pebble and mineralogical composition over several ice valleys, never occur in the proglacial zone. Their glacial ages cannot explain the occurrence oflarge angular origin can be readily seen in the longitudinal profile of boulders from the Urals and Central Siberia, which are the Quaternary thickness across the drift limit based on occasionally found in the central West Siberian low- numerous boreholes along the Ob and her left tribu- lands. These can probably be accounted for by the same taries. In the periglacial area Pre-Quaternary bottom mechanism as suggested for the Pechora Basin, i.e. by profiles, gently sloping parallel to the present-day fluvial superglacial transport ofstones frommountainous thalwegs, do not show any overdeepening (1–4 in borderlands towards central lowlands caused by the Fig. 5). Immediately north ofthe maximum glaciation ARTICLE IN PRESS V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311 1295

Fig. 5. Longitudinal profile ofburied valleys ofthe Ob system (from Astakhov, 1991). Symbols: 1—present thalwegs; 2—sub-Quaternary surface; 3—glacial drift limit; 4—profile of boreholes across the valley. Circled numbers: (1) source of Turgai river flowing to the Aral Sea; (2) source of Ubagan river; (3) mouth ofUbagan river; (4) mouth ofTobol river; (5) the Irtysh mouth; (6) the Ob mouth. Compare normal fluvial profiles of extraglacial valleys in the leftwith bumpy overdeepened trough ofglaciated area in the right ofthe figure.

(Astakhov, 1976, 1977, 1979). Lowland tills ofMesozoic provenance ofthe West Siberian Basin ( Fig. 4B) and erratic boulders found high in the Urals indicate that above the Arctic Circle inland ice covered summits more than 1 km high, whereas at 62–63N the trimline occurs at c. 500 m asl. In the western part ofthe Central Siberian Upland three till sheets with lowland erratics, 33 m thick altogether, were described on the table-like summit ofa dolerite monadnock 618 asl at 150 km from the drift limit (location 13 in Fig. 1, Fainer et al., 1976). The surface of the inland ice, which brought Mesozoic material from the northwest and deposited it atop the inselberg at loc. 13, Fig. 1, must have been higher than 700 m asl (Fainer et al., 1976) and at least 900 m upglacier in the Yenissei valley, south ofloc. 9 (Fig. 1), where till ofthe maximum glaciation was found below sea level (Zubakov, 1972). These observa- tions make the West Siberian ice sheet at least 500 m thick at 64N at 150 km from the drift limit and c. 1000 m thick at 65N, i.e. at 350 km upglacier from the margin ofthe maximum ice sheet ( Fig. 6). Fig. 6. Clast provinces in tills ofthe West Siberian maximum glaciation (from Sukhorukova et al., 1987, simplified). UR—Uralian Voronov (1964), basing on empirical profiles of province ofcrystalline and Palaeozoic erratics; CS—province of Antarctic and Greenland ice sheets and the known Central Siberian erratics: dolerites, carbonate rocks and sandstones; configuration ofthe driftlimit in Siberia, calculated the WS—province ofno Uralian and Central Siberian stones, with maximum thickness ofthe West Siberian ice sheet as predominating quartz, opokas, diatomites, siltstones, sandstones of 3.5 km on the Yamal Peninsula. Judging by the above West Siberian Mesozoic and Cenozoic formations. Taimyr boulders sporadically occur in both WS and CS provinces. Arrows are ice flow geological facts, this estimate seems to be realistic. To directions inferred from fabrics and erratics. Dispersal of Central overpower and divert southwards ice streams from the Siberian and Uralian erratics is not conformable with configuration of Putorana Plateau, as evidenced by the foreign tills east of the ice limit (bold line), orientation ofice-pushed ridges ( Fig. 1) and the Yenissei, the lowland ice dome must have been really with lowland tills deposited in the mountains (Fig. 4). thick and probably occupied the entire Kara Sea shelf. limit the buried bottom ofthe Ob valley becomes very irregular, plunging deep below sea level (5, 6 in Fig. 5), 4. Chronology of ice advances which cannot be explained by normal fluvial processes. The principal ice dome over the Kara Sea shelfand 4.1. European Russia adjacent lowlands only got generally (although not unanimously) accepted after remote sensing had re- 4.1.1. General stratigraphic framework vealed the pattern of ice pushed ridges and after foreign The modern chronological concepts ofMiddle Pleis- erratics found in the mountains were considered tocene glaciations in northern European Russia are ARTICLE IN PRESS 1296 V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311

Table 1 Correlations ofpre-Weichselian glacial events ofnortheastern European Russia with other regions

Astakhov, this paper Guslitser et al. (1986)+Velichko and Shick (2001) Central Russia Northern Germany Isotope stages

Sula interglacial Sula interglacial Mikulino interglacial Eem 5e Vychegda glacial Vychegda glacial Moscow glacial Warthe Kostroma interstadial Interstadial 6 Dnieper glacial Drenthe Interstadial 7 Rodionovo interglacial Cold stadial Wacken 8 Chekalin interglacial (Domnitz). 9 Pechora glacial Kaluga cold stage Fuhne 10 Chirva=Rodionovo Chirva interglacial Likhvin interglacial Holstein 11 Pechora glacial Pomus glacial Oka glacial Elster 12–14 Visherka interglacial Visherka interglacial Roslavl=Muchkap interglacial 15 Maximum glacial Beryozovka glacial Don glacial Cromer 16 Kolva transgression Tumskaya interglacial Ilyinka interglacial 17 Pre-Kolva glacial Kama glacial Likovo glacial 18

heavily dependent on the official stratigraphic scale of Dnieper ice advance and the younger till containing Central Russia, either directly, as in the Arkhangelsk Fennoscandian erratics—with the Moscow glaciation. Region ofprevailing Fennoscandian glaciers, or in The official regional stratigraphic scheme offers correla- disguise oflocal startigraphic labels, as in the Timan– tion units (climatostratigraphic horizons) replicating Pechora–Vychegda Region ofdominant ice advances central Russian climatoliths (Guslitser et al., 1986). from the northeast (Guslitser et al., 1986). Several glacials and interglacials were distinguished The stratigraphic cornerstones ofcentral European from litho- and pollen stratigraphy in parallel with the Russia are classical interglacial formations of biogenic old Central Russian stratigraphic scale, in which Saalian and limnic sediments at the Likhvin and Mikulino tills were separated by the very warm interglacial called stratotypes, indicating climates warmer than the present. Odintsovo (Roslavl) around Moscow and Rodionovo They were traditionally correlated to the Holsteinian on the Pechora. and Eemian. The most continuous till sheet is poorly However, in the 1970s it was discovered that the till of topographically expressed and for a long time has been the Don Lobe was older than the Dnieper till ofthe associated with the Dnieper maximum ofthe Fennos- Ukraine (Velichko et al., 1977; Velichko and Faustova, candian glaciation in the Ukraine, close to 48N, and 1986). The Don till is overlain by the Muchkap with the German Saale glaciation. Another, less interglacial strata with two characteristic deciduous extensive pre-Eemian glacial complex, mapped mainly optima and with remains oftypically Tiraspol (Cromer) north of54 N by its distinct glacial topography, is rodents. A similar fauna was described from the attributed to the Moscow glaciation, presumably stratotype sections ofthe Roslavl interglacial. This equivalent to the Warthe glaciation (Yakovlev, 1956; implied that the maximum glaciation on the Don was Goretsky et al., 1982). The till underlying the Likhvin separated from the Dnieper till by at least two marker horizon for a long time was thought to represent interglacials: Roslavl (Muchkap) and Likhvin. Thereby the oldest Oka (Elsterian) ice sheet ofmore limited an intra-Saalian interglacial had to be abandoned extent (Table 1). (Shick, 1989). The new stratigraphy was proven beyond The most problematic was the third interglacial any reasonable doubt when sediments with the Likhvin formation (Odintsovo, or Roslavl) found between two (Russian Holsteinian) floras were found atop the uppermost Middle Pleistocene tills. This sequence, Roslavl sequence (Biryukov et al., 1992). In the showing a characteristic pollen profile with two decid- stratigraphic scheme ofCentral Russia, presently used uous peaks, is notably different from both Likhvin and by the Russian Geological Survey, no early Saalian till Mikulino pollen successions. In formal stratigraphic or overlying intra-Saalian interglacial formation (OIS 8 schemes and in general maps ofthe Quaternary it was and 7) are mentioned. The only till sandwiched between for many years placed between the Moscow and the Likhvin and Mikulino formations is the Moscow maximum Dnieper glaciations (Krasnov, 1971; Gane- glacial complex correlated with OIS 6 which presumably shin, 1973), although there always were dissidents laterally merges with the Dnieper till ofthe Ukraine insisting on a much older age ofthe Roslavl strata. (Shick, 1989). This scheme with two separate Saalian glaciations was Still, there are researchers who accept the new pre- also applied to northern European Russia, where the Holsteinian Don–Muchkap cycle but also insist on two Novaya Zemlya till was associated with the maximum separate glaciations (the Dnieper and Moscow) between ARTICLE IN PRESS V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311 1297 the Likhvin and Mikulino warm stages (e.g. Sudakova The second shallow-water atlantic transgression was and Faustova, 1995). Recently an attempt to fill in the later supported by drillings in the Pechora catchment gap between the Moscow till and Likhvin strata has area, where two sandy formations with boreal fauna been undertaken by Velichko and Shick (2001). They were found alternating with diamictons. The upper suggest as climatochronological units 2 palaeosols and 2 boreal formation with Arctica islandica had been known loess units found directly on top of the Likhvin long ago from many sections above sea level as a lacustrine formation with Brasenia flora (Table 1). In counterpart ofthe Eemian, whereas the lower formation this new scheme they place the Dnieper glaciation at the with characteristic extinct mollusc Cyrtodaria jenisseae beginning ofOIS 6, implying that the maximum drift (angusta) was reported from beneath a thick diamicton limit in the Ukraine is somewhat older than the Moscow just below sea level (Zarkhidze, 1972). till within the same ice age. The maximum glaciation of A third marine sequence ofdeep-water silts, the so- the eastern Russian Plain is firmly outlined by the Don called Kolva formation, found well below sea level at the Lobe (probably OIS 16) reaching south to the 50th base ofthe Quaternary cover, contains mostly a parallel. Subarctic fauna with characteristic Propeamussium groenlandicum (Yakhimovich et al., 1973) and probably 4.1.2. Stratigraphy in the Arctic corresponds to unit m1 in the Lavrushin’s profiles The modern stratigraphic scheme first did not affect (Fig. 2). In some boreholes it is underlain by a the remote northern areas, where two Saalian tills diamicton. From the Kolva formation Gudina (1976) (Pechora and Vychegda) were conventionally correlated described an arctoboreal foraminifera with character- with the Dnieper and Moscow glaciations. These tills are istic Miliolinella pyriformis, which she correlated with separated by a warm interglacial with two climatic the Ob and Turukhan strata in West Siberia and with optima described at Rodionovo as a counterpart ofthe the marine Holsteinian. The latter correlation is at odds Odintsovo/Roslavl (2 in Fig. 1)(Guslitser et al., 1986; with palynological investigations which find Likhvin- Loseva et al., 1992; Duryagina and Konovalenko, 1993; type spectra stratigraphically higher than the Kolva Andreicheva et al., 1997). The situation is more formation (Fig. 2). In one borehole a diamicton was complicated in the Arctic, where inter-till marine discovered at the base ofthe Kolva formation ( Yakhi- formations are not readily correlated with the central movich et al., 1973). Russian stratotypes, or with terrestrial interglacials of the Middle and Upper Pechora catchment. Drilling 4.1.3. Stratigraphy in the Subarctic projects by the Geological Survey provided a wealth of Terrestrial interglacial sediments sandwiched between information on the structure of drift in the Arctic. Two tills are known mainly from three natural exposures in such borehole profiles supported by field observations the Pechora catchment area. The best studied is the (A and B in Fig. 1) reveal at least four independent Rodionovo section (2 in Fig. 1) with very compact, diamict sheets separated by four marine formations slated and slightly distorted peat up to 3.5 m thick (Fig. 2). The topmost glaciotectonised complex with contained in clay and silt within a sandy inter-till fluvial blocks ofmarine sand is obviously a deposit ofthe shelf- sequence. In the arboreal pollen spectra spruce, pine and based Weichselian glaciation (Astakhov et al., 1999; birch are dominant with admixture of Abies and Alnus Svendsen et al., this volume). It is separated from the (Loseva and Duryagina, 1973). Two climatic optima thickest till body g3 by sands with an Eemian fauna (m4 were inferred based on a small admixture of deciduous by Lavrushin et al., 1989). The g3 till lies above sea level tree pollen (Duryagina and Konovalenko, 1993), which and cannot be anything but a Saalian till. Just below sea led the referred authors correlate this sequence with the level it is underlain by lacustrine sediments with Likhvin Odintsovo/Roslavl interglacial ofCentral Russia. type pollen spectra merging seawards with marine In the Kipiyevo section (3 in Fig. 1) lacustrine strata formation m3 (Lavrushin et al., 1989). The two lower- with a similar pollen assemblage also contain large Unio most marine formations m1 and m2 and intervening till shells, thick coniferous logs with ‘formidable growth of sheets g1 and g2 cannot be correlated directly with any tree-rings’ and nuts of Ajuga reptans, which now lives ofthe southern stratotypes. only in oak forests some 400 km to the south (Guslitser The old view ofone boreal transgression separating and Isaychev, 1983). The Kipiyevo interglacial strata are two last glacial events (Ramsay, 1904) was challenged by dissected by ice wedge-casts and overlain by sediments Yakovlev (1956) and by geologists who reported the first containing a late Middle Pleistocene assemblage ofteeth drilling results in the White Sea basin (Biske and ofpied and grey lemming, similar to that foundbetween Devyatova, 1965). They maintained that two boreal the Likhvin lacustrine strata and the overlying till south transgressions with similar mollusc fauna occurred: the ofMoscow. The evolutionary level ofthe Kipiyevo Boreal transgression s. stricto, an arm ofthe Eemian sea, lemmings is somewhat higher than in the Likhvin and the so-called Northern transgression, which was stratotype, which made Guslitser and Isaychev (1983) conventionally correlated with an intra-Saalian event. suggest a Moscow–Warthe age for this rather archaic ARTICLE IN PRESS 1298 V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311

