late quaternary ice sheet history and dynamics in central ...311283/fulltext01.pdf · late...

Late Quaternary ice sheet history and dynamics in central and southern ScandinaviaTimothy F. Johnsen

Doctoral Thesis in Quaternary Geology at Stockholm University, Sweden 2010

Dissertations from the Department of Physical Geography and Quaternary Geology No 22

Late Quaternary ice sheet history and dynamics

in central and southern Scandinavia

Timothy F. Johnsen

Doctoral Dissertation 2010 Department of Physical Geography and Quaternary Geology

Stockholm University

To Mike

© Timothy F. Johnsen ISBN: 978-91-7447-068-0 ISSN: 1653-7211 Paper I: © Swedish Society for Anthropology and Geography Paper II: © The Boreas Collegium Layout: Timothy F. Johnsen (except for papers I and II) Cover photo: View from Handöl, Sweden to north-northeast over the shoreline of Lake Ånnsjön and to Mt. Åreskutan at background right (Timothy F. Johnsen, May 2005) Printed in Sweden by PrintCenter US-AB

Late Quaternary ice sheet history and dynamics in central and southern Scandinavia

Doctoral dissertation 2010 Timothy F. Johnsen Department of Physical Geography and Quaternary Geology Stockholm University

Abstract

Recent work suggests an emerging new paradigm for the Scandinavian ice sheet (SIS); one of a dynamically fluctuating ice sheet. This doctoral research project explicitly examines the history and dynamics of the SIS at four sites within Sweden and Norway, and provides results covering different time periods of glacial his-tory. Two relatively new dating techniques are used to constrain the ice sheet history: the optically stimulated luminescence (OSL) dating technique and the terrestrial cosmogenic nuclide (TCN) exposure dating tech-nique.

OSL dating of interstadial sediments in central Sweden and central Norway indicate ice-free conditions during times when it was previously inferred the sites were occupied by the SIS. Specifically, the SIS was absent or restricted to the mountains for at least part of Marine Isotope Stage 3 around 52 to 36 kyr ago. Inland portions of Norway were ice-free during part of the Last Glacial Maximum around 25 to 20 kyr ago.

Consistent TCN exposure ages of boulders from the Vimmerby moraine in southern Sweden, and their compatibility with previous estimates for the timing of deglaciation based on radiocarbon dating and varve chronology, indicate that the southern margin of the SIS was at the Vimmerby moraine ~14 kyr ago.

In central Sweden, consistent TCN ages for boulders on the summit of Mt. Åreskutan and for the earlier deglaciated highest elevation moraine related to the SIS in Sweden agree with previous estimates for the timing of deglaciation around 10 ka ago. These results indicate rapid decay of the SIS during deglaciation. Unusually old radiocarbon ages of tree remains previously studied from Mt. Åreskutan are rejected on the basis of incompatibility with consistent TCN ages for deglaciation, and incompatibility with established paleoecological and paleoglaciological reconstructions.

Altogether this research conducted in different areas, covering different time periods, and using compara-tive geochronological methods demonstrates that the SIS was highly dynamic and sensitive to environmental change. Keywords: Scandinavian ice sheet, ice sheet dynamics, luminescence dating, cosmogenic exposure dating, geochronology, moraine, interstadial, deglaciation, nunatak

5

Late Quaternary ice sheet history and dynamics

in central and southern Scandinavia

Timothy F. Johnsen

Department of Physical Geography and Quaternary Geology, Stockholm University, Sweden This doctoral thesis consists of a summary and four appended papers. The papers are listed below and are referred to as Paper I-IV in the summary.

Paper I: Johnsen, T.F., Alexanderson, H., Fabel, D., Freeman, S.P.H.T. 2009. New 10Be cosmogenic ages from the Vimmerby moraine confirm the timing of Scandinavian Ice Sheet deglaciation in southern Sweden. Geografiska Annaler: Series A, Physical Geography, 91: 113–120. – Reprinted with permission of the Swed-ish Society for Anthropology and Geography. Paper II: Alexanderson, H., Johnsen, T., Murray, A.S. 2010. Re-dating the Pilgrimstad Interstadial with OSL: a warmer climate and a smaller ice sheet during the Swedish Middle Weichselian (MIS 3)? Boreas, 39: 367–376. – Reprinted with permission of The Boreas Collegium. Paper III: Johnsen, T.F., Olsen, L., Murray, A., Submitted. OSL ages in central Norway confirm a MIS 2 interstadial (25-20 ka) and a dynamic Scandinavian ice sheet. Quaternary Science Reviews.

Paper IV: Johnsen, T.F., Fabel, D., Stroeven, A. High-elevation cosmogenic nuclide dating of the last de-glaciation in the central Swedish mountains: implications for the timing of tree establishment. Manuscript.

Timothy F. Johnsen

6

Introduction

Ice sheets are a crucial component of the function-ing of the Earth system (Oerlemans and van der Veen 1984). Their large size displaces vast areas of plants and animals (Robertsson 1994, Hewitt 2000) and changes the albedo and climate of the Earth (Manabe and Broccoli 1985, Ruddiman 2003). Tremendous quantities of water from the oceans are stored on land as ice, causing global sea levels to lower over 100 metres, and their massive weight depresses the surface of the Earth hundreds of metres which leads to flooding of coastal areas and the diversion of rivers (Lambeck and Chappell 2001). Immense quantities of meltwater can be discharged, disrupting ocean circulation and the climate system and resulting in sudden sea level rise (Fairbanks 1989). And, huge areas of the land-scape are altered and shaped by the erosional and depositional activity of ice sheets (Lundqvist 2002). In addition, our own past is closely linked to that of ice sheets and ice dynamics. Modern hu-mans evolved during the Late Quaternary, a period characterized by glaciations and rapid climatic and environmental shifts resulting in great changes in the distribution of organisms. The present genetic structure of populations, species and communities has been mainly formed by Quaternary ice ages (Hewitt 2000).

In recent time dramatic changes in the margins of the Greenland and Antarctic ice sheets have occurred including rapid but episodic glacier accel-eration and thinning from their marine-terminating sectors (Shepherd and Wingham 2007): e.g., the collapse of sections of the Larsen Ice Shelf in Ant-arctica (Rott et al. 1996), and loss of about 100 Gt yr-1 of mass from the the Greenland ice sheet (Shepherd and Wingham 2007). The rate of mod-ern changes of ice sheets is occurring faster than many scientists anticipated, and have made it easier to imagine dynamic glacier activity for the past. As well, a major change in our understanding of the dynamics of ice sheets occurred when present and past ice streams were recognized for their impor-tance in the mass balance of ice sheets and dy-namic behaviour (e.g., Bentley 1987, Stokes and Clark 2001, Alley et al. 2004). Decay of modern ice sheets along with global climate warming is set to cause global sea levels to rise considerably dur-ing this century and potentially displace tens of

millions of humans in coastal regions (IPCC 2007). These processes along with others related to global climate change remind us of how we as a species are linked to the activities of ice sheets that in turn at least partially reflect our own activity, and that our fate is tied to how the Earth and its ice sheets and climate will behave. Despite intensive scien-tific efforts it remains difficult to accurately predict the magnitude and rate of changes for the future, and this uncertainty is directly tied to our under-standing of the how the Earth system is operating and has operated. Answering important questions about the future climate and conditions on Earth are inextricably linked to our understanding of how the Earth has operated in the past. Thus, by under-standing the dynamics of past ice sheets, as in this doctoral research project, we will better predict how modern ice sheets will respond to climate change and affect society.

During the Quaternary Period multiple glaci-ations of varying spatial extent occurred in Scandi-navia starting in the Early Quaternary (Mangerud et al. 1996, Kleman et al. 2008) and with the first glaciation reaching the shelf-edge occurring ~1.1 Ma (Sejrup et al. 2000). As ice sheets are effective agents of erosion the best evidence for glaciations is from the most recent glaciation, the Weichselian, spanning from ~117 to 11.7 ka, and the reconstruc-tion of glacial history prior to the Last Glacial Maximum (LGM) is in many cases difficult and ambiguous (Fig. 1 and 2). Numerous deposits and landforms related to the last deglaciation dominate the landscape (Fredén 2002) while reconstruction of earlier ice sheet activity mostly relies on discov-ery and study of terrestrial sediments or landforms that have managed to survive being overrun by the Scandinavian ice sheet (SIS; e.g., Lagerbäck 1988, Robertsson and García Ambrosiani 1992, Kleman and Stroeven 1997, Olsen et al. 2001b, Hättestrand and Stroeven 2002, Lundqvist and Robertsson 2002, Heyman and Hättestrand 2006, Lokrantz and Sohlenius 2006), marine sediments (e.g., Sejrup et al. 1994, Baumann et al. 1995, Vorren and Laberg 1997), and ice sheet modelling (e.g., Holmlund and Fastook 1995, Kleman et al. 1997, Lambeck et al. 1998, Boulton et al. 2001, Siegert et al. 2001, Charbit et al. 2002, Näslund et al. 2003). With the exception of the last deglaciation, the timing of ice sheet advances and retreats is based mainly on correlation to ‘global’ continuous records such as


7

ice cores or marine sediment cores. This is partly due to the difficulty in dating glacial sediments, partly to the relative scarcity of datable deposits older than the LGM, and the age limit (~50 ka) of the popularly employed radiocarbon dating tech-nique. Altogether a general picture of the stadials and interstadials of the Weichselian glaciation has been produced although reconstructions of older stadials and interstadials are partly hypothetical (Lundqvist 1992, Mangerud 2004; Fig. 2). A prob-lem with correlation of more or less global records for climate change and inferred glacial response is that information on the local variation in timing, and regional leads or lags is missing. In order to fill this gap absolute dates for local and regional events must be acquired to build a more detailed glacial history and a better understanding of how the (gla-cial/climatic) system operates.

Lately, several studies have shown that the Mid- and Late Weichselian glacial history may be more complex than generally believed. For example, the

western margin of the ice sheet was highly dy-namic with multiple ice-free periods during the last 55 ka including around the LGM (Olsen et al. 2001a,b, 2002). Large moraine systems from the southeast portion of the ice sheet may be younger than previous estimates (Rinterknecht et al. 2006). There were possibly ice-free conditions around the LGM in southern Sweden (Alexanderson and Murray 2007) and southern Norway (Bøe et al. 2007). Based on dated stratigraphic sites and 2-D ice sheet modelling the MIS 4 ice sheet may not have persisted into the MIS 3 (Arnold et al. 2002), and more recent work suggests ice-free conditions during MIS 3 in Finland (Helmens et al. 2007a,b, Lunkka et al. 2008, Salonen et al. 2008) and Den-mark (Kjær et al. 2006) supported by numerous AMS radiocarbon dated mammoth throughout Scandinavia (Ukkonen et al. 2007, Wohlfarth 2010), although ice extents during MIS 3 are not agreed upon. Tree mega-fossils dated from high elevation areas in central Sweden suggest ice-free conditions as early as 17 cal. ka BP (Kullman 2002) although objected to (Birks et al. 2005). Modelling results of the Late Weichselian SIS (Lambeck et al. 1998) indicate that the ice sheet was much thinner than earlier estimates (Denton and Hughes 1981), while cosmogenic nuclide dat-ing results from central Norway indicate that the ice sheet was not thin enough to expose large areas of high elevation alpine land during the LGM in-terval (Goehring et al. 2008) as proposed by earlier workers (Nesje and Dahl 1990). Archaeological evidence suggests that humans may have inhabited Finland as far back as the Eemian interglacial (~120-130 ka; Schulz et al. 2002, cf. Pettitt and Niskanen 2005), Sweden >40 ka (Lundqvist 1964, Heimdahl 2006), and Arctic Russia ~35-40 ka (Pavlov et al. 2001).

Recent robust modelling of the nearby British-Irish ice sheet has produced a highly dynamic ice sheet with numerous binge/purge, advance/retreat (i.e., yo-yo) cycles dominated by ice streaming (Hubbard et al. 2009). Phases of predominant ice streaming activity coincide with periods of maxi-mum ice extent and are triggered by abrupt transi-tions from a cold to relatively warm climate, result-ing in major iceberg/melt discharge events (Hub-bard et al. 2009). The fjord-dominated landscape of Norway and its shelf are in no doubt a record of the dominance of ice streams in draining the SIS

Fig. 1: Chronostratigraphy of northwestern Europe (afterMangerud 1991), also showing the oxygen isotope stages, stadials and interglacials.

Timothy F. Johnsen

8

Fig. 2: General history of the Scandinavian ice sheet over the Weichselianglaciation. All maps are shown with modern sea level. Black dots are studyarea locations (Fig. 3). The Younger Dryas outline is according to Andersen et al. (1995) and Mangerud (2004), ages after Rasmussen et al. (2006). TheLGM outline is according to Mangerud (2004) and Vorren and Mangerud (2007), ages after Clark et al. (2009). Remaining stages are after Lundqvist(1992), Mangerud (2004), and Vorren and Mangerud (2007), isotope stageages from Martinson et al. (1987). The pre-LGM outlines are hypotheticalalthough using interpretations of known sites.


9

(Longva and Thorsnes 1997, Ottesen et al. 2005, Kessler et al. 2008, Kleman 2008; Fig. 3), and provide the setting for a potentially dynamic and responsive ice sheet. Dynamic ice sheet behaviour has been documented for Norway as well (Olsen et al. 2001a). Modelling of the SIS suggests that it is strongly sensitive to small-scale ice-sheet instabil-ity (Charbit et al. 2002). Perhaps the SIS (at least its western margin) behaved in an even more dy-namic manner than documented so far, similar to the British-Irish ice sheet.

Therefore, the objective of this doctoral thesis work is to improve our understanding of the history and dynamics of the SIS for different areas and covering different time periods. This will be achieved by employing two dating techniques, Optically Stimulated Luminescence dating (OSL) and Terrestrial Cosmogenic Nuclide (TCN) expo-sure dating. Thus, the main research questions stemming from this objective include:

� Was the behaviour of the SIS characterized as

stable or dynamic? For example, did its mar-gins wax and wane often as modelled for the British-Irish ice sheet, or infrequently?

� Was the ice sheet thick or thin during the Late Glacial? What was the vertical rate of deglacia-tion?

� Can OSL and TCN dating produce accurate re-sults and that are useful for improving our un-derstanding of the glacial history of the SIS? If so, when were areas deglaciated or when were they ice-free?

By answering these questions, this knowledge

can be used to constrain ice sheet models and guide further research into refining the history and dy-namics of the SIS.

Study areas

Two areas were selected for study (Fig. 2 and 3). The first area of study is in the Småland area of southern Sweden at Vimmerby and Lannaskede (Paper I), which is located in the southern portion of the former SIS. The second study area is the Jämtland-Trøndelag area of central Sweden and Norway, which includes three study sites located in this former central area of the SIS and west of the LGM ice-divide. Two of these sites contain sub-till

sediments and address ice-free periods/interstadials (Pilgrimstad and Langsmoen, Papers II and III), while the third addresses the vertical rate of glacia-tion during the Late Glacial (Mt. Åreskutan, Paper IV). Each site is the topic of one paper. All sites have complementary dating results either from the site or from nearby to allow comparison to multiple OSL or TCN dates.

Southern Sweden – Vimmerby moraine

Vimmerby: The South Swedish Upland lies be-tween the west and east coasts of southern Sweden and is 200-300 m asl. The crystalline bedrock forms an undulating landscape with isolated insel-bergs, shallow valleys and locally deep-weathered bedrock. The common occurrence of hummocky moraine in parts of the South Swedish Upland indicates widespread stagnation (dead ice) instead of active retreat during the last deglaciation (Björck and Möller 1987). The Vimmerby moraine is one of the few prominent features in the South Swedish Upland that is related to the former mar-gin of the decaying SIS (Agrell et al. 1976, Malm-berg Persson et al. 2007). The Vimmerby moraine site was selected for a number of reasons (Fig. 3). The timing and pattern of deglaciation in the South Swedish Upland is not well known (Lundqvist and Wohlfarth 2001). There is an opportunity to evalu-ate TCN dating results by comparison to other estimates for deglaciation from nearby (Lundqvist and Wohlfarth 2001), and prior to attempting to date moraines in areas where there is even less dating control (e.g., Paper IV). By dating the Vimmerby moraine, well-known deglacial chro-nologies from dated moraine systems from the west coast and detailed varve chronology on the east coast of Sweden can be better correlated across the South Swedish Upland (cf. Lundqvist and Wohlfarth 2001). As well, sandurs adjacent to the Vimmerby moraine have been dated thoroughly and give consistent OSL ages that are thousands of years older (~5-10 ka) than the estimate for degla-ciation of the Vimmerby moraine (Alexanderson and Murray 2007).

Central Sweden and Norway – Jämtland-Trøndelag area

Pilgrimstad: The Pilgrimstad stratigraphic site is an important site for paleoecological and paleogla-ciological reconstruction (Kulling 1945, Frödin

Timothy F. Johnsen

10

Fig. 3: Overview map with study site locations in southern Sweden (Vimmerby moraine) and in the Jämtland-Trøndelag area (Lansgmoen, Mt. Åreskutan, and Pilgrimstad). Locations of other sites from Finland and Sweden mentioned in the thesis are indi-cated, as well as several of the sites where bones of mammoth “M” have been found and dated to Mid- and early Late-Weichselian age. Curved large and small arrows are paleo-ice streams, solid line is the LGM ice margin, and the dashed line is the Younger Dryas ice margin (compiled from various sources by Olsen et al. 2001a). Ruunaa is just off the east edge of map at N 63°26' latitude.


11

1954, Lundqvist 1967, Robertsson 1988a,b, García Ambrosiani 1990, Wohlfarth 2010, Paper II; Fig. 3). The results from this critical central location, just east of the Scandinavian mountains and west of the former LGM ice-divide of the SIS, have implications for the SIS as a whole. It contains sub-till organic and minerogenic sediments, including lower sediments of proglacial sub-aquatic sedi-ments, and upper sediments representing a transi-tion from glaciofluvial to fluvial to lacustrine deposition (Lundqvist 1967), and possibly includ-ing some aeolian sediment (Robertsson 1988a,b). Paleoecological interpretation is of a cool, subarc-tic-arctic environment based on several proxies, mainly from the organic beds (Robertsson 1988a,b). A number of radiocarbon ages from organic-rich sediments have been at the limit of the technique and so have been considered unreliable (cf. Wohlfarth 2010). Consequently the site has been assigned an Early Weichselian age (MIS 5 a/c; >74 ka), based on pollen stratigraphic correla-tions with type sections in continental Europe. There is an opportunity to apply a complementary dating technique, OSL, on the sub-till sediments to determine the age of these sediments and ice-free conditions. And, detailed paleoenvironmental work has already been completed (Robertsson 1988a,b) and is ongoing (Wohlfarth et al. in prep.), so by accurately dating the sediments we can assign the paleoenvironmental inferences to the correct time period.

Langsmoen and Flora: Stratigraphic and geo-chronologic study of sub-till sediments from many sites throughout Norway has revealed that the western margin of the SIS behaved in a much more dynamic manner than previously believed (Olsen 1997, Olsen et al. 2001a,b, 2002). Four interstadi-als during the Middle and Late Weichselian glacia-tion mark periods of major ice retreat, and sub-till sediments from these periods have been identified and dated at many sites throughout Norway, indi-cating near-synchronous behaviour of the western margin of the ice sheet. Of these interstadials, the Trofors interstadial, which separates the LGM into two stadials, represents the strongest evidence for a dynamic ice sheet, and is the most difficult for the Quaternary community to accept. Thus, the Trofors interstadial requires more study to verify its exis-tence.

The Langsmoen and Flora sites contain sub-till sediments, are in central Norway (Trøndelag re-gion) ~110 km from the present coast, ~300 km from the LGM ice sheet outline (Fig. 2), 4 km from one another, and lie within the Nea River valley (relief ~300 m). Sediments from the Flora site represent an ice-proximal, ice-dammed glacio-lacustrine paleoenvironment, and sediments from the nearby Langsmoen site represent a fluvial del-taic ice-free paleoenvironment. Eight radiocarbon dates from fine-grained sediment within the Flora section give consistent ages of 20.9 ±1.6 cal. ka BP and indicate that these sediments were deposited during what is traditionally thought of as the LGM interval. The content of plant remains and other organics is not sufficiently high in the Langsmoen sub-till sediments to allow testing by radiocarbon dating. Thus, the Langsmoen stratigraphic site was selected for study as stratigraphically-related sedi-ments from the nearby Flora section give consis-tent radiocarbon dates that fall within the Trofors interstadial, and the sandy sediments of fluvial deltaic origin may be suitable for the OSL dating technique. Comparison of OSL and radiocarbon dating results will allow evaluation of the existence of the Trofors interstadial in this area of Norway, and an assessment of the radiocarbon dating of bulk sediment of low organic content, a method commonly, although not exclusively, used in the work of Olsen et al. (2001b).

Mt. Åreskutan: The county of Jämtland in cen-tral Sweden includes mountainous areas of moder-ate relief (~800 m) and adjacent low-relief (~100 m) rolling hill landscape with numerous lakes and peatlands. Within this area, Mt. Åreskutan is 1420 m asl and is largely isolated in its eastern position from other mountains in the Scandinavian moun-tain chain. The modern tree-line is at about 950-1050 m in this area. Remains of three tree species have been found at high-elevation alpine sites in central Sweden, principally Mt. Åreskutan (Fig. 3), that are hundreds of meters above modern tree-line, and dating to a time (as old as ~17 cal. ka BP, Kullman 2002) when it is commonly inferred that the sites were occupied by the SIS, and deglaciated 10.3 to 10.0 cal. ka BP (Lundqvist 2002). The existence of these tree remains, their elevation above modern tree-line, their species, and age po-tentially have tremendous implications for our understanding of the dynamics of the SIS, the pat-

Timothy F. Johnsen

12

tern and rate of migration of tree species, the loca-tion and role of refugia in re-establishing plants, and paleoclimatic conditions and variability. Con-sequently there has been debate on these data and what they represent (Birks et al. 2005, 2006, Kull-man 2005, 2006), but there has not been an attempt to directly assess these data against a complemen-tary dating technique, despite that the reliability of the ages is central in any accurate paleoglaciologi-cal or paleoecological reconstruction. Thus, study of the deglaciation of high-elevation sites in central Sweden, by using the TCN dating method, can allow determination of the timing of deglaciation, assessment of the thickness evolution of the ice sheet during the Late Glacial, and evaluation of radiocarbon results from high elevation tree re-mains.

