C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:17

1 The last Scandinavian Ice Sheet in northwestern : ice flow patterns

2 and decay dynamics

3 IGOR N. DEMIDOV, MICHAEL HOUMARK-NIELSEN, KURT H. KJÆR AND EILIV LARSEN

4 Demidov, I. N., Houmark-Nielsen, M., Kjær, K. H. & Larsen, E. 2006 (August): The last Scandinavian Ice Sheet 5 BOREAS in northwestern Russia: ice flow patterns and decay dynamics. Boreas, Vol. 35, pp. xxxÁxxx. Oslo. ISSN 0300- 6 9483.

7 Advance of the Late Weichselian (Valdaian) Scandinavian Ice Sheet (SIS) in northwestern Russia took place after 8 a period of periglacial conditions. Till of the last SIS, Bobrovo till, overlies glacial deposits from the previous 9 Barents and Kara Sea ice sheets and marine deposits of the Last Interglacial. The till is identified by its contents of 10 Scandinavian erratics and it has directional properties of westerly provenance. Above the deglaciation sediments, 11 and extra marginally, it is replaced by glaciofluvial and glaciolacustrine deposits. At its maximum extent, the last 12 SIS was more restricted in Russia than previously outlined and the time of termination at 18Á16 cal. kyr BP was 13 almost 10 kyr delayed compared to the southwestern part of the ice sheet. We argue that the lithology of the ice 14 sheets’ substrate, and especially the location of former proglacial lake basins, influenced the dynamics of the ice 15 sheet and guided the direction of flow. We advocate that, while reaching the maximum extent, lobe-shaped glaciers 16 protruded eastward from SIS and moved along the path of water-filled lowland basins. Ice-sheet collapse and 17 deglaciation in the region commenced when ice lobes were detached from the main ice sheet. During the 18 Lateglacial warming, disintegration and melting took place in a 200Á600 km wide zone along the northeastern 19 rim of SIS associated with thick Quaternary accumulations. Deglaciation occurred through aerial downwasting 20 within large fields of dead ice developed during successively detached ice lobes. Deglaciation led to the 21 development of hummocky moraine landscapes with scattered periglacial and ice-dammed lakes, while a sub- 22 arctic flora invaded the region. 23 Igor N. Demidov (e-mail: [email protected]), Russian Academy of Sciences, Karelian Research Centre, 24 Institute of Geology, Pushkinskaya Street 11, Petrozavodsk, Russia 185910; Michael Houmark-Nielsen, Institute of 25 Geology, University of Copenhagen, Øster Voldgade 10, DK-1350, Copenhagen K, Denmark; Kurt H. Kjær, 26 University of Copenhagen, Øster Voldgade 5Á7, DK-1350, Copenhagen K, Denmark; Eiliv Larsen, Geological 27 Survey of Norway, P.O. Box 3006, N-7002, Trondheim, Norway; received 6th December 2005, accepted 27th March 2006.

28 29 The eastern flank of the Scandinavian Ice Sheet (SIS) mode of advance and decay in time and space of 30 occupied large parts of the Arkhangelsk district of the the northeastern sector of the Scandinavian Ice 31 Russian European North during the Late Weichselian Sheet have until recently been poorly documented 32 (Valdaian). Although traces left by the SIS have been (Demidov et al. 2004). 33 recognized since the turn of the last century (Ramsay In this article, a revised version of Late Weichselian 34 1911), the ice flow dynamics and distribution of ice sheet flow dynamics, deglaciation pattern and the 35 landforms and sediments generated near the eastern development of landforms is presented from the time 36 limit of the maximum extent of SIS in northern of maximum extent of the Scandinavian Ice Sheet in 37 Russia are still subjects of debate (Svendsen et al. northwestern Russia until the beginning of the Holo- 38 2004). The duration and age of the glacial maximum cene. This was achieved through a critical analysis of 39 are significantly different compared with the ice previous investigations of the Quaternary stratigraphy 40 sheet’s termination nearer to the Atlantic continental and geomorphology supported by new field observa- 41 margin. It therefore appears that the peak glaciation tions and new age constraints. Data from the northern 42 in Russia was delayed by almost 10 kyr. In southwest Arkhangelsk Region (Fig. 1), including the Kanin 43 Scandinavia this state was reached some time between Peninsula, the seashore of the Kuloi Plateau and bluffs 44 32 and 25 kyr BP (Sejrup et al. 1994, 2003; Houmark- along the River, are synthesized with previously 45 Nielsen & Kjær 2003), when SIS eventually merged published data from the Severnaya Dvina and 46 with the British Isles Ice Sheet. On the northern rivers (Larsen et al. 1999; Lysa˚ et al. 2001; Kjær et al. 47 boundary, SIS coalesced with the Barents Sea Ice 2003) (Fig. 1). From 1996 to 2002, more than 100 key 48 Sheet around 22 kyr BP (Svendsen et al. 2004, and sections with Late Pleistocene deposits were investi- 49 references therein). The exact location of the merging gated by stratigraphical, sedimentological and palaeon- 50 is less evident, largely due to a lack of mappable tological methods along well-exposed cliff sites on the 51 landforms and the restricted stratigraphical control of seashore and river banks in the Arkhangelsk Region. 52 offshore sediments.UNCORRECTED Adding to these uncertainties, the By addressing these PROOF issues we close a gap in the

DOI 10.1080/03009480600781883 # 2006 Taylor & Francis C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:20

2 Igor N. Demidov et al. BOREAS 35 (2006)

Fig. 1. The topography of the Arkhangelsk Region, northwest Russia. Map shows alternative configurations of the maximum extent of the last Scandinavian Ice Sheet. Numbers 1Á48 refer to investigated sections and other localities in the region. 1/Tarkhanov; 2/Krynka; 3/Madakha; 4/Ostraya; 5/Pestsovaya; 6/Konushin; 7/Konushinskaya Korga; 8/Morzhovets Island; 9/Cape Kargovsky; 10/Cape Abramovsky; 11/Koida; 12/Megra; 13/Cape Tolstic; 14 and 15/Syomzha 1&2; 16/Zaton; 17/Bychie; 18/Kulogora; 19/Ezhuga; 20/Karpogory; 21/Shilega; 22/Verkola; 23/Olema; 24/Chelmokhta; 25/Lipovik; 26/Tomasha; 27/Erga; 28/Boltinskaya; 29/Yumizh; 30/Raibola; 31/Smotrakovka; 32/Osinovskaya; 33/Ust-Padenga; 34/Koleshka; 35/Pasva; 36/Nyandoma; 37/Mosha; 38/Telza; 39/Bobrovo; 40/Trepuzovo; 41/Psaryovo; 42/Chavanga; 43/Strelnya; 44/Kumzhevaya; 45/Babia; 46/Kachkovka; 47/Iokanga; 48/Oiva; 49/Tova. Compiled from the present study, Devyatova (1969, 1982), Armand (1969) (Kola Peninsula), Larsen et al. (1999), Atlasov et al. (1978), Arslanov et al. (1984) (Dvina lowland) and Lunkka et al. (2001).

53 knowledge about the Late Weichselian glaciation elevations from 10 to 80 m a.s.l. Deep troughs in the history along the fringes of the entire last SIS. lying more than 50 m below the present sea 54 level separate the Palaeozoic sedimentary platform 55 from the Precambrian crystalline basement of the 56 Setting Fennoscandian Shield. Apart from more than 100-m- 57 deep incised valleys, the Quaternary cover is 5Á60 m 58 The study area is located on the southern and eastern thick, except for uplifted regions where strongly 59 shores of the White Sea, along the Kanin Peninsula weathered bedrock and boulder fields are present 60 and the Kuloi Plateau, and it covers the middle part of (Fig. 2). Glacigenic deposits comprising till, fluvial 61 the Pinega River and the lower part of the Mezen River and lacustrine sediments of the Late Weichselian 62 (Figs 1, 2). According to Geology of USSR (1963) Scandinavian glaciation dominate subsurface expo- 63 rugged denudation plain composed of Palaeozoic sures in the western part of the discussed region, 64 sediments is dissected by the 100Á270 m high water whereas the areas to the east were glaciated by the 65 divides of the Kuloi Plateau and the Pokshenga High- Barents and the Kara ice sheets in Early and Middle 66 land, whereas the Kanin Ridge is composed of Late Weichselian time (Kjær et al. 2003; Larsen et al. 67 Precambrian metasediments and small amounts of 2006a). 68 igneous rocks (Figs 1, 2). The Severnaya Dvina, A comprehensive review of previous investigations 69 Pinega, Mezen and Kuloi rivers and the root of the in northwestern Russia has been provided by Demidov 70 KaninUNCORRECTED Peninsula occupy extensive depressions, with et al. (2004). PROOF Ramsay (1911) recognized three till units C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:22

BOREAS 35 (2006) Last Scandinavian Ice Sheet, NW Russia 3

Fig. 2. The main Quaternary geomorphological features of the Arkhangelsk region and the position of the largest extent of the Scandinavian Ice Sheet in the Late Weichselian. Compiled from Krasnov (1971), Ganeshin et al. (1980) and Lavrov (1991).