Table 2 Middle Pleistocene glacial chronology ofSiberia (after Arkhipov, 1989)

Glaciations/Interglaciations Arctic zone Subarctic zone Isotope stages Ages of marine formations, ka Ages of terrestrial formations, ka

Kazantsevo ESR 105710.5; 120713; 121.9; 134.8 TL 130725 5e Taz 6 Shirta ESR 170710 ESR 196.8720.6 7 TL 180740; 190736 Samarovo 8 Tobol ESR 306721 ESR 326.9; 396 9–11 TL 260756 to 390780

Late Shaitan TL 510765; 12–16 Early Shaitan 550(561)7110(140) Talagaika TL 6607180; 7407170 17 Mansi 18

fauna. They also describe another find of similar rodent 16079 and 170713 (Table 3). Thus, there are two remains in the Akis section 6 km downstream ofloc. 2, different sets of OSL dates: one is close to 150 ka and Fig. 1. At this site lemming teeth were collected from another is close to 200 ka. The consistency ofthe dates cross-bedded sand overlying the same Novaya Zemlya and the large measured dose rates, according to A. till as is at the base ofthe Kipiyevo sequence. The Murray, make these dates apparently reliable. The 50 ka evolutionary level ofthe Akis lemmings is only slightly difference in OSL ages of the two sets of samples higher than in the Lower Saalian beds ofthe Likhvin probably means that a Late Saalian glacier picked up a section, again with no progressive Late Pleistocene peat layer ca 200 ka old and deposited it as a stratiform morphotypes present. Guslitser and Isaychev (1983) raft together with younger glaciofluvial sand. A Saalian relate the Kipiyevo interglacial to the Saale–Warthe age ofthis sequence is supported by OSL dates of interglacial (OIS 7) and the underlying till to the ‘early 10978 from the top and 143712 ka from the bottom of Saale’. the overlying aeolian sand. A Late Saalian OSL date of Recent dating attempts by the Russian–Norwegian 15279 ka was also obtained from the sand between the PECHORA project confirm an OIS 6 age for the peat and the upper till at Rodionovo (Table 3, sample by surficial till ofthe Subarctic zone. The dated section is O. Maslenikova). located in present-day tundra on Seyda river, formally The age ofthe lower, ‘Novaya Zemlya’ till ofthe in the Arctic (4 in Fig. 1) but featuring only terrestrial Pechora basin is less certain. Its correlation depends on sediments. A compact peat layer 1 m thick and 300 m the age ofthe Rodionovo interglacial. OIS 7 suggested long, with forest pollen spectra even richer than in by local geologists by comparison with the central Rodionovo, is contained in a thin sand sheet at the base Russian Odintsovo–Roslavl sequence (Guslitser et al., ofthe 40 m thick stacked till sequence. The peat first 1986) looks like a miscorrelation, because the Roslavl yielded a finite radiocarbon date and was thought to strata around Moscow are certainly pre-Likhvin, or pre- represent the Middle Weichselian (Lodmashchelye sec- Holsteinian (see above). tion by Arslanov et al., 1987). However, more detailed sampling ofthis sequence by 4.1.4. How many ice advances? J.I. Svendsen and M. Henriksen provided much older This question heavily depends on the correlation of ages. An uninterrupted series ofsamples ofthe inter-till the rare interglacial sequences. It is interesting to peat analysed for U/Th ratio in the laboratory of St. compare the palaeontological characteristics ofthe Petersburg University yielded ages ofca 200 730 ka BP Rodionovo formation with the Chirva interglacial strata in the middle ofthe peat layer. Younger values obtained known from boreholes in the south of the Subarctic from the top and bottom of the peat are accounted for zone (Vychegda catchment area) and correlated by by postdepositional influx ofyounger uranium (analyst pollen with the Likhvin interglacial (Duryagina and Yu. Kuznetsov). OSL dating on quartz particles in the Konovalenko, 1993). These and other authors (Guslitser peat produced ages of180 713, 185712 and et al., 1986; Loseva et al., 1992; Andreicheva et al., 1997) 191737 ka, whereas the surrounding sand was dated presume that the Chirva predates the Rodionovo. to c. 144 ka (three datings) and once to 173775 ka. However, both formations contain almost identical Glaciofluvial and glaciotectonised sands in the upper botanical taxa, including indicative Likhvin species part ofthe overlying glacial complex sequences yielded Osmunda claytoniana and O. cinnamomea (Table 7 in OSL values of148 710, 149713, 152711, 156716, Duryagina and Konovalenko, 1993). Both pollen ARTICLE IN PRESS V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311 1299

Table 3 Optically stimulated luminescence dates by the PECHORA project according to A. Murray, the Nordic Laboratory for Luminescence Dating, Ris^, Denmark

Lab. no Sample no. Location Age (ka) Dose rate (Gy/ka) Paleo dose (Gy)

002527 99-1238 Rodionovo, above peat 15279 1.2070.04 18376 002526 99-1237 Rodionovo, below peat 334729 0.9770.04 323722 992529 98-3078 Seyda 1, aeolian 10978 1.7370.09 18978 992528 98-3077 Seyda 1, aeolian 143712 0.9570.05 13678 002540 99-4226 Seyda 1 sand in till 148710 1.4970.06 220710 002539 99-4225 Seyda 1, sand in till 149713 1.3870.05 206715 002541 99-4227 Seyda 1, sand in till 152711 1.50 230713 002542 99-4228 Seyda 1, sand in till 16079 1.74 27979 992512 98-3064 Seyda 1, sand in till 156716 1.8770.09 291724 992513 98-3065 Seyda 1, sand in till 170713 1.7970.09 305713 992515 98-3029 Seyda 1, sand below till 144714 1.7670.09 254718 992511 98-3063 Seyda 1, sand below till 144713 1.4470.07 207714 962523 95-0070 Seyda 1, sand below tilll 145742 2.15 289 992502 98-3035 Seyda 1, sand above peat 185712 1.8270.09 33674 992501 98-3033 Seyda 1, interglacial peat 180713 1.3970.07 251710 962524 95-0071 Seyda 1, interglacial peat 191737 1.96 347 962525 95-0072 Seyda 1, sand below peat 173715 1.68 271 012584 01-0197 Sangompan, below rhythmite 8275 1.6170.07 13174 012583 01-0196 Sangompan, below rhythmite 9377 0.7870.05 7272 022518 01-0190 Sangompan, below rhythmite 7776 1.5470.06 11878 022519 01-0191 Sangompan, below rhythmite 7275 1.2970.05 9374 012586 01-0134 Pyak-Yaha, above rhythmite 133711 1.5970.08 211714 012585 01-0133 Pyak-Yaha, above rhythmite 13878 1.5170.07 20875 012581 01-0115 Pichuguy-Yaha, above rhythmite 125710 1.7770.08 221714 012582 01-0116 Pichuguy-Yaha above rhythmite 13779 1.6670.08 22779 012579 01-0119 Pichuguy-Yaha, below rhythmite 197715 0.9170.06 17978 012580 01-0120 Pichuguy-Yaha, below rhythmite 192716 0.7870.05 15077 012548 00-0505 Aksarka 2, coversand 17.370.8 1.7070.07 8173 012547 00-0504 Aksarka 2, loess-like silt 19.870.9 2.1370.08 8173 002550 00-0503 Aksarka 2, sand with wood 9778 1.90 18479 002549 00-0500 Aksarka 2, sand with wood 84710 1.94 168716 002545 00-0450 Sopkay, sandur 167722 1.17 192723 002546 00-0451 Sopkay, sandur 96713 1.54 148717 012544 00-0418 Yerkata, Yamal, aeolian sand above till 5874 1.8470.07 10675 012543 00-0416 Yerkata, Yamal, aeolian sand above till 6275 1.8970.08 11877 012542 00-0414 Yerkata, Yamal, lacustrine sand above till 6875 1.6970.07 11576 012541 00-0413 Yerkata, Yamal, lacustrine sand above till 5974 1.9470.07 11475