Methods

The two dating methods used in this work (OSL, TCN) have the advantage of dating deposits that are directly related to the ice sheet, in addition to dating ‘ice-free’ depositional events rather than the invasion of plants or animals that radiocarbon is limited to. The age limit of these techniques greatly exceeds the radiocarbon technique, and potentially includes the entire Weichselian Glacial or older. In these respects the OSL and TCN dating techniques are complementary to the radiocarbon dating tech-nique; and, all of my sites have radiocarbon dating results or from nearby.

Samples were collected by me in the field fol-lowing standard procedures. Study also included sedimentological analysis, and geomorphological analysis using aerial photograph interpretation, GIS, and field mapping.

OSL dating

The optically stimulated luminescence (OSL) dat-ing technique (Papers II and III) determines the time elapsed since buried sediment grains were last exposed to daylight. It is a reliable chronological tool, reflected in its growing popularity in Quater-nary studies (Murray and Wintle 2000, Murray and Olley 2002, Duller 2004, Lian and Roberts 2006, Wintle 2008). However, the OSL dating technique may not always produce meaningful or good qual-ity results (e.g., Alexanderson and Murray 2009).

Sediments that were not adequately exposed to light during transport may give maximum ages due to the inheritance of a luminescence signal from a previous burial event. Likewise, in some geo-graphical areas the characteristics of the mineral (e.g., quartz) being measured may not be condu-cive to the OSL technique (Preusser et al. 2009, Alexanderson and Murray 2009, Steffen et al. 2009), for example, not satisfying the assumption that most of the luminescence signal is derived from what is termed the ‘fast’ component, or the signal is too weak/dim.

The OSL technique typically uses sandy sedi-ments; the same grain size available from sub-till sediment at the Pilgrimstad (Paper II) and Langsmoen (Paper III) sites. The sediments from these sites are potentially glacial or at least gla-cially-related, which means that the grains may not have been adequately exposed to light (i.e., the sediments were not fully bleached, or ‘zeroed’) prior to deposition (Fuchs and Owen 2008, Thrasher et al. 2009), and/or the quartz minerals may not be sensitive enough (e.g., through repeated transport and burial episodes) to produce adequate luminescence signals (Pietsch et al. 2008, Alex-anderson and Murray 2009). Thus, adjustments to the OSL protocols and additional tests were made to accommodate these potential issues and to get the most reliable dates possible; explained below. A key methodological decision was the collection and dating of multiple samples from a variety of lithofacies and positions within the sediments at each site, in order to evaluate the consistency of the results.

Defects and chemical impurities within miner-als such as quartz or feldspar can act as traps for free electrons (e.g. Aitken 1985, Preusser et al. 2009). These free electrons are produced from ionising radiation from radioactive elements in sediment and from cosmic rays. Luminescence results when some external stimulation (e.g., expo-sure to light) ejects electrons from traps to emit photons (light). The longer the mineral is exposed to ionising radiation, the greater the store of elec-trons trapped within the mineral and the stronger the luminescence signal, although the relative strength of this signal will vary aliquot-to-aliquot (an aliquot is a small portion of the sample placed on a small disc for measurements; Fig. 4). By comparing the natural luminescence of an aliquot


13

to the relationship between known doses given and the luminescence signal produced (i.e., a ‘growth curve’), the equivalent dose is determined (Fig. 4). The water content of the sediment over the entire burial history will impede ionising radiation and so must be considered in the final determination of the sediment dose rate. The cosmic ray contribution for shallow-depth samples is also considered. Then the luminescence age is calculated as the equivalent dose divided by the sediment dose rate.

To not expose the samples to light, they were taken in opaque plastic tubes and stored in black bags until opened under darkroom conditions for initial preparation. This included separating the sample into portions for OSL measurements, sedi-ment dose rate determination (gamma spectrome-try), and water content measurements. Samples

were wet-sieved below 250 μm and in some cases minerals magnetically separated using a Frantz magnetic separator. Chemical preparation to isolate quartz grains, including heavy liquid separation (2.62 g cm-3) in some cases to remove feldspars, treatment with 10% HCl for 5-30 minutes to re-move carbonates, 10% H2O2 for 15-30 minutes to remove organics, 10% and 38% HF for 60-120 minutes to remove everything but quartz and to etch the outer surface of the grains to remove the parts affected by alpha radiation, and 10% HCl again for 40 minutes to remove any fluorides that may have formed in the previous step. This was performed at the Nordic Laboratory for Lumines-cence Dating in Risø, Denmark, where the OSL measurements were also completed.

Fig. 4: (a) OSL samples were collected by hammering an opaque tube into sediments. (b) A Risø TL/OSL reader is open showingaliquots of a sample set on a numbered carrier. When in operation, the carrier rotates for a robotic lift to position individual aliquots to be irradiated, heated, or have the luminescence measured. The protocols are programmed into a sequence which is sent to a sec-ond computer that controls the reader. (c) Analysis of repeated measurements of the luminescence (inset) for an aliquot using differ-ent radiation doses results in a growth curve. The dose that would be equivalent to the natural luminescence is determined by inter-polation on this curve to derive the equivalent dose (ED). Results from repeat measurements of an identical dose (recycling points) and a zero dose (recuperation), along with other criteria, are used to assess the quality of measurements from each aliquot. Thisprocess is repeated until a population of equivalent doses can be analyzed to derive the equivalent dose for the sample. Finally, the age is derived by dividing the sediment dose rate by the equivalent dose, along with considering water content, sample depth, andother factors.

Timothy F. Johnsen

14

The samples were analysed using aliquots of quartz (63 to 250 μm) on Risø TL/OSL-readers (Fig. 4) equipped with calibrated 90Sr/90Y beta radiation sources (dose rate 0.14-0.35 Gy s-1), blue (470 ±30 nm; ~50 mW cm-2) and infrared (880 nm, ~100 mW cm-2) light sources, and detection was through 7 mm of U340 glass filter (Bøtter-Jensen et al. 2000). This system essentially automates the numerous repeated measurements, heatings and radiation doses given to multiple aliquots of each sample (i.e., the protocol). Analyses employed post-IR blue SAR-protocols (single-aliquot regen-erative-dose protocol, SAR; Murray and Wintle 2000, 2003; Banerjee et al. 2001), adapted to suit the samples based on internal methodological tests such as dose recovery and preheat experiments. The ‘post-IR blue’ portion of the protocol was added to minimize the contribution from any po-tential feldspar minerals that passed the physical and chemical treatments. A relatively high test-dose (~50 Gy) was necessary to get a statistically precise test signal since Swedish quartz is rela-tively dim (Alexanderson and Murray 2009). 100 s of illumination at 280° between cycles improved recuperation (response to zero dose), by emptying traps prior to a new radiation dose being given (Murray and Wintle 2003). Sediment dose rates were calculated from gamma spectrometry data (Murray et al. 1987) and included the cosmic ray contribution (Prescott and Hutton 1994) for shal-low-depth samples. Natural and saturated water content was measured either using a portion of sediment from the sample tube or using pF-rings (cylinder volumeters).

Blue-light stimulated luminescence signals have been shown to be made of a number of components (Bailey et al., 1997, Jain et al. 2003). If the initial part of the luminescence signal is not dominated by the fast component, inaccurate equivalent doses and ages may be produced (Choi et al. 2003, Tsu-kamoto et al. 2003) because of the differing char-acteristics of different signal components (Singa-rayer and Bailey 2003, Singarayer et al. 2004, Wintle and Murray 2006, Kitis et al. 2007). Simple component analyses of the continuous-wave OSL data from some aliquots was undertaken using SigmaPlot 10.0, based on the parameters and for-mulas of Choi et al. (2006). This allowed quantifi-cation of the fast, medium and slow components of the luminescence signal. The results of the compo-

nent analysis were used to select the peak and background portions of the luminescence signals (i.e., ‘channels’) that produced the best results for dose recovery experiments, and that gave a domi-nant fast component. Signal component analysis of some natural signals also showed the fast compo-nent to be dominant. The equivalent doses were then calculated in Risø Luminescence Analyst software and in Microsoft Excel. To be accepted aliquots had to pass rejection criteria for the recy-cling ratio, recuperation, equivalent dose error, and the signal had to be more than three sigma above the background. Decay and growth curves also had to be regular in shape. Ages were calculated using the mean and median of the equivalent dose popu-lation of accepted aliquots for each sample, as well as using the natural and saturated water contents. As well, a sensitivity analysis was completed to determine quantitatively which uncertainties have a larger effect on the age.

TCN dating

The terrestrial cosmogenic nuclide (TCN) exposure dating technique can allow determination of the amount of time that a rock surface has been ex-posed to cosmic radiation (e.g., how long ago an ice sheet deposited a boulder; Papers I and IV). Cosmic radiation causes the accumulation of 10Be in situ within quartz rock and the measurement of the concentration of 10Be against the production rate of 10Be for a given site of known latitude, ele-vation, topographic shielding, and sample thick-ness, provides determination of the exposure age (Lal 1991, Gosse and Phillips 2001). This tech-nique has proven useful in numerous studies of deglacial history and landform preservation (e.g., Phillips et al. 1997, Licciardi et al. 2001, Balco et al. 2002, Fabel et al. 2002, Clark et al. 2003, Rinterknecht et al. 2006).

The TCN dating technique is valuable to apply to the Vimmerby moraine (Paper I), and Mt. Åre-skutan and Mt. Snasahögarna (Paper IV) because glacially transported boulders are abundant in the area, and there is the opportunity to compare to results already available that used other dating techniques at or near these sites (i.e., altogether radiocarbon, varve chronology and OSL). Similar to my OSL methodology, a key decision was the collection and dating of multiple samples from each site. To minimize the risk of processes modi-


15

fying the exposure history of the boulder, such as cosmogenic nuclide inheritance (e.g. Briner et al. 2001), boulder exhumation, erosion or moraine deformation (Hallet and Putkonen 1994, Putkonen and Swanson 2003, Zreda et al. 1994), samples were from the tops of large, weathering-resistant boulders resting either on moraine crests (Papers I and IV) or bedrock (Paper IV). Boulder surfaces and adjacent ground were carefully inspected for indications of the rate and dominant mode of ero-sion. The height of quartz nodules and veins above adjacent softer rock within each boulder were used to estimate the rate of erosion. Altogether twelve samples were processed in the Glasgow Univer-sity-SUERC cosmogenic nuclide laboratory.

Using the CRONUS-Earth 10Be-26Al exposure age calculator (http://hess.ess.washington.edu), measured 10Be concentrations were converted to surface exposure ages. The different 10Be produc-tion rate scaling schemes were used to examine variation in the calculated age for each sample. The surface exposure ages were calculated using the ‘Lm’ scaling scheme which includes paleomag-netic corrections (Balco et al. 2008). Altogether ages were calculated by considering the sample thickness, latitude, weathering and erosion of the rock surface during exposure, the changing eleva-tion of the rock surface due to glacio-isostatic movement, and partial shielding of the rock surface from cosmic rays by topography and seasonal snow cover (Gosse and Phillips 2001).

Presentation of papers

My contributions to this project are stated below at the end of each paper summary, and in the ac-knowledgements section at the end of each of the four papers in the appendix.

For details of the motivations and background for each study site refer to the study area section above or the individual papers in the appendix.

Paper I: Vimmerby TCN

Johnsen, T.F., Alexanderson, H., Fabel, D., Freeman, S.P.H.T. 2009. New 10Be cosmogenic ages from the Vimmerby moraine confirm the timing of Scandinavian Ice Sheet deglaciation in southern Sweden. Geografiska Annaler: Series A, Physical Geography, 91: 113–120.

Samples for 10Be cosmogenic exposure dating were collected from six glacially transported, rounded to sub-rounded, weathering-resistant, granitic and quartzitic boulders (0.9-2.3 m b-axis; Fig. 5) that were resting on top of stable and level surfaces or on the broad crest of the Vimmerby moraine (Fig. 3 and 5). The six ages were internally consistent, ranging from 14.9 ±1.5 to 12.4 ±1.3 ka with a mean of 13.6 ±0.9 ka. Adjustments were made to these ages for the effects of surface erosion, snow burial and glacio-isostatic rebound, which causes the mean age to increase by only ~6% to ~14.4 ±0.9 ka. Thus, the southern margin of the SIS was at the Vimmerby moraine ~14 ka ago.

It is not clear which of the moraines west of the study area correlate with the Vimmerby moraine, and even less clear for moraines from the southeast portion of the ice sheet (northern Poland and the Baltic states). Nevertheless, the internal consis-tency of the six cosmogenic ages and their com-patibility with previous radiocarbon ages and varve chronology (Lundqvist and Wohlfarth 2001, Lundqvist 2002) indicate that the TCN (10Be) ex-posure dating technique works well for erratic boulders within this area and shows promise for further TCN exposure studies in southern Sweden. Sandur sediments adjacent to the moraine were recently thoroughly dated with OSL and provide consistent ages around the LGM (Alexanderson and Murray 2007). While this was stated in the paper it was not discussed; please see the discus-sion below.

My contribution to this work included boulder selection, sampling, and sample crushing by John-sen and Alexanderson, age calculations by Fabel and Johnsen, interpretations by Johnsen and co-authors, and writing mostly by Johnsen with input from co-authors.

Paper II: Pilgrimstad OSL

Alexanderson, H., Johnsen, T., Murray, A.S. 2010. Re-dating the Pilgrimstad Interstadial with OSL: a warmer climate and a smaller ice sheet during the Swedish Middle Weichselian (MIS 3)? Boreas, 39: 367–376.

Pilgrimstad, an important stratigraphic site for Weichselian history, was re-excavated to expose >4 m thick sub-till minerogenic and organic sedi-ments. The architecture and lithofacies of the sediments were described in detail. Eight units

Timothy F. Johnsen

16

were identified representing paleoenvironmental change from glaciofluvial to glaciolacustrine to lacustrine and back to fluvial or glaciofluvial depo-sition; somewhat glaciotectonized and crosscut by a clastic sandy dyke. Based on the sedimentologi-cal observations of this new section at Pilgrimstad, a single interstadial rather than two is favoured (Kulling 1967, Lundqvist 1967, Robertsson 1988a,b). As well, the sand within brecciated lacustrine sediments is likely reworked from adja-cent glaciofluvial sand through glaciotectonic processes, rather than having an aeolian origin (Robertsson 1988a,b).

Ten samples from the variety of lithofacies were collected for OSL dating. Single aliquots of

quartz were analysed using a post-IR blue SAR-protocol. Dose recovery tests were satisfactory at 1.05 ±0.04 (n = 21) with deconvolution results indicating that over 90% of the luminescence sig-nal is derived from the fast component. The OSL ages are internally consistent lying in the range 52-36 ka, except one from an underlying unit that is older; and compatible with existing radiocarbon ages, including two we measured with AMS. The mean of the OSL ages is 44 ±6 ka (n = 9). This places the interstadial sediments in the Middle Weichselian (MIS 3) and possibly corresponds to one or more of Greenland interstadials 17-10. The OSL ages cannot be assigned to the Early Weich-selian (MIS 5 a/c) as proposed by earlier workers

Fig. 5: (a) Deglaciation model for southern Sweden (Lundqvist 2002) with the newly-mapped Vimmerby moraine in the South Swedish Upland. Note that the moraine strikes across assumed ice-marginal lines and thus indicates a slightly different deglaciation pattern. (b) Geological map of the study area, emphasizing the end moraines of the Vimmerby moraine. TCN dating samples are from two sites, indicated by black circles. Map data from the National Quaternary geological database (Geological Survey of Swe-den, Permission 30-1730/2006). (c) Examples of boulders sampled from the crest of the moraine at Vimmerby and from atop a partly till-covered delta that is part of the moraine at Lannaskede. Modified from Paper I.


17

(Robertsson 1988a,b) for all reasonable adjust-ments to water content estimates and other parame-ters. These new ages indicate that, during the Mid-dle Weichselian, the climate was relatively warm and the SIS was absent or restricted to the moun-tains for at least part of MIS 3. This is supported by results from other recent studies completed in Fen-noscandia.

I provided significant input into all stages of this work including OSL sampling, preparation, sedimentological and stratigraphical work as well as writing. I analysed all the data, with some input from co-authors.

Paper III: Langsmoen OSL

Johnsen, T.F., Olsen, L., Murray, A. Submitted. OSL ages in central Norway confirm a MIS 2 interstadial (25-20 ka) and a dynamic Scandina-vian ice sheet. Quaternary Science Reviews.

Four samples were collected from sub-till sandy sediments within the Langsmoen stratigraphic site, as well as one from silty sand sediments in the stratigraphically-related Flora site. The local pa-leoenvironment was such that Langsmoen sedi-ments represent glaciofluvial/fluvial ice-distal environment followed by ice damming and the deposition of ice-proximal lacustrine sediments at the Flora site. These sites were in-turn buried by LGM-2 till and glaciotectonized.

OSL dose recovery tests were good at 1.06 ±0.03, n = 18, and signal component analysis indi-cated that over 90% of the luminescence signal was from the fast component. Both large and small aliquots were measured to see if there was a sig-nificant age difference caused by incomplete bleaching or other processes. OSL ages for all Langsmoen samples and for both large and small aliquots are consistent at 22.3 ±1.7 ka, n = 7. Study of modern river sediments in the region indicates that fluvial transported sediment can be bleached. The sample from ice-proximal glaciolacustrine sediment at the Flora site gave an apparent old age of ~100 ka likely reflecting incomplete bleaching of the sediments prior to deposition.

Eight radiocarbon ages of sediment from the Flora site gave consistent ages (20.9 ±1.6 cal. ka BP) that overlap within 1� with OSL ages from the nearby Langsmoen site. The similarity in age within and between these stratigraphically-related sites and using different geochronological tech-

niques strongly suggests that this area was ice-free around ~21 or 22 ka. This supports findings from other sites throughout Norway that indicate ice-free conditions ~25-20 ka, collectively termed the Tro-fors interstadial, an interstadial that divides what is generally thought of as the LGM into a two part LGM (Olsen 1997, Olsen et al. 2001a,b, 2002). The existence of the Trofors interstadial along with other interstadials during the Mid- and Late Weichselian (MIS 3 and MIS 2) indicates that not only the western margin, but the whole western part of the SIS, from the ice divide to the ice mar-gin was highly dynamic (Fig. 6). These large changes in the ice margin and accompanying drawdown of the ice surface would have affected the eastern part of the ice sheet as well.

For this work I was in charge of and performed all aspects: conception, acquiring funding, OSL-sampling and measurements, data analysis, and writing. Olsen completed the original stratigraphic and sedimentologic fieldwork years earlier, wrote this portion of the paper, and recommended strati-graphic sites for us to visit.

Paper IV: Åreskutan TCN

Johnsen, T.F., Fabel, D., Stroeven, A. High-elevation cosmogenic nuclide dating of the last deglaciation in the central Swedish mountains: implications for the timing of tree establish-ment. Manuscript.

TCN exposure ages of glacially transported boul-ders from the summit of Mt. Åreskutan (1420 m asl) in central Sweden are consistent (adjusted mean age = 10.6 ±0.6 ka, n = 3) and similar to lower elevation dates for deglaciation from the region. Thirty-five kilometres down-ice (1200 m asl; west) the highest elevation moraine in Sweden produced by the retreating SIS, the Snasahögarna SIS moraine, gave consistent TCN ages (sampled at 1125-1149 m asl; adjusted mean age = 12.0 ±0.6 ka, n = 3). The ages from both sites were adjusted for the effects of glacio-isostatic rebound and shielding by snow cover, and the effect of erosion was considered negligible. The difference in TCN ages between Mt. Åreskutan and Snasahögarna SIS moraine probably reflects a geographical differ-ence in the timing of deglaciation between sites as shown in detailed ice margin reconstructions for the area (Borgström 1989), and/or possibly differ-

Timothy F. Johnsen

18

ences in the actual historical snow depth between the sites.

Together the ages from both these sites, but par-ticularly for Mt. Åreskutan, clash with unusually old radiocarbon dates of the remains of three spe-cies of tree from the summit area (as old as ~17 cal. ka BP; Kullman 2002). We could not find a plausible hypothesis that accommodates both the radiocarbon and TCN ages from this site. The un-usually old radiocarbon ages are rejected on the basis of incompatibility with consistent TCN ages for deglaciation, and incompatibility with pa-leoecological and paleoglaciological reconstruc-tions. Nevertheless, the mere presence of tree re-mains, and of three different species, at this high elevation well above (400-500 m) modern-tree line indicates that there was considerable variation in the climate during the Holocene. The problem lies in reliably dating tree remains from high elevation in central Sweden. We suggest that contamination from calcareous bedrock or neutron production from lightning may have caused the age bias. Given these results it is strongly recommended that specimens for radiocarbon dating be thoroughly tested to ascertain possible sources of contamina-tion, and complementary dating techniques be employed before proposing radical changes to the paleoglaciological and paleoecological history.

High elevation areas deglaciated sometime be-tween ~12.0 and 10.6 ka coinciding approximately with the termination of the Younger Dryas cold interval (11.7 ka), giving a vertical rate of deglacia-tion ~50 m 100a-1. However, as this value is de-rived from data over a large area, it is averaging estimates over space and time; and thus local verti-cal rates of deglaciation can be higher. The vertical rate of deglaciation in the Mt Åreskutan area may have been as high as ~500 m 100a-1. Sometime after deglaciation Betula pubescens, Picea abies, and Pinus sylvestris, grew at 1360 m asl near the summit area of Mt. Åreskutan.

I was in charge of and performed all aspects of the project: conception, acquiring funding, TCN-sampling, data analysis, and writing, except for sample measurements which were completed by Fabel.

Discussion

The results from the four studies (Papers I-IV) directly address the objective and research ques-tions. These research topics include the history and geochronology of the SIS, the behaviour of the ice sheet (‘dynamic’ versus ‘stable’), the thickness and vertical rate of deglaciation of the ice sheet, and the application of two relatively new dating techniques for understanding glacial history. As explained in the introduction, there are strong motivations to better understand the history and dynamics of for-mer ice sheets – scientific, environmental and so-cial. A part of this knowledge-seeking is to have good tools (e.g., dating techniques) that produce reliable results to help confidently decipher natural archives. Thus, I will first discuss the quality of the TCN and OSL dating results before I discuss the implications of my results for our understanding of the history and dynamics of the SIS.