71 separated by shell-bearing sediments, i.e. on the Kanin Peninsula led Devyatova (1969), Aseev (1974) and 72 Peninsula, at the mouth of Mezen River and on the Demidov et al. (2004) to support the contention of 73 Kuloi Plateau seashore. Using the lithology of erratic Ramsay (1911) that the SIS reached the central axis of 74 boulders, Ramsay proposed that Scandinavia and the Kanin Peninsula during the Last Glacial Maximum 75 Novaya Zemlya were centres of glaciation. Two or (LGM) (Fig. 1). From the shores of Mezen Bay, the 76 three till beds have since been recognized in the Kanin boundary of the Late Weichselian SIS has been drawn 77 Peninsula area (Kalyanov & Androsova 1933; Lyutke- southward towards Severnaya Dvina via the central 78 vich 1947; Spiridonov & Jakovleva 1961). Spiridonov & part of the Kuloi Plateau (Krasnov 1971; Legkova & 79 Jakovleva (1961) also supported the idea that the upper Schukin 1972; Ganeshin et al. 1980), although Lavrov 80 till was of Scandinavian origin. This is confirmed by (1991) slightly modified these reconstructions using 81 Kjær et al. (2006) and by Larsen et al. (2006b). The aerial photographic interpretation. Edemsky (1931), 82 presence of Scandinavian boulders, young eskers and Devyatova (1982) and Filippov & Borodai (1987) 83 sandy kamesUNCORRECTED in the western parts of the Kanin suggested that the PROOF middle parts of the Pinega and C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:23

4 Igor N. Demidov et al. BOREAS 35 (2006)

84 Mezen rivers remained unglaciated after the Eemian erratics and other directional properties (Houmark- 85 (Mikulinian) interglacial, while others argued that SIS Nielsen et al. 2001; Kjær et al. 2001, 2003, 2006; Larsen or the Kara Sea Ice Sheet reached the area either et al. 2006a, b). Till of successive advances of SIS during the Early Weichselian (Krasnov 1971) or at the contain erratic clasts from the Fennoscandian Shield 86 end of the Weichselian (Aseev 1974; Lavrov 1991; such as Precambrian granites, gneisses, basic rocks and Lavrov & Potapenko 2005). quartzite from the KareliaÁKola province; in particu- 87 lar, Kola Peninsula nepheline syenites are found 88 dispersed in the northern part of the Arkhangelsk 89 Methodology district. Reddish, Late Proterozoic sandstone from the 90 southern seashore of the Kola Peninsula also occurs in 91 Our geomorphological interpretation is based on till deposited by the Scandinavian Ice Sheet, and in the 92 analyses of topographic and geological maps of scale fine-gravel fraction such till is almost devoid of lime- 93 1:100 000 and 1:200 000 and satellite images of scale stone and dominated by crystalline rock fragments 94 1:500 000 and 1:1000 000 (Zatulskaya 1984a, b, c). Field followed by Lower Palaeozoic sandstone, mudstone observations in selected areas served as ground control. and shale. 95 In Fig. 2 we have distinguished between end moraines, Luminescence dating was carried out at the Nordic 96 hummocky moraine, kames, sandar, glacial lake basins, Laboratory for Luminescence Dating, Risø, Denmark 97 bedrock or weathered rock, bedrock cliff, valley trains and the results are listed in Table 1. Doses were 98 and plateaus. measured using quartz and a Single Aliquot Regener- 99 River and coastal sections were logged and facies ated dose protocol (Murray & Wintle 2000). Dose rate 100 associations were interpreted in terms of depositional analysis is based on high resolution gamma spectro- 101 environment. Diamictons were described, classified metry (Murray et al. 1987). Calculations used a 102 and interpreted following common guidelines (cf. saturation water content of 25Á30%. These are typical 103 Kru¨ger & Kjær 1999; Kjær et al. 2001). Directional for such sandy sediments, but two samples had 104 properties of till units include striations on subtill measured water contents of 6Á7% lower than this. 105 bedrock and boulder pavements and, most frequently, Using these values would only reduce the ages by 106 data from clast fabric analyses. Samples of 25 clasts 5Á6%, and since these lower saturated values are 107 were obtained from a subhorizontal surface c. 25/25 considered unlikely, we have chosen to use the higher 108 cm and only clasts with a long-to-intermediate axis range for all samples, with an assigned uncertainty of 109 ratio ]/1.5 and a length between 0.6 and 6 cm were 9/4%. It is assumed that the sediments were saturated 110 measured. Eigenvectors (V ) and eigenvalues (S and 1 1 for more than 95% of the time after burial (Larsen et 111 S ) for each sample were computed using the program 3 al. 1999; Houmark-Nielsen et al. 2001; Kjær et al. 112 SpheriStat, version 2.0 (Pangaea Scientific) based on 2003). We base this assumption on the inference that 113 Mark (1973, 1974). According to Anderson & Ste- the sediments were frozen during most of the Weichse- 114 phensen (1971), clasts are considered to have statistical lian; in periods of thaw they would have remained in 115 preferred orientation if the S and S eigenvalues are 1 3 the saturated groundwater zone until sub-recent ero- 116 ]/0.52 and 5/0.17, respectively. In order to verify the sion by modern rivers lowered the local groundwater 117 significance of eigenvectors and test for bimodal 118 distribution of clasts, contoured diagrams were pre- table. The sample burial depth also affects the age, 119 pared following the approach by Kamb (1959). because of attenuation of cosmic rays by overburden 120 When unit boundaries comprise angular unconfor- sediments. The lifetime-averaged burial depth is, of 121 mities indicating glaciotectonic deformation such as course, not known. Although the present depths 122 shearing, folding and thrusting, a deformational chron- beneath the surface do not necessarily equal the 123 ology including the direction of glacier movement is averaged burial depths over time, we have assumed a 124 added to the stratigraphic record (Houmark-Nielsen & constant burial depth of 3.9 m. Had we used observed 125 Kjær 2003). Fine-gravel counts were carried out in the depths (some of which can be deduced from Figures 3 126 3Á8 mm fraction and split into different provenance- and 7), the ages would only have changed, on average, 127 dependent rock types, which enables discrimination of by 0.3%. This mean change is small compared to the 128 tills from glaciers that originated on the Fennoscandian typical overall age uncertainty of Â/10%, as is the 129 Shield to the west from those originating on the standard deviation of this change. All uncertainties 130 Palaeozoic platform and Mesozoic uplifts to the east resulting from these possible variations in burial depths 131 and northeast (Kjær et al. 2001). areB/55% of the total uncertainty in the ages, and the 132 Relative chronological control was obtained by using ages critical of the dating of the LGM (those between 133 the Eemian marine sediments as a lower marker bed 30 and 12 kyr) are changed byB/1000 years. In 134 (Devyatova 1982; Larsen et al. 1999; Funder et al. summary, the possible uncertainties arising from var- 135 2002) and by distinguishing tills of Scandinavian origin iations in water contents and sample depths are 136 from older Weichselian tills of Barents and Kara considered unlikely to add significantly to the overall 137 SeaUNCORRECTED origin by their content of provenance-dependent uncertainties PROOF given in Table 1. C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:27

BOREAS 35 (2006) Last Scandinavian Ice Sheet, NW Russia 5

Table 1. Luminescence (OSL) and cosmogenic exposure dates (CED) from the eastern part of the White Sea and central Pinega River, northwestern Russia.

No. Fig. 1/ Strat. Pos. Locality Risø code Sample no. Age (kyr) Dated material Dose rate W.C. %

1/Below Tarkhanov river 001020 00421 809/6 Glaciolacustrine 1.78 23 1/Below Tarkhanov river 001021 00423 879/8 Glaciolacustrine 1.89 25 1/Beyond Tarkhanov river 011001 01403 719/4 Glaciolacustrine 1.37 27 1/Below Tarkhanov river 011002 01405 589/3 Glaciolacustrine 1.62 31 1/Beyond Tarkhanov river 011030 01404 639/5 Glaciolacustrine 1.52 32 4/Beyond Ostraya Hill 011014 01508 509/4 Lacustrine sand 2.44 30 5/Beyond Pestsovaya Hill 011006 01500 729/4 Fluvial sand 1.66 33 6/Beyond Cape Konushin 001015 00-410 949/6 Glaciofluvial 1.29 24 7/Beyond Konushinskaya Korga 001011 00-406 669/7 Glaciolacustrine sand 2.26 11 8/Below Morzhovets island 001035 00441 19.19/1.5 Glaciofluvial sand 0.83 20 8/Above Morzhovets island 001036 00442 11.19/0.9 Glaciofluvial sand 1.38 21 9/Above Cape Kargovsky 993828 99422 12.79/1 Fluvial sand 1.81 14 10/Below Cape Abramovsky 001031 00437 20.99/1.7 Fluvial sand 1.19 27 11/Above Koida river 001033 00439 11.49/0.7 Lacustrine sand 1.81 17 11/Below Koida river 001032 00438 319/3 Fluvial sand 1.15 27 11/Below Koida river 001034 00440 19.99/1.4 Fluvial sand 1.27 23 49/Below Tova river 001037 00444 779/8 Marine sand 0.91 28 13/Above Cape Tolstik 993817 99410 10.69/0.7 Fluvial sand 1.50 18 13/Above Cape Tolstik 993818 99411 10.89/0.8 Fluvial sand 1.43 15 15/Above Syomzha 001038 99404 10.99/0.8 Fluvial sand 1.90 18 15/Above Syomzha 001039 99405 12.69/1.0 Fluvial sand 1.72 19 18/Below Kulogora 993832 99426 779/7 Glaciofluvial sand 1.40 18 18/Above Kulogora 993833 99427 15.09/1.6 Aeolian sand 1.21 18 19/Beyond Ezhuga river 001043 99429 26.69/1.9 Glaciolacustrine sand 1.11 18 20/Beyond Karpogory 993838 99432 1069/8 Glaciofluvial sand 0.98 22 20/Beyond Karpogory 993839 99433 1009/7 Glaciolacustrine sand 1.72 12 Cosmogenic exposure dates 1/Beyond Tarkhanov river 00418 51.59/3 Quarts vein 1/Beyond Tarkhanov river 00419 55.09/4 Quarts vein 1/Above Tarkhanov river 00420 18.09/1.6 Quarts vein

138 Radiocarbon dating of plant macro fossils was done for Radiometric Dating, University of Trondheim, 14 139 at the AMS facility at the A˚ ngstro¨m Laboratory, Norway (Larsen et al. 1999). The C age is calibrated 140 University of Uppsala, Sweden and the Laboratory to calendar years after the calculations of Kitagawa &

Fig. 3. StratigraphicUNCORRECTED logs from selected sections along the northern shores of the Kuloi Plateau. PROOF C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:28

