Introduction

Our collaborative Swedish-American-Russian project seeks to clarify the long-debated northern extent of the Kara Sea Ice Sheet during the Last Glacial Maximum (LGM, approximately 20,000 years BP) and earlier glacial intervals (see eg Velichko et al., 1984; Grosswald, 1998). Although substantial research efforts in the Kara Sea region in recent years have shown limited glacial ice in the southern Kara Sea, little data is available on the north (see eg. Svendsen et al., 1999; Alexanderson et al., 2001; Forman et al., 2002; Polyak et al., 2002; Stein et al., 2002; Lokrantz et al., 2003). Ongoing uncertainty in the configuration of the Kara Sea Ice Sheet and the adjoining Barents Sea Ice Sheet has resulted in striking discrepancies among ice sheet models during the LGM and earlier cycles. These discrepancies hinder understanding of global changes in ice distribution and sea level through time. The still poorly known paleogeographic conditions also limit numerical climate model simulations of past environmental change. To help resolve these problems, we are performing focused geomorphic, stratigraphic, and geochronologic studies of Severnaya Zemlya, a high Arctic archipelago in the northeastern Kara Sea at the rim of the Arctic Ocean (figure 1). Our studies are evaluating the recorded limits and timing of glacial and deglacial events and associated changes in relative sea level.

Our team is concentrating on one field region, southern October Revolution Island (figure 1). Russian studies of this region in the 1970s and 1980s showed an especially rich archive of glacial and marine deposits associated with more than one glacial event (see eg Bolshiyanov and Makeyev, 1995). However, these studies were not able to definitively determine the genesis and age of the deposits and the sequence of events. Our group will resolve the history of this region through additional studies using recent improvements in Quaternary dating methods and refinements in conceptual models of sedimentation. Our field efforts in 2002 and 2003 are focusing on solving two key questions:

• What are the magnitude and timing of glacier loading on Severnaya Zemlya?
• Was Severnaya Zemlya an ice cap/ice-sheet nucleation area during the LGM or earlier periods of ice expansion?

2002 field results

During our first field season in the summer of 2002 we collected data to constrain the age and elevation of raised beaches that occur up to about 135 m.a.s.l. and the age and genesis of a 40 m thick sediment sequence of alternating raised marine sediments and glacial tills in the Ozernaya River Valley. More information about the project background and 2002 field results can be found in Ingólfsson et al. (2002) and Möller et al. (2003).

2002-2003 laboratory results

Laboratory work focused on dating samples collected in the field. Radiocarbon dates on 13 samples of marine molluscs and whalebone from the beaches and the sediment sequence, yielded reservoir-corrected ages > 37 ka (ka = thousand years BP), some as old as > 49 ka. Given that tiny amounts of modern carbon contamination can cause “infinitely” aged shell and bone samples to yield “finite” ages in the > 37 ka range, we consider all of these ages as minimum estimates. Even the lowest beaches at 12 m.a.s.l. yielded ages > 40 ka. These radiocarbon results, combined with the lack of evidence for “young” beaches near the modern shoreline, rule out the possibility of substantial ice-sheet loading in the northeastern Kara Sea during the LGM. Thus our data support the basic interpretations of Bolshiyanov and Makeyev (1995) and Raab et al. (2000) for limited LGM ice. Nevertheless, at least one very large glacial loading event predating the LGM must have occurred given that raised beaches occur up to around 135 m.a.s.l. Even more pre-LGM glacial events are clearly recorded in the sediment sequence along the Ozernaya River. Here, the three glacial tills mapped in 2002 (T-1, T-2, and T-3) are separated by three raised marine deposits. Preliminary optically stimulated luminescence dates suggest that the oldest marine sediments (M-1) are >250-400 ka old, possibly older than 700 ka.

2003 field results

Field efforts in 2003 focused on testing and refining the results of the 2002 season, particularly to date more accurately the older-than-expected sediments. New samples of marine mollusks, sediment, and bone were collected from beaches and sediment sections in the Ozernaya River Valley region for optically stimulated luminescence and amino acid dating. More intensive studies of existing sediment sections and explorations of new sections are being used to assemble a more complete sedimentary history (photo 1). Much work was performed on the main sections at Site 29 (named by Bolshiyanov and Makeev, 1995) that were visited in 2002 as well.

Raised beaches

Ancient raised beaches occur between approx. 12–135 m.a.s.l., with the most prominent beach sequences at approximately 12, 30–40, 85, 100, 130, and 140 m.a.s.l. (figure 2). Additional samples collected in 2003 from the higher beaches will permit more accurate age dating to determine whether the overall beach sequence represents a response to a single glacier unloading event or to multiple ones. Our hypothesis is that at least two events are represented. The most prominent shoreline at around 30–40 m.a.s.l. might demarcate a younger event that extends down to at least 12 m. The 30–40 m shoreline can be traced in the field, as well as on satellite and air-photos, for tens of km. In varies from a broad beach to an erosional escarpment of few meters. Geochronological samples from many of the prominent beach sequences are being dated to test our hypothesis.