diagrams feature five zones and two climatic optima sequence in boreholes 8-y, 5-y, 3-y and 71 (Fig. 7)to with an admixture ofdeciduous trees, both, unlike the the Chirva interglacial, but the marine formation at the older Visherka horizon, contain no Tertiary relicts and same level in the SE boreholes 754 and 755 is correlated only rare exotic (Balkan-Caucasus) elements such as with the younger Rodionovo interglacial. Consequently, Betula sect. Costatae, Picea sect. Omorica, P. sect. the overlying till in the NW boreholes is thought to be Strobus. The slightly richer floristic compostion ofthe the Pechora till (OIS 8 in their correlation), but its Chirva strata is easily explained by the more southerly lateral counterpart in the SE is referred to the younger position ofthe studied site. As mentioned above, the Vychegda till (OIS 6) (Fig. 7). Neither the lithological rodent assemblages above the Kipiyevo (Rodionovo) composition ofthe tills, nor the landscape features strata are very similar to those found on top of the support such a differentiation. In the profile described Likhvin type sequence. The only OSL date available by Lavrushin et al. (1989) slightly farther to the east (A from the base of the Rodionovo peat gave a fairly old in Fig. 1) the lateral change from the Likhvin lacustrine value 334729 ka, more appropriate for OIS 9 or 11 than formation to the marine formation m3 (Fig. 2), for OIS 7 (sample by O. Maslenikova, Table 3). suggested by the authors, seems more logical. The most important is the stratigraphic position of The Chirva and Rodionovo interglacials have never the interglacial formations in borehole profiles of the been found in superposition (Figs. 2 and 8). Therefore, it Arctic zone (A, B and C in Fig. 1). Loseva et al. (1992) appears that the uppermost Middle Pleistocene inter- relate the pollen spectra ofthe terrestrial inter-till glacial, the Rodionovo on the Pechora, is identical with ARTICLE IN PRESS 1300 V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311

Fig. 7. Middle Pleistocene formations in boreholes of profile C (Fig. 1) (from Loseva et al., 1992, simplified). Letters indicate correlation with regional climatostratigraphic horizons assumed by Loseva et al. (1992): v—Visherka interglacial, pm—Pomus glacial, W—Chirva interglacial; pW— Pechora glacial; r—Rodionovo interglacial, vW—Vychegda glacial. Broken lines is alternative correlation suggested by this author. the Chirva interglacial ofthe Vychegda catchment area. thought as Elsterian in age. There is no lithostrati- Many interglacial sequences covered by the upper till in graphic evidence ofa laterally consistent Pomus glacial the Subarctic zone, and lying just below sea level in the complex. Arctic, probably belong to the same interglacial interval, Whether the Chirva/Rodionovo interglacial corre- most likely represented by the marine formation with sponds to OIS 9 or to OIS 11 is uncertain, but its Cyrtodaria angusta. The Cyrtodaria strata are in places stratigraphic postion makes it the best candidate for found up to 70 m asl (Zarkhidze, 1972), which gives a correlation with the Russian Holsteinian, i.e. the rough idea ofthe large isostatic depression caused by the Likhvin s. stricto. The preceding Visherka interglacial thick preceding ice sheet. sequence, containing Tertiary relics such as Liquidambar In the compromise scheme by Velichko and Shick and Pterocarya plus exotic Tsuga and Ilex, not known (2001), accepting the different Chirva and Rodionovo from younger interglacials, consequently would better interglacials, the Pechora glaciation is not correlated correspond to the Roslavl/Muchkap strata overlying the anymore with the Dnieper or other maximum ice Don till ofthe maximum glaciation, as suggested by advance ofCentral Russia. Instead it is shifteddown- Velichko and Shick (2001) (Table 1). wards to OIS 10, probably in order to accommodate at In this respect sections close to the glacial drift limit OIS 12 another glaciation called the Pomus by Guslitser are ofparticular interest. The best picture ofthe et al. (1986) (Table 1). This results in the correlation of geological structure ofthe Pechora–Volga interfluve the mighty Pechora glaciation ofNE provenance with area is given by the geotechnical drilling data described the thin Kaluga loess on top ofthe Likhvin stratotype by Stepanov (1974, 1976). Two ofhis profiles crossing sequence. However, the Pechora till, widely observed in each other at right angles are presented in Fig. 8. The many Subarctic sections and covered by the distinctly surficial till g3 correlated by Stepanov with the Dnieper interglacial Rodionovo formation, is still the most glaciation is covered by only one interglacial alluvial salient feature of the northern drift. Therefore, it better formation of the modern valleys with fresh-water correlates with a central Russian till covered by the molluscs, rich diatom flora and chracteristic Mikulino Likhvin strata, the Oka till, which has been always pollen spectra. A modern area ofconcentration ofthe ARTICLE IN PRESS V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311 1301