6 Igor N. Demidov et al. BOREAS 35 (2006)

141 van der Plicht (1998). Rock samples for in situ Glaciofluvial sand and gravel beds overlie the 10 142 cosmogenic Be surface exposure dating (CED) were Scandianvian till or they constitute its lateral equiva- collected and treated after the procedures described in lent. Deposits also show ice-wedge casts and other Linge et al. (2006). Ages are expressed in thousands of periglacial structures overlain by fluvial deposits con- 143 years (kyr) and indicated in stratigraphical logs and taining plant detritus. These sediments yield ages of 13 144 listed in Fig. 9 and Table 1. and 11 kyr (Table 1). Often the examined sections are covered by Holocene peat. Because the presence of 145 Scandinavian till can be demonstrated in sections 146 Maximum extent and age of the last found at Cape Kargovsky and westward, we assume 147 Scandinavian Ice Sheet the limit of SIS to be located near the east bank of the 148 mouth of the Kuloi River (Figs 1, 3). Even though the 149 The last SIS covered the northern coastal areas of the topography is rather flat and glacial landforms are 150 Kuloi Plateau and reached the northwestern part of the blurred by peat bogs, an 8-km-long chain of end- 151 Kanin Peninsula and the central part of the Pinega moraines and hills that cross the Kuloi River about 30 152 River (Figs 1, 2). From these areas, stratigraphical logs km southwest of the Cape favours our proposed 153 and cross sections containing information on glacial maximum position (Fig. 2). 154 deposits with Scandinavian provenance are compiled 155 and used as a prominent marker. Detailed knowledge Western coast of the Kanin Peninsula 156 on the last SIS has previously been presented from the 157 Severnaya Dvina area (Larsen et al. 1999; Kjær et al. Demidov et al. (2004) supported earlier views on the 158 2001; Lysa˚ et al. 2001), where till of that ice sheet was limit of SIS based on the distribution of Scandinavian 159 named Bobrovo till. The Volodga area to the south of erratics, young eskers and hummocky moraines by the Arkhangelsk Region was accounted for by Lunkka Ramsay (1911), Devyatova (1969) and Aseev (1974), as indicated in Figs 1 and 2. However, our present 160 et al. (2001), and areas to the east that were not overridden by the ice sheet are described by Houmark- understanding suggests the distribution of the Scandi- 161 Nielsen et al. (2001). navian till to be restricted only to the section near 162 Tarkhanov southeast of the Cape Kanin Nos. This is 163 under the assumption that optically stimulated lumi- 164 Northern coast of the Kuloi Plateau nescence (OSL)-dated glaciofluvial and glaciolacus- 165 trine sediments with ages older than 50 kyr have been 166 Eight sections from the northwestern corner of the well bleached (Fig. 1, sites 4Á7, Table 1). These 167 Kuloi Plateau to the Mezen Bay coast at the root of the deposits overlie a till plain from earlier Weichselian 168 Kanin Peninsula comprise a stratigraphical cross sec- glaciations at Konushin and other coastal sections at 169 tion through a segment of the marginal zone of the last the western and northern Kanin Peninsula. 170 SIS (Figs 1, 3). Marine Eemian sediments underlie At the Tarkhanov cliff section (Fig. 1, site 1) a 40 to 171 Weichselian glacial deposits in the river valleys drain- 45-m-high terrace of Weichselian sediments is banked 172 ing the northwestern part of the Kuloi Plateau upon the Precambrian metasediments that are inclined 173 (Geology of the USSR 1963) just as sections with up to 908 (Fig. 5). On top of the exposure the Kanin 174 similar setting are known from the opposite shore on Ridge is more or less flat and almost without Qua- 175 the Kola Peninsula (Armand et al. 1969). Previous ternary deposits. Strongly weathered bedrock, tors and 176 detailed stratigraphical studies from the Mezen Bay are fields of residual boulders are widespread, but occa- 177 adopted from Kjær et al. (2003). sionally 10 to 15-m-high kame-like hills composed of 178 Till of former Weichselian glaciations and fluvial bedded silt and sand occur. In the beach terrace, 179 deposits underlie glacial deposits of Scandinavian marine Eemian sediments are overlain by glacial and 180 origin. Ages range between 77 and 21 kyr (Table 1). glaciofluvial deposits from the Early and Middle 181 Scandinavian till unconformably overlies these sedi- Weichselian dated to 87, 80 and 58 kyr (Table 1). 182 ments and may have been glaciotectonically deformed. Boulders of crystalline Scandinavian rocks occur both 183 The basal part of the till often shows sheared and on the present-day beach and on the deflation surface 184 folded lenses of Quaternary deposits. Subtill and of the 35-m-high terrace. A brown sandy till unit less 185 intertill glaciodynamic structures, striations on clasts than 2 m thick is preserved in patches below the 186 and clast fabric orientation indicate glacier movements deflation surface (Fig. 6). Clast fabric orientations 187 from westerly and southwesterly directions (Fig. 4, sites indicate ice movement from the NW (Fig. 4). The till is 8, 9, 11). Occasionally, ice-dammed lake sediments of rich in Scandinavian rock fragments and has a sharp 188 similar stratigraphic position contain dropstones of erosional contact to the underlying lacustrine sand. 189 Scandinavian origin (Fig. 3, sites 8Á13, 15, 49). The till According to the lithology of clasts, this till unit is 190 ranges in thickness from 2 to 6 m and the clast content equivalent to the Bobrovo till deposited by the last SIS 191 is characterized by a mixture of local bedrock and rock further south. Cosmogenic nuclide dating of quartz 192 fragmentsUNCORRECTED of westerly provenance. veins on PROOF the bedrock surface at different altitudes C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:29

BOREAS 35 (2006) Last Scandinavian Ice Sheet, NW Russia 7

Fig. 4. Clast fabric diagrams from selected sections stereographically plotted on the lower hemisphere.

193 indicates ages of exposure of 18.6 kyr below c. 50 m a.s.l., Kanin Ridge reaches altitudes of 100Á180 m a.s.l., and 194 whereas at adjacent inland altitudes around 100Á115 m because older sediments lack till cover at sites 2Á7, 195 a.s.l. ages indicate exposure of around 51 and 56 kyr we assume that this part of the Kanin Peninsula was 196 (Linge et al. 2006). OSL dates of sediments in the kame ice-free during the Late Weichselian. 197 hills at altitudes above 130 m on the inland plateau a few 198 kilometres northward gave OSL ages of 71 and 63 kyr Central Pinega River 199 (Table 1, Fig. 5). Aerial photographs show a ridge 200 winding the NW-tip of the Kanin Peninsula at altitudes Demidov et al. (2004) proposed that the Pinega ice lobe 201 about 40Á50 m a.s.l. along an escarpment in the bedrock. covered a significant part of the PinegaÁMezen 202 We interpret this ridge to be a type of terminal moraine watershed. However, new data from sections at Ez- 203 associated with the latest glacier cover over Kanin huga, Karpogory and Verkola (Fig. 1, sites 19, 20, 22 204 Peninsula. and Fig. 7) suggest that the limit should be drawn 205 Thus, the upper till in the Tarkhanov section was about 50 km further to the west. The thickness of the 206 most probably deposited by the Scandinavian Ice Quaternary cover in the Pinega lowland is limited 207 Sheet, the flow direction of which was either deflected to less than 5Á10 m. Sediments with Eemian marine 208 by the advancing Barents Ice Sheet from the north or shells lacking a till cover are widespread in surface 209 the direction reflects local advance-phase flow influ- exposures (Fillipov & Borodai 1987). Near Karpogory, 210 enced by the steep slope bedrock topography. Our data a section lying south of the pronounced end-moraine 211 suggest that the last SIS reached an altitude up to 60Á ridges crossing area from NE to SW (Figs 1, 2), a 212 70 m a.s.l. and moved along the White Sea slope of the red-brownish clayey till with Scandinavian and local Kanin Ridge beforeUNCORRECTED 18 kyr ago. Because the top of the clasts is present. Fabric PROOF analyses in the till (Fig. 4) and C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:32

8 Igor N. Demidov et al. BOREAS 35 (2006)

channels that occupy the northwestern part of the area (Zatulskaya 1984a,b). Glaciolacustrine sand at Ezhuga (site 19), with an age of 26.69/1.9 kyr and lying above clast-bearing mud (Fig. 7, Table 1), does not provide conclusive evidence as to whether the last SIS did override this area. West of the end-moraine belt, i.e. east of Kulogora (Fig. 1, site 18), till of Scandinavian provenance rests on silt and sand with an age of 779/7 kyr containing marine shells (Fig. 7, Table 1). The till is covered by aeolian sand dated to 159/1.6 kyr. Clast fabric indicates ice movement from the westÁsouthwest (Fig. 4). Our data from the central part of the Pinega River suggest that the limit of the last SIS in this area is correct compared to previous estimates (Figs 1, 2, 7), but the stratigraphic signifi- cance of the Scandinavian erratics in this area is limited. Till with a Scandinavian clast signature is present in most of the area, and such diamicts are situated both above and below sandy beds of re-deposited marine shells of Eemian age.

