Ice sheets normally develop in high mountainous areas under cool, moist conditions and later invade the surrounding lowlands (Ives et al., 1975; Flint, 1971). The last Scandinavian ice sheet forming under such conditions invaded the surrounding shelves and parts of north-western Russia (figure 1). However, north of mainland Russia and Siberia, ice sheets also initiated on exposed shelf areas. Soviet scientists already introduced this theory in the 1950’s when they found traces of buried ice on land in northern Russia and Siberia, indicating growth of ice sheets on the shelf and movement of ice from the present sea into the lowland further south (Yakovlev, 1956). Even though well established within the Soviet Union, the theory was considered with strong scepticism in the western countries until American scientists proposed the idea and promoted the ice-shelf hypothesis as being responsible for ice sheet formation at the beginning of the last ice-age (Denton and Hughes, 1981). Despite their subordinate sizes (2–3 million km²), the Eurasian ice sheets were an important factor in the climate system (Siegert et al., 2001). As one of the largest and most shallow areas on earth, the Eurasian shelf constitutes a climate-sensitive platform because it is the final destination for warm Atlantic surface water to the Arctic and freshwater from rivers draining the continent.

Build-up of glacier ice on the Eurasian shelf is initiated when ice sheets are developed elsewhere on land, thereby lowering the global sea level. Simultaneously, permanent sea ice over the Barents and Kara Seas thickens into a continuous ice-sheet ice-shelf system including land-based glaciers. Further thickening results in grounding of ice shelves and dispersal of glacier ice from the shelf. Following this concept, ice sheets from the Kara Sea and the Barents Sea reached the coastal areas of northern Russia during the last glacial period (Yakovlev, 1956, Denton and Hughes, 1981, Grosswald, 1998 and Svendsen et al., 1999). Reconstruction of these ice sheets in terms of extent and timing is important because waxing and waning are intimately related to climate change, and because the ice sheets serve as feedback mechanism to the climate system (Clark et al., 1999).

Fieldwork 2002

The expedition in 2002 was firstly focused on collecting samples for exposure dating along the Timan Ridge in the eastern part of our study area. This strongly eroded mountain ridge reaches approx. 300 m.a.s.l. and can be traced for more than 700 km from south to north, with continuity on the Kanin Peninsula. In the second half of the field session we worked on the northern coast of the Kanin Peninsula between the Madakhá and Krinka Rivers. The main objective of the field investigations in 2002 was to test the event-stratigraphical model erected during the last 6–7 years using a new dating technique (Kjær et al., 2002). Finally we also sought to gain new knowledge on the huge Quaternary sections found on the northern coast of Kanin Peninsula.

Following preparations in St. Petersburg and Arkhangelsk our expedition left towards the Timan Ridge on 21 June . This year, transportation was exclusively provided by a Mi-8 helicopter from the Vaskovo domestic airport in Arkhangelsk, with which a competent crew brought us to most designated destinations despite difficult ground conditions. After eight days in the field with continuous helicopter support, we were brought to the mouth of the Madakhá River on northern Kanin to work along the coastal cliff. On 11 July the camp was moved to the Krinka River by helicopter, which also retrieved us 22 July on the return way to Arkhangelsk.

TEDAP – preliminary results

Timan Exposure DAting Project (TEDAP) – Many uncertainties are still associated with the reconstruction of the ice sheets described above. However we used a new dating technique to test and expand our hypothesis on the chrono-logy of the glacial events. On the basis of our current knowledge it seems that the southernmost part of the Timan Ridge was never covered with ice during the last ice age. The central part was in a short time covered by a local Timan ice cap whereas the northern part was firstly covered by the Timan ice cap and later by the Kara Sea ice sheet. Towards the northwest, the cape of the Kanin Peninsula was transgressed by at least three ice sheets: first the Timan ice cap, then the Kara Sea ice sheet, and finally the Scandinavian ice sheet. Thus the Timan Ridge was occupied by ice over successively longer periods going from south to north. This also implies that bedrock of the ridge was exposed to cosmic rays from the atmosphere over progressively longer periods from north towards the south.