Fig. 8. Borehole profiles ofMiddle Pleistocene tills at Lake-Chusovskoye, Pechora–Kama interfluve, N 61 (from Stepanov, 1974, simplified). 1—diamicton; 2—gravel; 3—sand; 4 laminated silt and clay; 5—loess-like silt; 6—bedrock; 7—borehole. Broken line shows intersection oftwo perpendicular profiles. Till units are numbered upwards in the succession g1, g2, g3 irrespectively oftheir possible age. fossil flora can be found in southern Germany and Kama river (Yakhimovich et al., 1973). The overlying g3 Poland. The same is true ofthe Mikulino type flora of till should be the Uralian counterpart ofthe Vychegda central Russia. till. The underlying dark-grey till g2, low in Uralian The interglacial formation between g3 and the boulders, contains mostly clasts ofsedimentary underlying g2 till (Fig. 8) contains a flora similar to rocks and in this respect is not different from the the Likhvin assemblage with indicator plants such as Pechora till. Osmunda claytoniana, Pinus sect. Strobus, Picea sect. The lower inter-till formation (between g2 and g1) Omorica, Tilia tomentosa which have a modern area of shows pollen spectra of mixed forests with very few concentration in the western foothills of the Alps, pollen of Carpinus, Corylus and exotic trees. The high according to L. Tyurina. The plant macrofossils percentage ofpyrite grains is typical forthe oldest identified by P.I. Dorofeyev are typical for the Singil interglacial ofthe eastern Russian Plain ( Stepanov, floras ofthe Russian ‘Mindel-Riss’ ( Stepanov, 1976). A 1974). Ifthe lower interglacial sequence, locally labeled single find offorest elephant Palaeoloxodon sp. is known the Solikamsk formation, can be correlated with from a similar sequence on Kolva river, tributary to the the Roslavl/Muchkap strata, then the underlying ARTICLE IN PRESS 1302 V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311 overconsolidated g1 till (the Kama till ofStepanov) Siberia. Therefore, in distinguishing interglacials one should be a counterpart ofthe Don till ofsouthern has to rely on N–S shifts of very broad biogeographical Russia. zones, which would demand hundreds ofmeticulously The alternative interpretation puts the Solikamsk studied sections. In literature one can usually see interglacial into OIS 17–19 and the Kama till into OIS generalisations ofa continental scale based on a handful 20 (Zubakov, 1992), which makes the Kama till ofsites and, which is even worse, with European labels contemporaneous to the Mansi till ofWest Siberia attached to totally unrelated objects, i.e. atlantic terms presently attributed to the Matuyama chron (Volkova applied to a non-atlantic environment. This certainly and Babushkin, 2000). Ifthe latter interpretation is hinders development ofan independent Siberian Pleis- correct, then the maximum glaciation ofnortheastern tocene startigraphy. Many geologists, exasperated by Russia is represented by g2 till ofStepanov’s profile the difficulties encountered in Siberia, took to elaborat- (Fig. 8). The limit ofthis maximum glaciation shown in ing local stratigraphic scales by exclusively chronometric the left lower corner of Fig. 1 was for many years methods which normally should only be used for long- attributed to the Dnieper (early Saalian) glaciation distance correlation. This poor practice, exemplified by (Krasnov, 1971; Ganeshin, 1973) which later proved to the popular misuse ofradiocarbon dating, led to be wrong (Velichko et al., 1977; Shick, 1989). corruption ofthe traditional stratigraphic nomencla- The correlation between the Arctic region ofmarine ture, which is not reliable anymore, being permanently transgressions and the Upper Pechora–Vychegda catch- reshuffled after each laboratory ‘discovery’ (Astakhov, ments remains problematic. A formal comparison says 2001). To lesser degree the same happens with the that the three main glacial complexes underlying the Middle Pleistocene, although the limited number of Upper Pleistocene ofthe Arctic ( Fig. 2) probably available chronometric methods helps to keep the old correspond to the three tills ofthe Pechora–Kama stratigraphic terminology in a better shape. interfluve (Fig. 8). However, g1, g2 and g3 in both Two stratigraphic markers, identified in key sections profiles are just diamict units numbered by this author ofthe West Siberian basin, over several decades have and cannot be viewed as synchronous with similarly been the cornerstones ofglacial history and Quaternary designated tills in the Arctic profiles. Actually, it is not mapping. These are the interglacial formations of the easy to find in the Arctic Pleistocene counterparts to all Kazantsevo transgression in the Arctic and the Tobol warm intervals ofthe southern record. Two boreal alluvium in central and southern West Siberia. The transgressions are more or less satisfactorily reflected in Kazantsevo formation consists mostly of shallow-water the warm Mikulino and Likhvin floras ofthe south, facies containing rich arctoboreal mollusc fauna with whereas there is no evidence to synchronise the cool characteristic boreal species such as Arctica islandica Kolva marine strata with the Visherka or Solikamsk indicating water temperatures 4–8C above the present- interglacial formations with their exotic plant remains. day Kara Sea temperature (Sachs, 1953; Troitsky, 1975). It is possible that the Kolva interglacial is missing in the This formation is conventionally correlated with the terrestrial record ofthe Subarctic zone. This question Eemian strata ofthe Boreal transgression in European will stay open until more reliable means oflong-distance Russia. This correlation seems to be confirmed in correlation are found. Yenissei Siberia by several ESR dates on marine shells The above briefanalysis ofthe available stratigraphic in the range 108–134 ka (Sukhorukova, 1999). The main data allows to conclude that there are at least three drawback ofthis marker horizon is its being limited to readily recognisable glacial complexes and major pre- the Arctic, where it is overlain by till and often badly Weichselian ice advances in northeastern European glaciotectonised. Russia, which certainly does not preclude their subdivi- The marker significance ofthe Kazantsevo formation sion into minor glacial stages. The fourth ice advance is somewhat diluted by the fact, that, as in European preceding the Kolva transgression in the Arctic might Russia, there are distinct traces ofanother boreal tentatively be correlated with the Kama till ofthe transgression. First, boreal mollusc shells and foramini- Pechora–Volga intefluve. fera have been locally found in Middle Pleistocene diamictons (Arkhipov and Matveyeva, 1964; Kaplyans- 4.2. Siberia kaya and Tarnogradsky, 1975). Second, there are inter- till strata with rare boreal molluscs (Zubakov, 1972). 4.2.1. General stratigraphic situation Third, some Kazantsevo marine sequences contain The Siberian glacial chronology is more arbitrary due shells ofextinct species Cyrtodaria jenisseae (angusta) to the formidable size of the country and poor (Sachs, 1953; Kind and Leonov, 1982, see also discus- applicability ofpollen analysis. Unfortunately the sion on Taimyr in Svendsen et al., this volume) and dominant boreal forests have a monotonous composi- probably belong to earlier interglacials. tion with practically no deciduous trees presently The Tobol strata are much better preserved in the growing, except sparse lime in southwestern West area oftheir classic occurrence along the transverse Ob ARTICLE IN PRESS V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311 1303 and lower Irtysh, mostly between 62Nand56N, i.e. Semeika varves are related to the Tobol climatostrati- just within the drift limit and in the periglacial area. graphic horizon s.stricto (Volkova and Babushkin, They are washed, diagonally bedded quartz sands with 2000). Correlation ofthe Tobol horizon with the thick lenses ofgreen-bluish clayey silts, partly covered Holsteinian is apparently supported by TL dates of by till ofthe maximum glaciation. Literature on the 300775 and 313780 ka and ESR date 306720 ka on a Tobol strata is extensive, especially regarding palaeon- Corbicula shell. However, in the periglacial zone tological questions. A comprehensive overview of Corbicula shells yielded ESR ages 285, 219 and 174 ka geological data is given in a collection ofpapers edited (Arkhipov, 1989). Later the palaeomagnetic excursion by Arkhipov (1975), and also in numerous works by Biwa-II-Semeika along with the ESR date of396 ka palaeobotanists. were reported from the basal part of the Tobol alluvium Pollen spectra, as always in Siberia, are not very with Corbicula shells (Volkova and Babushkin, 2000). characteristic. Sequences in the present taiga zone Anyway, there are two independent interglacial mainly show a predominance ofarboreal Betula and markers, one ofmarine and another ofterrestrial origin rich herb assemblages with a minor coniferous compo- that can be seen in West Siberian exposures. The Arctic nent, which is common for the Siberian forest-steppe. marker, ifidentified correctly, should provide the upper Some successions reveal two coniferous peaks below and boundary ofthe Middle Pleistocene tills. The Subarctic above the main forest-steppe phase. The interglacial marker separates surficial Middle Pleistocene tills from nature of the formation is clear from abundant fresh- pre-Holsteinian glacial events known only from glacial water molluscs, especially Corbicula fluminalis (tibeten- deposits ofburied valleys. sis) presently living only in Central Asia. However, a rather cool and relatively humid climate is indicated by 4.2.2. Stratigraphy in the Arctic the rich macrofossil flora with exotic aquatic ferns The till overlying the Kazantsevo strata in most cases Azolla interglacialica, Selaginella selaginoides, Salvinia can be safely correlated with the Weichselian, the natans, etc. This assemblage, known as the ‘Flora ofthe underlying till being related to OIS 6 (Svendsen et al., Diagonal Sands’, is similar to the Singil flora ofthe this volume). These ice advances are distinguished by Russian Plain and has for several decades been the main their geographic distribution, because the post-Kazant- indicator ofthe so-called ‘Siberian Mindel-Riss’ and an sevo glaciation is limited to the Arctic, whereas the pre- argument for its correlation with the Likhvin and Kazantsevo ice advance reached far beyond the Arctic terrestrial Holsteinian. Later Azolla remains were also Circle. The main problem is that in many cases either found in Upper Pleistocene sediments of the Lower Ob marine strata with a boreal fauna are older than the (Arkhipov et al., 1977). Kazantsevo s. stricto, or the Eemian interglacial is The mammal fauna is the most controversial issue, represented by non-marine facies. The first case is probably due to the all-pervading redeposition of known from the Pupkovo section in the Yenissei valley osteological material by the huge laterally migrating (loc. 9 in Fig. 1), where marine strata with interglacial rivers. More frequentare finds ofbones ofthe Tiraspol pollen spectra and rare finds ofa boreal fauna,including (Cromer) mammals. There is only one known site with Arctica islandica, are sandwiched between two tills well forest fauna of the Singil type represented by Palaeolox- beyond the limit ofLate Pleistocene glaciation ( Zuba- odon, Megaloceros, Bison, Eolagurus, Arvicola, etc. In kov, 1972). Troitsky (1975), who insisted on the Eemian many places the ‘diagonal sands’ contain mammal age of this marine formation, had therefore to extend species ranging from pre-Tiraspol to typical Late the Weichselian ice limit along the Yenissei as far south Pleistocene lemming faunas. The explanation is that as 64N, which is refuted by periglacial evidence the ‘diagonal sands’ may be deposited at the same (Astakhov et al., 1986). A similar situation is in the topographic level by slow meandering rivers during central part ofthe West Siberian basin close to the different interglacials. The Tobol formation proper, Arctic Circle, but well beyond the currently accepted according to Arkhipov (1975), chronologically ranges Weichselian ice limit (Svendsen et al., this volume). In from the second half of Mindel to the Mindel-Riss the Samburg borehole a thick marine formation with an interval, i.e. belongs to two interglacials. arctoboreal fauna was found beneath a Middle Pleisto- This interpretation was supported by a wedge of cene till (Zubakov, 1972). glaciolacustrine varves (the Semeika formation) found Non-marine Eemian is found on the Lower Ob, where in the middle ofthe Tobol formation on the Irtysh, two main stratigraphic concepts have been competing. where it is also overlain by till ofthe maximum The classical concept relates most ofsurficial diamictons Samarovo glaciation. The lower alluvial sequence, called and varved sequences to glacial or glaciomarine forma- the Talagaika formation, is thought to represent a pre- tions ofthe Middle Pleistocene ( Lazukov, 1970; Holsteinian interglacial (Kaplyanskaya and Tarno- Zubakov, 1972). According to Zubakov, this interpreta- gradsky, 1974). In the present stratigraphic scheme only tion is based on pollen spectra characteristic ofsouthern the Corbicula sands between the Samarovo till and the taiga found in sand infills incised into the thick varved ARTICLE IN PRESS 1304 V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311