Fig. 5. Sketch of the section at Tarkhanov, northwest shores of the Kanin Peninsula. Late Pleistocene deposits are banked upon Position of the last SIS in the Arkhangelsk region the bedrock cliff. Luminescence dates (OSL) of sediments and cosmogenic exposure dates (CED) on bedrock are indicated. Inferred Since Scandinavian till is clearly absent in sections glacier limit is shown. along the coastal area of the western Kanin Peninsula, the ice margin continued from Cape Kanin Nos probably across the eastern White Sea floor southward 213 age of the glaciofluvial and glaciolacustrine sediments to Cape Kargovsky (Figs 1, 2). Cape Kargovsky is the 214 below and above (Fig. 7) suggest that the till has been easternmost point in the Mezen Bay area where 215 deposited by an ice sheet from northÁnorthwest in the Bobrovo till is observed (Kjær et al. 2001, 2003). 216 Early Weichselian or possibly in the Late Saalian. On From this point, the ice margin ran southward through 217 the left side of the Ezhuga River, marine interglacial the water divide between the Kuloi and Mezen Rivers 218 sediments resting on the Moscowian (Late Saalian) and crossed the valley of the Kuloi River near the mouth of the Soyana River. The ice front turned 219 Scandinavian till are, in turn, covered by a re-deposited westward from the mouth of the Soyana River and 220 diamicton containing both local and Scandinavian touched the steep, 50 to 80-m-high, northeastern cliffs 221 boulders (Fig. 7). of the Kuloi Plateau, the northeastern part of which is 222 Devyatova (1969, 1982) suggests the limit of the Late deprived of Quaternary deposits. Discontinuous belts 223 Weichselian SIS to have been located near the Karpo- of sandy end-moraine hills constitute the western 224 gory village; we propose that the limit is possibly boundary of outwash plains that guided meltwater 225 marked by the many end-moraine ridges and meltwater towards the east (Krasnov 1971; Legkova & Schukin 1972). During the maximum extent, the 180-m-high south- eastern part of the Kuloi Plateau was ice free (Edemsky 1931); we propose that the ice margin only rimmed the plateau. From the south, an ice tongue moved eastward along the Pinega River lowland to altitudes of 30Á60 m a.s.l., and the ice front covered only the middle part of the Pinega River (Figs 1, 8). The ice front ran along the terminal moraine belt. It expanded to the south and crossed the Pinega River and found its limit north of the of Shilega and Karpogory, where Eemian marine sediments with whale bones are not covered by till (Devyatova 1982). The ice margin continued along a discontinued chain of hills and ridges south- westward to join terminal formations which truncate Fig. 6. Till with Scandinavian clasts and boulders underlying the the Pokshenga Hills from the north and west between deflationUNCORRECTED surface, Tarkhanov section, Kanin Peninsula. sites 21 andPROOF 24 (Figs 1, 2) (Arslanov et al. 1984). The C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:34

BOREAS 35 (2006) Last Scandinavian Ice Sheet, NW Russia 9

Fig. 7. Stratigraphic logs from selected sections on the middle parts of the Pinega River.

226 maximum limit is connected with the limit of SIS in the grounds of the Kola Peninsula (Ha¨ttestrand 2006). 227 Severnaya Dvina, which has been supported for similar Although we are unable to exclude this possibility, 228 reasons by Larsen et al. (1999) and Atlasov et al. there is compelling evidence for an extensive proglacial 229 (1978). meltwater drainage system associated with the sug- 230 In our present configuration of the last SIS, the data gested ice-sheet configuration in northwest Russia. 231 suggest a smaller distribution of glacier ice and 232 consequently its eastern boundary is moved westward Age of the maximum extent of the last SIS 233 by 20Á120 km compared to previous reconstructions. 234 We recognize that the western coast of the Kanin OSL dates above and below the Bobrovo till in the 235 Peninsula, the River Kuloi and the eastern part of the Arkhangelsk region, supplemented by dates from Lake 236 Kuloi Plateau, and areas situated eastward and south- Beloe in the Volodga Region from similar stratigraphic 237 ward of the Karpogory village in the middle Pinega levels (Lunkka et al. 2001), provide an age frame for River, were ice-free during the last glaciation, but may 238 the maximum ice extent (Fig. 9). Subtill dates range have been within the reach of previous Scandinavian 239 from 32 to 15.99/1.2 kyr, whereas dates from above the Ice Sheets. Lowland areas were occupied by large, 240 Bobrovo till range from 17.29/1.3 kyr to 11 kyr. partly ice-dammed lakes (Figs 8, 10AÁD). An almost 241 Allowing the ice sheet to settle at the maximum straight-crested 130-km-long belt of hills can be traced 242 position for at least 1 kyr, the assumed age of this from the mouth of the Pyoza River, southward through 243 event lies between 18 and 16 kyr9/1 kyr. Sediments the Kuloi River and Mezen water divide. It was 244 outside the former ice margin indicate that damming of proposed to mark an end-moraine zone (Zatulskaya 245 substantial lowland areas by ice-marginal lakes com- 1984c; Lavrov 1991) and according to Demidov et al. (2004) it depicted the maximum extent of SIS. How- menced about 4 kyr before and lasted at least 1 kyr 246 ever, at exposures close to the mouth of the Pyoza after the maximum event. Even though the maximum 247 River we have not identified any end-moraine ridges or extent was most likely attained at different times, the 248 Scandinavian till within this belt; on the contrary, this uncertainties in the data sets from the Kanin Peninsula, 249 straight ridge merely reflects the strike of gently Kuloi and Pinega areas are too large to indicate any 250 eastward dipping bedrock. significant geographical difference in age. The new 251 A fundamental assumption in our reconstruction is dates from the northern part of the Arkhangelsk 252 that ice sheets deposit till or leave a tectonic imprint district have not added significantly to previous 253 that indicates ice advance, hence the distribution of a estimates based on data restricted to the Severnaya 254 till sheet or glaciotectonic deformation can be used to Dvina area (Larsen et al. 1999). In addition, exposure 255 reconstruct the extent of the ice sheet. In our opinion, dating on the Kanin suggests that SIS began to retreat 256 it is likely that wet-based ice lobes extended eastward from the northwestern part of Peninsula as early as 257 during the LGM ice-sheet build-up. However, frozen 18.6 kyr ago, and therefore neither contradicts this age 258 bed conditions could have been regained in the estimate nor provides a better accuracy. An average age 259 marginal zone when ice lobes reached higher and of the maximum extent of the eastern part of the 260 permafrozen ground east of the White Sea basin, as Scandinavian Ice Sheet placed between 18 and 16 kyr is 261 seems to be theUNCORRECTED case with the ice cover on the high at least 3Á5 kyr andPROOF perhaps as much as 7Á9kyr C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:35

10 Igor N. Demidov et al. BOREAS 35 (2006)

The limitation of ages from melting ice facies only gives a rough estimate of the duration and spatial develop- ment of sedimentation during the deglaciation (Fig. 9). Ages of sediments and plant detritus from deglaciation basins indicate retreat of the active margin of SIS, downwasting, thermo-karst activity and immigration of a sub-arctic flora from c. 15 kyr and about 5 kyr ago onward.

Mezen Bay ice lake Since the work of Ramsay (1911), many authors have suggested that the wide valley of the Kuloi River was the main discharge system of ice-dammed lakes either during the last maximum of SIS or during the final stages of deglaciation (Yakovlev 1956). An ice-dammed lake developed in the Mezen Bay area (Figs 8, 10A, lake 1) and was fed from the melting ice sheet in the White Sea and from the Mezen and Pyoza Rivers draining ice-free periglacial areas (Houmark-Nielsen et al. 2001). The level of the ice lake was about 15 m a.s.l. and was controlled by a threshold located on the lowermost part of the Kanin Peninsula on the water divide between the White Sea basin and the Chyosha Bay. Present-day valleys of the Chizha and Chyosha rivers (Fig. 1) crossing the peninsula are filled with 5Á8 m of fluvial sediments. The valley continues as a submarine canyon eastward at depths around 15Á30 m below sea level. It is more than 20 km length, up to 1 km in width and up to 30 m in depth. The ice-dammed lake discharged to the Chyosha Bay. Laminated clay on the bottom of this lake is located at altitudes of 5Á17 m a.s.l. (Kjær et al. 2003) and can be seen in the section from sites 13 and 15 (Figs 1, 3). The elevated river terraces observed by Zekkel (1939), with altitudes of 5Á6, 8Á10, 20 m a.s.l. on the lower Mezen and Pyoza rivers, may represent falling water level in the lake with time. A terrace at 15 m on the Kanin Peninsula may correspond to one of these stages.

Fig. 8. Palaeogeographical reconstruction of the eastern flank of Scandinavian Ice Sheet at its maximum extent. Numbers refer to ice Kuloi ice lake dammed lakes. 1/Mezen Bay ice lake; 2/Kuloi ice lake; 3/Pinega The ice-dammed Lake Kuloi coexisted with an ice- ice lake; 4/Severnaya Dvina ice lake. dammed lake in the Mezen Bay until SIS lost contact with the Cape Kanin Nos, which led to discharge of the 262 younger than previously proposed by Arslanov et al. lake into the Barents Sea. The Kuloi ice lake filled the 263 (1970, 1971). valley and the adjacent lowland of the Kuloi River 264 (Figs 8, 10AÁB; lake 2). This lake was no more than 265 15Á20 m deep and it drained from a threshold in the 266 Deglaciation drainage pattern northeast into the Mezen Bay ice lake. According to 267 Zekkel (1939), the highest terraces of the Kuloi River 268 Palaeogeographical reconstructions depict the position are composed of coarse sand and have an altitude of 269 of the retreating ice margin, the distribution of ice- 20Á25 m a.s.l. near the entrance to the lake. Combined 270 dammed lake basins and the drainage directions of with the mapping by Lavrov (1991), this indicates that 271 meltwater from the time of the largest extent of SIS the lake level was about 30 m a.s.l. with a gradual drop 272 (Fig. 8) until the Allerød interstadial (Fig. 10AÁD). to 16 m. As deglaciation proceeded, the ice front 273 Ice-dammed lake and river terrace levels are based on retreated from the Lower Kuloi River, which inevitably 274 theUNCORRECTED present-day topography and geological mapping. led to disappearancePROOF of the Kuloi ice lake (Fig. 10B). C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:36

BOREAS 35 (2006) Last Scandinavian Ice Sheet, NW Russia 11

Fig. 9. Plot of luminescence, radiocarbon and cosmogenic exposure ages, northwest Russia (1 standard derivation). Dated sediments and plant remnants are from below and above the Bobrovo till deposited by the last Scandinavian Ice Sheet. Dated extramarginal deposits relate to phases of ice damming or deglaciation. Compiled from the present study, Larsen et al. (1999), Lunkka et al. (2001) and Kjær et al. (2003). 1/Dates published by Larsen et al. (1999); 2/conventional date published by Atlasov et al. (1978); 3/dates published by Lunkka et al. (2001).