TCN dating quality

Both papers that use the TCN dating technique (Paper I and IV) demonstrate that the TCN dating technique can produce consistent results for the timing of local deglaciation. In Paper I, ages for the Vimmerby moraine were consistent with a 900 year standard deviation for six boulders with a mean of 13.6 ka – which is 7% standard deviation of the mean. In other words, these results are of high precision. These results are considered excel-lent for this dating technique and for the dating of moraines (Fabel, pers. comm.). Studies of moraines elsewhere have included problems of nuclide in-heritance which produces old ages, or problems of boulder exhumation, erosion or moraine deforma-tion (e.g., Fabel et al. 2006, Briner et al. 2001, Hallet and Putkonen 1994, Putkonen and Swanson 2003, Zreda et al. 1994) which produces young ages. The six boulders selected from the Vimmerby moraine do not appear to suffer from any of these processes as the TCN dates are highly consistent with each other and with estimates for the deglacia-tion of the Vimmerby moraine from the radiocar-bon dated varve chronology (Lundqvist and Wohl-farth 2001). Thus, these TCN dating results are considered both precise and accurate, although the uncertainties with the TCN dates are typically large compared to other dating techniques like radiocar-bon dating. These results are highly promising and


19

Fig. 6: A refined general history of the Scandinavian ice sheet over the Mid-and Late-Weichselian glaciation (cf. Fig. 2). All maps are shown with modernsea level. Black dots are study area locations (Fig. 3). See text for explana-tion.

Timothy F. Johnsen

20

may mean that other moraine systems in southern Sweden can be dated successfully (cf. Larsen et al. 2010). In Paper IV, TCN samples were collected from groups of three erratics from two mountain sites. The first site was from the summit of Mt. Åreskutan, and the second was from a high eleva-tion moraine of the SIS on Mt. Snasahögarna. Dates were exceptionally consistent at each site; i.e., very high precision. The percentage standard deviation of the mean of the TCN dates for the three erratics at each site is only 1%. Note that for the Vimmerby moraine study, twice as many errat-ics were used in deriving this statistic. Neverthe-less, the very high precision of the ages at each site is considered outstanding. The difference in age between the sites likely reflects the geographical difference in the relative timing of deglaciation as revealed in detailed mapping of the retreating SIS margin (Borgström 1989). Similar to the Vim-merby moraine study, it appears that these samples do not suffer significantly from any processes that could cause ages to appear too young or old, as the TCN dates are highly consistent at each site and with estimates for deglaciation from lower eleva-tion sites in the area. However, as acknowledged in the paper, the estimate of the snow depth covering the boulders over the entire exposure history may be important at these mountain sites and is an esti-mate that may even vary between the Mt. Åresku-tan and Mt Snasahögarna sites. This snow depth estimate along with the inherent uncertainties of the TCN dating technique (Gosse and Phillips 2001) mean that while the radiocarbon dates from the summit of Mt. Åreskutan can be rejected (Pa-per IV), the accuracy of the value for the vertical rate of deglaciation is a rough estimate.

OSL dating quality

Similar to results from the TCN dating technique, both papers that use the OSL dating technique (Paper II and III) demonstrate that this technique can produce consistent results. Nine out of ten dates from the Pilgrimstad site gave fairly consis-tent ages (44 ±6 ka, n = 9; standard error of the mean of 14%), despite the ages being derived from variety of lithofacies representing different sedi-mentary environments and transport histories. The tenth date was from the lowest stratigraphic unit in the excavation and made of coarse gravels that were likely ice-proximal, giving an erroneous old

age likely due to incomplete bleaching of the lumi-nescence signal. Thus this tenth age is a maximum age for the deposition of overlying sediments. The good correspondence of the nine ages indicates that incomplete bleaching of the sediments was not a problem. Put another way, if incomplete bleaching were a significant process we would not expect sediments from a variety of lithofacies to have the same degree of incomplete bleaching to result in consistent ages. In terms of the OSL dating tech-nique these are considered consistent results of adequate precision. A mean age of 44 ±6 ka (n = 9; 14% standard deviation of the mean) for the Pil-grimstad sediments agrees with the bulk of previ-ous age determinations from the site and sites in northern and central Sweden, which fall between ~60 and ~35 ka (Wohlfarth 2010).

Fewer samples for OSL dating were collected from the Langsmoen site (n = 4; Paper III). How-ever, unlike for the Pilgrimstad site both large and small aliquots were measured (for three of the samples) to look for any differences in dose that may reflect incomplete bleaching of the sediments (Murray and Olley 2002). As well, a modern flu-vial sediment sample was collected from the region to evaluate modern bleaching potential. Both large and small aliquots gave consistent OSL ages (22.3 ±1.7 ka, n = 7; standard deviation of the mean of 8%) for sub-till glaciofluvial/fluvial sediments at the Langsmoen stratigraphic site, and an apparent old age (~100 ka) for a poorly bleached sample of glaciolacustrine sediment at the nearby strati-graphically-related Flora site. The apparent old age for the Flora site sediment is expected for ice-proximal glaciolacustrine sediments due to incom-plete bleaching (Alexanderson and Murray 2009). The consistency of the ages from Langsmoen sediments indicate that they were likely completely bleached prior to deposition, and are of good preci-sion. Eight radiocarbon ages of bulk sediment in the Flora section are fairly consistent (20.9 ±1.6 cal. ka BP; standard deviation of the mean of also 8%) and overlap within 1� with OSL ages from the nearby Langsmoen site. As sediments at these two sites are expected to have formed around the same time based on stratigraphic study, and the ages overlap between both sites, both the OSL and ra-diocarbon ages are considered to be accurate. This raises the credibility of using radiocarbon dating of bulk sediments; a technique that has been used to


21

study interstadial sites throughout Norway (Olsen 1997, Olsen et al. 2001a,b, 2002).

Incomplete bleaching results in an apparent age overestimation, and is fairly common in glacial settings (Fuchs and Owen 2008, Alexanderson and Murray 2009). As sediments from both Langsmoen and Pilgrimstad were glacial or glacially-related, incomplete bleaching was addressed, and in differ-ent ways. The age of multiple samples from differ-ent lithofacies and over a vertical range of the sediments from both sites was arguably the best approach to address incomplete bleaching. As stated earlier, it is not reasonable to expect sedi-ments that represent different environments and transport histories to produce consistent ages if incomplete bleaching is significant. As fewer sam-ples were available from the Langsmoen site, measurement of both large and small aliquots was undertaken to see if smaller populations of grains would give a significantly younger age due to in-complete bleaching (Murray and Olley 2002, Duller 2008); ages were similar. As well, a sample of modern fluvial sediment was measured and shown to not have a significant residual signal. Thus, incomplete bleaching is not an important phenomena for sediments at both these sites; if it were then the ages would be considered maximum ages.

Specific tests and analyses were conducted to ensure the OSL results were of good quality. Sediments were bleached, given a known labora-tory radiation dose, and then the equivalent dose was measured and compared to the given labora-tory dose (i.e., a dose recovery experiment); results were satisfactory (i.e., close to unity). Deconvolu-tion of the luminescence signals was completed to see what portions of the signal produced fast, me-dium and slow components and to select integra-tion limits (‘channels’) that best isolate the fast component and produce the best dose recovery results; these channels were also used for equiva-lent dose determinations. Values for the recycling ratio and recuperation, and the shape of the lumi-nescence signal and growth curves were all as-sessed and used to isolate aliquots that had good OSL characteristics (Fig. 4). A sensitivity analysis was completed at each site to understand the rela-tive importance of different variables and their uncertainties in the calculation of the age; and revealed that water content estimates were not

highly important. Ages were not biased by varia-tion in the sediment dose rate as these did not cor-relate. The sediments from both sites had rather dim signals, probably indicating that the sediments experienced few transport and burial cycles; cycles that would help ‘sensitize’ the quartz mineral and make it more capable of holding charge (Pietsch et al. 2008, Alexanderson and Murray 2009). This resulted in lower signal-to-noise ratios, requiring more measurements of the equivalent dose as a higher number of aliquots did not meet the rejec-tion criteria.

Applying the TCN and OSL techniques

Within this work the TCN and OSL techniques provided fruitful results. One important difference is that while the TCN dating technique was used to determine the timing of local deglaciation, the OSL dating technique was used to estimate the timing of ice-free periods; parts of interstadials that likely spanned thousands of years. Like TCN dat-ing technique, the OSL dating technique produces ages with typically large uncertainties. Thus, gen-erally speaking, the radiocarbon technique is better to use when precision is important. However, at the Pilgrimstad site (Paper II) a number radiocarbon results were at or very near the limit of the radio-carbon technique (~50 ka), while at the Flora site (Paper III) radiocarbon dates were from bulk sedi-ment samples and produced controversial ages. Thus the reliability of the radiocarbon results at both these sites was in question and benefitted from comparison to an alternative dating technique like OSL. Note that radiocarbon dates at Mt. Åre-skutan were also questionable (Paper IV) and so warranted the use of an alternative although less precise dating technique like TCN exposure dating. For glacial research, OSL dating will be mostly limited to use in valley positions where sandy sediments may be available. It clearly is useful in the study of interstadial sediments and where the lower precision is tolerable, or at least as a com-plementary dating technique. Within Sweden, in-completely bleached sediments and low-sensitivity quartz are a common problem in OSL dating, and appears to be related to the degree of sediment reworking and the source bedrock, with the Dala sandstone providing high-sensitivity quartz (Alex-anderson and Murray 2009). In this doctoral re-search, TCN dating was limited to surface expo-

Timothy F. Johnsen

22

sure dating of glacially transported boulders as I was most interested in the deglaciation history of the ice sheet. However, TCN dating can also be used in paleoglaciology to measure the pattern and depth of glacial erosion, and dating the burial of sediments (Fabel and Harbor 1999).

Late Quaternary history and dynamics of the SIS

Results from the four papers have improved our understanding of the history, geochronology, and dynamics of the SIS. TCN dating results from the Vimmerby moraine (Paper I) were in agreement with earlier estimates of the timing of deglaciation and thus confirm the age of this feature and the former position of the SIS in southern Sweden. New mapping of the Vimmerby moraine com-pleted by the Swedish Geological Survey (Malm-berg Persson et al. 2007) has required some altera-tion in the pattern of deglaciation for this time period (Fig. 5). These results suggest that prior to the formation of the Vimmerby moraine (~14 ka), the mean ice retreat rate in eastern portion of southern Sweden was rather high at >150 m a-1 (Lundqvist and Wohlfarth 2001). The local retreat rate in the Vimmerby area was very approximately 70 m a-1, while north of the moraine a concentra-tion of varve data indicates the retreat rate was 135 m a-1 (Wohlfarth et al. 1998).

TCN dating results from Mt. Snasahögarna and the summit of Mt. Åreskutan (Paper IV) contribute to the three-dimensional understanding of the ice sheet history. Results were in agreement with local estimates for deglaciation while controversial ra-diocarbon ages of high elevation tree mega-fossils (Kullman 2002) were rejected. The TCN results indicate that the summit of Mt. Åreskutan (1420 m asl) was covered by the SIS during the Late Gla-cial. High elevation areas may have deglaciated around the termination of the Younger Dryas inter-val (~11.7 ka) and shortly before adjacent valley bottoms. The vertical rate of deglaciation in the Mt Åreskutan area may have been as high as ~500 m 100a-1. Thus, the SIS thinned rapidly (cf. Krabill et al. 2000). Clay-varve studies in valleys indicate that the margin of the SIS retreated horizontally 350 m a-1 (mean; Lundqvist 1973). Earlier studies suggested that the ice sheet during the LGM may not have covered large areas of mountainous ter-rain in central and south Norway delineated by the

geography of alpine blockfields (Nesje and Dahl 1990). Extrapolation of this upper ice sheet limit to central Sweden would have meant that numerous mountain tops, including Mt. Åreskutan, were ice-free during the LGM. This notion seemed like a distinct possibility when considering the results from the Langsmoen site in central Norway for ice-free conditions sometime from 25-20 ka (Paper III). However, the TCN dating results from the summit of Mt. Åreskutan (Paper IV) do not support this hypothesis. Recent TCN study in central and south Norway indicates that the apparent trend in the lower limit of blockfields mapped by Nesje et al. (1988) appears to represent an englacial thermal limit of the SIS, rather than the upper limit of SIS during the LGM (Goehring et al. 2008; cf. Ballan-tyne and Hall 2008). A site near the Swedish-Norwegian border ~195 km south of Mt. Åreskutan indicates that the LGM ice sheet was above 1460 m asl and rapid deglaciation commenced approxi-mately coinciding with the termination of the Younger Dryas cold interval (i.e., similar to results from Paper IV; Goehring et al. 2008).

Papers II and III present evidence of ice-free pe-riods during times when it has been assumed the sites were covered by the SIS (Mangerud 2004; Fig. 2). For the Pilgrimstad site, it was inferred that the Pilgrimstad interstadial sediments belonged to the MIS 5a/c based mostly upon pollen stratigra-phy (Robertsson 1988a,b), while OSL dates and the bulk of radiocarbon dates indicate ice-free conditions during MIS 3 (Paper II, Wohlfarth 2010). The extent of the MIS 3 ice sheet has not been agreed upon and several different ice-sheet scenarios exist (Wohlfarth 2010 and references therein). The central location of Pilgrimstad east of the Scandinavian mountain range and west of the LGM ice-divide has implications for the ice sheet as a whole. The findings from the Pilgrimstad site mean the SIS was very small and at least restricted to the Scandinavian mountain range during at least a part of MIS 3; smaller in extent than proposed for the possibly correlative Ålesund interstadial and perhaps similar in extent to the older Brørup or Odderade interstadials (Fig. 2), or similar to the minimum MIS 3 outline according to Arnold et al. (2002; Fig. 6). The Pilgrimstad interstadial may correlate with the pre-Hattfjelldal interstadial of Olsen et al. (2001a; cf. Bø and Austnes interstadi-als at the western coast, Andersen et al. 1981,


23

1983, Larsen et al. 1987, Mangerud et al. 2003, 2010) which occurred ~55-45 ka, and perhaps is a more reasonable correlation than with the younger Ålesund interstadial (~35 ka). A very recent review of published and unpublished TL/OSL, 14C and U-series dates for Sweden, and that includes OSL results from Paper II, suggests that central and northern Sweden were ice-free during the early and middle part of MIS 3 and that southern Sweden remained ice-free until ~25 cal. ka BP (Wohlfarth 2010). Greenland ice core data indicates that the MIS 3 stage had a mild interstadial climate marked by numerous rapid switches between brief cold and longer warm times (i.e., Dansgaard/Oeschger events; GRIP 1993).

The confirmation of the Trofors interstadial in central Norway (Paper III) indicates that the west-ern marine- and ice streaming- influenced portion of the ice sheet experienced a two-part LGM (LGM-1 and LGM-2) separated by this interstadial that occurred sometime between 25 and 20 ka. If results from other interstadial sites throughout Norway are considered reliable (Olsen et al. 2001a,b, 2002), the Trofors interstadial is a re-gional phenomenon (Fig. 3; cf. Andøya interstadial at the northwest coast, Vorren et al. 1988). The LGM ice sheet had the largest extent of all stadials during the Weichselian glacial period (Mangerud 2004), yet the ice sheet margins behaved dynami-cally during this stadial. The ice sheet margin in central Norway was at least 110 km inland during the Trofors interstadial but later grew in extent to reach the Norwegian shelf (Olsen et al. 2001a) and thickened to cover mountain tops in central Swe-den (Paper IV). This indicates quite dynamic con-ditions. This required average ice margin retreat and advance rates of less than 200-400 m per year (Paper III).

The Trofors interstadial may have extended into the South Swedish Highland as indicated by con-sistent OSL ages (~19-25 ka; Alexanderson and Murray 2007). However, these ages are of sandur sediments that are adjacent to the younger-dated Vimmerby moraine (~14 ka; Paper I). A hypothesis to explain the results from both dating results could include the following events: (1) formation of san-dur system during the Trofors interstadial, (2) cold-based ice conditions and readvance over the san-durs without modification of the sandurs or deposi-tion of the erratics on top of the sandurs, (3) retreat

of the ice sheet and the formation of the Vimmerby moraine. However, the moraine is found in inter-fluves and may be cross-cut by the sandurs, mean-ing that the sandurs may be younger. Alternatively, no or only small moraines were deposited in the valleys. Possibly the sandur and Vimmerby mo-raine formed during the Trofors interstadial, and then during readvance and retreat of the ice sheet during LGM-2 the dated boulders were deposited on top of the moraine. A similar formation for relict moraines in the Swedish mountains has been suggested (Fabel et al. 2006). Alternatively, the OSL ages may be incorrect. Further research is needed to solve this mystery.

Together, the limited ice sheet extent during the Pilgrimstad interstadial and the existence of the Trofors interstadial require the SIS to behave in a more dynamic manner than previously thought. Field science is limited by the discovery and accu-rate dating of interstadial sediments that are un-common. Thus, ice sheet models play a potential role in filling the gaps in our field-based knowl-edge of ice sheet history and dynamics. At the same time those models must be constrained by field results for the geography and timing of degla-ciation and ice-free intervals. Recent robust model-ling of the nearby British-Irish ice sheet has pro-duced a highly dynamic ice sheet with numerous advance/retreat cycles dominated by ice streaming (Hubbard et al. 2009). Arguably modelling has exceeded the practicalities of field science as dy-namic ice sheet conditions mean that it is difficult to find sediments from earlier ice-free periods that have survived subsequent, and multiple, glacial erosion events. It is also difficult to have the reso-lution to distinguish between successive events in the field. Thus, relying on field data alone could lead to a biased less-dynamic view of the behav-iour of the SIS or any paleo-ice sheet. Even the recent dynamic activity of the Greenland and Ant-arctic ice sheets (e.g., ice margins and ice streams) has surprised many scientists and reshaped our view of what may be possible for ice sheets both past and future. Both field data and modelling have their limitations, but our knowledge will be best advanced through employing both and in an inte-grated manner.

Timothy F. Johnsen

24

A refined SIS history

Incorporation of results from newer studies can result in a more detailed picture of the history and dynamics of the ice sheet. Compiling such results into a coherent picture is difficult given large gaps in our spatial and temporal knowledge, the large scale of study (northern Europe), the regionally asynchronous behaviour of the ice sheet, and com-peting ideas on this history. Nevertheless, such reconstructions have been proposed previously (e.g., Denton and Hughes 1981, Lundqvist 1992, Holmlund and Fastook 1995, Kleman et al. 1997, Lambeck et al. 1998, 2010, Boulton et al. 2001, Arnold et al. 2002, Mangerud 1991, 2004, Svend-sen et al. 2004; Fig. 2), and I attempt to present a slightly more refined reconstruction of the Weich-selian Glacial from MIS 4 to the Younger Dryas drawing from the literature, and motivated from my own research (Fig. 6). The ages of the stadials and interstadials are approximate, and the outlines of the ice sheet are partly hypothetical and based on limited field data.

Given the strong results in favour of the Trofors interstadial (Paper III), ice sheet reconstructions for this interstadial and neighbouring stadials are in-corporated (LGM-1 and LGM-2; Olsen 1997, Ol-sen et al. 2001a). As this is one of most difficult interstadials to accept, the other older and less controversial interstadials proposed by Olsen 1997 and Olsen et al. (2001a) and other workers (see references below) are included as well. Altogether these comprise, from oldest to youngest, the Kar-møy/MIS 4 maximum extent stadial, Bø/Austnes/ Pre-Hattfjelldal interstadial, Skjonghelleren stadial, Ålesund/Sandnes/Hattfjelldal-1 interstadial, Rogne stadial, Hamnsund/Hattfjelldal-2 interstadial, LGM-1 stadial, Trofors interstadial, LGM-2 sta-dial, and the Younger Dryas cold interval. The following is a summary of the information com-piled to produce Figure 6. The Karmøy/MIS 4 Maximum extent stadial outline is according to Mangerud (2004), age from Lambeck et al. (2010). The Bø/Austnes/Pre-Hattfjelldal interstadial out-line is quite hypothetical but if ice-free conditions at Pilgrimstad (Paper II) correlate with this inter-stadial, then the ice sheet would be restricted to the Scandinavian mountain range perhaps similar to the MIS 3 minimum ice sheet outline that proposed by Arnold et al. (2002); age from Olsen et al. (2001a). The Skjonghelleren stadial outline is too

hypothetical to resolve on a map, but its margins may have approached those of the MIS 4 stadial (Lambeck et al. 2010); age from (Mangerud et al. 2010). The Ålesund/Sandes/Hattfjelldal-1 intersta-dial outline and age is according to Olsen (2001a; cf. Andersen et al. 1981, Mangerud et al. 1981, 2003, 2010). The ice extent during the last part of this interstadial is assumed to have been small. For example, the ice thickness inferred from glacioi-sostatic depression in the west and northwest dur-ing this period was probably only 50-70 % of that of the YD interval (Olsen 2010). The Rogne stadial outline is according to Olsen et al. (2001a), with name from Mangerud et al. (1981) and age from Mangerud et al. (2010). During this stadial the ice sheet may have extended into Denmark to corre-spond with the Klintholm advance (Houmark-Nielsen et al. 2005, Houmark-Nielsen 2010). The Hamnsund/Hattfjelldal-2 interstadial outline and ages are according to Olsen et al. (2001a; cf. Valen et al. 1996). The LGM-1 outline for the western margin is according to Olsen et al. (2001a), and the southern margin from Houmark-Nielsen and Kjær (2003), age from Olsen et al. (2001a). The Trofors interstadial outline is according to Olsen et al. (2001a), ages from Olsen et al. (2001a) and Paper III. The LGM-2 outline is according to Mangerud (2004) and Vorren and Mangerud (2007), age from Olsen et al. (2001a). Note that at 19 ka the ice sheet had started to retreat in the south and in Denmark in the southwest (e.g., Houmark-Nielsen and Kjær 2003; Ehlers et al. 2004) so this age es-timate is too young for this portion of the ice sheet. However, the ice margin reached its last maximum position at the mouth of the Norwegian Channel as late as ~19 ka according to, e.g., Nygaard et al. (2007), which clearly indicate the asynchronous and highly dynamic behaviour of the ice sheet. Finally, the Younger Dryas outline is according to Andersen et al. (1995) and Mangerud (2004), ages after Rasmussen et al. (2006).