One of the most fascinating components of the raised marine sediment history is the sudden death of a pod of 9 or 10 Narwhal whose closely spaced and mostly complete skeletons occur at about 50 m.a.s.l. (first described by Bolshiyanov and Makeev, 1995 and also studied in 2002). The skeletons do not lie on a beach but in deeper water sediments, which suggests that all died when sea ice in the fjord became too thick and sealed their breathing holes. Additional geochronologic samples collected from the Narwhal site and from apparently correlative sediments across the river will provide a more precise age on the death of the whales while simultaneously yielding more insight into the timing and magnitude of glacial unloading. New stratigraphic sites in the region are also improving our understanding of the geomorphic history of what was once a fjord-shaped estuary but which now lies well a.s.l. A more complete spatial view will improve our understanding of the origins of the sediments and their environmental history.

Glacial Stratigraphy

In both 2002 and 2003 much effort was put into studying the stratigraphy along the Ozernaya River, particularly at Site 29. In 2002, we mapped three marine units, divided by three glacial tills. In 2003, we mapped much more of the lowermost marine unit (M-1) and extended it down to a fourth till (T-0) near the level of the riverbed. We also studied the uppermost diamicton (T-3) in detail, showing its glacial origin.

A brief description of the full sequence is as follows (figure 3). The lowermost till (T-0) is overlain by clayey-to-silty marine sediments and shallow-marine stratified sand (M-1) with in situ mollusc shells. Unit M-1 is tectonized by a glacier which advanced southwards from inland Severnaya Zemlya towards the Kara Sea, depositing till T-1. Clayey-to-sandy marine sediments (M-2) in turn overlie this till, containing some driftwood, in situ mollusks, and sporadic dropstones. Clast fabrics from the above-lying till (T-2) show that also this glacier advance was southwards into the Kara Sea. The uppermost marine unit (M- 3) is a sandy sublittoral deposit and beach gravels. The beach gravels are covered in turn by a thin till (T-3).

Tentative interpretation

There is a valuable and complex Late Quaternary glacial record on southern October Revolution Island, as shown by the presence of four diamictons separated by marine sediments in the Ozernaya River Valley, and multiple raised beaches on the adjacent uplands between about 12–135 m.a.s.l.

Till T-0 is immediately overlain by a coarsening upward marine sequence (M-1) capped by thick shallow water sands up to at least 65 m.a.s.l. These occurrences suggest a substantial ice-loading event (T-0) followed by a period of glacial isostatic recovery (M-1). Preliminary optically stimulated luminescence and electron spin resonance dating suggests an age > 250 ka for M-1, implying that T-0 was deposited before that time. The age of the T-0 glacial event (and all subsequent events) will be better constrained with our ongoing dating program. Another goal of that program is to help tie the stratigraphic record to the raised beach sequence, thereby allowing us to better estimate the magnitude of ice loading for one or more events.

Till T-1 does not follow the till T-0 pattern, as it is overlain by a marine unit (M-2) lacking a clear coarsening upward sequence. Preliminary optically stimulated luminescence and electron spin resonance dating suggests an age ≥ 180 ka for M-2, implying that T-1 was deposited before that time. Additional dating will help test whether the lack of coarse sediments in the upper portions of M-2 are indicative of incomplete glacial isostatic recovery. Glacial tectonism of the uppermost portion of the M-1 unit underneath T-1 shows clear southward ice flow.

Till T-2 is immediately overlain by a coarsening upward marine sequence (M-3) capped by beach gravels at > 80 m. This elevation implies a substantial ice loading event and subsequent isostatic recovery. Clast fabric in T-2 shows southward ice flow. Preliminary optically stimulated luminescence and electron spin resonance dating suggests an age ≥ 90 ka for M-3, implying that T-2 was deposited before that time. The loading and flow directions imply that local ice caps grew and expanded (nucleation area) and probably coalesced with ice from Taymyr and Novaya Zemlya to form a Kara Sea Ice Sheet.

Till T-3 at the top of the sequence may have formed from a single glacial event or multiple events. We cannot rule out a LGM event as forming all or part of this till.

Radiocarbon ages > 40 ka in the lowest raised beaches at 12 m combined with the lack of “young” raised beaches near the present shoreline show that ice loading during the LGM was not substantial.

Future work

The glaciation history is intriguing and the age uncertainty will be narrowed considerably in the coming year with our geochronological programme involving multiple dating methods. Optically stimulated luminescence ages will be tested against those generated from electron spin resonance and amino acids (A/I-D/L and Aspartic Acid techniques). Some paleomagnetic samples will also be analyzed. A solid geochronologic framework will allow comparisons with the stratigraphy of surrounding regions, including the Taymyr Peninsula to the south (eg, Alexandersson et al., 2002; Hjort et al., 2003).