275 When discharge from the Pinega and Dvina ice Severnaya Dvina ice lake 276 lakes occurred through the Kuloi River, a 3-km-wide During the maximum extent of SIS, an ice-dammed 277 meltwater valley developed (Fig. 10 D). 278 lake filled the upper parts of the Severnaya Dvina 279 River basin (Fig. 8, lake 4). The level of the lake Pinega ice lake 280 reached elevations around 130Á120 m a.s.l. (Atlasov et al. 1978; Arslanov et al. 1984) and runoff was directed 281 An ice-dammed lake occupied the middle Pinega River southward into the Volga basin (Kvasov 1975; Lunkka 282 during the SIS maximum (Fig. 8, lake 3). The level of et al. 2001). As deglaciation proceeded and the ice front 283 this lake was at 80 m a.s.l. and was controlled by a melted into the middle lower Dvina basin, a spillway 284 threshold of the water divide between the River Ezhuga 285 and River Ezhuga Zyranskaya, with discharge east- opened eastward and the lake level dropped to about 286 ward into the Mezen basin (Figs 8, 10A). These ancient 80Á90 m a.s.l.; discharge was redirected into the Mezen 287 ice-lake terraces, with an altitude of 80Á90 m a.s.l., are Bay through the middle part of the Pinega and Mezen 288 known from both the Severnaya Dvina and Pinega rivers (Fig. 10A). 289 River lowlands (Atlasov et al. 1978; Arslanov et al. When the ice melted away from the mouth of the 290 1984). As the glacier retreated, the Severnaya Dvina Ezhuga River, the lake level dropped from 80 to 50 m 291 ice-dammed lake merged with the Pinega ice lake (Fig. a.s.l. and a new 50-m threshold developed on the water 292 10A), both discharging into the Mezen basin. Even- divide of the PokshengaÁPukshenga rivers, while run- 293 tually, the water level dropped from 80 to 50 m a.s.l. off was directed into the Mezen River (Fig. 10B). This 294 and most Quaternary deposits were washed away from possibly occurred about 15Á15.79/1.2 kyr ago, as 295 the banks of the Pinega River (Fig. 10B). The Pinega indicated by dates of lacustrine deposits underlying 296 ice lake practically ceased to exist around 159/1.6 kyr subaerial sediments in the Chelmokhta sections at an ago, but the Pinega River maintained its discharge, altitude 12 m a.s.l. (Fig. 1, site 24 and Figs 9, 11) (Larsen et al. 1999). At Raibola, further upstream in 297 now supplied with meltwater from the Severnaya Dvina ice lake. As the ice melted further away, the Dvina basin (Fig. 1, site 30), fluvial sand resting on 298 discharge carved a deep spillway at altitudes of 18Á20 laminated clays at an altitude of 56 m a.s.l. shows OSL 299 m a.s.l. on the Kuloi RiverÁPinega River watershed ages of 17.2 and 14.0 kyr (Larsen et al. 1999). We 300 (Fig. 10D). The ages of events are based on dates of conclude that the lake level dropped from 80 to 50 m 301 aeolian sediments lying at altitudes about 30 m a.s.l. in a.s.l. in the Dvina basin about 15 kyr ago. 302 the former lake basin at the Kulogora village on the As the active ice margin retreated from the mouth of 303 PinegaÁKuloiUNCORRECTED water divide (Figs 1, 7, site 18). the Pinega River, thePROOF level dropped from 50 to 18 m C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:39

12 Igor N. Demidov et al. BOREAS 35 (2006)

Fig. 10. Palaeogeographical reconstructions picturing four stages (AÁD) of Lateglacial ice retreat, deglaciation and development of ice-dammedUNCORRECTED lakes and run-off patterns in northwest Russia. PROOF C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:42

BOREAS 35 (2006) Last Scandinavian Ice Sheet, NW Russia 13

Fig. 11. Stratigraphic logs from seven sections at Chelmokhta, Severnaya Dvina River (Fig. 1, site 24). Sediments above Bobrovo till indicate aerial downwasting, inversion of landscape and deposition under thermokarst conditions. Luminescence and calibrated 14C ages are from Larsen et al. (1999).

304 a.s.l., water discharged through the lower Pinega River Chelmokhta, an example of aerial downwasting 305 into the Kuloi River basin and the Dvina ice lake At Chelmokhta (Fig. 1, site 24) the ice locally 306 ceased to exist (Fig. 10C). This probably occurred 307 about 13.7 kyr ago, as indicated by subaerial deposi- moving from northerly directions ceased to flow 308 tion at Chelmokhta (Fig. 1, site 24 and Figs 9, 11) and the development of constantly subsiding ice- 309 (Larsen et al. 1999). Later, a new Severnaya Dvina ice confined depressions caused the sedimentation of a 310 lake appeared near the mouth of present river at 18Á20 variety of supraglacial, mostly waterlain, sediments 311 m a.s.l. (Fig. 10D). Discharge from the lake was into found up to about 20 m above the present river level 312 the Mezen Bay through the rivers Pinega and Kuloi (Fig. 11). The sediments are exposed in the river 313 until the time when the ice front retreated from the banks of the present Dvina River 175 km upstream 314 southern part of the (Fig. 10D). This from Arkhangelsk. Seven sections along a 1-km-long 315 probably occurred immediately before the Allerød segment of the river were documented. The sedimen- event, as indicated in the Psaryovo section (Fig. 1, tary successions were all influenced by mass move- site 41) where plant remains with an age 13.2 kyr occur ment and gravitational sliding during inversion of the 316 in fluvial sand at 25 m a.s.l. (Larsen et al. 1999). Pollen topography due to melting of stagnant ice beneath 317 studies indicate Older Dryas and Allerød ages of (Lysa˚ et al. 2001). 318 varved and homogenous silt and clay known from Overlying basal till with Scandinavian provenance 319 boreholes north of the Psaryovo section (Fig. 1, site 41) components of four major depositional events can be 320 ranging from altitudes of 34 m a.s.l. to zero (Bara- detected. Lowermost, the basal till gradually changes 321 novskaya et alUNCORRECTED. 1977; Pleshivtseva 1977). into stratified diamict PROOF interbedded with sorted sand C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:44

14 Igor N. Demidov et al. BOREAS 35 (2006)

322 and mud showing upward-decreasing clast content, velocity. Corridors of rapidly flowing ice separated by 323 all deposited in local depressions. At some sites, slower flowing ice constituted ice-divide zones. 324 imbricate gravel or cross-bedded sand with an ero- Our results suggest that glacier tongues and lobes, 325 sional lower contact suggests deposition in meltwater probably less than 300 m in thickness, not only at the 326 streams, and this possibly signals the lowering of the maximum glaciation itself but also during different 327 base level caused by a drop in the water table in the phases of deglaciation protruded the general northÁ 328 Dvina ice lake about 15 kyr ago. Laterally restricted south trending ice margin and flowed rapidly eastward 329 and strongly deformed thin beds of sand and mud through shallow depressions on the low relief north- 330 overlie these deposits. The sediments show normal as west Russian plain. This hypothesis, which still has to 331 well as inverse grading, and combined with laminated be thoroughly tested, is based on theoretical models of 332 mud they indicate deposition by sediment gravity similar dynamic behaviour of terrestrially based ice 333 flows and fall out from suspension. Sand could either streams as predicted by Stokes & Clark (1999) and 334 have been blown by wind or transported by melt- Boulton et al. (2001). Previously, Lagerlund (1987) had 335 water into small ice-dammed ponds and lakes. Plant introduced the concept of ‘outlet surges’, in which 336 remains, such as lenses of peat, twigs and leaves, minor depressions in the overall pre-ice-advance topo- 337 indicate the development of a vegetation cover on the graphy of the circum-Baltic lowlands determined the 338 clastic debris covering the stagnant ice. Shells of distribution of proglacial lakes, whose water bodies and 339 Anodonta indicate the immigration of freshwater sediments acted at gateways for rapid flowing ice. 340 molluscs into the lakes. The flora includes species Similar prerequisites, including little or no topographic 341 such as Betula, Salix, Populus, Larix and Picea (Lysa˚ control, and special bed lithologies for the location of 342 et al. 2001). potentially fast ice-flow corridors have recently been 343 Finally, deposition of rhythmically laminated mud revived within the Laurentian Ice Sheet (Stokes & 344 with occasional diamict lenses, dropstones and beds Clark 2003). 345 of sand was followed by an episode of erosion caused In areas along the periphery of SIS elsewhere, former 346 by a sudden lowering of the local water level. This is terrestrially based ice streams had supposedly been 347 indicated by deposition of thick sheets of bedded flow operating (Clark et al. 2003). These areas coincide with 348 diamict and massive sand containing boulders, rip-up a region of sedimentary bedrock and thick Quaternary 349 clasts of peat and bedded lacustrine sand with leaves deposits surrounding the crystalline Fennoscandian 350 and twigs. Because plant remains from the underlying Shield. The largest extent of the SIS was reached 14 351 units range in age between 11.8 and 10.5 cal. Ckyr earlier and under colder climatic conditions in south- 352 (Larsen et al. 1999), we estimate that the final western Scandinavia and the Baltic compared to 353 collapse of the dead ice took place in the early northwestern Russia. Because similar glaciation sce- 354 Holocene. Possibly at the same time, the White Sea narios including fast and laterally confined ice flow was inundated by arctic marine waters. Radiocarbon occurred around the whole SIS, factors such as the ages of 93309/120 yr BP on marine shells indicate presence of water-filled basins and deformable sub- 355 that the ocean waters inundated the White Sea stratum must have played a major role. From the around 11 kyr ago (Koshechkin et al. 1977) and northern North Sea to the circum-Baltic region, 356 subsequent regression began at the onset of the rapidly flowing ice streams bounded by slow flowing 357 Boreal period about 9100 years ago (Baranovskaya inter-stream areas dominate the pattern of Late 358 et al. 1977). Weichselian glaciation. With the exception of the 359 main advance inter-stream event in Denmark (22Á19 360 kyr BP), the Norwegian Channel Ice Stream (26 kyr 361 Ice-sheet dynamics and flow pattern BP; Sejrup et al. 1994), the Kattegat Ice Stream (29Á25 362 kyr BP) and the Young Baltic Ice Streams (19Á15 kyr 363 The ice-sheet configuration and flow pattern along the BP), all advanced from SIS across marine waters or ice- 364 eastern flank of the last SIS in northwest Russia is dammed lakes. These advances show flow patterns and 365 illustrated in Fig. 12. The SIS invaded southern and left behind morphological features similar to those of 366 central Finland less the 25 kyr ago, as indicated by terrestrial-based ice streams (Boulton et al. 2001; Clark 367 dating of mammoth tusks pre-dating deposition of till et al. 2003 and references therein; Houmark-Nielsen & 368 by the last SIS (Ukkonen et al. 1999; Lunkka et al. Kjær 2003). 369 2001). In the following 5Á9 kyr, the glacier front had We therefore propose that also in northwest Russia 370 advanced at an average velocity of at least 10Á12 km ice-dammed lakes and water-saturated fine-grained 371 per hundred years about 500Á800 km towards the east proglacial sediments in topographic depressions could 372 and southeast. The maximal position in northwest have caused rapid outlets from the periphery of the 373 Russia was reached at 19Á15 kyr ago (Fig. 12). main ice sheet. Moreover, glacier flow velocities were 374 Ice-sheet flow through large water-filled bedrock probably reduced and the ice considerably thinner in 375 depressions on the rim of the Baltic shield and the areas of elevated Pre-Quaternary bedrock. In these 376 PalaeozoicUNCORRECTED platform to the east probably enhanced the inter-lobe PROOF areas, third-order flow divides sensu Boulton C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:45