It is not clear which of the stadials the Ristinge advance (Houmark-Nielsen 2010) would correlate to; possibly the early part of the Karmøy/MIS 4 Maximum extent stadial? Note that the Bø and Austnes may not correlate directly with each other as the Bø interstadial may be slightly older than the Austnes interstadial (Mangerud et al. 2010). Fur-thermore, the pre-Hattfjelldal interstadial correlates most likely with the Austnes interstadial, but the


25

pre-Hattfjelldal interstadial is less accurately dated so it may well include or overlap with the Bø inter-stadial as well (Olsen et al. 2001a). The very recent review of absolute ages of Mid-Weichselian inter-stadial deposits for Sweden suggests that northern and central Sweden was ice-free from ~60 to 35 cal. ka BP, and southern Sweden from ~40 to 25 cal. ka BP (Wohlfarth 2010). If this is true, depend-ing on the real age of events, it may require a smaller ice sheet for the Rogne stadial and possibly the Skjonghelleren stadial (Fig. 6).

Are the rates of ice sheet change reasonable in a glaciological sense? Without ice sheet modelling, only a back-of-the-envelope evaluation can be made by assuming that it could take approximately 5 ka for a LGM-sized ice sheet to retreat to the mountains with more rapid retreat in the Baltic, and perhaps 10 ka for the ice sheet to grow to LGM-size from the Scandinavian mountains. Ice growth is precipitation-limited and so much slower than deglaciation (Näslund and Wohlfarth 2008). The smaller an ice sheet is during a given intersta-dial, a disproportionally large amount of time is needed to re-grow the ice sheet for the following stadial. The timing and extent of the various stadi-als and interstadials are approximate, and large uncertainties exist. Nevertheless, during the MIS 2 the growth phase from the Trofors interstadial into LGM-2 may be problematic for the east and south portions of the ice sheet if southern Sweden and the Baltic were deglaciated during the Trofors interstadial. The minimum extent of the Trofors interstadial can not be too small or too long in duration as more time would be needed to build the ice sheet thickness for the margin to advance to the LGM-2 position. This is not limited for the west marine-proximal portion of the ice sheet (Paper IV). During the MIS 3 the advance into the Rogne stadial may again be problematic for the east and south portions of the ice sheet, and if the Klintholm stadial in Denmark correlates with this stadial. As potentially about 10 ka spans between the Bø/Austnes/Pre-Hattfjelldal interstadial and the following Skjonghelleren stadial, a large ice sheet could have existed during this stadial (cf. Lambeck et al. 2010).

Despite uncertainties in the actual outlines of the ice sheets or the accuracy of the ages, the most striking feature of Figure 6 is that there are many stadials and interstadials over the Mid- and Late

Weichselian (cf. Fig. 2). Thus, a dynamic ice sheet is revealed. Further study of interstadial sediments, relict moraines (e.g., Fabel et al. 2006), and state-of-the-art modelling of the SIS constrained by field data will help fill gaps in our knowledge and may potentially reveal an even more dynamic and com-plex ice sheet history (cf. Hubbard et al. 2009).

Recommendations for future research

The study of the history and dynamics of any ice sheet is a very broad research area and so I limit my recommendations to those that are most related to this doctoral research.

As emphasized, field study of interstadial sites must continue. Many interstadial sites have already been discovered in Scandinavia (e.g., Robertsson and García Ambrosiani 1992, Olsen et al. 2001b, Lundqvist and Robertsson 2002, Lokrantz and Sohlenius 2006) and await further study using complementary dating techniques like OSL. Many relict moraines await more detailed study as well. However, reconstructions of the ice sheet must be complemented by state-of-the-art ice sheet models that rely less on global continuous data, and more on regional field data. These models will help fill the still rather large gaps in our understanding of the SIS history.

I wholeheartedly recommend taking multiple samples for dating from study sites as this appears to be the best way to address potential issues with any dating technique. Where possible, using more than one dating technique would be fruitful as well; even so-called ‘negative’ results can contribute to new understanding and developments in geochro-nological techniques. If I only got positive results, I would have half the skill and knowledge I now possess. Perhaps completing a regional reconnais-sance-level study would be reasonable to discover sites where the technique is ‘working’ prior to doing intensive study (cf. Alexanderson and Murray 2009). I also strongly recommend that students or researchers work in the dating lab with their samples especially for OSL dating, and that such arrangements are made early in a research project. For OSL samples consider setting up a preparation laboratory at your home institution where samples can be partitioned and wet-sieved in dim light. This will likely expedite the processing of submitted samples. When using TCN dating for study of deglaciation I recommend multiple sam-

Timothy F. Johnsen

26

ples of glacial erratics from each site especially in areas where cold-based ice or minimal glacial ero-sion had occurred, as nuclide inheritance may be a problem.

At the very least consult with experienced users of the dating technique about field procedures and interesting approaches, collect more samples than thought are needed, and collect and submit samples early in a research program. For instance it never occurred to me that it would be valuable to OSL-date till at Langsmoen, but after more discussions with other workers I realized it would have been worthy. Networking and reaching out to other workers around you and around the world will enlarge your experience and provide creative ideas, knowledge and opportunity.

Building directly on the doctoral results from the four study sites I recommend the following. The consistency of TCN ages from the Vimmerby moraine, and OSL ages from adjacent sandur thou-sands of years older is as yet unexplained (Alex-anderson and Murray 2007, 2009). Finding simi-larly-aged sediments elsewhere in the South Swed-ish Highland would be a further evaluation of the OSL results and improve our understanding of the geographical extent of the Trofors interstadial. Future research efforts should focus on more de-tailed mapping of the various moraine systems in southern Sweden, and employ an integrated dating approach (e.g. radiocarbon dating of the first estab-lishment of vegetation, 10Be dating of erratic boul-ders, and OSL dating of glaciofluvial and other sediments). Current work is being conducted by Per Möller of Lund University, and Larsen et al. (2010). I suspect that LIDAR aerial surveys will revolutionize geomorphic mapping and the pattern of deglaciation.

Many interstadial sites throughout Scandinavia could potentially benefit from OSL dating. Part of the challenge in formerly glaciated landscapes is finding sediments that do not have too dim quartz or that do not suffer from incomplete bleaching. For this reason I recommend a reconnaissance-level study be conducted prior to intensive study at selected sites where preferably there is some age-control. The age of multiple samples from different lithofacies and over a vertical range of the sedi-ments is arguably the best approach to addressing incomplete bleaching, and an approach I recom-mend for probably any stratigraphic study where

appropriate material and adequate resources are available.

Field data on deglaciation is mostly from the outer margins of ice sheets to generate map-views of deglaciation. A significant contribution to de-glaciation studies would be to examine the three-dimensional (thickness evolution) of ice sheets, as I have done in Paper IV and like studies by Landvik et al. (2003), Paus et al. (2006) and Goehring et al. (2008). This important data would provide impor-tant vertical and subsequent volumetric constraints on ice sheet evolution. The most valuable data would be from high elevation sites of the interior of ice sheets, but may also provide the greatest challenge to acquire meaningful TCN results due to frost-weathering and possible nuclide inheri-tance processes for high elevation erratic boulders. As a compromise, mountainous sites closer to the margin of the ice sheet may provide more fruitful results.

Conclusions

Study using the TCN and OSL dating techniques from four sites located in southern Sweden and central Scandinavia has improved our understand-ing of the history and dynamics of the SIS. TCN dating of the Vimmerby moraine has confirmed earlier estimates for the formation of this important landform in the South Swedish Upland around 14 ka. The Pilgrimstad interstadial in central Sweden belongs to MIS 3 rather than to MIS 5a/c and gla-ciers were restricted to the Scandinavian mountain range during at least part of MIS 3. The existence of the Trofors interstadial was confirmed for cen-tral Norway, an ice-free period that occurred ~25-20 ka and separates the LGM into two parts. Fur-ther study is needed to confirm whether southern Sweden was ice-free during this interstadial. High elevation tree remains of Late Glacial age in cen-tral Sweden have been rejected in favour of consis-tent TCN dating results. High elevation areas de-glaciated coinciding approximately with the termi-nation of the Younger Dryas interval. Altogether these results indicate dynamic behaviour for the SIS. Promising results for TCN and OSL dating were achieved for these sites and future research will benefit from application of these techniques to decipher the history and dynamics of the SIS.


27

Acknowledgements

The summary chapter of the thesis was conceived and written by myself with feedback from Helena Alexanderson, Jan Lundqvist, Lars Olsen, Mona Henriksen, and Stefan Wastegård.

Many people helped and inspired me over the course of my doctoral studies. Foremost, I am very thankful to my supervisor Helena Alexanderson, and co-supervisors Arjen Stroeven and Jan Lundqvist – just simply excellent people to work with. Collaborations with Helena, Arjen, Jan, Lars Olsen, Derek Fabel and Andrew Murray were en-joyable and very fruitful. I was inspired by discus-sions with the above people and many others in-cluding: Jonas Bergman, Barbara Wohlfarth, Har-ald Sveian, Hilary Birks, Ann-Marie Robertsson, Terri Lacourse, Robert Lagerbäck, Ingmar Borgström, Martina Hättestrand, Clas Hättestrand, Johan Kleman, Stefan Wastegård, Jan Risberg, Sven Karlsson, Jakob Heyman, Jan-Pieter Buylaert, Dimitri Vandenberghe, Damian Steffen, and others. Daniel Veres, Jakob Heyman, Marie Koitsalu, Helena Alexanderson and Jonas Bergman assisted in the fieldwork and provided enjoyable company.

I am grateful to the many organizations that fi-nancially supported my research activities: Swed-ish Society for Anthropology, Gerard De Geer fund, Carl Mannerfelt fund, Royal Swedish Acad-emy of Sciences, Lars Hiertas remembrance fund, Swedish Tourist Association, and the Geological Survey of Sweden.

Last but not least, I am once again amazed by the tremendous support and patience of my partner Tina who accompanied me on this journey. I love you so much.

My apologies if I neglected to acknowledge you here.

Summary in Swedish

Den skandinaviska inlandsisen har ansetts vara stor, tjock och ganska stabil under stora delar av den senaste istiden (Weichselistiden, 117 000 - 11 700 år sedan). Undersökningar som gjorts de senaste åren ger däremot en helt ny bild av inlandsisen: en mer aktiv is som snabbt växlat i storlek. I min doktorsavhandling har jag särskilt studerat den skandinaviska inlandsisens historia

och dynamik (hur isen ändras med tiden) på fyra platser i Sverige och Norge som också representerar olika tidpunkter under den senaste istiden. Jag har använt mig av två ganska nya metoder för åldersbestämning för att få hållpunkter för nedisningshistorien: optiskt stimulerad luminiscensdatering (OSL) och kosmogen exponeringsdatering.

På flera platser i Skandinavien finns avlagringar från så kallade interstadialer, faser av den senaste istiden då det var mer eller mindre isfritt. Pilgrimstad i Jämtland, Sverige och Langsmoen i Trøndelag, Norge är två sådana platser men vilka tidpunkter representerar de? OSL-datering av nio prover från avsättningarna i Pilgrimstad tyder på att det var isfritt där för ca 50 000 - 38 000 år sedan. Detta är betydligt senare än vad man tidigare antagit. Utifrån analyser av pollen i Pilgrimstad-avsättningarna har man trott att just den här isfria fasen ägde rum i början av den senaste istiden, för ca 90 000 - 75 000 år sedan. Mina resultat placerar istället den här händelsen i mitten av Weichsel-istiden och det betyder att den skandinaviska inlandsisen då måste ha varit mycket liten och bara funnits i fjällen, eller kanske varit helt försvunnen. Klimatet var antagligen också förhållandevis varmt. Resultat från andra undersökningar både i Skandinavien och i Finland stödjer mina slutsatser.

Mina OSL-dateringar av avlagringarna i Langsmoen bekräftar att delar av Norge var isfria för ca 22 000 år sedan, under den så kallade Trofors-interstadialen. Detta är under en tid, det senaste nedisningsmaximat, då den skandinaviska inlandsisen traditionellt ansetts ha varit som störst under de senaste dryga hundratusen åren.

Exponeringsdatering av flyttblock ger tidpunkten för när den senaste inlandsisen smälte bort från ett område. Sex flyttblock från Vimmerbymoränen i Småland, södra Sverige ger samstämmiga resultat och visar att isen försvann därifrån för ca 14 000 år sedan. Detta stämmer bra med tidigare uppskattningar av tidpunkten för isavsmältningen som baserats på kolfjortondatering och lervarvskronologi.

Tre flyttblock från toppen av Åreskutan i Jämtland ger likadana exponeringsåldrar och daterar isens avsmältning där till ca 11 000 år sedan, medan isen försvann lite tidigare längre västerut på Snasahögarna, enligt

Timothy F. Johnsen

28

exponeringsdatering av tre flyttblock därifrån. Detta är i överensstämmelse med tidigare resultat och visar att den skandinaviska inlandsisen snabbt smälte bort och sjönk ihop under den senaste isavsmältningen. Mina resultat innebär också att man inte kan acceptera de ovanligt gamla kolfjortondateringar av trädrester från Åreskutan som gjorts tidigare. Träd på Åreskutan för upp till 17 000 år sedan går inte ihop med mina dateringar av isavsmältningen och med rekonstruktioner av is- och vegetationsutbredning.

Sammantaget visar min forskning, som utförts i olika områden, för olika tidsperioder och med olika metoder, att den skandinaviska inlandsisen var mycket dynamisk och känslig för miljöförändringar. Detta leder till en omvärdering av vår uppfattning av hur forna och nutida istäcken reagerar, och av vår förståelse för sambanden mellan istäcken och olika miljöförändringar.

References Agrell, H., Friberg, N., Oppgården, R. 1976. The Vimmerby

line – an ice-margin zone in north-eastern Småland. Svensk Geografisk Årsbok, 52: 71-91.

Aitken, M.J. 1985. Thermoluminescence dating. Academic, London.

Alexanderson, H., Murray, A. 2007. Was southern Sweden ice free at 19-25 ka, or were the post LGM glacifluvial sedi-ments incompletely bleached? Quaternary Geochronology, 2: 229-236.

Alexanderson, H., Murray. A.S. 2009. Problems and potential of OSL dating Weichselian and Holocene sediments in Sweden. Quaternary Science Reviews, in press.

Alley, R.B., Anandakrishnan, S., Dupont, T.K., Parizek, B.R. 2004. Ice streams--fast, and faster? Comptes Rendus Phy-sique, 5: 723-734.

Andersen, B.G., Nydal, R., Wangen, O.P., Østmo, S.R. 1981. Weichsel before 15,000 years B.P. at Jæren – Karmøy in southwestern Norway. Boreas, 10: 297-314.

Andersen, B.G., Sejrup, H.P., Kirkhus, L. 1983. Eemian and Weichselian deposits at Bø on Karmøy, SW Norway: A pre-liminary report. Norges geologiske undersøkelse, 380: 189- 201.

Andersen, B.G., Lundqvist, J., Saarnisto, M. 1995. The Younger Dryas margin of the Scandinavian ice sheet—an introduction. Quaternary International, 28: 145-146.

Arnold, N.S., van Andel, T.H., Valen, V. 2002. Extent and Dynamics of the Scandinavian Ice Sheet during Oxygen Iso-tope Stage 3 (65,000-25,000 yr B.P.). Quaternary Research, 57: 38-48.

Bailey, R.M., Smith, B.W., Rhodes, E.J. 1997. Partial bleach-ing and the decay form characteristics of quartz OSL. Radia-tion Measurements, 27: 123-136.

Balco, G., Stone, J.O.H., Porter, S.C., Caffee, M.W. 2002. Cosmogenic nuclide ages for New England coastal mo-raines, Martha’s Vineyard and Cape Cod, Massachusetts, USA. Quaternary Science Reviews, 21: 2127-2135.

Balco, G., Stone, J.O.H., Lifton, N.A., Dunai, T.J. 2008. A complete and easily accessible means of calculating surface exposure ages or erosion rates from 10Be and 26Al measure-ments. Quaternary Geochronology, 3: 174-195.

Ballantyne, C.K., Hall, A.M. 2008. The altitude of the last ice sheet in Caithness and east Sutherland, Northern Scotland. Scottish Journal of Geology, 44: 169-181.

Banerjee, D., Murray, A.S., Bøtter-Jensen, L., Lang, A. 2001. Equivalent dose estimation using a single aliquot of po-lymineral fine grains. Radiation Measurements, 33: 73-94.

Baumann, K.H., Lackschewitz, K.S., Mangerud, J., Spielha-gen, R.F., Wolf-welling, T.C.W., Henrich, R., Kassens, H. 1995. Reflection of Scandinavian ice sheet fluctuations in Norwegian Sea sediments during the past 150,000 years. Quaternary Research, 43: 185-197.

Bentley, C.R. 1987. Antarctic ice streams: a review. Journal of Geophysical Research, 92: 8843-8858.

Birks, H.H., Larsen, E., Birks, H.J.B. 2005. Did tree-Betula, Pinus and Picea survive the last glaciation along the west coast of Norway? A review of the evidence, in light of Kullman (2002). Journal of Biogeography, 32: 1461-1471.

Birks, H.H., Larsen, E., Birks, H.J.B. 2006. On the presence of late-glacial trees in western Norway and the Scandes: a fur-ther comment. Journal of Biogeography, 33: 376-378.

Björck, S., Möller, P. 1987. Late Weichselian Environmental History in Southeastern Sweden during the Deglaciation of the Scandinavian Ice Sheet. Quaternary Research, 28: 1-37.

Bøe, A.G, Murray, A., Dahl, S.O. 2007. Resetting of sedi-ments mobilised by the LGM ice-sheet in southern Norway. Quaternary Geochronology, 2: 222-228.

Borgström, I. 1989. Terrängformerna och den glaciala utveck-lingen i södra fjällen. Meddelande från Naturgeografiska institutionen vid Stockholms universitet, Nr A 234, 133 p. (Ph.D. thesis).

Bøtter-Jensen, L., Bulur, E., Duller, G.A.T., Murray, A.S. 2000. Advances in luminescence instrument systems. Radia-tion Measurements, 32: 523-528.

Boulton G.S., Dongelmans P., Punkari M., Broadgate M. 2001. Palaeoglaciology of an ice sheet through a glacial cy-cle: the European ice sheet through the Weichselian. Qua-ternary Science Reviews, 20: 591-625.

Briner, J.P., Swanson, T.W., Caffee, M. 2001. Late Pleisto-cene cosmogenic Cl-36 glacial chronology of the south-western Ahklun Mountains, Alaska. Quaternary Research, 56: 148-154.

Charbit, S., Ritz, C., Ramstein, G. 2002. Simulations of Northern Hemisphere ice-sheet retreat: sensitivity to physi-cal mechanisms involved during the Last Deglaciation. Qua-ternary Science Reviews, 21: 243-265.

Choi, J.H., Murray, A.S., Cheong, C.S., Hong, D.G., Chang, H.W. 2003. The resolution of stratigraphic inconsistency in the luminescence ages of marine terrace sediments from Ko-rea. Quaternary Science Reviews, 22: 1201-1206.

Choi, J.H., Duller, G.A.T., Wintle, A.G. 2006. Analysis of quartz LM-OSL curves. Ancient TL, 24: 9-20.


29

Clark, P.U., Brook, E.J., Raisbeck, G.M., Yiou, F., Clark, J. 2003. Cosmogenic Be-10 ages of the Saglek Moraines, Torngat Mountains. Labrador Geology, 31: 617-620.

Clark, P.U., Dyke, A.S., Shakun, J.D., Carlson, A.E., Clark, J., Wohlfarth, B., Mitrovica, J.X., Hostetler, S.W., McCabe, A.M. 2009. The Last Glacial Maximum. Science 325: 710-714.

Denton, G.H., Hughes, T.J. (eds.). 1981. The Last Great Ice Sheets. Wiley, New York, NY.

Duller, G.A.T. 2004. Luminescence dating of Quaternary sediments: recent advances. Journal of Quaternary Science, 19: 183-192.

Duller, G.A.T. 2008. Single-grain optical dating of Quaternary sediments: why aliquot size matters in luminescence dating. Boreas, 37: 589-612.

Ehlers, J., Eissmann, L., Lippstreu, L., Stephan, H.J., Wansa, S. 2004. Pleistocene glaciations of North Germany. In: Ehlers, J., Gibbard, P.L. (eds.), Quaternary Glaciations – Extent and chronology, 135-146. Elsevier, Amsterdam.

Fabel, D., Harbor, J. 1999. The use of in-situ produced cos-mogenic radionuclides in glaciology and glacial geomor-phology. Annals of Glaciology, 28: 103-110.

Fabel, D., Stroeven, A.P., Harbor, J., Kleman, J., Elmore, D., Fink, D. 2002. Landscape preservation under Fennoscandian ice sheets determined from in situ produced 10Be and 26Al. Earth and Planetary Science Letters, 201: 397-406.

Fabel, D., Fink, D., Fredin, O., Harbor, J., Land, M., Stroeven, A.P. 2006. Exposure ages from relict lateral moraines over-ridden by the Fennoscandian ice sheet. Quaternary Re-search, 65: 136-146.

Fairbanks, R.G. 1989. A 17 000-year glacio-eustatic sea-level record: influence of glacial melting rates on the younger dryas event and deep ocean circulation. Nature, 342: 637-642.

Fredén, C. (ed.). 2002. Sveriges Nationalatlas: Berg och jord. 3rd ed. 208 p.

Frödin, G. 1954. De sista skedena av Centraljämtlands glaciala historia. Geographica, 24.

Fuchs, M., Owen, L.A. 2008. Luminescence dating of glacial and associated sediments: review, recommendations and future directions. Boreas, 37: 636-659.

García Ambrosiani, K. 1990. Pleistocene stratigraphy in cen-tral and northern Sweden. University of Stockholm, De-partment of Quaternary Research, Report 14.

Goehring, B.M., Brook, E. J., Linge, H., Raisbeck, G. M., Yiou, F. 2008. Beryllium-10 exposure ages of erratic boul-ders in southern Norway and implications for the history of the Fennoscandian Ice Sheet. Quaternary Science Reviews, 27: 320-336.

Gosse, J.C., Phillips, F.M. 2001. Terrestrial in situ cosmogenic nuclides: theory and application. Quaternary Science Re-views, 20: 1475-1560.

GRIP (Greenland Ice-Core Project Members). 1993. Climate instability during the last interglacial period recorded in the GRIP ice core. Nature, 364, 203-207.