Fig. 9. Principal profile across the Ob valley on the Arctic Circle south ofEarly Weichselian Sopkay moraines (not to horizontal scale). Russian– Norwegian PECHORA project. 1—loess-like silt; 2—diamicton; 3—silt/clay rhythmite changing into varves; 4—glacifluvial sand and fine gravel; 5— laminated sand with peat lenses. Open dots are OSL (optically stimulated luminescence) datings, blackened dots are radiocarbon dates; age values are indicated in kiloyears BP. Note on southern bank non-finite radiocarbon dates and four OSL ages 125–138 ka from interglacial sand overlying rhythmite ofmid-Weichselian age by Arkhipov et al. (1977). rhythmite at the Pyak-Yaha section ofthe southern only one till, contains an assemblage ofarctoboreal bank ofthe Ob in the present forest-tundra ( Fig. 9). foraminifera of ‘Miliolinella pyriformis zone’. The thick Only thin lenses ofdiamict and gravel materials which diamictic sequence positioned between the two inter- sometimes occur on the surface suggest a Late glacial marine formations in this scheme belongs to OIS Pleistocene glaciation (Lazukov, 1970). In contrast, the 6–8. Arkhipov et al. (1994) offer TL-dates on core concept ofthe ‘young stratigraphy’, based mostly on samples of153 715 ka for their ‘Kazantsevo formation’ sparse radiocarbon dates obtained in different sections, and 246723, 306726, 366731, 370731 ka for the ‘Ob ascribes a Late Pleistocene age to all units ofstratified strata’. and non-stratified drift observed in natural sections. In The latest results from the Russian–Norwegian the latter scheme the Kazantsevo marker is represented PECHORA project clearly refute the ‘young stratigra- by sand with boreal foraminifera identified in boreholes phy’. Weichselian tills and related glaciolacustrine just below sea level (Arkhipov et al., 1977). This concept sediments are found only on the northern bank of the is currently accepted in the official stratigraphic scheme Ob river (Fig. 9). The fluvial or deltaic sands with forest (Volkova and Babushkin, 2000). pollen spectra incised into the thick varved sequence of Arkhipov et al. (1994), applying their radiocarbon- the southern bank yielded four OSL dates of 125 to based ‘young stratigraphy’ to the entire Arctic Pleisto- 138 ka (samples 115, 116, 133 and 134, Pyak-Yaha, in cene, suggest a Holsteinian age for another interglacial Table 3), together with non-finite radiocarbon dates formation at 100–200 m bsl, the so-called ‘Ob marine supporting an Eemian age ofthese sands. Glaciofluvial strata’. This older interglacial formation, underlain by sands beneath the varved sequence, as well as on high ARTICLE IN PRESS V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311 1305 interfluves ofthe Uralian piedmont, are OSL dated to c. lower marine unit and a mid-Weichselian age for the 200 ka (samples 119 and 120 Pichuguy-Yaha in Table 3), upper marine formation (Kind and Leonov, 1982). The which is consistent with a Middle Pleistocene age ofthe latter interpretation is not only palaeontologically last ice advance on the Arctic Circle (Fig. 9). The dubious, it is also at odds with the latest results ofthe underlying Middle Pleistocene till and terrestrial sand QUEEN program which do not support a Late with peat are stratigraphically above the Arkhipov’s Pleistocene glaciation that far east (Svendsen et al., ‘Kazantsevo’ with TL-date of153 ka, which means that 2004). The traditional interpretation ofthis sequence in the TL method underestimates the age, as compared to which the earlier transgression is thought to be OSL dating. The better reliability ofOSL datings has Holsteinian (e.g. Gudina, 1976) suggests a Saalian age been independently confirmed by U/Th dating ofthe for the Murukta till, which agrees well with its Seyda peat and by the above pollen data. Therefore, the widespread occurrence in Central Siberia. sand with boreal foraminifera below sea level, Eemian by Arkhipov et al. (1977), is most likely Middle 4.2.3. Stratigraphy in the Subarctic Pleistocene, probably Holsteinian. Best identifiable in this zone is the Samarovo till ofthe Consequently, the deep lying Ob interglacial forma- maximum glaciation named after the settlement close to tion must be pre-Holsteinian, similar to the Kolva the Irtysh mouth, where the famous thick sequence of formation of the Pechora Basin with the same for- stacked tills and Palaeogene opoka rafts has long been aminifera assemblage (Gudina, 1976). This implies that known (Shatsky, 1965). The additional diamicton unit the thickest (more than 100 m) till sequence, sandwiched on top ofthe Samarovo till north ofthe Irtysh mouth is between two interglacial marine formations, is not thought to represent the penultimate Taz glaciation. The Saalian but older. Another important implication is Samarovo till, resting on the Tobol alluvium, is getting that, accepting the traditional correlation ofthe Ob thinner upstream and disappears south ofthe Semeika strata with the Tobol alluvium, the latter must be pre- village (5 in Fig. 1). North ofthis point a couple of Holsteinian. However, it is possible that the Tobol additional tills beneath the Tobol formation are known alluvium corresponds in the Arctic to the first marine from boreholes. The lowermost interglacial formation formation with boreal foraminifera lying just bsl, i.e. to (the Talagaika alluvium) north of63 N is overlain by a the ‘Kazantsevo strata’ by Arkhipov et al. (1977, 1994). double diamict formation up to 70 m thick called the In this case the lower marine formation, the Ob strata, Shaitan till which is thought to correspond to the may correlate with the Talagaika or even older Semeika glaciolacustrine clay on the Irtysh (Volkova interglacial deposits ofthe south (see below). and Babushkin, 2000). The lowermost till (the Mansi Middle Pleistocene tills ofthe eastern glaciated Arctic till) is found beneath the Talagaika interglacial at the are poorly studied. The thick fine-grained diamictons bottom ofa drill well north ofthe driftlimit on the directly underlying the Kazantsevo marine strata with Irtysh (Arkhipov, 1989). boreal fauna were related to the marine or glaciomarine There are two main problems with the till count in Sanchugovka formation with a sparse, predominantly this zone. The first is connected with the Taz glaciation Arctic fauna (Sachs, 1953; Lazukov, 1970; Zubakov, which is thought to predate the Eemian and deposit a till 1972) until Kaplyanskaya and Tarnogradsky (1975) oflimited distribution on top ofthe Samarovo glacial thoroughly investigated the type section and found that complex. Originally the Taz till was mapped in the upper the formation consisted of basal tills with rafts of marine reaches ofriver Taz, where it is separated fromthe sediments. The diamictons contain not only an Arctic underlying Samarovo strata by sands with ambiguous fauna, but also shells of boreal and Cretaceous molluscs. palaeontological characteristics. Later such a till was The Sanchugovka till is believed to represent OIS 6, distinguished all over West Siberia. Although no although Arkhipov (1989) thinks that below sea level interglacial formation with abundant organics has ever there are real marine strata with Arctic foraminifera of been found between the Samarovo and Taz till, the OIS 7 underlain by tills ofOIS 8. alleged Shirta interglacial, called ‘interstadial’ by more One ofthe rare sections in which two interglacial cautious geologists, persists in the regional stratigraphic marine formation are separated by the Murukta till can schemes (Volkova and Babushkin, 2000). Interglacial be seen is Novorybnoye at the mouth ofKhatanga river organics are in general rare in Siberian inter-till (10 in Fig. 1). The upper marine unit, which is not formations. This led Lazukov (1970) to suggest that all covered by till, is commonly interpreted as a deposit of till sheets along the Ob river are deposited by the the Eemian transgression, whereas the lower marine maximum Samarovo glaciation, whereas (Arkhipov formation, containing foraminifera of the Miliolinella 1989; Arkhipov et al., 1978) tried to subdivide this pyriformis zone, and in another section farther to the sequence by means ofthermoluminescence dating (A in west also Cyrtodaria angusta, is thought to represent the Fig. 10). Siberian equivalent ofthe Holsteinian. The alternative According to Arkhipov the upper sand and silt in stratigraphic model suggests an Eemian age for the Kormuzhikhantka section on the Belogorye Upland ARTICLE IN PRESS 1306 V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311

Fig. 10. Key sections ofMiddle Pleistocene tills ofWest Siberia. A—Kormuzhikhantka section at locality 6 in Fig. 1 (from Arkhipov, 1989; Arkhipov et al., 1978, simplified), B—Bakhtinsky Yar section at locality 8 in Fig. 1 (Arkhipov and Matveyeva, 1964), C—Khakhalevsky Yar section supplemented by a borehole at locality 7 in Fig. 1 (Levina, 1964). 1—loess-like silt; 2—soft diamicton; 3—compact diamicton; 4—varved clay; 5—laminated silt; 6—sand; 7—peat and gyttja; 8—cryoturbations; 9—fresh-water mollusks; 10—TL samples.

(loc. 6 in Fig. 1) is Eemian, and the overlying soft formation with rare boreal shells (loc. 9 in Fig. 1). diamicton is therefore an Early Weichselian till (Fig. Zubakov (1972) relates the upper till to the last Middle 10A). The till below the Eemian TL values belongs to Pleistocene glacial event called the Yenissei glaciation, the Taz glaciation ofOIS 6, and the till next down- OIS 6. This event is supposed to be identical with the wards—to the Samarovo glaciation ofOIS 8. The Taz glaciation ofWest Siberia. Isayeva (1963) acknowl- lowermost till in this sequence is related to the Late edged this till as a counterpart ofher second (Nizhnyaya Shaitan glaciation ofOIS 10–12 ( Arkhipov, 1989). This Tunguska) belt ofend moraines running around the correlation does not contradict the latest PECHORA Putorana Plateau between the drift limit and the Late results in the Arctic Ob area. However, taking into Pleistocene Onyoka moraines. The Nizhnyaya Tungus- account the too young TL dates in the Arctic zone, this ka moraines in the northeast merge with the Murukta lowermost till might be even older. As to the uppermost moraines, which in the stratigraphic scheme ofCentral Kormuzhikhantka till, this soft, mantle-like diamicton, Siberia are positioned above the Eemian level (Isayeva very low on pebbles, without structures ofice flow or et al., 1986). This seems to be a miscorrelation because, shear planes at the base, is neither traceable regionally, according to the QUEEN results, the Early Weichselian nor associated with other glacigenic sediments. There- maximum is represented by the youngest belt ofthe fore, it should be better viewed as a solifluction bed, or a Onyoka moraines sensu Isayeva (1963) (see Taimyr flowtill derived from residual Middle Pleistocene ice, but discussion in Svendsen et al., this volume). Therefore, not as a signature ofa Late Pleistocene ice advance both Yenissei and Murukta tills should belong to OIS 6. (Zolnikov, 1990). The latter interpretation fits the The till underlying the Pupkovo marine strata (loc. 9 PECHORA data indicating no Late Pleistocene glacia- in Fig. 1) is traditionally correlated with the Samarovo tion on the Ob along the Arcticle Circle (Fig. 9). glaciation ofOIS 8, which is thought to have reached the Another problem is the correlation ofthe Ob-Irtysh drift limit also on the Yenissei (Zubakov, 1972). ice advances with the Central Siberian events. As was However, it is more likely that the earlier boreal pointed out, south ofthe Late Pleistocene ice limit on transgression, as on the Pechora and Ob, relates to the the Yenissei there are two tills separated by a marine Holsteinian or OIS 11, thus making the lower till in the ARTICLE IN PRESS V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311 1307