A relatively new dating technique can actually measure the age that bedrock has been exposed to atmospheric cosmic rays. The obtained exposure ages will then give the minimum duration for the time elapsed since the ice melted away in the area. When applied to bedrock in previously glaciated regions, the method has several advantages over other dating techniques: First of all, it provides a mean of dating glacially polished surfaces directly, particularly important for settings devoid of organic material suitable for radiocarbon dating, or where the deposits are older than the range of the radiocarbon method. Secondly, it provides ages in calendar years, and it can be used on deposits or surfaces several hundreds of thousand years old. Implemented on the Timan Ridge, a series of bedrock samples taken along the ridge and its extension on Kanin Peninsula is expected to show decreasing ages of the exposed bedrock.

Cosmogenic nuclides are isotopes that are produced when cosmic rays interact with the nucleus of an atom. The main production of cosmogenic nuclides is in the atmosphere, where the flux of cosmic ray particles is high, and the particles have high energy. However, cosmogenic nuclides are also produced in situ in the lithosphere, i.e. in rock surfaces exposed to cosmic rays. When the geographical and altitudinal position of a rock surface is known, the concentration of cosmogenic nuclides provides a measure of the time the surface has been exposed for cosmic rays (hence the term “exposure dating”). Common cosmogenic nuclides produced in the upper few metres of the lithosphere are Beryllium-10 (¹⁰Be) and Aluminium-26 (²⁶Al). Both are radioactive isotopes and can be used as clocks, both through their predicted accumulation with time and through their radioactive decay. Usually samples of 20 to 40 g of pure quartz provide sufficient amounts of ¹⁰Be and ²⁶Al to determine their concentration and exposure age by AMS (accelerator mass spectrometry). Quartz is the preferred material as it is ubiquitous and resists chemical weathering.

Nine exposure age estimates from three different areas on the Kanin Peninsula have been obtained so far (figure 3). Until more dates can be obtained the data set is somewhat inconclusive. The three northernmost samples are from erratic boulders, whereas the rest are from in situ bedrock. This means basically that the northernmost samples can be fitted into our glaciation model provided that the boulders derive from the two oldest Weichselian advances. This cannot be tested at present. The easternmost three samples fit our glaciation model, being somewhat younger than the last Barents/Kara advance. The spread in ages in the three westernmost samples could be explained by either the 18 ka date being too young, or by it being correct, and allowing that the Scandinavian ice sheet penetrated a little more eastwards than assumed (figure 1 and 3). If correct, this  means that the two sites giving old ages were covered by, but not eroded by the Scandinavian ice sheet. At present we consider the first explanation to be the most likely.

The section on northern Kanin – preliminary results

Without doubt, the most spectacular section in the Arkhangelsk region is found on the northern shores of the Kanin Peninsula (figure 1 and photo 2). The section stretches over 35–40 kilometres where the coast meets the Barents Sea with an east-west orientation and is typically 40 metres high, although it occasionally rises to 80 metres. The combination of tidal coastal erosion and permafrozen sediments results in strong erosion and well-exposed sections. In 2001 we collected a photo mosaic of the section between Madakhá and Krinka – covering nearly 20 kilometres – which this year was used to document the structural architecture of the different sedimentary units (figure 5). In addition, 32 detailed sedimentary logs were recorded from critical areas along the section together with sampling for luminescence dating (OSL). Previous work by Ramsay (1904) and Spiridonov and Yakovleva (1961) suggested that both marine, lacustrine and several till units are present in the section. However, using a modern structural approach it turned out that a major part of the section is built up of a repetitive succession of the same till sheet and interbed (figure 5 and photo 2). This means that a glacier from the Barents Sea probably dislocated Eemian sediments and till from a western-northwestern direction. The section is capped by a marine and lacustrine basin discordantly overlain by a till unit. If our preliminary hypothesis proves to be correct, northern Kanin shows an unprecedented scale of glaciotectonic dislocation including low-angle thrust and nappe structures with a lateral displacement between 20–40 kilometres, which have never been reported before in the literature.