BOREAS 35 (2006) Last Scandinavian Ice Sheet, NW Russia 15

Fig. 12. Reconstruction of glacier flow patterns, marginal position and ice-dammed lakes during the maximum extent of the eastern flank of the Scandinavian Ice Sheet 18Á17 kyr BP. Numbers IÁV indicate the corridors of fast-flowing ice. Compiled from the present study with data from Aseev (1974), Gey & Malakhovsky (1998), Ekman & Iljin (1995) and Lunkka et al. (2001).

377 et al. (2001) could have been located. Along the (Fig. 12, IIA) moved to the northeast along the White northeastern part of the ice sheet in Russia, the Barents Sea inlet, flanked to the north by the Kola Peninsula 378 SeaÁKanin fast-flowing ice moved from the west and highland and to the east and south by the Kuloi 379 northwest and rimmed the ice divide zone located on Plateau, where it reached heights of 150Á220 m a.s.l. 380 the central part of the Kola Peninsula (Fig. 12, I). The The Dvina and Pinega ice tongues (Fig. 12, IIB, IIC) 381 ice occupied the northern part of the White Sea and flowed from the northwest along Severnaya Dvina and 382 reached the northwestern point of the Kanin Peninsula Pinega lowlands along with fast-flowing ice in the 383 where the Scandinavian Ice Sheet merged with the Onega River depression (Fig. 12, IID), which also had 384 Barents Sea Ice Sheet. its origin in the White Sea. Further southward, other 385 South of the Kola Peninsula, the White Sea ice corridors of rapid ice flow in the SIS moved from 386 flowed from the west along the White Sea depression more northerly directions and occupied the depres- 387 and dispersed into a number of ice lobes in the sions and adjacent areas of the lakes Onega (III), 388 Arkhangelsk regionUNCORRECTED (Fig. 12, II). The Kuloi ice tongue Ladoga (IV) and ChudskoePROOF (V). Ice divides were C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:51

16 Igor N. Demidov et al. BOREAS 35 (2006)

Fig. 13. Reconstruction of the degradation along the eastern flank of the last Scandinavian Ice Sheet. Vast regions of aerial downwasting are indicated east of the BøllingÁAllerød ice marginal position. Data from the present study are compiled with data from Aseev (1974), Gey & Malakhovsky (1998), Ekman & Iljin (1995), Lunkka et al. (2001) and Ha¨ttestrand (2006).

389 situated along bedrock highs of 150Á300 m a.s.l. positions and recessive stages of the SIS (Krasnov (Aseev 1974). 1971; Aseev 1974), but our data do not challenge this 390 hypothesis. The sediment successions at Chelmokhta 391 (Fig. 11) indicate deposition in an environment of 392 Decay of the ice sheet aerial downwasting of buried stagnant glacier ice that 393 was subjected to thermo-karst processes. The ice along 394 Event-stratigraphic models in the region often relate the lower-middle Dvina basin had been detached LateglacialUNCORRECTED climatic variations to the ice-marginal from the PROOF active ice sheet. Subsidence and sagging C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:53

BOREAS 35 (2006) Last Scandinavian Ice Sheet, NW Russia 17

395 caused the development of kettle holes with space From the Allerød interstadial and onwards, decay of 396 for deposition of a variety of sedimentary facies and the the active ice margin changed to frontal-type deglacia- 397 re-sedimentation of unstable soils and remnants of the tion. The ice sheet rested on crystalline bedrock and 398 pioneer vegetation by gravitational slumping and was not enriched with debris to the same degree as had 399 sliding. The sub-arctic flora arrived in the region at been the case further east, where it flowed over 400 the beginning of the Allerød period (Baranovskaya et deformable sediments. The relatively thin and lobe- 401 al. 1977; Wohlfarth et al. 2002, 2004), while buried shaped ice front quickly melted during the ameliorated 402 glacier ice was melting causing inversion of the land- climate of the Allerød. The glacier retreated about 403 scape, as was also the case in the area northeast of the 150Á200 km from the Neva terminal belt to the Young 404 Arkhangelsk district which was covered by an older Dryas margin in c. 800 years (Fig. 13). Supraglacial 405 Weichselian ice sheet from the BarentsÁKara Seas sediments rarely occur in this area, but fields of 406 (Tveranger et al. 1995). From the eastern North Sea drumlins are widespread, as are long and prominent 407 across the southern Baltic to northwest Russia, Late eskers systems (Ekman & Iljin 1995). Under both 408 Weichselian streamlined terrains separated by belts of types of deglaciation, the large ice-dammed lakes in 409 terminal moraines possibly generated by narrow and the Baltic, the White Sea and the Ladoga and Onega 410 rapidly flowing ice lobes are overprinted by landforms Lakes assisted calving and increased the rate of 411 that relate to aerial downwasting (cf. Boulton et al. deglaciation. 412 2001). In Russia, an inter-BøllingÁAllerød terminal 413 belt (Neva stage) separates a region of dead ice relief 414 generated by downwasting to the east and to the south Conclusions 415 and another with evidence of frontal deglaciation to 416 the west and north (Fig. 13). The terminal zone 417 stretches from the Baltic Sea south of the Gulf of . Till deposited by the last SIS (Bobrovo till) is 418 Finland via the southern shores of Lake Ladoga and identified by its stratigraphic position, its contents 419 through the lake Onega into the and further of Scandinavian erratics and its directional proper- 420 out into the White Sea. Further north the terminal belt ties indicating ice flow from westerly directions. 421 encircled the highland of the Kola Peninsula. Between . The eastern flank of the Late Weichselian Scandi- 422 the maximum position of SIS and the Neva stage 2 navian Ice Sheet reached its maximal position in 423 termini, a several hundred thousand km large zone of northwest Russia some time between 20 and 15 kyr 424 downwasting contains a complex dead ice relief, ago; in the Arkhangelsk region this probably including fields of kame-like hills and hummocky occurred around 18Á16 kyr ago. This position was 425 moraine. Over large areas glaciolacustrine sediments reached up to 10 kyr after the maximum extent in 426 may blur the original shape and size of the dead ice the southwestern part of SIS. 427 landforms. . At the maximum, the ice margin stretched from the 428 We assume that long and thin lobes of fast-flowing northwest part of the Kanin Peninsula across the 429 ice became detached from the main ice sheet once the bottom of the White Sea southward to the mouth of 430 ice-sheet profile along the flow paths had become flat Kuloi River. It touched the eastern part of the Kuloi 431 enough. Thus, the ice lobes were no longer fed by the Plateau and covered the lower and middle parts of 432 ice sheet and the flow ceased. The marginal zone now the Pinega River and turned westward into the 433 became subjected to the melting of large dead ice Severnaya Dvina River lowland. 434 masses. In northwest Russia, detachment of peripheral . In the zone between the former margin of SIS and 435 debris-rich lobes from the flowing ice began c. 16 15 Á the Neva terminal belt, aerial downwasting caused 436 kyr ago and lasted perhaps until the beginning of the the inversion of landscapes. Deglaciation sediments 437 Allerød interstadial (Fig. 13). Bølling sediments rarely are composed of gravity flows of diamictic material, 438 occur in this zone probably because the larger part of lacustrine mud with plant debris and aeolian and 439 the area was covered by unstable dead ice. The age of fluvial sand, all often strongly deformed by slump- 440 the oldest organic remnants in sediments from lakes ing, sliding and periglacial frost-thaw processes. 441 across the downwasting zone belongs to the Allerød Deposition lasted until the beginning of the Holo- 442 interstadial (Ekman & Iljin 1995; Davydova et al. cene and was accompanied by the slow melting of 443 1998). Deposition of organic debris in small lakes on buried and stagnant ice and the migration of a sub- 444 the watersheds covered by dead ice began as late as the arctic flora and fauna. 445 Preboreal and Boreal (Demidov & Lavrova 2001; 446 Wohlfarth et al. 2002, 2004). Huge areas in the 447 peripheral parts of the SIS therefore stagnated rapidly, Acknowledgements. Á Invaluable help during fieldwork on the eastern 448 fields of debris-covered dead ice spent 4Á7 kyr of shores of the White Sea and along Pinega River in the Arkhangelsk region was given by J. K. Nielsen, formerly at the Geological 449 degradation as the climate was cold and permafrost Museum, Copenhagen, Denmark, and by Øystein Jæger and Maria 450 still prevailed and because thick overburden retarded Jensen, Geological Survey of Norway, Trondheim, Norway and 451 heat transfer,UNCORRECTED thereby delaying downwasting. Michael Tyagushkin from PROOF Petrozavodsk, Russia. We thank Andrew C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:53