Yenissei bluffs Elsterian in age. The official stratigraphic 4.2.4. How many ice advances? scheme keeps all these strata in the Bakhta superhorizon The most pressing stratigraphic problem concerns the (Volkova and Babushkin, 2000) exemplified by the age ofthe two surficial tills—the Taz and Samarovo— Bakhtinsky Yar section (B in Fig. 10, loc. 8 in Fig. 1). overlying the Tobol horizon. They are traditionally There is no indications ofwarm palaeoclimates in the related to OIS 6 and 8 correspondingly (Table 2). The inter-till sediments traditionally correlated with the alternative that both belong to OIS 6 is unlikely, because ‘Shirta interglacial’ (or interstadial) (Arkhipov and glacial landforms are not known atop of the Samarovo Matveyeva, 1964). This sequence is a rare case ofthe till covered by loess-like silts up to 15 thick. Ifthe Tobol lower till overlying interglacial alluvium with a bone of horizon proves to be pre-Holsteinian, and real inter- Alces latifrons. The latter is characteristic ofthe Tiraspol glacial organics are found between the Taz and (Cromer) mammalian complex (Arkhipov and Matveye- Samarovo tills, then the Taz till may relate to the va, 1964; Arkhipov, 1975), which implies that the lower Saalian (OIS 6) and the Samarovo till to the Elsterian. till in the Yenissei sections might well be Elsterian. The latter option was considered by Arkhipov based Upstream ofBakhtinsky Yar a thicker till formation on TL and ESR datings. He concluded that the Tobol was found below the Bakhta strata in a borehole formation s. stricto might be pre-Holsteinian and reaching 342 bsl. This so-called Lebed till is the lower- therefore the Samarovo glaciation might be Elsterian. most member ofthe official stratigraphic scheme of However, this alternative would upset the traditional Central Siberia (Isayeva et al., 1986). correlation pattern and therefore not desirable for the Another key section is Khakhalevsky Yar close to the time being (Arkhipov, 1989). Anyway, the great thick- drift limit (C in Fig. 10, loc. 7 in Fig. 1). It represents the ness ofthe tills underlying the penultimate interglacial in Turukhan alluvium (Arkhipov and Matveyeva, 1964), all major buried valleys suggests that Elsterian or earlier or Panteleyeva formation (Zubakov, 1972), sandwiched ice sheets ofWest Siberia were very large. It is quite between two tills. The coarse channel alluvium overlying possible that several ice advances are still hidden in the the lower till grades upwards into floodplain silts with 70–100 m thick diamict sequences ofthe buried valleys, gyttja and peat lenses, changing farther upwards into where no good interglacial formations have been proglacial varves distorted by the maximum ice advance. distinguished so far. The latest results of deep coring The succession is capped by till and glaciofluvial sand. in Arctic West Siberia show that assemblages of Pollen spectra ofthe channel alluvium show an upward arctoboreal foraminifera (previously perceived as the increase of pollen of coniferous forests with predomi- Holsteinian ‘zone of Miliolinella pyriformis’) occur at nance of Picea and Pinus sibirica. In the floodplain silts three different stratigraphic levels (Volkova and Ba- they are gradually replaced by Betula dominated park- bushkin, 2000). This makes correlation ofthe Arctic tills lands with abundant herbs, Ericaceae and Selaginella with the terrestrial record in the south even more selaginoides (Levina, 1964). By these and other similar difficult. spectra the Turukhan alluvium is correlated with the The above data allow to infer four major pre-Eemian Tobol alluvium ofthe Ob and Irtysh ( Arkhipov, 1975). ice advances in the Siberian record (Table 2). The oldest However, pollen spectra ofthe preceding Talagaika Mansi till reflects a more restricted ice sheet. In the interglacial are not much different. Therefore, it cannot present official stratigraphic scheme of West Siberia be excluded that the maximum glaciation on the (Volkova and Babushkin, 2000) this ice advance is Yenissei correlates not with OIS 8, as suggested by related to the Matuyama chron ofreverse polarity, i.e. is Arkhipov for the Samarovo glaciation, but with the thought to be older than 700 ka, although Arkhipov preceding cold stages OIS 10 or 12. (1989) preferred to place it at OIS 18. The thickest The latter option is even more likely for the Central Shaitan tills separated by stratified drift with Arctic Siberian Upland, where the poorly preserved tills ofthe foraminifera are certainly pre-Holsteinian but within the maximum glaciation are in a stark contrast with the Bruhnes chron, their more precise correlation with the inner belt oftopographically expressive Nizhnyaya European record being premature. The correlation with Tunguska moraines (Isayeva, 1963). There are authors OIS 12–16 suggested by Arkhipov (Table 2) is only who correlate the maximum glaciation limit ofthe based on very old and therefore hardly reliable TL eastern Central Siberia with the thick sequence of dates. The maximum glaciation (OIS 8, or 10, or 12) diamictons and stratified drift found by boreholes at probably produced the thickest (up to 3.5 km) ice sheet 240 m bsl and lower on the bottom ofthe overdeepened that grew over the Kara Sea shelfand eventually Yenissei valley at Lebed (loc. 14 in Fig. 1). By its overrode nearby mountain ranges up to an altitude of position under the sedimentary complex ofthe max- 1 km. The last Middle Pleistocene ice sheet called the imum Yenissei glaciation the Lebed double till sequence Taz (or Yenissei, or Murukta), judging by the fresh is similar to the Shaitan tills on the Ob (Volkova and glacial landscapes, was formed in OIS 6. It was thick Babushkin, 2000). However, the Lebed tills might even enough to cover almost the same area as the maximum be pre-Elsterian in age. glaciation and flow southwards over the low mountains. ARTICLE IN PRESS 1308 V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311

5. Conclusions Editing efforts by J. Ehlers, A. Raukas and an anonymous referee have considerably improved the (1) The huge ice sheets, much larger than the Weichse- text. The author offers his sincere thanks to the lian ones, at least 4 times covered the Russian mentioned persons and institutions. mainland and adjacent shelves prior to the Eemian. (2) The main centre ofice accumulation was the Kara Sea shelf, additional sources of inland ice being Fennoscandia and the Barents Sea shelfin Eur- References opean Russia, and the Putorana Plateau in Siberia. The Ural Mountains played mostly the passive role Andreicheva, L.N., 1992. Basal tills ofthe European Northeast and ofan orographic barrier in the ice flow pattern. their Lithostratigraphic Significance (Osnovnye moreny Yevro- (3) The extent and geological work ofthe coalesced ice peiskogo Severo-Vostoka i ikh litostratigraficheskoye znacheniye). sheets imply an Antarctic type ofglaciation with ice Nauka, St. Petersburg, 125pp. (in Russian). Andreicheva, L.A., Nemtsova, G.M., Sudakova, N.G., 1997. Middle thickness up to 3.5 km. Unlike Antarctic and most Pleistocene Tills ofthe North and Centre ofthe Russian Plain ofNorth Atlantic ice sheets, North Russian (Sredniepleistotsenovye moreny severa i tsentra Russkoi ravniny). continental glaciers acted on predominantly soft Uralian Branch ofRussian Academy ofSciences, Yekaterinburg, and perennially frozen substrate, which was deeply 83pp. (in Russian). affected by pervading glaciotectonism. Andreyeva, S.M., 1978. The Zyryanka glaciation in northern Middle Siberia. Izvestiya Akademii Nauk SSSR, Geography 5, 72–79 (in (4) The thickest tills are found beneath interglacial Russian). formations similar to the Holsteinian, suggesting Andreyeva, S.M., Isayeva, L.L., 1974. Interrelations ofice sheets of that the most extensive ice sheets are pre-Saalian Putorana and Anabar disperal centres at Samarovo and Taz ice not only in European Russia, but probably also in ages. Bulleten Komissii po izucheniyu chetvertichnogo perioda, Siberia. Moscow 41, 69–74 (in Russian). Arkhipov, S.A., 1971. The Quaternary Period ofWestern SIBERIA (5) During the maximum (Cromerian) glaciation and (Chetvertichny period v Zapadnoi Sibiri). Nauka, Novosibirsk, subsequent ice ages (OIS 16 to 10) the influence of 331pp. (in Russian). the Fennoscandian ice dome was limited in north- Arkhipov, S.A., (Ed.), 1975. The Tobol Horizon ofthe Siberian ern European Russia. The Fennoscandian ice sheet, Pleistocene (Tobolsky gorizont sibirskogo pleistotsena). Nauka, however, culminated during the penultimate glacia- Novosibirsk, 96pp. (in Russian). Arkhipov, S.A., 1989. A chronostratigraphic scale ofthe glacial tion (OIS 6), when shelfsources ofinland ice were Pleistocene ofthe West Siberian North. In: Skabichevskaya, N.A. less powerful than in pre-Holsteinian times. (Ed.), Pleistotsen Sibiri. Stratigrafia i mezhregionalnye korrelatsii. (6) The drift limit in Northern Russia is time-trans- Nauka, Novosibirsk (in Russian). gressive, being certainly pre-Holsteinian in north- Arkhipov, S.A., Matveyeva, O.V., 1964. The Anthropogene ofthe eastern Europe, probably OIS 8 in West Siberia and Southern Yenissei Depression (Antropogen yuzhnoi okrainy Yeniseiskoi depressii). Institute ofGeology and Geophysics, getting older east ofthe Yenissei. Siberian Branch ofAcademy ofSciences ofthe USSR, Novosi- (7) There are several unsolved questions concerning birsk, 128pp. (in Russian). correlation ofthe marine transgressions with Arkhipov, S.A., Andreyeva, S.M., Zemtsov, A.A., Isayeva, L.L., terrestrial interglacial events. The reliability ofthe Mizerov, B.V., Fainer, Yu.B., 1976. Terrestrial ice sheets and the correlation ofRussian pre-Eemian interglacials topography. In: Timofeyev, D.A. (Ed.), Problemy ekzogennogo reliefoobrazovaniya, Vol. 1. Nauka, Moscow, pp. 7–89 (in with their western European counterparts notice- Russian). ably decreases northwards and eastwards partly Arkhipov, S.A., Votakh, M.R., Golbert, A.V., Gudina, V.I., Dovgal, because ofthe insufficientdata available, and also L.A., Yudkevich, A.I., 1977. The Last Glaciation in the Lower Ob due to the fading biotic signal of the climatic River Region. (Posledneye oledeneniye Nizhniego Priobya). fluctuations. Nauka, Novosibirsk, 215pp. (in Russian). Arkhipov, S.A., Panychev, V.A., Shelekhova, T.G., Shelkoplyas, V.N., 1978. Glacial Geology ofthe Belogorsk Upland, the West Siberian Plain, the Lower Ob region. Siberian Branch ofAcademy of Acknowledgements Sciences ofthe USSR, Novosibirsk, 132pp. Arkhipov, S.A., Isayeva, L.L., Bespaly, V.G., Glushkova, O., 1986. This work is part ofthe current Russian–Norwegian Glaciation ofSiberia and North-East USSR. Quaternary Science Reviews 5 (1–4), 463–474. PECHORA project supported mostly by the Norwegian Arkhipov, S.A., Levchuk, L.K., Shelkoplyas, V.N., 1994. Stratigraphy Research Council. It is also a contribution to the and geological structure ofthe Quaternary in the Lower Ob- European Science Foundation program QUEEN. OSL Yamal-Taz region ofWest Siberia. Geologia i geofizika 6, 87–104 dating was performed by A. Murray at the Nordic (in Russian). Laboratory for Luminescence Dating, Ris^, Denmark. Arslanov, Kh.A., Lavrov, A.S., Potapenko, L.M., Tertychnaya, T.V., Chernov, S.B., 1987. New data on geochronology and paleogeo- Radiocarbon and U/Th dates have been obtained in the graphy ofthe Late Pleistocene and Early Holocene in the northern Geochronological Laboratory ofSt. Petersburg Uni- Pechora Lowland. In: Punning, J.-M.K, Ivanova, I.K., Kind, N.V., versity, Russia, under the guidance ofKh. Arslanov. Chichagova, O.A. (Eds.), Novye dannye po geokhronologii ARTICLE IN PRESS V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311 1309