Hallet, B., Putkonen, J. 1994. Surface dating of dynamic landforms: young boulders on aging moraines. Science, 265: 937-940.

Hättestrand, C., Stroeven, A. 2002. A preglacial landscape in the centre of Fennoscandian glaciation: geomorphological evidence of minimal Quaternary glacial erosion. Geomor-phology, 44: 127–143.

Heimdahl, J. 2006. Spår av en mellanpaleolitisk befolkning i Sverige? Förnyad granskning av Mårtensöfynden 1938 och 1964. (Traces of a Middle Palaeolithic population in Swe-den? A renewed assessment of the finds from Mårtensön in 1938 and 1964.) Fornvännen 101. Stockholm.

Helmens, K.F., Bos, J.A.A., Engels, S., van Meerbeeck, C.J., Bohncke, S.J.P., Renssen, H., Heiri, O., Brooks, S.J., Seppa, H., Birks, H.J.B., Wohlfarth, B. 2007a. Present-day tem-peratures in northern Scandinavia during the last glaciation. Geology, 35: 987-990.

Helmens, K.F., Johansson, P.W., Räsänen, M.E., Alexander-son, H., Eskola, K.O. 2007b. Ice-free intervals at Sokli con-tinuing into Marine Isotope Stage 3 in the central area of the Scandinavian glaciations. Geological Society of Finland Bulletin, 79: 17-39.

Helmut, R., Skvarca, P., Nagler, T. 1996. Rapid Collapse of Northern Larsen Ice Shelf, Antarctica. Science, 271: 788-792.

Hewitt, G. 2000. The genetic legacy of the Quaternary ice ages. Nature, 405: 907-913.

Heyman, J., Hättestrand, C. 2006. Morphology, distribution and formation of relict marginal moraines in the Swedish mountains. Geografiska Annaler, 88: 253-265.

Holmlund, P., Fastook, J. 1995. A time dependent glaciologi-cal model of the Weichselian ice sheet. Quaternary Interna-tional, 27: 53-58.

Houmark-Nielsen, M., Krüger, J., Kjær, K. 2005. De seneste 150.000 år i Danmark. Geovidden - Geologi og Geografi, 2: 1-20.

Houmark-Nielsen, M. 2010. Extent, age and dynamics of Marine Isotope Stage 3 glaciations in the southwestern Bal-tic Basin. Boreas, 39: 343-359.

Houmark-Nielsen, M., Kjaer. K.H. 2003. Southwest Scandi-navia, 40-15 kyr BP: palaeogeography and environmental change. Journal of Quaternary Science, 18: 769-786.

Hubbard, A., Bradwell, T., Golledge, N., Hall, A., Patton, H., Sugden, D., Cooper, R., Stoker, M. 2009. Dynamic cycles, ice streams and their impact on the extent, chronology and deglaciation of the British–Irish ice sheet. Quaternary Sci-ence Reviews, 28: 758-776.

IPCC. 2007. Climate Change 2007: Mitigation of Climate Change. Contribution of Working Group III to the Fourth Assessment Report of the Intergovernmental Panel on Cli-mate Change. B. Metz, O.R. Davidson, P.R. Bosch, R. Dave, L.A. Meyer (eds.). Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA., 851 p.

Jain, M., Murray, A.S., Bøtter-Jensen, L. 2003. Characterisa-tion of blue-light stimulated luminescence components in different quartz samples: implications for dose measure-ments. Radiation Measurements, 37: 441-449.

Kessler, M.A., Anderson, R.S., Briner, J.P. 2008. Fjord inser-tion into continental margins driven by topographic steering of ice. Nature Geoscience, 1: 365-369.

Kjær, K.H., Lagerlund, E., Adrielsson, L., Thomas, P. J., Murray, A., Sandgren, P. 2006. The first independent chro-

Timothy F. Johnsen

30

nology for Middle and Late Weichselian sediments from southern Sweden and the island of Bornholm. GFF, 128: 209-220.

Kitis, G., Polymeris, G.S., Kiyak, N.G. 2007. Component-resolved thermal stability and recuperation study of the LM-OSL curves of four sedimentary quartz samples. Radiation Measurements, 42: 1273-1279.

Kleman, J. 2008. Geomorphology: Where glaciers cut deep. Nature Geoscience, 1: 343-344.

Kleman, J., Stroeven, A.P. 1997. Preglacial surface remnants and Quaternary glacial regimes in northwestern Sweden. Geomorphology 19: 35-54.

Kleman J., Hättestrand, C., Borgström I., Stroeven, A. 1997. Fennoscandian paleoglaciology reconstructed using a glacial geological inversion model. Journal of Glaciology, 43: 283-299.

Kleman, J., Stroeven, A.P., Lundqvist, J. 2008. Patterns of Quaternary ice sheet erosion and deposition in Fennoscandia and a theoretical framework for explanation. Geomorphology, 97: 73-90.

Krabill, W., Abdalati, W., Frederick, E., Manizade, S., Martin, C., Sonntag, J., Swift, R., Thomas, R., Wright, W., Yungel, J. 2000. Greenland Ice Sheet: High-Elevation Balance and Peripheral Thinning. Science, 289: 428-430.

Kulling, O. 1945. Om fynd av mammut vid Pilgrimstad i Jämtland. Sveriges Geologiska Undersökning C 473. 61 p.

Kulling, O. 1967. Yttrande med anledning av J. Lundqvists föredrag ‘Submoräna sediment i Jämtland’. Geologiska Föreningens i Stockholm Förhandlingar, 89: 123-125.

Kullman, L. 2002. Boreal tree taxa in the central Scandes during the Late-Glacial: implications for Late-Quaternary forest history. Journal of Biogeography, 29: 1117-1124.

Kullman, L. 2005. Reply - On the presence of late-glacial trees in the Scandes. Journal of Biogeography, 32: 1499-1500.

Kullman, L. 2006. Late-glacial trees from arctic coast to alpine tundra: response to Birks et al. 2005 and 2006. Journal of Biogeography, 33: 376-378.

Lagerbäck, R. 1988. The Veiki moraines in northern Sweden – widespread evidence of an Early Weichselian deglaciation. Boreas, 17: 469-486.

Lal, D. 1991. Cosmic-ray labeling of erosion surfaces: in situ nuclide production rates and erosion models. Earth and Planetary Science Letters, 104: 424-439.

Lambeck, K., Purcell, A., Zhao, J., Svensson, N.-O. 2010. The Scandinavian Ice Sheet: from MIS 4 to the end of the Last Glacial Maximum. Boreas, 39: 410-435.

Lambeck, K., Chappell, J. 2001. Sea level change through the last glacial cycle. Science, 292: 679-686.

Lambeck, K., Smither, C., Johnston. P. 1998. Sea-level change, glacial rebound and mantle viscosity for northern Europe. Geophysical Journal International, 134: 102-144.

Landvik, J.Y., Brook, E.J., Gualtieri, L., Raisbeck, G., Salvigsen, O., Yiou, F. 2003. Northwest Svalbard during the last glaciation: Ice-free areas existed. Geology, 31: 905-908.

Larsen, E., Gulliksen, S., Lauritzen, S.-E., Lie, R., Løvlie, R., Mangerud, J. 1987. Cave stratigraphy in western Norway: Multiple Weichselian glaciations and interstadial vertebrate fauna. Boreas, 16: 267-292.

Larsen, N.K, Linge, H., Hilkansson, L, Fabel, D. 2010. Rapid disintegration of the marine-based parts of the Scandinavian Ice sheet during the last deglaciation. 29th Nordic Geologi-cal Winter Meeting, Oslo, January, 2010. Abstract, p. 108.

Lian, O.B, Roberts, R.G. 2006. Dating the Quaternary: pro-gress in luminescence dating of sediments. Quaternary Sci-ence Reviews, 25: 2449-2468.

Licciardi, J.M., Clark, P.U., Brook, E.J., Pierce, K.L., Kurz, M.D., Elmore, D., Sharma, P. 2001. Cosmogenic He-3 and Be-10 chronologies of the late Pinedale northern Yellowstone ice cap, Montana, USA. Geology, 29: 1095-1098.

Lokrantz, H., Sohlenius, G. 2006. Ice marginal fluctuations during the Weichselian glaciation in Fennoscandia, a literature review. SKB technical report no. TR-06-36. Swedish Nuclear Fuel and Waste Management Co., Stockholm.

Longva, O., Thorsnes, T. 1997. Skagerak in the past and the present - an integrated study of geology, chemistry, hydrography and microfossil ecology. Norges geologiske undersøkelse, Special Publication, 8: 1-100.

Lundqvist, G. 1964. Interglaciala avlagringar i Sverige. Sveriges geologiska undersökning C 600. 60 p.

Lundqvist, J. 1967. Submoräna sediment i Jämtlands län. Sveriges Geologiska Undersökning C 618. 267 p.

Lundqvist, J. 1973. Isavsmältningens förlopp i Jämtlands län. Sveriges Geologiska Undersökning C 681. 187 p.

Lundqvist, J. 1992. Glacial stratigraphy in Sweden. Geological Survey of Finland, Special paper, 15: 43-59.

Lundqvist, J. 2002. Weichsel-istidens huvudfas. In: Fredén, C. (ed.), Berg och jord. Sveriges Nationalatlas. p. 124-135.

Lundqvist, J., Robertsson, A.-M. 2002. Istider och mellanis-tider. In: Fredén, C. (ed.), Sveriges Nationalatlas: Berg och jord. 3rd ed. p. 120-124.

Lundqvist, J., Wohlfarth, B. 2001. Timing and east–west correlation of south Swedish ice marginal lines during the Late Weichselian. Quaternary Science Reviews, 20: 1127-1148.

Lunkka, J.P., Murray, A., Korpela, K. 2008. Weichselian sediment succession at Ruunaa, Finland, indicating a Mid-Weichselian ice-free interval in eastern Fennoscandia. Bo-reas, 37: 234-244.

Malmberg Persson, K., Persson, M., Lindén A.G. 2007. Is-randstråket Vimmerbymoränen mellan Knivshult och Van-stad i nordöstra Småland. Sveriges Geologiska Undersökn-ing Rapport 2007, 7.

Manabe, S., Broccoli, A.J. 1985. The influence of continental ice sheets on the climate of an Ice Age. Journal of Geo-physical Research, 90: 2167-2190.

Mangerud, J. 1991. The last interglacial/glacial cycle in Northern Europe. In: Shane, L.C.K., Cushing, E.J. (eds.), Quaternary Landscapes. University of Minnesota Press, Minneapolis, p. 38-75.

Mangerud, J. 2004. Ice sheet limits on Norway and the Nor-wegian continental shelf. In: Ehlers, J., Gibbard, P. (eds.), Quaternary Glaciations—Extent and Chronology. Vol. 1. Europe. Elsevier, Amsterdam. p. 271-294.


31

Mangerud, J., Jansen, E., Landvik, J. 1996. Late Cenozoic history of the Scandinavian and Barents Sea ice sheets. Global and Planetary Change, 12: 11-26.

Mangerud, J., Gulliksen, S., Larsen, E., Longva, O., Miller, G.H., Sejrup, H.-P., Sønstegaard, E. 1981. A Middle Weich-selian ice-free period in Western Norway: The Ålesund In-terstadial. Boreas, 10: 447-462.

Mangerud, J., Løvlie, R., Gulliksen, S., Hufthammer, A.-K., Larsen, E., Valen, V. 2003. Paleomagnetic correlations be-tween Scandinavian Ice-Sheet fluctuations and Greenland Dansgaard–Oeschger Events, 45,000–25,000 yrs B.P. Qua-ternary Research, 59: 213-222.

Mangerud, J., Gulliksen, S., Larsen, E. 2010. 14C-dated fluc-tuations of the western flank of the Scandinavian Ice Sheet 45–25 kyr BP compared with Bølling–Younger Dryas fluc-tuations and Dansgaard–Oeschger events in Greenland. Bo-reas, 39: 328-342.

Martinson, D.G., Pisias, N.G., Hays, J.D., Imbrie, J., Moore Jr., T.C., Shackleton, N.J. 1987. Age dating and the orbital theory of the ice ages: development of a high-resolution 0 to 300,000-yr chronostratigraphy. Quaternary Research, 27: 1-29.

Murray, A.S, Olley, J.M. 2002. Precision and accuracy in the optically stimulated luminescence dating of sedimentary quartz: a status review. Geochronometria, 21: 1-16.

Murray, A.S., Wintle, A.G. 2000. Luminescence dating of quartz using an improved single-aliquot regenerative-dose protocol. Radiation Measurements, 32: 57-73.

Murray, A.S., Wintle, A.G. 2003. The single aliquot regenera-tive dose protocol: potential for improvements in reliability. Radiation Measurements, 37: 377-381.

Murray, A.S., Marten, R., Johnson, A., Martin, P. 1987. Analysis for naturally occurring radionuclides at environ-mental concentrations by gamma spectrometry. Journal of Radioanalytical and Nuclear Chemistry Articles, 115: 263-288.

Näslund, J.-O., Wohlfarth, B. (eds.). 2008. Fennoscandian paleo-environment and ice sheet dynamics during Marine Isotope Stage (MIS) 3. Swedish Nuclear Fuel and Waste Management Co, Report R-08-79, p. 1-52.

Näslund J.-O., Rodhe L., Fastook J.L., Holmlund, P. 2003. New ways of studying ice sheet flow directions and glacial erosion by computer modelling – examples from Fenno-scandia. Quaternary Science Reviews, 22: 245-258.

Nesje, A., Dahl, S.O. 1990. Autochthonous block fields in southern Norway: implications for the geometry, thickness, and isostatic loading of the Late Weichselian ice sheet. Journal of Quaternary Science, 5: 225-234.

Nesje, A., Dahl, S.O., Anda, E., Rye, N. 1988. Block fields in southern Norway: significance for the Late Weichselian ice sheet. Norsk Geologisk Tidsskrift, 68: 149-169.

Nygård, A., Sejrup, H.P., Haflidason, H., Lekens, W.A.H., Clark, C.D., Bigg, G.R. 2007. Extreme sediment and ice discharge from marine based ice streams; new evidence from the North Sea. Geology, 35: 395-398.

Oerlemans, J., van der Veen, C.J. 1984. Ice sheets and climate. Reidel, Dordrecht. 217 p.

Olsen, L. 1997. Rapid shifts in glacial extension characterise a new conceptual model for glacial variations during the Mid

and Late Weichselian in Norway. Norges Geologiske Un-dersøgelse Bulletin, 433: 54-55.

Olsen, L. 2010. A buried late MIS 3 shoreline in northern Norway – implications for ice extent and volume. Norges geologiske undersøkelse. Bulletin 450: 1-14.

Olsen, L., Sveian, H., Bergstrøm, B., Selvik, S.F., Lauritzen, S.-E., Stokland, Ø, Grøsfjeld, K. 2001a. Methods and strati-graphies used to reconstruct Mid- and Late Weichselian pa-laeoenvironmental and palaeoclimatic changes in Norway. Norges geologiske undersøkelse Bulletin, 438: 21-46.

Olsen, L., Van der Borg, K., Bergstrøm, B., Sveian, H., Lauritzen, S.-E., Hansen, G. 2001b. AMS radiocarbon dat-ing of glacigenic sediments with low organic carbon content – an important tool for reconstructing the history of glacial variations in Norway. Norwegian Journal of Geology, 81: 59-92.

Olsen, L., Sveian, H., Van der Borg, K., Bergstrøm, B., Broekmans, M. 2002. Rapid and rhythmic ice sheet fluctua-tions in western Scandinavia 15-40 kya – a review. Polar Research, 21: 235-242.

Ottesen, D., Rise, L., Knies, J. Olsen, L., Henriksen, S. 2005. The Vestfjorden-Trænadjupet palaeo-ice stream drainage system, mid-Norwegian continental shelf. Marine Geology, 218: 175- 189.

Paus, A., Velle, G., Larsen, J., Nesje, A., Lie, Ø. 2006. Late-glacial nunataks in central Scandinavia: Biostratigraphical evidence for ice thickness from Lake Flåfattjønn, Tynset, Norway. Quaternary Science Reviews, 25: 1228-1246.

Pavlov, P., Svendsen, J.I., Indrelid, S. 2001. Human presence in the European Arctic nearly 40,000 years ago. Nature, 413: 64-67.

Pettitt P., Niskanen, M. 2005. Neanderthals in Susiluola cave, Finland, during the last interglacial period? Fennoscandia archaeological, XXII: 79-87.

Phillips, F.M., Zreda, M.G., Evenson, E.B., Hall, R.D., Chadwick, O.A., Sharma, P. 1997. Cosmogenic Cl-36 and Be-10 ages of Quaternary glacial and fluvial deposits of the Wind River Range, Wyoming. Geological Society of Amer-ica Bulletin, 109: 1453-1463.

Pietsch, T.J., Olley, J.M., Nanson, G.C. 2008. Fluvial transport as a natural luminescence sensitiser of quartz. Quaternary Geochronology, 3, 365-376.

Prescott, J.R., Hutton, J.T. 1994. Cosmic ray contributions to dose rates for luminescence and ESR dating: large depths and long-term time variations. Radiation Measurements 23, 497-500.

Preusser, F., Chithambo, M.L., Götte, T., Martini, M., Ram-seyer, K., Sendezera, E.J., Susino, G.J., Wintle, A.G. 2009. Quartz as a natural luminescence dosimeter. Earth-Science Reviews, 97: 196-226.

Putkonen, J., Swanson, T. 2003. Accuracy of cosmogenic ages for moraines. Quaternary Research, 59: 255-261.

Rasmussen, S.O., Andersen, K.K., Svensson, A.M., Steffen-sen, J.P., Vinther, B.M., Clausen, H.B., Siggaard-Andersen, M.L., Johnsen, S.J., Larsen, L.B., Dahl-Jensen, D., Bigler, M., Rothlisberger, R., Fischer, H., Goto-Azuma, K., Hans-son, M.E., Ruth, U. 2006. A new Greenland ice core chro-nology for the last glacial termination. Journal of Geophysi-cal Research-Atmospheres, 111: 1-16.

Timothy F. Johnsen

32

Rinterknecht, V.R., Clark, P.U., Raisbeck, G.M. , Yiou, F., Bitinas, A., Brook, E., Marks, J.L., Zelcs, V., Lunkka, J.-P., Pavlovskaya, I.E., Piotrowski, J.A., Raukas, A. 2006. The last deglaciation of southeastern sector of the Scandinavian Ice Sheet. Science, 311: 1449-1452.

Robertsson, A.-M. 1988a. Biostratigraphical studies of inter-glacial and interstadial deposits in Sweden. Ph.D. thesis, Department of Quaternary Research Report 10, University of Stockholm.

Robertsson, A.-M. 1988b. A reinvestigation of the biostrati-graphy of the till-covered sediments at Pilgrimstad and Öje, central Sweden. In: Robertsson, A.-M. (ed.), Biostrati-graphical studies of interglacial and interstadial deposits in Sweden. p. 1-53. University of Stockholm, Department of Quaternary Research, Stockholm.

Robertsson, A.-M. 1994. Vegetational history. In: Fredén, C. (ed.), Geology, National Atlas of Sweden. Almqvist & Wik-sell International, Stockholm, 208 p.

Robertsson, A.-M., García Ambrosiani, K. 1992. The Pleisto-cene in Sweden – a review of research, 1960–1990. Sveriges Geologiska Undersökning Ca 81, 299-306.

Rott, H., Skvarca, P., Nagler, T., 1996. Rapid Collapse of Northern Larsen Ice Shelf, Antarctica. Science, 271: 788-792.

Ruddiman, W.F. 2003. Orbital insolation, ice volume, and greenhouse gases. Quaternary Science Reviews, 22: 1597-1629.

Salonen, V.-P., Kaakinen, A., Kultti, S., Miettinen, A., Eskola, K.O., Lunkka, J.P. 2008. Middle Weichselian glacial event in the central part of the Scandinavian Ice Sheet recorded in the Hitura pit, Ostrobothnia, Finland. Boreas, 37: 38-54.

Schulz, H.-P., Eriksson, B., Hirvas, H., Huhta, P., Jungner, H., Purhonen, P., Ukkonen, P., Rankama, T. 2002. Excavations at Susiluola Cave. Suomen Museo, 109: 5-45.

Sejrup, H.P., Haflidason, H., Aarseth, I., King, E., Forsberg, C.F., Long, D., Rokoengen, K. 1994. Late Weichselian gla-ciation history of the northern North Sea. Boreas, 23: 1-13.

Sejrup, H.P., Larsen, E., Landvik, J., King, E.L., Haflidason, H., Nesje, A. 2000. Quaternary glaciations in southern Fen-noscandia: evidence from southwestern Norway and the northern North Sea region. Quaternary Science Reviews, 19: 667-685.

Shepherd, A., Wingham, D. 2007. Recent Sea-Level Contribu-tions of the Antarctic and Greenland Ice Sheets. Science, 315,: 1529-1532.

Siegert, M.J., Dowdeswell, J.A., Hald, M., Svendsen, J.-I. 2001. Modelling the Eurasian Ice Sheet through a full (Weichselian) glacial cycle. Global and Planetary Change, 31: 367-385.

Singarayer, J.S., Bailey, R.M. 2003. Further investigations of the quartz optically stimulated luminescence components using linear modulation. Radiation Measurements 37, 451-458.

Singarayer, J.S., Bailey, R.M.,Ward, S., Stokes, S. 2004. Assessing the completeness of optical resetting of quartz OSL in the natural environment. Radiation Measurements, 40: 13-25.

Steffen, D., Preusser, F., Schlunegger, F. 2009. OSL quartz age underestimation due to unstable signal components. Quaternary Geochronology, 4: 353-362.

Stokes, C. R, Clark, C.D. 2001. Palaeo-ice streams. Quater-nary Science Reviews, 20: 1437-1457.

Svendsen, J.I, Alexanderson, H., Astakhov, V.I., Demidov, I., Dowdeswell, J.A., Funder, S., Gataullin, V., Henriksen, M., Hjort, C., Houmark-Nielsen, M., Hubberten, H.W., Ingólfsson, Ó., Jakobsson, M., Kjær, K.H., Larsen, E., Lok-rantz, H., Lunkka, J.-P., Lyså, A., Mangerud, J., Matiouch-kov, A., Murray, A., Möller, P., Niessen, F., Nikolskaya, O., Polyak, L., Saarnisto, M., Siegert, C., Siegert, M. J., Spiel-hagen, R.F., Stein, R. 2004. Late Quaternary ice sheet his-tory of northern Eurasia. Quaternary Science Reviews, 23: 1229–1271.