18 Igor N. Demidov et al. BOREAS 35 (2006)

452 Murray (Nordic Laboratory for Luminescence Dating, Risø, and the Demidov, I. N., Houmark-Nielsen, M., Kjær, K. H., Larsen, E., 453 Department of Earth Sciences, University of Aarhus, Denmark), who Lysa˚, A., Funder, S., Lunkka, J.-P. & Saarnisto, M. 2004: Valdaian 454 carried out the luminescence dating and readily discussed the results. glacial maxima in the Arkhangelsk district of Northwest Russia. In 455 Henriette Linge provided the cosmogenic exposure dates and Britta Ehlers, J. & Gibbard, P. L. (eds.): Quaternary Glaciations Á Extent 456 Munch and Christian Hagen (Geological Institute, University of and Chronology. Part I: Europe, 321Á336. Elsevier, Amsterdam. 457 Copenhagen) gave the drawings their finish. Finally, we are greatly Devyatova, E. I. 1969: Degradation of Valdai glaciation and indebted to the two reviewers, Mona Henriksen (University of lateglacial history of Baltic and White Sea. In Gerasimov, I. P. Bergen) and Clas Ha¨ttestrand (Stockholm University), who put (ed.): Poslednii lednikovyi pokrov na severo-zapade evropeiskoi time-consuming and tedious effort into raising the standards of the chasti SSSR, 231Á241. Nauka, Moscow (in Russian). manuscript. Devyatova, E. I. 1982: Late Pleistocene environments as related to 458 human migrations in the Severnaya Dvina basin and in Karelia. 156 pp. Karelia, Petrozavodsk (in Russian). 459 Edemsky, M. B. 1931: Geological investigation in the basins of the 460 References Rivers Pinega, Kuloi and Mezen at 1929. Trudy geologicheskogo 461 muzeja Akademii Nauk SSSR T.VIII,71Á122 (in Russian). 462 Anderson, T. W. & Stephensen, M. A. 1971: Tests for randomness of Ekman, I. & Iljin, V. 1995: Deglaciation, the Younger Dryas end 463 directional against equatorial and bimodal alternatives. Depart- moraines and their correlation in Russian Karelia and adjacent 464 ment of Statistics, Stanford University, Technical Report 5,19pp. areas. In Ehlers, J. & Gibbard, P. L. (eds.): Glacial Deposits in 465 Armand, A. D., Armand, N. N., Grave, M. K., Evzerov, V. Ya. & North-East Europe, 195Á209. A. A. Balkema, Rotterdam. 466 Lebedeva, R. M. 1969: Generalized scheme of Quaternary Filippov, V. V. & Borodai, L. V. 1987: Key sections of the Mikulino 467 (Anthropogen) deposits of Kola Peninsula in the light of the new marine deposits in the north of the Mezen hollow. Vestnik 468 data. In Grave, M. K. & Koshechkin, B. I. (eds.): Main Problems of Leningradskogo Gosudarstvennogo Universiteta, Geologia, Geogra- 469 Geomorphology and Anthropogen Stratigraphy of the Kola Penin- fia 21,72Á74 (in Russian). 470 sula, 7Á24. Nauka, Leningrad (in Russian). Funder, S., Demidov, I. & Yelovicheva, Ya. 2002: Hydrography and 471 Arslanov, Kh., Auslender, V., Gromova, L., Zubkov, A. & Khomu- mollusc faunas of the Baltic and White SeaÁNorth Sea seaway in 472 tova, V. 1970: Palaeogeography and absolute age of the Valdaian the Eemian. Palaeogeography, Palaeoclimatology, Palaeoecology 473 glacial maximum in the Lake Kubenskoe area. Doklady Academii 184, 275Á304. 474 Nauk SSSR 195, 1395Á1398 (in Russian). Ganeshin, G. S., Krasnov, I. I., Marinov, N. A. & Murzaeva, V. E. 475 Arslanov, Kh. A., Voznyachuk, L. N., Velichkevich, F. Yu, Kurerova, (eds.) 1980: Map of Quaternary Deposits of Eurasia. Scale 476 L. V. & Petrov, G. S. 1971: Age of the maximal stage of the last 1:5000000. Ministerstvo geologii SSSR, VN II zarubezhgeologia 477 glaciation on the watershed of the Zapadnaya Dvina and Dnepr (in Russian). 478 rivers. Doklady Academii Nauk SSSR 196, 161Á164 (in Russian). Geology of USSR 1963: Arkhangelsk and Vologda Districts and Komi 479 Arslanov, Kh. A., Levina, N. B., Ostanin, V. E., Baranova, V. N., ASSR. Vol. 2. Part 2,1Á1079. Nedra, Moscow. 480 Smirnova, V. M. & Tertychnaya, T. V. 1984: On geochronology and Gey, V. P. & Malakhovsky, D. B. 1998: On the age and expansion of stratigraphy of the Late Pleistocene and Early Holocene in the maximal late Pleistocene glacial thrust in western part of the 481 basin of the Severnaya Dvina river. Vestnik Leningradskogo Vologda area. Izvestiia Russkogo geograficheskogo obschestva1, 482 Gosudarstvennogo Universiteta, Geologia, Geografia 12,57Á66 (in 43Á53 (in Russian). Russian). 483 Ha¨ttestrand, C. 2006: The glacial morphology of Kola Peninsula and 484 Aseev, A. A. 1974: Ancient Continental Glaciations of Europe. 319 pp. adjacent areas in Murmansk Region, Russia. Journal of Maps 485 Nauka, Moscow (in Russian). v2006,30Á42. 486 Atlasov, R. R., Bukreev, V. A., Levina, N. B. & Ostanin, V. E. 1978: Houmark-Nielsen, M., Demidov, I., Funder, S., Grøsfjeld, K., Kjær, 487 Relief characteristics of the marginal zone of Valdaian glaciation K. H., Larsen, E., Lavrova, N., Lysa˚, A. & Nielsen, J. K. 2001: 488 on OnegaÁVaga interfluve and in Vaga river valley. In: Kraevye Early and Middle Valdaian glaciations, ice-dammed lakes and 489 obrazovania materikovych oledeneniy. Naukova Dumka, Kiev, 30Á periglacial interstadials in northwest Russia: new evidence from the 490 38 (in Russian). Pyoza River area. Global and Planetary Change 31, 215Á237. 491 Baranovskaya, O. F., Pleshvtseva, E. S. & Travina, M. A. 1977: Houmark-Nielsen, M. & Kjær, K. H. 2003: Southwest Scandinavia, 492 Stratigraphy of marine late Pleistocene and Holocene sediments in 40Á15 kyr: palaeogeography and environmental change. Journal of 493 the lower reaches of Severnaya Dvina River Á by key sections. In Quaternary Science 18, 769Á786. 494 Biske, G. S. (ed.): Chetvertichnaya stratigrafiya i paleogeografiya 495 Severa Evropeiskoi chasti SSSR, 111Á118. KFAN SSSR, Petroza- Kalyanov, V. P. & Androsova, V. P. 1933: Geographical observation 496 vodsk (in Russian). on the Kanin. Zemlevedenie 25,16Á25 (in Russian). 497 Boulton, G. S., Donglemans, P., Punkari, M. & Broadgate, M. 2001: Kamb, W. B. 1959: Ice petrofabric observations from Blue Glacier, 498 Palaeoglaciology of an ice sheet through a glacial cycle: the Washington, in relation to theory and experiment. Journal of 499 European ice sheet through the Weichselian. Quaternary Science Geophysical Research 64, 1891Á1909. Reviews 20, 591Á625. Kitagawa, H. & van der Plicht, J. 1998: A 40,000 year varve 500 14 501 Clark, C. D., Evans, D. J. A. & Piotrowski, J. A. 2003: Palaeo-ice chronology from lake Sigetsu, Japan: extension of the C 502 streams: an introduction. Boreas 32,1Á3. calibration curve. Radiocarbon 40, 515Á550. 503 Davydova, N. N., Subetto, D. A. & Khomutova, V. I. 1998: Late Kjær, K. H., Demidov, I., Houmark-Nielsen, M. & Larsen, E. 2001: 504 Pleistocene stage in the evolution of small lakes in Russia’s North- Discrimination between easterly- and westerly-flowing Valdaian ice 505 West. In Rumyantsev, V. A. (ed.): The History of Pleistocene Lakes streams in Arkhangelsk region, Northwest Russia. Global and 506 of the East European Plain, 166Á193. Nauka, St. Petersburg (in Planetary Change 31, 201Á214. 507 Russian). Kjær, K. H., Demidov, I. N, Larsen, E., Murray, A. & Nielsen, J. K. 508 Demidov, I. N. & Lavrova, N. B. 2001: Composition of Quaternary 2003: Mezen Bay Á a key area for understanding Weichselian 509 cover in the basin of Vodla River (Eastern Karelia) and peculia- glaciations in northern Russia. Journal of Quaternary Science 18, 510 rities of vegetation development at lateglacial and postglacial time. 73Á93. 511 In Chervyakov, O. V., Chervyakov, O. V., Antipin, V. K. & Loginov, Kjær, K. H., Larsen, E., Funder, S., Demidov, I., Jensen, M., 512 K. K. (eds.): National Park ‘Vodlozersky’: Natural Diversity and Ha˚kansson, L. & Murray, A. 2006: Eurasian ice sheet interaction 513 Cultural Heritage,49Á60. Kar RC RAN, Petrozavodsk (in in northwestern Russia throughout the late Quaternary. Boreas 35, 514 Russian).UNCORRECTED000Á000 PROOF (this issue). C:/3B2WIN/temp files/SBOR43_S100.3d[x] Tuesday, 23rd May 2006 14:39:55