chetvertichnogo perioda. Nauka, Moscow, pp. 101–111 (in pleistotsena Severo-Vostoka yevropeiskoi chasti Rossii). Nauka, Russian). St. Petersburg, 125pp. (in Russian). Astakhov, V.I., 1974a. A unique monument ofcontinental glaciation Fainer, Yu.B., Borisov, V.A., Gayntsev, F.M., 1976. Glacial sediments in the Urals. Doklady Academii Nauk SSSR 219 (3), 683–685 (in on inselbergs ofthe Central Siberian Plateau (as exemplified Russian). by Bolshaya Tundrovaya mountain). Bulleten Komissii Astakhov, V.I., 1974b. The ice-margin features and some problems of po izucheniyu chetvertichnogo perioda, Moscow 46, 1–21 (in the Pleistocene paleogeography ofthe Upper Pechora basin. Russian). Bulleten Komissii po izucheniyu chetvertichnogo perioda, Acad. Ganeshin, G.S. (Ed.), 1973. Map ofQuaternary Deposits ofthe USSR, Sci. USSR 41, 56–62 (in Russian). Scale 1:2500000. VSEGEI, Leningrad, 16 sheets. Astakhov, V.I., 1976. Geological evidence ofa centre ofthe Pleistocene Goretsky, G.I., Chebotareva, N.S., Shick, S.M. (Eds.), 1982. Moscow glaciation on the Kara shelf. Doklady Akademii Nauk SSSR 231 Ice Sheet ofEast Europe (Moskovsky lednikovy pokrov Vostoch- (5), 1178–1181 (in Russian). noi Yevropy), Nauka, Moscow, 235pp. (in Russian, English Astakhov, V.I., 1977. The Kara centre ofthe Pleistocene glaciation as summary). reconstructed from ancient moraines of West Siberia. In: Materialy Gornostay, B.A., 1990. The Peri-Timan glacial disturbances. Bulleten glatsiologicheskikh issledovaniy, Vol. 30. Academy ofSciences of Komissii po izucheniyu chetvertichnogo perioda, Moscow 59, the USSR, Moscow, pp. 60–69 (in Russian, English summary). 152–155 (in Russian). Astakhov, V.I., 1979. New data on the latest activity ofKara-shelf Gudina, V.I., 1976. Foraminifers, stratigraphy and palaeozoogeogra- glaciers in West Siberia. In: Sibrava,$ V. (Ed.), IGCP Project phy ofthe marine Pleistocene in the northern USSR (Foraminifery, 73\1\24 ‘Quaternary glaciations in the Northern Hemisphere’ stratigrafia i paleozoogeografia morskogo pleistotsena Severa Prague, Vol. 5, pp. 21–31. SSSR). Nauka, Novosibirsk, 1976, 126pp. (in Russian). Astakhov, V.I., 1991. The fluvial history ofWest Siberia. In: Starkel, Guslitser, B.I., 1973. On the origin ofthe boulder clay in northern Cis- L., Gregory, K.J., Thornes, J.B. (Eds.), Temperate Paleohydrol- Uralia. In: Trudy, Vol. 16. Institute ofGeology, Academy of ogy. Wiley, London, pp. 381–392. Sciences ofthe USSR, Komi Branch, Syktyvkar, pp. 3–19 (in Astakhov, V., 1997. Late glacial events in the central Russian Arctic. Russian). Quaternary International 41/42, 17–25. Guslitser, B.I., Isaychev, K.I., 1983. The age ofthe Rogovaya Astakhov, V., 2001. Stratigraphic framework for the Upper Pleisto- formation of the Timan-Pechora region as determined by fossil cene ofthe Russian Arctic: changing paradigms. Global and pied lemmings. Bulleten Komissii po izucheniyu chetvertichnogo Planetary Change 31/1-4, 281–293. perioda, Moscow 52, 58–72 (in Russian). Astakhov, V., 2003. Pleistocene ice limits in Russian northern Guslitser, B.I., Loseva, E.I., Lavrov, A.S., Stepanov, A.N., 1986. lowlands. In: Ehlers, J., Gibbard Ph. (Eds.), Extent and Chronol- Timan–Pechora–Vychegda region. In: Krasnov, I.I., Zarrina, Ye.P. ogy ofGlaciations. Europe, Vol. 1. Elsevier, Amsterdam (in press). (Eds.), Resheniye 2 Mezhvedomstvennogo stratigraficheskogo Astakhov, V.I., Isayeva, L.L., 1988. The ‘Ice Hill’: an example of soveshchania po chetvertichnoi sisteme Vostochno-Yevropeiskoi retarded deglaciation in Siberia. Quaternary Science Reviews 7, platformy. VSEGEI, Leningrad (in Russian). 29–40. Isayeva, L.L., 1963. Traces ofQuaternary glaciation in the north- Astakhov, V.I., Isayeva, L.L., Kind, N.V., Komarov, V.V., 1986. On western part ofthe Central Siberian Upland. Izvestiya Akademii geological and geomorphological criteria ofsubdivision ofglacial Nauk SSSR, geology 2, 90–98 (in Russian). history in the Yenissei North. In: Velichko, A.A., Isayeva, L.L. Isayeva, L.L., Kind, N.V., Laukhin, S.A., Kolpakov, V.V., Shofman, (Eds.), Chetvertichnye oledeneniya Srednei Sibiri. Nauka, Mos- I.L., Fainer, Yu.B., 1986. The stratigraphic scheme ofthe cow, pp. 18–28 (in Russian). Quaternary ofCentral Siberia. In: Velichko, A.A., Isayeva, L.L. Astakhov, V.I., Kaplyanskaya, F.A., Tarnogradsky, V.D, 1996. (Eds.), Chetvertichnye oledeneniya Srednei Sibiri. Nauka, Mos- Pleistocene permafrost of West Siberia as a deformable glacier cow, pp. 4–17 (in Russian). bed. Permafrost and Periglacial Processes 7, 165–191. Kaplyanskaya, F.A., Tarnogradsky, V.D., 1974. The Middle and Astakhov, V.I., Svendsen, J.I., Matiouchkov, A., Mangerud, J., Lower Pleistocene ofthe Lower Irtysh area (Sredny i nizhny Maslenikova, O., Tveranger, J., 1999. Marginal formations of the pleistotsen nizovyev Irtysha). Nedra, Leningrad, 160pp. (in last Kara and Barents ice sheets in northern European Russia. Russian). Boreas 28 (1), 23–45. Kaplyanskaya, F.A., Tarnogradsky, V.D., 1975. Origin ofthe Ber, A.G., 1948. On the direction ofice flow in the Peri-Polar Urals Sanchugovka Formation and the problem ofinterrelation of during the maximum glaciation. In: Edelstein, Ya.S., Gerasimov, glaciations and marine transgressions in the north ofWest Siberia. I.P. (Eds.), Materialy po geomorfologii Urala, Vol. 1. Publishing In: Zubakov, V.A. (Ed.), Kolebaniya urovnya Mirovogo Okeana v House ofMinistry ofGeology ofthe USSR, Moscow–Leningrad, pleistotsene. Geographical Society ofthe USSR, Leningrad, pp. pp. 324–327 (in Russian). 53–95 (in Russian). Biryukov, I.P., Agadzhanyan, A.K., Valuyeva, M.N., Velichkevich, Karpukhin, S.S., Lavrov, A.S., 1974. Flow directions and location of F.Yu., Shick, S.M., 1992. The Quaternary deposits ofthe Roslavl the contact ofthe Dnieper ice sheets on the Russian Plain. Doklady stratotype area. In: Velichko, A.A., Shick, S.M. (Eds.), Stratigrafia Akademii Nauk SSSR 216 (1), 158–161 (in Russian). i paleogeografia chetvertichnogo perioda Vostochnoi Yevropy. Kind, N.V., Leonov, B.N. (Eds.), 1982. The Anthropogene ofTaimyr Institute ofGeography, Russian Academy ofSciences, Moscow, (Antropogen Taimyra). Nauka, Moscow, 184pp. (in Russian). pp. 152–180 (in Russian). Krasnov, I.I. (Ed.), 1971. Map ofthe Quaternary Deposits ofthe Biske, G.S., Devyatova, E.I., 1965. Pleistocene transgressions in European USSR and Adjacent Regions, Scale 1:1 500 000. northern Europe. Trudy NIIGA, Leningrad 143, 155–176 (in VSEGEI, Leningrad, 16 sheets. Russian, English summary). Kuznetsova, L.A., 1971. Pleistocene ofthe Pechora Cis-Uralia Chernov, G.A., 1974. On the Quaternary and geomorphology of (Pleistotsen Pechorskogo Priuralya). Kazan University, 123pp. (in Vangyr district ofthe Peri-Polar Urals. Bulleten Komissii po Russian). izucheniyu chetvertichnogo perioda, Moscow 42, 66–81 (in Lavrov, A.S., Nikiforova, L.D., Potapenko, L.M., 1986. The dynamics Russian). ofthe Pleistocene ice sheets, vegetation and climate in the northeast Duryagina, D.A., Konovalenko, L.A., 1993. Palynology ofthe ofthe European USSR. In: Sidnev, A.V., Nemkova, V.K. (Eds.), Pleistocene ofthe Northeastern European Russia (Palinologia Novye materialy po paleogeografii i stratigrafii pleistotsena. ARTICLE IN PRESS 1310 V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311