Thrasher, I.M., Mauz, B., Chiverrell, R.C., Lang, A. 2009. Luminescence dating of glaciofluvial deposits: A review. Earth-Science Reviews, 97: 145-158.

Tsukamoto, S., Rink, W.J., Watanuki, T. 2003. OSL of tephric loess and volcanic quartz in Japan and an alternative proce-dure for estimating De from a fast OSL component. Radia-tion Measurements, 37: 459-465.

Ukkonen, P., Arppe, L., Houmark-Nielsen, M., Kjær, K.H., Karhu, J.A. 2007. MIS 3 mammoth remains from Sweden - implications for faunal history, palaeoclimate and glaciation chronology. Quaternary Science Reviews, 26: 3081-3098.

Valen, V., Larsen, E., Mangerud, J., Hufthammer, A.K. 1996. Sedimentology and stratigraphy in the cave Hamnsundhel-leren, western Norway. Journal of Quaternary Science, 11: 185-201.

Vorren, T.O., Laberg, J.S. 1997. Trough Mouth Fans – Pa-laeoclimate and ice-sheet Monitors. Quaternary Science Re-views, 16: 865-882.

Vorren T.O., Mangerud J. 2007. Istider kommer og går. In: Ramberg I.B., Bryhni I., Nøttvedt A. (eds.), Landet blir til. 2nd ed. Norsk Geologisk Forening, Trondheim. p. 478-531.

Vorren, T.O., Vorren, K.-D., Alm, T., Gulliksen, S., Løvlie, R. 1988. The last deglaciation (20,000 to 11,000 B.P.) on Andøya, northern Norway. Boreas, 17: 41-77.

Wintle, A.G. 2008. Luminescence dating: where it has been and where it is going. Boreas, 37: 471-482.

Wintle, A.G., Murray, A.S. 2006. A review of quartz optically stimulated luminescence characteristics and their relevance in single-aliquot regeneration dating protocols. Radiation Measurements, 41: 369-391.

Wohlfarth, B. 2010. Ice-free conditions in Sweden during Marine Oxygen Isotope Stage 3? Boreas, 39: 377-398.

Wohlfarth, B., Björck, S., Possnert, G., Holmquist, B. 1998. A 800-year long, radiocarbon-dated varve chronology from south-eastern Sweden. Boreas, 27: 243-257.

Zreda, M., Phillips, F.M., Elmore, D. 1994. Cosmogenic 36Cl accumulation in unstable landforms 2. Simulations and measurements on eroding surfaces. Water Resources Re-search, 30: 3127-3136.

Paper I

© The authors 2009Journal compilation © 2009 Swedish Society for Anthropology and Geography 113

NEW 10BE COSMOGENIC AGES FROM THE VIMMERBY MORAINE CONFIRM THE TIMING OF SCANDINAVIAN ICE SHEET DEGLACIATION IN SOUTHERN SWEDEN

BYTIMOTHY F. JOHNSEN1, HELENA ALEXANDERSON1,2, DEREK FABEL3 AND

STEWART P.H.T. FREEMAN4

1Department of Physical Geography and Quaternary Geology, Stockholm University, Sweden2 Department of Plant and Environmental Sciences,Norwegian University of Life Sciences, Ås, Norway

3Department of Geographical and Earth Sciences, University of Glasgow, UK4SUERC-AMS, Scottish Universities Environmental Research Centre, East Kilbride, UK

Johnsen, T.F., Alexanderson, H., Fabel, D. and Freeman,S.P.H.T., 2009: New 10Be cosmogenic ages from the Vimmerbymoraine confirm the timing of Scandinavian Ice Sheet deglaciation in southern Sweden. Geogr. Ann. 91 A (2): 113 120

ABSTRACT. The overall pattern of deglaciation ofthe southern part of the Scandinavian Ice Sheet hasbeen considered established, although details ofthe chronology and ice sheet dynamics are less wellknown. Even less is known for the south SwedishUpland because the area was deglaciated mostly bystagnation. Within this area lies the conspicuousVimmerby moraine, for which we have used the ter-restrial cosmogenic nuclide (10Be) exposure datingtechnique to derive the exposure age of six glaciallytransported boulders. The six 10Be cosmogenicages are internally consistent, ranging from 14.9 ±1.5 to 12.4 ± 1.3 ka with a mean of 13.6 ± 0.9 ka. Ad-justing for the effects of surface erosion, snow bur-ial and glacio-isostatic rebound causes the meanage to increase only by c. 6% to c. 14.4 ± 0.9 ka. The10Be derived age for the Vimmerby moraine is inagreement with previous estimates for the timing ofdeglaciation based on radiocarbon dating andvarve chronology. This result shows promise forfurther terrestrial cosmogenic nuclide exposurestudies in southern Sweden.

Key words: terrestrial cosmogenic nuclide (10Be) exposure dating,deglaciation, Scandinavian Ice Sheet, Vimmerby moraine, Sweden, south Swedish Upland

IntroductionAccurate reconstructions of past ice sheets areneeded to better understand their contributions tochanges in climate, sea level, and solid Earth geo-physics. Ice sheet models play a central role in thiseffort but are too frequently poorly constrained byfield data, especially for the interior areas of

former ice sheets. The deglaciation pattern of theScandinavian Ice Sheet is generally well recon-structed, but the absolute timing and the dynamicsof the retreating ice margin are less well known(Lundqvist and Wohlfarth 2001). Several recentstudies have shown that the Late Weichselian gla-cial and deglacial history may be more complicat-ed than generally believed. For example, the west-ern margin of the ice sheet was very dynamic withmultiple ice-free periods during the last 40 ka in-cluding around the Last Glacial Maximum(LGM) (Olsen et al. 2001a,b, 2002); large mo-raine systems from the southeast portion of the icesheet may be younger than previous estimates(Rinterknecht et al. 2006); there were possiblyice-free conditions around the LGM in southernSweden (Alexanderson and Murray 2007) andsouthern Norway (Bøe et al. 2007); and treemega-fossils dated from high elevation areas incentral Sweden suggest ice-free conditions as ear-ly as 17 ka cal. ka BP (Kullman 2002).

In southern Sweden, the deglaciation history isbest known in the coastal areas where the LateWeichselian ice margin actively retreated and canbe traced by conspicuous moraines (west coast) orby patterns of varved clay deposition (east coast).Above the highest Late Weichselian coastline is thesouth Swedish Upland, an area with a poor degla-cial chronology. The correlation between the westand east coasts, across the south Swedish Upland,is problematic due to a lack of continuous geo-logical and geomorphological evidence and be-cause the chronologies are based on different tech-niques.

The Vimmerby moraine (Agrell et al. 1976), is

TIMOTHY F. JOHNSEN, HELENA ALEXANDERSON, DEREK FABEL AND STEWART P.H.T. FREEMAN

© The authors 2009Journal compilation © 2009 Swedish Society for Anthropology and Geography114

Fig. 1. (A) Location of the Vimmerby moraine on the south Swedish Upland and in relation to the standard deglaciation model of Sweden(Lundqvist 2002). Note that the moraine strikes across assumed ice marginal lines and thus indicates a slightly different deglaciationpattern. (B) Geological map of the study area, emphasizing the end moraines of the Vimmerby moraine. Samples are from two sites,indicated by black circles. Adapted from the National Quaternary geological database (Geological Survey of Sweden, Permission 301730/2006)

NEW COSMOGENIC AGES FROM THE VIMMERBY MORAINE


a discontinuous ice-marginal zone, at least 100-km-long, lying in the eastern part of the upland(Fig. 1); it is a distinctive feature since ice-marginalfeatures reflecting active ice are very scarce in thispart of the upland. New mapping of the moraineshows that it strikes across the tentative ice-mar-ginal lines (isochrones) of the standard deglacia-tion model, and this indicates that the deglaciationpattern emerging from this new mapping is differ-ent from the deglaciation pattern previously as-sumed (Lundqvist and Wohlfarth 2001; Lundqvist2002). Thus, dating of this feature would efficientlyfill a gap in our knowledge of the deglacial historyin the south Swedish Upland and would be usefulto compare to other deglaciation chronologies fromthe region.

We have used and present results of one of thefirst applications of the terrestrial cosmogenic nu-clide (10Be) exposure dating technique in southernSweden with the aim of improving our understand-ing of the chronological history of the decay of theScandinavian Ice Sheet, and for comparison to ra-diocarbon and varve chronology.

Study area and previous researchThe south Swedish Upland (57–58˚N, 13–16˚E) isthe highest area in southernmost Sweden and is sit-uated 200–300 m above present sea level (Fig. 1).The crystalline bedrock forms an undulating land-scape with isolated inselbergs, valleys and deep-weathered bedrock. The study area in the easternpart of the upland (Fig. 1a) is situated above theLate Weichselian highest shoreline and the surfacecover is dominated by till (cover moraine, drum-lins, hummocky moraine), glaciofluvial deposits(valley fills, deltas) and peat (Fig. 1b). The commonoccurrence of hummocky moraine in parts of thesouth Swedish Upland indicates widespread stag-nation (dead ice) instead of active retreat (e.g.Björck and Möller 1987); stagnation discouragesthe formation of end moraines. The Vimmerby mo-raine (Agrell et al. 1976; Lindén 1984; MalmbergPersson 2001; Persson 2001; Malmberg Persson etal. 2007) is thus an exception in the area. It consistsof small end moraines and partly till-covered ice-marginal glaciofluvial deposits, and separates anice-proximal landscape with thicker till cover fromone with thin and discontinuous till. Distally, san-durs fill the river valleys down to the highest coast-line (115–130 m a.s.l.; Malmberg Persson et al.2007).

The Vimmerby moraine is believed to reflect a

still-stand and/or readvance in the general ice-mar-gin retreat during the last deglaciation (e.g. Malm-berg Persson 2001). According to the most currentaccepted deglaciation model for southern Sweden,which uses a combination of calibrated radiocar-bon ages, partly radiocarbon-dated clay-varvechronology, and geomorphology (Lundqvist andWohlfarth 2001), this still-stand and/or readvancehappened between 14.0 and 13.3 cal. ka BP, roughlycorresponding to Greenland Interstadial 1 (Björcket al. 1998). A maximum age for deglaciation in thestudy area is provided by a well dated site 200 kmsouth, where AMS radiocarbon dating of leaf frag-ments and twigs gives ages for early vegetation es-tablishment and a minimum age of deglaciationthere around 15.1–14.4 cal. ka BP (Davies et al.2004). A minimum deglaciation age for the studyarea, c. 10.0 cal. ka BP (Lindén 1999; calibrated byus using OxCal v4.0.5 and IntCal04 atmosphericcurve; Ramsey 2007; Reimer et al. 2004), is givenby a radiocarbon date of early organic sedimenta-tion in a kettle hole within the moraine.

MethodologySamplingThe accumulation of in situ produced terrestrialcosmogenic 10Be in quartz exposed to cosmic radi-ation provides a means of determining the amountof time the rock has been at or near the ground sur-face (Lal 1991; Gosse and Phillips 2001). Thistechnique has proven useful in numerous studies ofdeglacial histories and landform preservation (e.g.Balco et al. 2002; Clark et al. 2003; Fabel et al.2002; Licciardi et al. 2001; Phillips et al. 1997;Rinterknecht et al. 2006). Unlike the radiocarbondating technique that gives the age of events fol-lowing deglaciation (e.g. migration and establish-ment of vegetation followed by deposition andpreservation of plant remains in a basin), the 10Be-dating technique can give a direct age of deglacia-tion.

In this study, we collected quartz-rich samplesfrom glacially transported granitic and quartziticboulders on features belonging to the Vimmerbymoraine: an end moraine at Vimmerby and a partlytill-covered marginal delta at Lannaskede (Figs 1b,2; Table 1). To minimize the risk of processes mod-ifying the exposure history of the boulder, such ascosmogenic nuclide inheritance (e.g. Briner et al.2001), boulder exhumation, erosion or moraine de-formation (Hallet and Putkonen 1994; Putkonenand Swanson 2003; Zreda et al. 1994), we sampled



the tops of large (0.9–2.3 m b-axis) rounded to sub-rounded, weathering-resistant (granitic and quartz-itic) boulders. Boulders were resting on top of sta-ble and level surfaces or on the broad crest of themoraine. In total six boulders were processed in theGlasgow University–SUERC cosmogenic nuclidelaboratory.

Measurements and calculationsAll samples were processed for 10Be from quartzfollowing procedures based on methods modifiedfrom Kohl and Nishiizumi (1992) and Child et al.(2000). Approximately 20 g of pure quartz wasseparated from each sample, purified, spiked withc. 0.25 mg 9Be carrier, dissolved, separated by ion

Fig. 2. Sampled boulders at the till covered ice marginal delta at Lannaskede and at an end moraine close to Vimmerby. Both sites arepart of the Vimmerby moraine



chromatography, selectively precipitated as hy-droxides, and oxidized. AMS measurements werecarried out at the SUERC AMS Facility. Measured10Be/9Be ratios were corrected by full chemistryprocedural blanks with 10Be/9Be of <3 × 10 15. In-dependent measurements of AMS samples werecombined as weighted means with the larger of thetotal statistical error or mean standard error. We cal-culated the analytical uncertainty by assuming thatthe uncertainties in AMS measurement and Be car-rier are normal and independent, adding them inquadrate in the usual fashion (e.g. Bevington andRobinson, 1992). The resulting analytical uncer-tainties range from 5 to 8% (Table 1). All 10Be con-centrations were converted to exposure ages by us-ing a production rate linked to a calibration data setusing a 10Be half-life of 1.5 Ma.

Measured 10Be concentrations were convertedto surface exposure ages using the CRONUS-Earth10Be–26Al exposure age calculator version 2 (http://hess.ess.washington.edu), assuming no prior ex-posure and no erosion during exposure. The resultsfor the different 10Be production rate scalingschemes used by the online calculator yielded agesthat vary by less than 2%. The quoted surface ex-posure ages (Table 1) are for the ‘Lm’ scalingscheme which includes palaeomagnetic correc-tions (Balco et al. 2008).

Physical factors influencing 10Be agesVarious physical factors can affect the accuracy ofthe exposure age calculations such as: (1) morainedegradation (Putkonen and Swanson, 2003); (2)

variation in the initial conditions of the populationof boulders delivered to the site, for example, cos-mogenic nuclide inheritance (e.g. Briner et al.2001); (3) the weathering and erosion of the rocksurface during exposure that removes in situ pro-duced terrestrial cosmogenic nuclides from thesampled surface; (4) the partial shielding of therock surface from cosmic rays by seasonal snowcover; (5) the partial shielding of the rock surfacefrom cosmic rays by vegetation or by burial underwater; or (6) the changing elevation of the rock sur-face due to glacio-isostatic movement (Gosse andPhillips 2001). Apart from inheritance, all thesefactors cause 10Be ages to appear too young andwithout adjusting for them 10Be ages will be mini-mum ages. On the other hand, if inheritance is dom-inating, 10Be ages will give maximum ages.

Glacio-isostatic rebound changes the elevationof the sample site during exposure, which in thecase of uplift will cause air pressure to decreaseover time. This means that the cosmic ray flux willvary through time. Usually this can be corrected forby using a nearby relative sea-level curve and inte-grating the changes in 10Be production related torelative sea-level changes over time. However, thenearest sea-level curves are from the east coast ofSweden where before 9.5 ka cal. ka BP water levelswere influenced by the isolated Baltic Ice Lake andAncylus Lake. As air pressure changes were likelymore influenced by changes in sea level than locallake levels, the portion of the water level curve priorto 9.5 ka cal. ka BP does not accurately reflect thechanges in air pressure. Fortunately, detailed mod-elling of shoreline displacements in south-central

Table 1. Summary of terrestrial cosmogenic nuclide (10Be) exposure data.

Altitude Shielding Thickness* [10Be]† ExposureSample Lab ID (m a.s.l.) Lat. (°N) Long. (°E) factor correction (104 atom/g) agee (kyr)‡

S1 b1722 208 57.3947 14.9039 1.0000 0.975 7.61 ± 0.43 12.4 ±1.3 (0.7)S3 b1723 211 57.3866 14.8983 0.9809 0.975 8.98 ± 0.47 14.9 ±1.5 (0.8)S4 b2474 217 57.3808 14.8787 0.9731 0.967 8.00 ± 0.41 13.4 ±1.3 (0.7)S7 b1727 145 57.6695 15.8057 1.0000 0.975 7.62 ± 0.40 13.3 ±1.3 (0.7)S8 b1434 136 57.6700 15.8064 0.9761 0.975 7.99 ± 0.42 14.4 ±1.4 (0.7)S9 b1808 140 57.6688 15.8031 1.0000 0.967 7.51 ± 0.58 13.2 ±1.5 (1.0)

a Calculated using a rock density of 2.7 g/cm3 and an effective attenuation length for production by neutron spallation of 160 g/cm2.b Measured at SUERC AMS relative to NIST SRM with a nominal value of 10Be/9Be = 3.06 x 10-11 (Middleton et al. 1993). Uncertaintiespropagated at ±1σ level including all known sources of analytical error.c Exposure ages calculated using the CRONUS Earth 10Be 26Al exposure age calculator version 2 (http://hess.ess.washington.edu) assuming no prior exposure and no erosion during exposure. The quoted values are for the ‘Lm’ scaling scheme which includes palaeomagnetic corrections (Balco et al. 2008). Uncertainties are ±1σ (68% confidence) including 10Be measurement uncertainties and a 10Beproduction rate uncertainty of 9%, to allow comparison with ages obtained with other methods. Values in parentheses are uncertaintiesbased on measurement errors alone, for sample to sample comparisons.



Sweden and the evolution of the Baltic Sea sincethe LGM has been completed (Lambeck et al.1998). As part of this work water-level curves wereproduced showing both the inclusion and exclusionof ice and land damming. In other words, the water-level curve that excludes the effect of damming iseffectively the sea-level curve that should representchanges in air pressure. Therefore we used thenearest modelled water level curve from Oskars-hamn, c. 55 km south of Vimmerby. The highestcoastline using the Oskarshamn water-level curveis 130 m and is similar to the highest coastline forthe Vimmerby moraine at 115–130 m (MalmbergPersson et al. 2007). Since the highest coastline el-evations are similar between these two sites we didnot apply a scaling factor.

Results and discussionGlacially transported boulder 10Be agesThe six apparent exposure ages range from 14.9 ±1.5 to 12.4 ± 1.3 ka with a mean of 13.6 ± 0.9 ka(with uncertainty at 1σ standard deviation of theages; Table 1).

Boulder exhumation and cosmogenic nuclideinheritance do not appear to be significant issuessince our ages were consistent for boulders fromstable and level surfaces and from the broad crestof the moraine. There were also no indications inthe field of significant erosion or weathering of theboulder surfaces. If we assume a reasonable boul-der-surface constant erosion rate of 1 mm/ka, themean apparent exposure age increases by c. 1%(Table 1). A medium density (0.3 g/cm3) snow cov-er of 0.3 m thickness on top of the boulders for fourmonths a year would similarly increase the meanexposure age by only 1% (Wastenson 1995; Gosseand Phillips 2001). The effect of shielding by theburial from water is not considered as both sites areabove the Late Weichselian highest coastline andabove any local lakes. Also, the shielding effectfrom vegetation is less than 1% and so is not con-sidered (Plug et al. 2007).

The integrated production rate from isostatic re-bound accounts for a c. 4% increase in the exposureage within the study area. Another factor that af-fects air pressure, in addition to elevation changes,is the presence of the nearby ice sheet (Staiger et al.2007); however, since the southern margin of theScandinavian ice sheet retreated rapidly this effectis unlikely to have persisted for long enough to af-fect the calculated exposure ages. Altogether, theseeffects give an adjusted mean exposure age for the

six boulder samples of c. 14.4 ± 0.9 ka (an increaseof c. 6% from the unadjusted exposure age).

Comparisons with other data from the areaThe mean apparent cosmogenic exposure age of thesix boulders of 13.6 ± 0.9 ka is within the previouslyestimated deglaciation age range for the area of14.0–13.3 cal. ka BP (Lundqvist and Wohlfarth 2001and references therein) and within uncertainties tothe 10Be age adjusted for the effects of erosion,snow and glacio-isostatic rebound of 14.4 ± 0.9 ka.Thus, given the similarity between the results of thedifferent approaches used (radiocarbon dating,varve chronology, and now 10Be dating), we areconfident that the deglacial age for the Vimmerbymoraine is c. 14 ka and possibly slightly older if weconsider the adjusted mean 10Be age. This resultalso demonstrates that the 10Be dating approachworks well in this area for boulders from morainesand till surfaces, and may be a useful technique forother areas in southern Sweden. It is not yet under-stood why optically stimulated luminescence agesof glaciofluvial sediments associated with the Vim-merby moraine are many thousands of years older(Alexanderson and Murray 2007, submitted).

Confident correlation of the Vimmerby moraineto moraines in the west half of southern Sweden re-mains problematic because (1) there is not a largedifference in the ages of moraines as the overall rateof deglaciation was relatively fast compared to theresolution of the dating methods, and (2) at LakeVättern located in central southern Sweden, ice-marginal positions for different time periods were insimilar positions (Lundqvist and Wohlfarth 2001).Thus, the Vimmerby moraine may correlate with theTrollhättan moraine, or with the Berghem or Levenemoraines (Fig. 1a; Lundqvist and Wohlfarth 2001).Tracing the ice margin east of the Vimmerby mo-raine is more complicated because (1) ice positionswithin the Baltic Sea basin are not well known, and(2) the rate of deglaciation in the southeast portion ofthe ice sheet (northern Poland and the Baltic states)was also relatively fast compared to the resolution ofthe dating methods (Rinterknecht et al. 2006).