BOREAS 35 (2006) Last Scandinavian Ice Sheet, NW Russia 19

515 Koshechkin, B. I., Devyatova, E. I., Kagan, L. Ya. & Punning, Ya. Ramsay, W. 1911: Beitra¨ge zur Geologie der Halbinsel Kanin. Fennia 516 M. 1977: Lateglacial marine transgression in Onezhskoe Belom- 21,1Á66. 517 orie, Onega Bay of White Sea. In Biske, G. S. (ed.): Quaternary Sejrup, H. P., Haflidason, H., Aarseth, I., King, E., Forsberg, C. F., 518 Stratigraphy and Palaeogeography of North of European Part of Long, D. & Rokoengen, K. 1994: Late Weichselian glaciation 519 USSR,5Á16. KFAN SSSR, Petrozavodsk (in Russian). history of the northern North Sea. Boreas 23,1Á13. 520 Krasnov, I. I. (ed.) 1971: Map of the Quaternary Deposits of the Sejrup, H. P., Larsen, E., Haflidason, H., Berstad, I. M., Hjelstuen, B. 521 European Part of the USSR. Scale 1:1.5000000. Ministerstvo O., Jonsdottir, H., King, E. L., Landvik, J., Longva, O., Nyga˚rd, 522 geologii SSSR. A., Ottesen, D., Raunholm, S., Rise, L. & Stalsberg, K. 2003: 523 Kru¨ger, J. & Kjær, K. H. 1999: A chart for field description and Configuration, history and impact of the Norwegian Channel Ice 524 genetic interpretation of glacial diamicts and associated sediments Stream. Boreas 32,18Á36. 525 Á with examples from Greenland, Iceland and Denmark. Boreas Spiridonov, M. A. & Jakovleva, S. V. 1961: Quaternary deposits of 526 28, 386Á402. the Kanin Peninsula shore and River Pyoza basin. Materialy po 527 Kvasov, D. D. 1975: Late Quaternary History of Large Lakes and chetvertichnoi geologii i geomorfologii SSSR 3,75Á89 (in Russian). 528 Inland Seas of Eastern Europe. 278 pp. Nauka, Leningrad. Stokes, C. R. & Clark, C. D. 1999: Geomorphological criteria for 529 Lagerlund, E. 1987: An alternative Weichselian glaciation model, identifying Pleistocene ice streams. Annals of Glaciology 28,67Á74. 530 with special reference to the glacial history of Ska˚ne, south Stokes, C. R. & Clark, C. D. 2003: Laurentide ice streaming on the 531 Sweden. Boreas 16, 433Á459. Canadian Shield: a conflict with the soft bedded ice stream 532 Larsen, E., Lysa˚, A., Demidov, I., Funder, S., Houmark-Nielsen, M. paradigm? Geology 4, 347Á350. & Kjær, K. H. 1999: Age and extent of the Scandinavian Ice Sheet 533 Svendsen, J. I., Alexanderson, H., Astakhov, V. I., Demidov, I., in north-west Russia. Boreas 28, 115Á132. 534 Dowdeswell, J. A., Funder, S., Gataullin, V., Henriksen, M., Hjort, 535 Larsen, E., Kjær, K. H., Demidov, I., Funder, S., Grøsfjeld, K., C., Houmark-Nielsen, M., Hubberten, H. W., Ingo´lfsson, O´ ., 536 Houmark-Nielsen, M., Jensen, M., Linge, H. & Lysa˚, A. 2006a: Jakobsson, M., Kjær, K. H., Larsen, E., Lokrantz, H., Lunkka, J. Late Pleistocene glacial and lake history of northwestern Russia. 537 P., Lysa˚, A., Mangerud, J., Matioushkov, A., Murray, A., Mo¨ller, Boreas 35, 000Á000 (this issue). 538 P., Niessen, F., Nikolskaya, O., Polyak, L., Saarnisto, M., Siegert, 539 Larsen, E., Kjær, K. H., Jensen, M., Demidov, I., Ha˚kansson, L. & C., Siegert, M. J., Spielhagen, R. F. & Stein, R. 2004: Late 540 Paus, Aa. 2006b: Early Weichselian palaeoenvironments recon- Quaternary ice sheet history of northern Eurasia. Quaternary structed from a mega-scale thrust-fault complex, Kanin Peninsula, 541 Science Reviews 23, 1229Á1271. northwestern Russia. Boreas 35, 000Á000 (this issue). 542 Tveranger, J., Astakhov, V. & Mangerud, J. 1995: The margin of the Lavrov, A. S. 1991: Map of the Quaternary Deposits. Sheet Mezen 543 last BarentsÁKara ice sheet at Markhida, Northern Russia. Scale 1: 1000000. VSEGEI, Leningrad (in Russian). 544 Quaternary Research 44, 328Á340. Lavrov, A. S. & Potapenko, L. M. 2005: Neopleistocene of North-East 545 Ukkonen, P., Lunkka, J. P., Ljunger, H. & Donner, J. 1999: New of the Russian Plain. 222 pp. Aerogeologia, Moscow (in Russian). 546 radiocarbon dates from Finnish mammoths indicating large ice- 547 Legkova, V. G. & Schukin, L. A. 1972: Belts of terminal formations free areas in Fennoscandia during the Middle Weichselian. Journal in north-western part of Arkhangelsk district. In Goretsky, G. I., 548 of Quaternary Science 14, 711Á714. 549 Pogulyaev, D. I. & Shik, S. M. (eds.): Kraevye obrazovania Velichko, A. A., Kononov, Y. M. & Faustova, M. A. 1997: The Last 550 materikovyh oledenenii, 202Á205. Nauka, Moscow (in Russian). Glaciation on Earth: size and volume of ice sheets. Quaternary 551 Linge, H., Larsen, E., Kjær, K. H., Demidov, I., Brook, E. J., 10 International 41/42,43Á51. 552 Raisbeck, G. M. & Yiou, F. 2006: Cosmogenic Be exposure Wohlfarth, B., Filimonova, L., Bennike, O., Bjorkman, L., Brunn- 553 dating across Early to Late Weichselian ice-marginal zones in 554 northern Russia. Boreas 35, 000Á000 (this issue). berg, L., Lavrova, N., Demidov, I. & Possnert, G. 2002: Late- 555 Lunkka, J.-P., Saarnisto, M., Gey, V., Demidov, I. & Kiseleva, V. glacial and early Holocene environmental and climatic change from Lake Tambichozero in southeastern Russian Karelia. Qua- 556 2001: The area and timing of the Last Glacial maximum in the ternary Research 58, 261Á272. 557 Valdaian (Weichselian) cold stage in Vologda and adjacent areas of 558 the NW Russian Plain. Global and Planetary Change 31, 407Á426. Wohlfarth, B., Schwark, L., Bennike, O., Filimonova, L., Tarasov, P., 559 Lysa˚, A., Demidov, I., Houmark-Nielsen, M. & Larsen, E. 2001: Late Bjorkman, L., Brunnberg, L., Demidov, I. & Possnert, G. 2004: 560 Pleistocene stratigraphy and sedimentary environment of the Unstable early Holocene climatic and environmental conditions in 561 Arkhangelsk area, northwest Russia. Global and Planetary Change northwestern Russia derived from multidisciplinary study of a lake- sediment sequence from Pichozero, southeastern Russian Karelia. 562 31, 179Á199. 563 Lyutkevitch, E. M. 1947: The history of Kanin peninsula in The Holocene 14, 732Á746. 564 Quaternary time. Izvestia VGO 79,8Á73 (in Russian). Yakovlev, S. A. 1956: The Fundamentals of the Quaternary Geology of 565 Mark, D. M. 1973: Analysis of axial orientation data including till the Russian Plain Á Stratigraphy. 314 pp. GOSGEOLTEHIZDAT, Moscow (in Russian). 566 fabrics. Geological Society of America Bulletin 84, 1369Á1374. 567 Mark, D. M. 1974: On the interpretation of till fabric. Geology 2, Zatulskaya, T. Yu. 1984a: State Map of the Quaternary Deposits. 568 101Á104. Sheet Q-38-XXV-XXVI. Pinega. Scale 1:200000. List Q-38-XXV- 569 Murray, A. S., Marten, R., Johnston, A. & Martin, P. 1987: Analysis XXVI. Pinega. VSEGEI, Leningrad (in Russian). 570 of naturally occurring radionuclides at environmental concentra- Zatulskaya, T. Yu. 1984b: State Map of the Quaternary Deposits. 571 tions by gamma spectrometry. Journal of Radioanalytical and Sheet Q-38-XXXI-XXXII. Pochezer’e. Scale 1:200000. List Q-38- 572 Nuclear Chemistry 115, 263Á288. XXV-XXVI. Pochezer’e. VSEGEI, Leningrad (in Russian). 573 Murray, A. S. & Wintle, A. G. 2000: Luminescence dating of quartz Zatulskaya, T. Yu. 1984c: State Map of the Quaternary Deposits. 574 using an improved single-aliquot regenerative-dose protocol. Sheet Q-38-XXI-XXII. Mezen. Scale 1:200000. List Q-38-XXI- 575 Radiation Measurements 32,57Á73. XXII. Mezen. VSEGEI, Leningrad (in Russian). 576 Pleshivtseva, E. S. 1977: Changes of paleogeography conditions in Zekkel, Ya. D. 1939: Geological investigations in the areas of the North-Dvina lowland at late-postglacial time. Priroda i hozyaistvo White Sea Zimnii coast, River Kuloi basin and lower River Mezen. Severa 6,39Á47 (in Russian). Trudy Severnogo geologicheskogo upravlenia 3,10Á82 (in Russian). UNCORRECTED PROOF AUTHOR’S QUERY SHEET

Author(s): Igor N. Demidov et al. SBOR 43 Article title: Article no:

Dear Author

Some questions have arisen during the preparation of your manuscript for typesetting. Please consider each of the following points below and make any corrections required in the proofs.

Please do not give answers to the questions on this sheet. All corrections should be made directly in the printed proofs.

UNCORRECTED PROOF