Academy ofSciences ofthe USSR, Bashkir Branch, Ufa,pp. 69–78 stratigrarfii i korrelatsii pliotsenovykh i pleistotsenovykh otlozhe- (in Russian). niy. Academy ofSciences ofthe USSR, Bashkir Branch, Ufa,pp. Lavrushin, Yu.A., Chistyakova, I.A., Gaidamanchuk, A.S., Golubev, 62–85 (in Russian). Yu.K., Vasilyev, V.P., 1989. The structure and composition ofthe Sudakova, N.G., Faustova, M.A., 1995. Glacial history ofthe Russian glacial palaeoshelfsediments in Bolshezemelskaya Tundra. In: Plain. In: Ehlers, J., Kozarski, S., Gibbard, Ph. (Eds.), Glacial Lavrushin, Yu.A. (Ed.), Litologiya kainozoiskikh otlozheniy. Deposits in North-East Europe. Balkema, pp. 151–156. Geological Institute, Academy ofSciences ofthe USSR, Moscow, Sukhorukova, S.S., 1999. Late Pleistocene palaeogeography ofNorth- pp. 3–51 (in Russian). west Siberia. Journal ofGeological Sciences. Anthropozoic, Czech Lazukov, G.I., 1970. The Anthropogene ofthe Northern HalfofWest Geological Survey, Prague 23, 37–42. Siberia, Stratigraphy (Antropogen severnoi poloviny Zapadnoi Sukhorukova, S.S., Kostyuk, M.A., Podsosova, L.L., Babushkin, Sibiri, stratigrafiya). Moscow University, 322pp. (in Russian). A.Ye., Zolnikov, I.D., Abramov, S.A., Goncharov, S.V., 1987. Levina, T.P., 1964. Quaternary pollen spectra from the proglacial zone Tills and Dynamics ofGlaciations ofWest Siberia (Moreny i ofthe Samarovo ice sheet (the Yenissei catchment). In: Sachs, dinamika oledeneniy Zapadnoi Sibiri). Nauka, Novosibirsk, V.N., Khlonova, A.F. (Eds.), Sistematika i metody izucheniya 159pp. (in Russian). iskopayemykh pyltsy i spor. Nauka, Moscow, pp. 208–217 (in Svendsen, J.I., Alexanderson, H., Astakhov, V.I., Demidov, I., Russian). Dowdeswell, J.A., Funder, S., Gataullin, V., Henriksen, M., Hjort, Loseva, E.I., Duryagina, D.A., 1973. Results ofintegrated study ofthe C., Houmark-Nielsen, M., Hubberten, H.W., Ingolfsson,! O.,! key section ofPleistocene sediments on the Middle Pechora close to Jakobsson, M., Kjær, K.H., Larsen, E., Lokrantz, H., Lunkka, village Rodionovo. In: Trudy, Vol. 16. Institute ofGeology, J.P., Lysa,( A., Mangerud, J., Matioushkov, A., Murray, A., Academy ofSciences ofthe USSR, Komi Branch, Syktyvkar, pp. Moller,. P., Niessen, F., Nikolskaya, O., Polyak, L., Saarnisto, M., 20–34 (in Russian). Siegert, C., Siegert, M.J., Spielhagen, R.F., Stein, R., 2004. Late Loseva, E.I., Duryagina, D.A., Andreicheva, L.N., 1992. The Middle Quaternary ice sheet history ofNorthern Eurasia. Quaternary Pleistocene ofcentral Bolshezemelskaya Tundra. In: Trudy, Vol. Science Reviews, this issue (doi:10.1016/j.quascirev.2003.12.008). 75. Institute ofGeology, Russian Academy ofSciences, Komi Troitsky, S.L., 1975. The Modern Antiglacialism (Sovremenny Branch, Syktyvkar, pp. 113–123 (in Russian). antiglatsialism): A Critical Essay. Nauka, Moscow, 163pp. (in Matveyeva, G.V., 1967. Results ofstudy ofglacial pebbles in Middle Russian). Timan. Trudy VSEGEI, New Series, Nedra, Moscow 145, 292–301 Urvantsev, N.N., 1931. Quaternary glaciation ofTaimyr. Bulleten (in Russian). Komissii po izucheniyu chetvertichnogo perioda, Moscow 3, 23–42 Obruchev, V.A., 1931. Traces ofthe Ice Age in northern and central (in Russian). Asia. Bulleten Komissii po izucheniyu chetvertichnogo perioda Varsanofieva, V.A., 1933. On the traces ofglaciation in the Northern Leningrad 3, 43–120. Urals. Trudy Komissii po izucheniyu chetvertichnogo perioda, Potapenko, L.M., 1974. New data on ice limits ofthe Moscow age in Leningrad 3, 81–105 (in Russian). the Upper Vychegda catchment area. Bulleten Komissii po Velichko, A.A., Faustova, M.A., 1986. Glaciations in the East izucheniyu chetvertichnogo perioda, Moscow 41, 63–68 (in European Region ofthe USSR. Quaternary Science Reviews 5, Russian). 447–461. Ramsay, W., 1904. Beitrage. zur Geologie der recenten und pleistoc- Velichko, A.A., Shick, S.M. (Eds.), 2001. Middle Pleistocene glacia- anen. Bildungen der Halbinsel Kanin. Fennia 21 (7), 70. tions in Eastern Europe (Oledeneniya sredniego pleistotsena Ryzhov, B.V., 1974. On origin ofboulder clays ofthe Severnaya Sosva Vostochnoi Yevropy). GEOS, Moscow, 160pp. (in Russian). catchment area, the Urals. Litologia i poleznye iskopayemye 1, Velichko, A.A., Udartsev V, .P., Markova A, .K., Morozova T, .D., 145–151 (in Russian). Pevzner M, .Ya., Gribchenko, Yu.N., Sychova, S.A., 1977. New Sachs, V.N., 1953. The Quaternary Period in the Soviet Arctic ideas on the age ofthe Dnieper and Don Lobes ofice sheets ofthe (Chetvertichny period v Sovietskoi Arktike). Leningrad–Moscow, Russian Plain. Izvestiya Akademii Nauk SSSR, Geography 6, 627pp. (in Russian). 25–36 (in Russian). Savelyev, A.A., 1966. The planation surface of the Polar Urals and Volkova, V.S., Babushkin, A.Ye., 2000. The Unified Regional some questions ofthe glacial history. In: Popov, A.I., Yenokian, Stratigraphic Scheme ofthe Quaternary ofthe West Siberian Plain V.S. (Eds.), Geologiya kainozoya Severa Yevropeiskoi chasti (Unifitsirovannaya regionalnaya stratigraficheskaya skhema chet- SSSR. Moscow University, pp. 73–84 (in Russian). vertichnykh otlozheniy Zapadno-Sibirskoi ravniny). Explanatory Shatsky, S.B., 1965. The glacially detached blocks in the Quaternary at note. SNIIGGiMS, Novosibirsk, 64pp. (in Russian). Yurty Yeutskie on river B. Yugan and in the vicinity ofKhanty- Voronov, P.S., 1951. New data on the glaciation and Quaternary Mansiysk. In: Sachs, V.N. (Ed.), Osnovnye problemy izucheniya sediments ofthe central Pai-Hoi. In: Sbornik statei po geologii chetvertichnogo perioda. Nauka, Moscow, pp. 206–217 (in Arktiki, Vol. 2. Institute ofGeology ofthe Arctic, Leningrad, Russian). pp. 84–92 (in Russian). Shick, S.M., 1989. Problems oflong distance correlation ofQuaternary Voronov, P.S., 1964. To the method ofpalaeo- and mellogeographic deposits in view ofnew data on the Pleistocene stratigraphy ofthe reconstruction ofgeometry ofcontinents and ice sheets. In: European USSR. In: Trudy, Vol. 657. Institute ofGeology and Izvestiya, Vol. 5. Geographical Society ofthe USSR, pp. 370–382 Geophysics, Academy ofSciences ofthe USSR, Siberian Branch, (in Russian). Nauka, Novosibirsk, pp. 118–126 (in Russian). Yakhimovich, V.L., Nemkova, V.K., Semyonov, I.N., 1973. The Sirin, N.A., 1947. On traces oftwo glaciations in the Peri-Polar Urals. Stratigraphy ofPliocene–Pleistocene deposits ofthe Timan-Uralian Bulleten Komissii po izucheniyu chetvertichnogo perioda, Moscow region and their correlation along Cis-Uralia (Stratigrafiya 10, 155–170 (in Russian). pliotsen-pleistotsenovykh otlozheniy Timano-Uralskoi oblasti I Stepanov, A.N., 1974. Stratigraphy and sedimentary environments of ikh korrelatsiya po Preduralyu). Nauka, Ioscow," 100pp. (in the Upper Cenozoic ofthe Pechora–Kama interfluve. Resume of Russian). Ph.D. thesis, Geological Faculty, Moscow University, 34pp. (in Yakovlev, S.A., 1956. The Fundamentals ofthe Quaternary Russian). Geology ofthe Russian Plain (Osnovy geologii chetvertichnykh Stepanov, A.N., 1976. Pliocene (?)–Pleistocene deposits ofthe otlozheniy Russkoi ravniny). Gosgeoltekhizdat, Moscow, 314pp. Pechora–Kolva interfluve. In: Yakhimovich, V.L. (Ed.), Voprosy (in Russian). ARTICLE IN PRESS V. Astakhov / Quaternary Science Reviews 23 (2004) 1285–1311 1311

Zakharov, Yu.F., 1968. Exotectonic disturbances in the sedimentary Zemtsov, A.A., 1973b. Mineralogical composition ofthe Quaternary cover ofWest Siberia. Geologiya i geofizika, Novosibirsk 6, sediments and palaeogeographical questions in the north ofWest 148–155 (in Russian). Siberia. Izvestiya Vysshikh uchebnykh zavedeniy, geology and Zarkhidze, V.S., 1972. The Padimei formation of the western and exploration, Moscow 6, 49–55 (in Russian). central Timan-Uralian Region. In: Yakhimovich, V.L. (Ed.), Zolnikov, I.D., 1990. The Kormuzhikhantka formation as evidence of Voprosy stratigrafii i korrelatsii pliotsenovykh i pleistotsenovykh the Ob impounding in the Late Pleistocene. In: Trudy, Vol. 759. otlozheniy severnoi i yuzhnoi chastei Preduralya, Vol. 1. Academy Institute ofGeology and Geophysics, Academy ofSciences ofthe ofSciences ofthe USSR, Bashkir Branch, Ufa,pp. 56–66 (in USSR, Siberian Branch, Novosibirsk, pp. 52–57 (in Russian). Russian). Zubakov, V.A., 1972. Recent Sediments ofthe West Siberian Lowland Zarrina, Ye.P., Krasnov, I.I., Tarnogradsky, V.D., 1961. Map ofthe (Noveishie otlozheniya Zapadno-Sibirskoi nizmennosti). Nedra, Quaternary Deposits ofthe West Siberian Lowland, Scale 1:1 500 Leningrad, 312pp. (in Russian). 000. Gostoptekhizdat, Moscow, 6 sheets. Zubakov, V.A., 1992. The Caspian transgressions as an indicator of Zemtsov, A.A., 1973a. Petrographic composition ofpebbles and the interglacial–glacial transition. In: Kukla, G.J., Went, E. (Eds.), palaeogeographical questions in the north ofWest Siberia. Start ofa Glacial. NATO ASI Series, Vol. 13, Springer, Berlin– Izvestiya Akademii Nauk SSSR, geography 2, 80–90 (in Russian). Heidelberg, pp. 253–271.