Conclusion10Be ages for the Vimmerby ice-marginal zone of13.6 ± 0.8 (14.4 ± 0.9 adjusted) ka are in agreementwith previous estimates for the timing of deglacia-tion based on radiocarbon dating and varve chro-nology. Thus, the southern margin of the Scandi-



navian Ice Sheet was at the Vimmerby moraine c.14 ka ago. It is not clear which of the moraines westof the study area correlate with the Vimmerby mo-raine, and even less clear for moraines from thesoutheast portion of the ice sheet (northern Polandand the Baltic states). Nevertheless, the internalconsistency of the six 10Be ages and their compat-ibility with previous radiocarbon ages and varvechronology indicate that the terrestrial cosmogenicnuclide (10Be) dating technique works well for er-ratic boulders within this area. This result showspromise for further terrestrial cosmogenic nuclideexposure studies in southern Sweden. Future re-search efforts should focus on more detailed map-ping of the various moraine systems in southernSweden, and employ an integrated dating approach(e.g. radiocarbon dating of the first establishmentof vegetation, 10Be dating of erratic boulders, andoptically stimulated luminescence dating of gla-ciofluvial and other sediments).

AcknowledgementsWe thank Maria Miguens-Rodriguez and HenrietteLinge for chemistry and preparation of AMS tar-gets. Jan Lundqvist and Arjen Stroeven providedfruitful discussions, and Jakob Heyman assisted inthe fieldwork. Two anonymous reviewers and Hil-dred Crill helped improve the manuscript. Fundingwas provided by a grant from the Geological Sur-vey of Sweden (no. 60–1356/2005).

The contributions by the co-authors included:project conception by Alexanderson, boulder se-lection and sampling by Johnsen and Alexander-son, samples crushed by Alexanderson andJohnsen and chemically processed by Fabel, AMSmeasurements by Freeman, age calculations byFabel and Johnsen, interpretations by Johnsen,Fabel and Alexanderson, and writing mostly byJohnsen with input from Alexanderson and Fabel.

Timothy F. Johnsen, Department of Physical Geo-graphy and Quaternary Geology, Stockholm Uni-versity, SE-10691 Stockholm, SwedenE-mail: [email protected]

Helena Alexanderson, Department of PhysicalGeography and Quaternary Geology, StockholmUniversity, SE-10691 Stockholm, Sweden and De-partment of Plant and Environmental Sciences,Norwegian University of Life Sciences,P.O. Box 5003, N-1432 Ås, NorwayE-mail: [email protected]

Derek Fabel, Department of Geographical andEarth Sciences, East Quadrangle, Main Building,University of Glasgow, Glasgow, G12 8QQ, UKE-mail: [email protected]

Stewart P.H.T. Freeman, SUERC-AMS, ScottishUniversities Environmental Research Centre, EastKilbride, G75 0QF, UK

ReferencesAgrell, H., Friberg, N. and Oppgården, R., 1976: The Vimmerby

line an ice margin zone in north eastern Småland. SvenskGeografisk Årsbok, 52: 71 91.

Alexanderson, H. and Murray, A., 2007: Was southern Swedenice free at 19 25 ka, or were the post LGM glacifluvial sediments incompletely bleached? Quaternary Geochronology,2: 229 236.

Alexanderson, H. and Murray, A., submitted: Why doesn’t OSLwork well for deglacial sediments in Sweden? QuaternaryGeochronology

Balco, G., Stone, J.O.H., Porter, S.C. and Caffee, M.W., 2002:Cosmogenic nuclide ages for New England coastal moraines,Martha’s Vineyard and Cape Cod, Massachusetts, USA. Quaternary Science Reviews, 21: 2127 2135.

Balco, G., Stone, J.O.H., Lifton, N.A. and Dunai, T.J., 2008: Acomplete and easily accessible means of calculating surfaceexposure ages or erosion rates from 10Be and 26Al measurements. Quaternary Geochronology, 3: 174 195.

Bevington, P. R. and Robinson, D. K., 1992: Data reduction anderror analysis for the physical sciences. McGraw Hill, NewYork. 328 p.

Björck, S. and Möller, P., 1987: Late Weichselian EnvironmentalHistory in Southeastern Sweden during the Deglaciation ofthe Scandinavian Ice Sheet. Quaternary Research, 28: 1 37.

Björck, S., Walker, M.J.C., Cwynar, L., Johnsen, S., Knudsen,K.L., Lowe, J.J., Wohlfarth, B. and INTIMATE members,1998: An event stratigraphy for the last termination in theNorth Atlantic region based on the Greenland ice core record:a proposal by the INTIMATE group. Journal of QuaternaryScience, 13: 283 292.

Bøe, A.G., Murray, A.S. and Dahl, S.O., 2007: Resetting of sediments mobilised by the LGM ice sheet in southern Norway.Quaternary Geochronology, 2: 222 228. doi: 10.1016/j.quageo.2006.05.031

Briner, J.P., Swanson, T.W. and Caffee, M., 2001: Late Pleistocene cosmogenic Cl 36 glacial chronology of the southwestern Ahklun Mountains, Alaska. Quaternary Research,56: 148 154.

Child, D., Elliott, G., Mifsud, C., Smith, A.M. and Fink, D., 2000:Sample processing for earth science studies at ANTARES.Nuclear Instruments and Methods in Physics Research Section B, Beam Interactions with Materials and Atoms, 172:856 860.

Clark, P.U., Brook, E.J., Raisbeck, G.M., Yiou, F. and Clark, J.,2003: Cosmogenic Be 10 ages of the Saglek Moraines, Torngat Mountains. Labrador Geology, 31: 617 620.

Davies, S.M., Wohlfarth, B., Wastegård, S., Andersson, M., Blockley, S. and Possnert, G., 2004: Were there two Borrobol Tephras during the early Lateglacial period: implications for tephrochronology? Quaternary Science Reviews, 23: 581 589.

Fabel, D., Stroeven, A.P., Harbor, J., Kleman, J., Elmore, D. andFink, D., 2002: Landscape preservation under Fennoscandian



ice sheets determined from in situ produced 10Be and 26Al.Earth and Planetary Science Letters, 201: 397 406.

Gosse, J.C. and Phillips, F.M., 2001: Terrestrial in situ cosmogenic nuclides: theory and application. Quaternary Science Reviews, 20: 1475 1560.

Hallet, B. and Putkonen, J., 1994: Surface dating of dynamic landforms: young boulders on aging moraines. Science, 265: 937940.

Kohl, C.P. and Nishiizumi, K., 1992: Chemical isolation of quartzfor measurement of in situ produced cosmogenic nuclides,Geochimica Cosmochimica ACTA, 56: 3586 3587.

Kullman, L., 2002: Boreal tree taxa in the central Scandes duringthe Late Glacial: implications for Late Quaternary forest history. Journal of Biogeography, 29: 1117 1124.

Lal, D., 1991: Cosmic ray labeling of erosion surfaces: in situ nuclide production rates and erosion models. Earth and Planetary Science Letters, 104: 424 439.

Lambeck, K., Smither, C. and Johnston, P., 1998: Sea levelchange, glacial rebound and mantle viscosity for northern Europe. Geophysical Journal International, 134: 102 144.

Licciardi, J.M., Clark, P.U., Brook, E.J., Pierce, K.L., Kurz, M.D.,Elmore, D. and Sharma, P., 2001: Cosmogenic He 3 and Be10 chronologies of the late Pinedale northern Yellowstone icecap, Montana, USA. Geology, 29: 1095 1098.

Lindén, A.G., 1984: Some ice marginal deposits in the east central part of the south Swedish Upland. Sveriges GeologiskaUndersökning, C 805.

Lindén, A.G., 1999: Översiktlig dokumentation av Lannaskedeplatån. Sveriges Geologiska Undersökning JRAP 99001. 26 p.

Lundqvist, J., 2002: Weichsel istidens huvudfas. In: Fredén, C.(ed.): Berg och jord. Sveriges Nationalatlas. 124 135.

Lundqvist J. and Wohlfarth, B., 2001: Timing and east west correlation of south Swedish ice marginal lines during the LateWeichselian. Quaternary Science Reviews, 20: 1127 1148.

Malmberg Persson, K., 2001: Beskrivning till jordartskartan 6ENässjö SO. Description to the Quaternary map. Sveriges Geologiska Undersökning Ae 145 (in Swedish).

Malmberg Persson, K., Persson, M. and Lindén, A.G., 2007: Israndstråket Vimmerbymoränen mellan Knivshult och Vanstad i nordöstra Småland. Sveriges Geologiska Undersökning.Rapport 2007, 7.

Middleton, R., Brown L., Dezfouly Arjomandy, B., and Klein, J.,1993: On 10Be standards and the half life of 10Be. Nuclear Instruments and Methods in Physics Research B, 82: 399 403.

Olsen, L., Sveian, H., Bergstrøm, B., Selvik, S.F., Lauritzen, S. E.,Stokland, Ø. and Grøsfjeld, K., 2001a: Methods and stratigraphies used to reconstruct Mid and Late Weichselianpalaeoenvironmental and palaeoclimatic changes in Norway.Norges geologiske undersøkelse Bulletin, 438: 21 46.

Olsen, L., Van der Borg, K., Bergstrøm, B., Sveian, H., Lauritzen,

S. E. and Hansen, G., 2001b: AMS radiocarbon dating of glacigenic sediments with low organic carbon content an important tool for reconstructing the history of glacial variationsin Norway. Norwegian Journal of Geology, 81: 59 92.

Olsen, L., Sveian, H., Van der Borg, K., Bergstrøm, B. and Broekmans, M., 2002: Rapid and rhythmic ice sheet fluctuations inwestern Scandinavia 15 40 kya a review. Polar Research,21: 235 242.

Persson, M., 2001: Beskrivning till jordartskartan 6F VetlandaSV. (Description of the Quaternary map). Sveriges Geologiska Undersökning Ae 147.

Phillips, F.M., Zreda, M.G., Evenson, E.B., Hall, R.D., Chadwick,O.A. and Sharma, P., 1997: Cosmogenic Cl 36 and Be 10ages of Quaternary glacial and fluvial deposits of the WindRiver Range, Wyoming. Geological Society of America Bulletin, 109: 1453 1463.

Plug, L.J., Gosse, J.C., McIntosh, J.J. and Bigley, R., 2007: Attenuation of cosmic ray flux in temperate forest. Journal ofGeophysical Research, 112, F02022. doi: 10.1029/2006JF000668

Putkonen, J. and Swanson, T., 2003: Accuracy of cosmogenicages for moraines. Quaternary Research, 59: 255 261.

Ramsey, C.B., 2007: Deposition models for chronologicalrecords. Quaternary Science Reviews, 27: 42 60.

Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W.,Bertrand, C., Blackwell, P.G., Buck, C.E., Burr, G., Cutler,K.B., Damon, P.E., Edwards, R.L., Fairbanks, R.G., Friedrich, M., Guilderson, T.P., Hughen, K.A., Kromer, B., McCormac, F.G., Manning, S., Bronk Ramsey, C., Reimer, R.W.,Remmele, S., Southon, J.R., Stuiver, M., Talamo, S., Taylor,F.W., van der Plicht, J. and Weyhenmeyer, C.E., 2004:INTCAL04 terrestrial radiocarbon age calibration, 0 26 calkyr BP. Radiocarbon, 46: 1029 1058.

Rinterknecht, V. R., Clark, P. U., Raisbeck, G. M., Yiou, F., Bitinas, A., Brook, E., Marks, J. L., Zelcs, V., Lunkka, J. P., Pavlovskaya, I. E., Piotrowski, J. A. and Raukas, A., 2006: Thelast deglaciation of southeastern sector of the ScandinavianIce Sheet. Science, 311: 1449 1452.

Staiger, J. Gosse, J., Toracinta, R., Oglesby, B., Fastook, J. andJohnson, J.V., 2007: Atmospheric scaling of cosmogenic nuclide production: Climate effect. Journal of Geophysical Research, 112, B02205. doi:10.1029/2005JB003811

Wastenson, L. (ed.) 1995: National Atlas of Sweden: Climate,Lakes and Rivers. SNA. 176 pp.

Zreda, M., Phillips, F.M. and Elmore, D., 1994: Cosmogenic36Cl accumulation in unstable landforms 2. Simulations andmeasurements on eroding surfaces. Water Resources Research, 30: 3127 3136.

Manuscript received Aug. 2008, revised and accepted Febr. 2009.

Paper II

Re-dating the Pilgrimstad Interstadial with OSL: a warmer climate and asmaller ice sheet during the Swedish Middle Weichselian (MIS 3)?

HELENA ALEXANDERSON, TIMOTHY JOHNSEN AND ANDREW S. MURRAY

BOREAS Alexanderson, H., Johnsen, T. & Murray, A. S. 2010 (April): Re-dating the Pilgrimstad Interstadial with OSL: awarmer climate and a smaller ice sheet during the Swedish Middle Weichselian (MIS 3)? Boreas, Vol. 39,pp. 367–376. 10.1111/j.1502-3885.2009.00130.x. ISSN 0300-9483.

Pilgrimstad in central Sweden is an important locality for reconstructing environmental changes during the lastglacial period (the Weichselian). Its central location has implications for the Scandinavian Ice Sheet as a whole.The site has been assigned an Early Weichselian age (marine isotope stage (MIS) 5 a/c; 474 ka), based on pollenstratigraphic correlations with type sections in continental Europe, but the few absolute dating attempts so farhave given uncertain results. We re-excavated the site and collected 10 samples for optically stimulated lumines-cence (OSL) dating frommineral- and organic-rich sediments within the new Pilgrimstad section. Single aliquots ofquartz were analysed using a post-IR blue single aliquot regenerative-dose (SAR) protocol. Dose recovery testswere satisfactory and OSL ages are internally consistent. All, except one from an underlying unit that is older, lie inthe range 52–36 ka, which places the interstadial sediments in the Middle Weichselian (MIS 3); this is compatiblewith existing radiocarbon ages, including two measured with accelerator mass spectrometry (AMS). The mean ofthe OSL ages is 44�6 ka (n=9). The OSL ages cannot be assigned to the Early Weichselian for all reasonableadjustments to water content estimates and other parameters. The new ages suggest that climate was relatively mildand that the Scandinavian Ice Sheet was absent or restricted to the mountains for at least parts of MIS 3. Theseresults are supported by other recent studies completed in Fennoscandia.

Helena Alexanderson (e-mail: [email protected]), Department of Plant and Environmental Sciences,Norwegian University of Life Sciences, P.O. Box 5003, NO-1432 As, Norway, and Department of Physical Geo-graphy and Quaternary Geology, Stockholm University, SE-106 91 Stockholm, Sweden; Timothy Johnsen (e-mail:[email protected]), Department of Physical Geography and Quaternary Geology, Stockholm University,SE-106 91 Stockholm, Sweden; Andrew S. Murray (e-mail: [email protected]), Nordic Laboratory for Lumine-scence Dating, Department of Earth Sciences, Aarhus University, Risø DTU, DK-4000 Roskilde, Denmark; received10th February 2009, accepted 6th October 2009.

Pilgrimstad in central Sweden (Fig. 1) is an importantsite for reconstructing environmental changes duringthe Weichselian in Scandinavia, since it is situated closeto the former ice divide of the Scandinavian Ice Sheetand contains subtill organic and minerogenic sediments(Fig. 2). The site has been investigated and described bya number of authors over the past �70 years (mainlyKulling 1945; Frodin 1954; Lundqvist 1967; Robertsson1988a, b; Garcıa Ambrosiani 1990), but the absolutechronology of the site is still poorly known. In thisstudy, we present and evaluate results of optically sti-mulated luminescence (OSL) dating that place the Pil-grimstad Interstadial in marine isotope stage (MIS) 3.

OSL has been used successfully to date Weichseliandeposits in, for example, Russia (Svendsen et al. 2004;Thomas et al. 2006), Greenland (Hansen et al. 1999;Adrielsson & Alexanderson 2005), the Himalayas (Spen-cer & Owen 2004) and New Zealand (Preusser et al.2005), but in Scandinavia results have been of varyingquality (Kjær et al. 2006; Alexanderson & Murray 2007;Lagerback 2007; Houmark-Nielsen 2008). Because ofthis, and because our results are controversial with respectto previous age determinations of the site, in this articlewe focus on the methodology and reliability of the OSLages, while the palaeoglaciological and palaeoecologicalimplications of the dates will be discussed elsewhere.

The Pilgrimstad site

Kulling (1945) recognized three separate series of sandand gravel within the subtill sediments at Pilgrimstad.The lowermost series was interpreted as deposited in aproglacial sub-aquatic environment, while the uppertwo series represent a transition from glacifluvial to flu-vial to lacustrine deposition and contain fine-grainedminerogenic and organic material (Lundqvist 1967). Awell-sorted sandy bed within the lacustrine sedimentshas been interpreted as an aeolian deposit (Robertsson1988a, b). The sediments were exposed at the surface forsome time before the most recent ice advance. Detaileddescriptions of the stratigraphic units and a review ofinterpretations are available in Lundqvist (1967).

The palaeoecological interpretation as a cool, sub-arctic–arctic environment is based on several proxies,mainly from the organic beds. According to pollen andcoleoptera, open herb–shrub vegetation was followedby a forest border setting during a climatic optimum(Robertsson 1988a, b). This warm interval wassucceeded by a colder climatic phase with periglacialconditions and a subsequent phase of herb–shrubvegetation (Robertsson 1988b). Both diatoms andinsects record deposition in a nutrient-rich lake (Lund-qvist 1967; Robertsson 1986). The insect fauna also

DOI 10.1111/j.1502-3885.2009.00130.x r 2009 The Authors, Journal compilation r 2009 The Boreas Collegium

indicates a cool, arctic–subarctic climate, consistentwith the finds of mammoth (Mammuthus primigenius),reindeer (Rangifer tarandus) and elk (Alces alces)(Lundqvist 1967).

The site has been assigned an Early Weichselian(MIS 5 a/c) age based on a correlation of the palaeo-environmental reconstruction with type sections incontinental Europe and represents the type section forthe local Pilgrimstad Interstadial (Kulling 1967; cf. alsoJamtland Interstadial; Lundqvist 1967; Lundqvist &Miller 1992). Robertsson (1988a, b), for example, cor-related the organic-rich beds with one or possibly twoEarly Weichselian interstadials (Brørup and/or Odder-ade) depending on how the climatic cooling in the mid-dle of the section is interpreted – as a phase within oneinterstadial or as separating two different interstadials.

So far, there have been few absolute dating attemptsand these have given uncertain results. According to arecent evaluation of all radiocarbon dates from Pil-grimstad, 10 samples that are acceptable from a qualityperspective range between 59 and 46 cal. ka BP in age(Wohlfarth 2009) (Fig. 3). Moreover, a single thermo-luminescence measurement resulted in 60 ka for thesandy unit within the organic beds (Garcıa Ambrosiani1990) and one U/Th measurement provided an age of38�4 ka (Heijnis in Robertsson & Garcıa Ambrosiani1992). Recently, Ukkonen et al. (2007) also re-dated a

0 200 400100 Kilometers

40°E30°E

20°E

20°E

10°E

10°E0°

70°N

65°N

55°N

65°N

60°N

55°N

SWEDEN

FINLA

ND

NORW

AY

DENMARK

ESTONIA

LITH

UANIALA

TVIA

RUSSIA

Pilgrimstad

Sokli

RuunaaHitura

0-600 m a.sl. >600 m a.s.l. mammoth 38-32 ka LGM margin

SkåneFig. 1. Location map of Pilgrimstad. Mam-moth locations fromUkkonen et al. (2007), LastGlacial Maximum (LGM) margin from Svend-sen et al. (2004). Other sites that also indicate ice-free conditions during MIS 3 are shown: Sokli(�50ka; Helmens et al. 2007a, b), Ruunna(50–25ka; Lunkka et al. 2008), Hitura (deglacia-tion 62–55ka; Salonen et al. 2008) and severalsites in Skane (39–24ka; Kjær et al. 2006).

Fig. 2. Photograph of the upper part of the new section at Pilgrim-stad (see Fig. 3 for comparisons). D–H represent the lithologicalunits.

368 Helena Alexanderson et al. BOREAS

mammoth molar from Pilgrimstad to 34–29 cal. ka BP.The exact stratigraphic position of the mammoth re-mains is not clear, but based on Kulling’s (1945) andRobertsson’s (1988a) stratigraphic descriptions, itseems to derive from the lower part of the organic bedsanalysed by Robertsson (1988a).

These ages are all younger than the age estimate basedon pollen stratigraphy; a possible correlation of Pilgrim-stad with the Moershoofd Interstadial (46–44kaBP)(Behre & van der Plicht 1992) has therefore been pro-posed but considered less likely (Robertsson 1988a).

Setting

The study site is located near Pilgrimstad in Jamtland,central Sweden (Fig. 1) and is situated in an abandonedgravel pit located at the edge of a valley floor.Most of thepreviously described sections have been mined or cov-ered by colluvium or fill. A small hill in the southern partof the pit was the focus of our excavation (Fig. 2). Thetop (including the till cover) had been removed pre-viously, and for a short time the section had been covered

by blast stone (J. Lundqvist, pers. comm. 2007). Thus, weadjusted the sample depths by adding 1m, correspondingto the general till thickness in the area. The position ofthe investigated section is 62157.50N, 15101.10E and itselevation 300ma.s.l., as measured by a Garmin GPSVista C and checked against topographic maps.

The beds in the excavated section were correlated tothe stratigraphies presented by Kulling (1945) and Ro-bertsson (1988b). Robertsson’s original section was si-tuated adjacent to and at right angles to our newsection; her stratigraphy is therefore best comparable tothe new stratigraphy.

Methods

Sampling, preparation and measurement

Fieldwork was conducted in September 2006 and in July2007. Five OSL samples were collected during eachcampaign at the levels marked in Fig. 3 (see also Table 1).The samples were taken in opaque plastic tubes andstored in black bags until opened under darkroom

Unit A. Sandy gravel withlenses of sand and silt.

Unit H. Yellow sand, well-sorted, homogeneous. Wedge-like structure. Contains rip-upsof gyttja.

Unit C. Organic-rich sand.Compact, brecciated.

Unit B. Varved clay and silt.

Unit E. Sandy gyttja, smallerbrecciated pieces than in D.

Unit G. Laminated and massivesand, deformed in lower part.Yellowish to rusty colour.

Unit F. Brown sandy gyttja withoccasional pebbles <15 cm.

B Sediment description

39.2±2

43±5

45±4

48±8

39±3

38±3

>40

OSL(ka)

52±4

46±3

74±5

49±4

36±3

C(ka BP)

All radiocarbon ages in right column from Wohlfarth (2009)except for 34-29 cal. ka (Ukkonen et al. 2007), TL age fromGarcía Ambrosiani (1990). These ages are from other sectionsat the Pilgrimstad site.

Uncertain stratigraphic position.

60 ka TL

A Pilgrimstad section

x

x

x

x

xx

x

x

x

x

G

F

H

D

E

B

A

H

C

344

345

346

347

348

318 319

320

321

322

m1

SGy

SGy

SiGyb

CSil

CoGmm

GySb

Sm

Sm/Sl

SiSlSm

SiSm SiSm

Sm(ng)

NWSE

NE

SSil

~46~48

~47~50 ~52

~59~52~55

~52

~47

Unit D. Grey sandy silty gyttja.Compact, brecciated.

C Chronology

34-29

Correlated ages(cal. ka BP)

Fig. 3. A. Sketch of the new section at Pilgrimstad, lithology and position of OSL and 14C samples. B. Brief sediment description. C. Results ofOSL and 14C age determination.

BOREAS A warmer climate and a smaller ice sheet during the Swedish MIS 3? 369

conditions at Stockholm University (2006) and the Nor-wegian University of Life Sciences (2007), where the in-itial preparation was completed. Final preparation,including heavy liquid separation (2.62g/cm3) to removefeldspars (081318–22 only), treatment with 10% HCl for5–30min, 10% H2O2 for 15–30min, 38% HF for60–120min and 10% HCl again for 40min, was done atthe Nordic Laboratory for Luminescence Dating, wherethe OSL measurements were also undertaken.

The samples were analysed using large aliquots ofquartz (180–250 mm) on Risø TL/OSL readers equip-ped with calibrated 90Sr/90Y beta radiation sources(dose rate 0.14–0.35Gy/s), blue (470�30 nm;� 50mW/cm2) and infrared (880 nm, �100mW/cm2) light sour-ces, and detection was through 7mm of U340 glassfilter (Bøtter-Jensen et al. 2000). Analyses employedpost-IR blue SAR protocols (Murray & Wintle 2000,2003; Banerjee et al. 2001) adapted to suit the samplesbased on dose recovery and preheat experiments (firstbatch: preheat 2601C for 10 s, cut-heat 2201C; secondbatch 2401/2001C). A relatively high test-dose (�50Gy)was necessary to obtain a statistically precise test signal,and 100 s of illumination at 2801C between cycles im-proved recuperation (response to zero dose).

We calculated the dose rates from gamma spectro-metry data (Murray et al. 1987) (Table 2) and includedthe cosmic ray contribution (Prescott & Hutton 1994).Natural and saturated water content was measuredwith pF rings (cylinder volumeters) (Table 1). To ac-count for water content changes through time duemainly to compaction, especially for the organic-richsediments, we applied a simple three-stage model to allour samples (Table 3). The mean water content ð�wÞsince time of deposition was then calculated as:

�w ¼ w1 � t1 þ w2 � t2 þ w3 � t3 ð1ÞTwo samples were collected for radiocarbon dating.

Small twigs (unknown species) were picked out anddated by AMS 14C at the Lund University Radio-carbon Dating Laboratory.

Data analysis

Simple component analysis of the continuous waveOSL data from some aliquots was undertaken usingSigmaPlot 10.0 based on the parameters and formulasof Choi et al. (2006). The results of the componentanalysis are discussed further below (Fig. 4), but basedon such information from dose recovery measurements(Fig. 5) we chose channels 1–2 (first 0.16 s) and channels4–6 (0.32–0.56 s) as peak and background integrationlimits for all aliquots. The equivalent doses were thencalculated in Risø Luminescence Analyst 3.24 (ex-ponential curve fitting) and in Microsoft Excel. To beaccepted, aliquots had to pass the following rejectioncriteria: recycling ratio within 20% of unity, recupera-tion o5%, equivalent dose error o50% and signalmore than 3s above the background. Decay andgrowth curves also had to be regular. Ages were calcu-lated using the mean and median of the equivalent dosepopulation of accepted aliquots for each sample, as wellas using the natural and saturated water contents.

We also did a sensitivity analysis to determinequantitatively which uncertainties have the largesteffect on age. Ages were recalculated after adjustmentswere made to each of the parameters in turn, usingreasonable estimates of uncertainty for each parameter.The estimated uncertainties (1 SD) used in these cal-culations were: depth below surface �1m, elevation�50m, grain size �10%, water content (gammaand beta) �10%, dose rate gamma �5%, dose ratebeta �5%, internal dose rate �30%, density �10%,cosmic ray contribution �5% and beta source cali-bration �2%.

Results

Sedimentology and stratigraphy

We distinguished eight units in the new section at Pil-grimstad (shown and briefly described in Figs 2 and 3).

Table 1. Pilgrimstad OSL sample properties and settings. Listed in order of sample number. For lithofacies codes and stratigraphic unitnumbers, see Fig. 3.

Sample Sampling depth� (m) Stratigraphic unit Lithofacies code Water content (weight %) Organiccontent (%)

w1 w2 w3 w

061344 1.2 G Sm 30 29.5 9.1 28 2.2061345 1.8 H SiSm 30 29.4 5.2 27 3.9061346 2.4 F SGy 150 41.9 16.9 72 9.1061347 2.9 H Sm/Sl 40 35.7 8.5 34 2.0061348 3.1 H Sm/Sl 35 32.8 8.0 31 1.9081318 0.5 H SiSm 35 32.8 5.5 31 1.4081319 0.5 G SiSl 40 38.0 4.9 35 1.4081320 2.1 H Sm 35 34.6 3.8 32 1.2081321 4.1 C GySb 35 34.2 9.0 32 1.1081322 6.3 A Sm(ng) 35 32.7 6.2 31 2.7

�Depth used in calculation was sampling depth plus 1m; see text for explanation.


The sediments have been tectonized, as indicated bybrecciation and deformation structures. Within thelimited exposure available to us, the sediments never-theless seem to have a pancake stratigraphy, with theexception of unit H, which cuts units E, F and G.

Unit A in the new section correlates with the ‘lime-stone-rich pebbly gravel series’ of Kulling (1945), whileunits B–H correspond to his ‘silt-stratified sand series’.We interpret these sediments as showing an environ-mental succession from glacifluvial (unit A) to glacila-custrine (unit B) to lacustrine (units C, D, E, F) and

back to fluvial or glacifluvial (unit G), in line with pre-vious interpretations. The sandy unit H might representthe aeolian sand of Robertsson (1988a, b). However,the cross-cutting relationship with the other units in-dicates formation after the deposition of units F and G,and suggests a possible glacitectonic rather than aeoliangenesis. We therefore interpret unit H as a clastic dykeformed subglacially during the Late Weichselian iceadvance (cf. Larsen & Mangerud 1992; Linden et al.2008). The sand is thus likely reworked sand from unitG, and the gyttja clasts are rip-ups from the surround-ing beds. This has implications for the interpretation ofthe organic beds as belonging to one or two events, andfor the ages of unit H, as discussed below.

OSL characteristics and ages

The major sample properties, settings and results are lis-ted in Tables 1–3 and are shown in Figs 3 and 7. Thequartz from Pilgrimstad is insensitive, which necessitatedcareful selection of peak and background channels tobest isolate the fast component (Fig. 4). For the 21 ac-cepted dose recovery experiments (out of 27), the averageproportions of the net component signal to the total netsignal used for dose calculation were 92�8% (fast),8�7% (medium) and 0.2�0.3% (slow). When thesechannels were selected for peak and background integra-tion, the resulting signal was dominated by the fastcomponent. Doses calculated with this channel selectiongave good dose recovery (1.05�0.04, n=21) (Fig. 5)and demonstrated that the SAR protocols used were ableaccurately to recover a known dose administered beforeany heating. The signals were not close to saturation(Fig. 6).

Equivalent doses range between 89 and 145Gy anddose rates between 2.5 and 3.1Gy/ka (Table 4). Thenine upper samples are 52–36 ka, while the lowermost isolder (74 ka) (Table 4 and Fig. 7). OSL ages from unitH average 44�7 ka (n=5), from unit G 41�3 ka(n=2) and single ages from unit F and C are 48�8 kaand 46�3 ka, respectively. The sensitivity analysis

Table 2. Summary of radionuclide concentrations measured with high-resolution gamma spectrometry on the Pilgrimstad OSL samples. Betaand gamma dose rates refer to dry material; for water contents and final dose rates, see Tables 1 and 3.

Sample 238U (Bq/kg) 226Ra (Bq/kg) 232Th (Bq/kg) 40K (Bq/kg) Beta (Gy/ka) Gamma (Gy/ka)

061344 39�6 42.1�0.7 35.4�0.7 642�12 2.18�0.04 1.25�0.04061345 34�5 42.8�0.6 34.7�0.5 668�9 2.14�0.04 1.15�0.04061346 178�10 100.8�1.2 45�0.9 675�13 3.19�0.07 1.84�0.09061347 33�5 49.1�0.7 37�0.5 715�10 2.38�0.04 1.37�0.04061348 28�3 48.5�0.5 37�0.4 716�7 2.36�0.03 1.36�0.04081318 40�8 40.2�0.8 34.3�0.8 663�15 2.22�0.05 1.24�0.04081319 27�6 38.5�0.7 33.2�0.7 670�12 2.17�0.04 1.21�0.03081320 29�5 31.9�0.5 30.7�0.5 680�12 2.15�0.04 1.14�0.03081321 41�10 100.3�1.3 35.1�0.9 649�15 2.50�0.07 1.65�0.09081322 42�7 74.4�0.9 30.7�0.7 614�11 2.26�0.05 1.39�0.06

Table 3. Water-content modelling to calculate average water contentsince the time of deposition, accounting for compaction and environ-mental changes. For values, see Table 1.

Assumptions Three-stage hydrologicalevolution

The sediments have never beendrier than at the time ofsampling, and the natural watercontent is a minimum watercontent. (The samples weretaken within a large gravel pit inwhich the groundwater tablehas been lowered.)

The saturated water content isthe maximum water content forthe sediments in their presentstate. (It is limited by porosity.)

The porosity of minerogenicsediments did not changesignificantly with compaction.

Most of the compaction tookplace early in the sediments’history due to continuedsedimentation and, later,pressure from the ice cover.

1. Lake/fluvial stage before theice advance (sediments loose andsaturated).(a) Water content w1 is thesaturated value rounded up to thenearest 5 or 0 for minerogenicsamples and an assumed value of150% for the organic sample(061346), derived from youngsamples with similar organiccontent.(b) Duration is �30% of the timesince deposition (t1=0.3).

2. Ice cover (sediments compactedand saturated).(a) Water content w2 is thesaturated value.(b) Duration is �60% of the timesince deposition (t2=0.6).

3. After deglaciation (sedimentscompacted and drier).(a) Water content w3 is thenatural value.(b) Duration is �10% of the timesince deposition (t3=0.1).


demonstrates that the calculated age is most sensitiveto changes in the equivalent dose, beta source calibra-tion and dose rate, followed by water content (Fig. 8).

New radiocarbon ages from twigs are 39.2�2 and440 14C kaBP (Table 5).

Discussion

OSL and sediments

From a sedimentological perspective, we can identifytwo potential problems for OSL-dating at this site: (1)the glacifluvial deposit (unit A) may be incompletelybleached and (2) we cannot properly estimate the ori-ginal mean water content of the relatively organic-richunit F.

Incomplete bleaching results in an apparent ageoverestimation, and is fairly common in glacial settings(e.g. Fuchs & Owen 2008). As we have used only largealiquots for measurements, it is difficult to infer any-thing about incomplete bleaching from the dose dis-tribution of sample 081322, which is from unit A. Incombination with the stratigraphic position of unit A(lowest), we thus consider the OSL age from unit A asproviding a maximum age for the overlying organicbeds. The good correspondence of ages (n=9) from allthe other beds (derived from various depositional set-tings) suggests that there are no problems with in-complete bleaching for those.

Organic-rich deposits tend to be fairly loose at thetime of deposition and may have very high water con-tent, i.e. up to several hundred percent. With time theywill become compacted due to sediment loading and theoverriding ice sheet; the water content will change sig-nificantly, so that what is measured today is not

Growth curve using standard data

0.0

0.5

1.0

1.5

2.0

2.5

0 200 400 600 800 1000 1200Dose (s)

Lx/T

x

Growth curve using derived fast component data

0.0

0.5

1.0

1.5

2.0

2.5

0 200 400 600 800 1000 1200Dose (s)

Lx/T

x

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%A

B

C

0 1 2 3 4 5Stimulation time (sec)

% o

f bul

k si

gnal

0

500

1000

1500

2000

2500

3000

3500

4000

Bul

k si

gnal

natural signal

modeled signal

fast

medium

slowba

ckgr

ound

Sample 061347

given dose

given dose

Sample 061347

Sample 061347

Recycling = 1.13Dose recovery = 0.90

Recycling = 1.11Dose recovery = 0.92

Fig. 4. Data derived from the 27 dose recovery experiments allowedselecting peak and background channels that provided theoverall best recycling and dose recovery and a dominance of the fastcomponent after background subtraction. A. Signal componentanalysis of sample 061347 showing the natural signal which wasde-convoluted into fast, medium and slow components (left axis),and the natural and modelled (sum) signal (right axis). Also shownare the peak and background portions of the signal used (0–0.16 sand 0.32–0.56 s). Note that only the first 5 s of the total length of thestimulation (40 s) are shown. B. The upper growth curve usesthe standard data (bulk signal) with the selected channels for thesame aliquot. C. The lower growth curve is produced by using thederived (de-convoluted) fast component signal data only. Theinsignificant difference between the growth curves indicates thatthe channel selection effectively isolated the fast component andthat the selection of peak and background channels provided theoptimum recycling and dose recovery.

0

2Fre

quen

cy

4

0-0.

1

0.2-

0.3

0.4-

0.5

0.6-

0.7

0.8-

0.9

1-1.

1

1.2-

1.3

1.4-

1.5

1.6-

1.7

1.8-

1.9

Ratio of measured to given dose

Mean = 1.05Std err = 0.04

n = 21

Fig. 5. Histogram of dose recovery tests, excluding aliquots that didnot pass rejection criteria. A mean close to unity indicates that thesamples can be used for OSL-dating with our analytical protocols.


representative of the mean water content since time ofdeposition (cf. Alexanderson et al. 2008). An under-estimation of the mean water content results in OSL agesthat appear too young, and vice versa. In our case, theOSL ages from the surrounding minerogenic beds canprovide some constraints, since these sediments do notsuffer to the same degree from water-content variations.

The sandy unit H is the youngest, since it cross-cutsunits E, F and G. The OSL ages from unit H shouldthus be minimum ages for the organic deposits. How-ever, as mentioned above, re-interpretation of the unitas a clastic dyke with remobilized material implies thatthe sand in unit H could be contemporaneous with unitG and that the OSL ages do not represent the timing ofthe formation of the dyke. It also implies that the se-paration into units E and F is secondary.

Based on the stratigraphic information and on theOSL ages, we consider the OSL ages from units B to Has representing one event, while unit A is older. TheOSL ages from units B–H are internally consistent, i.e.all lie in the range 52–36 ka, with a mean of 44�6 ka;this places the interstadial sediments in the MiddleWeichselian (MIS 3; 58–24 ka). Taken at face value, the74 ka age from unit A gives the timing of the precedingdeglaciation, but we cannot rule out the possibility thatthe sediment suffers from some incomplete bleaching,and that the true deposition age is younger.

What is required to make these OSL ages older(MIS 5)?

To capture uncertainties in the ages conservatively, Fig.7 and Table 4 show the OSL ages calculated under avariety of assumptions. Even when considering thebroadest age range for each sample, all upper nine agesfall within the Middle Weichselian (MIS 3) and no rea-sonable adjustments in assumptions can collectivelybring their ages to the Early Weichselian (MIS 5 a/c,474 ka). As the sensitivity analysis showed the calcu-lated age to depend most on changes in the equivalentdose and various factors influencing the dose rate (Fig.8), we tested what those factors would have to be toobtain �90 ka ages from these samples. To get earlyWeichselian ages from the current data we would needto double the equivalent doses or half the dose rates; thelatter for example by using water contents 4100% byweight, or a combination of these. We consider it un-likely that any of these parameters is in error to thisdegree.

Comparison with other chronologies and records

A mean age of 44�6 ka (n=9) for the Pilgrimstad se-diments agrees with the bulk of previous age determi-nations from the site, which fall between 60 and 45 ka

0

1

2

3

0 100 200 300 400 500 600

500

1000

1500

2000

2500

3000200 10 30 40

Stimulation time (s)

Pho

ton

coun

ts /

0.16

s

Sample 061345

Regenerative dose (Gy)

Sen

sitiv

ity c

orre

cted

OS

L

Fig. 6. Single-aliquot regenerative-dose (SAR) OSL growth curve fora single aliquot of sample 061345. The equivalent dose (ED) (opendiamond) is obtained by interpolating the corrected natural signal(open triangle) on the growth curve. Repeated dose points (opensquares) and a zero dose point (open circle) are also shown. (Inset)The natural decay curve.

Table 4. Pilgrimstad OSL sample results and ages. Listed in order of sample number.

Sample Age (ka) Equivalent dose (Gy) n Dose rate (Gy/ka) Recycling ratio

Mean Median Dry1 Saturated2

061344 43�5 37 36�4 43�5 116�12 18 2.72�0.10 1.14�0.24061345 45�4 44 37�4 46�4 118�10 19 2.62�0.10 1.06�0.26061346 48�8 49 32�5 39�7 134�23 18 2.82�0.09 1.06�0.31061347 52�4 50 42�4 53�4 145�10 16 2.77�0.10 1.08�0.25061348 49�4 49 40�3 50�4 139�10 22 2.83�0.10 1.03�0.20081318 38�3 36 30�3 38�3 101�7 28 2.69�0.10 1.17�0.09081319 39�3 34 29�3 40�3 97�7 31 2.51�0.09 1.07�0.12081320 36�3 35 27�2 36�3 89�7 32 2.52�0.09 1.05�0.17081321 46�3 44 37�3 47�3 143�6 30 3.10�0.13 1.10�0.12081322 74�5 72 58�4 75�5 202�9 30 2.74�0.11 1.07�0.12

1Age calculated with water content at time of sampling.2Age calculated with saturated water content.


(Fig. 3) (Wohlfarth 2009). These radiocarbon dates areconsidered to be of acceptable quality, although closeto the methodological limit (Wohlfarth 2009). Themean OSL age also agrees with the finite of our twoAMS 14C ages (39.2�2 14C kaBP; Table 5) and is notcontradicted by the non-finite one.

Oxygen isotope curves from the Greenland Ice Sheetprovide estimates of temperature variations duringthe Weichselian in the North Atlantic region (e.g.Johnsen et al. 2001). If the sediments at Pilgrimstad areof MIS 3 age, as the OSL and other ages suggest,we would expect a correlation with one of the warmestand/or longest interstadials during this time – withduration long enough to allow for the subarctic–arcticflora and fauna to become established. Taking theOSL ages at face value leads to a correlation with

Greenland interstadials 12–10 (47–41 ka), but con-sidering the uncertainties and the actual spread in ages,interstadials 17–10 are also possible options (cf. Walkeret al. 1999).

This is in agreement with Wohlfarth (2009), whotentatively places the Pilgrimstad site within Greenlandinterstadials 17–12. Hattestrand (2008) proposes a si-milar age shift of the Tarendo II Interstadial in north-ern Sweden, i.e. from the Early Weichselian (MIS 5 a)to the Middle Weichselian (MIS 3). Our OSL data,from a critical location in the former central area of theice sheet, thus support the documented ice-free condi-tions reported elsewhere in Fennoscandia during MIS 3(e.g. Olsen et al. 2001; Arnold et al. 2002; Kjær et al.2006; Helmens et al. 2007a, b; Ukkonen et al. 2007;Lunkka et al. 2008; Salonen et al. 2008) (see also Fig. 1).

Change in parameter (%)

30

35

40

45

50

55

–100 –80 –60 –40 –20 0 20 40 60 80 100

Cha

nge

in a

ge (

ka)

Fig. 8. Sensitivity analysis showing by howmuch the age will change (percentage change) ina given parameter for sample (061)344(age=43ka). Steeper slopes indicate that theage is more sensitive to changes in the givenparameter. The age is most sensitive to changesin the equivalent dose, beta source calibrationand dose rate, followed by water content; verylarge changes in these parameters are needed tomake the age consistent with MIS 5. The rangeof x-axis parameter values corresponds to ourestimate of 2 SD in that parameter. Similar re-sults were obtained for the other samples.

0A

B

10

20

30

40

50

60

70

80

90318,H 345,H 320,H 347,H 348,H 344,G 319,G 346,F 321,C 322,A

Sample and stratigraphic unit

Agr

(ka

)

0.60.81.01.21.41.6

Rec

yclin

g ra

tio

Fig. 7. A. Graphical summary of ages for allsamples. B. Corresponding recycling ratios.OSL ages are shown in stratigraphic order withyoungest ages to the left and oldest to the right.H–A refers to the stratigraphic units in Fig. 3.Age calculated from the population of acceptedaliquots using the mean and median, and fornatural water content at sampling (dry) and sa-turated water content (range shown by greysquare). Error bars represent 1 SD. The SD ofmean of ages excludes the older, strati-graphically lower sample (081)322. Note thatthe water content for organic-rich sample(061)346 has been adjusted to compensate forcompression. Marine isotope stages (MIS) in-dicated on the right.


Implications for Weichselian history

Moving the age of the Pilgrimstad Interstadial fromEarly to Middle Weichselian (MIS 5 a/c to MIS 3) hasimplications for the understanding of glacial historyand environmental change in Scandinavia during theWeichselian. An ice-free Pilgrimstad at �50–40ka re-quires that the Scandinavian Ice Sheet at that time wasrestricted, possibly limited to the highest mountains (cf.Arnold et al. 2002), i.e. much smaller than previouslybelieved (Lundqvist 2002; Mangerud 2004; Houmark-Nielsen 2007) (Fig. 1). It also implies that the ice sheetmust have expanded rapidly thereafter to reach south-eastern Denmark just prior to 30ka (the Klintholm ad-vance; Houmark-Nielsen & Kjær 2003; Ukkonen et al.2007).

Although the vegetation reconstruction for Pilgrim-stad (Robertsson 1988b) still largely holds true,recent studies indicate that forest may not have been asclose as suggested (Helmens et al. 2007a). Nevertheless,a reconciliation of the Pilgrimstad record with con-temporaneous records of tundra in northern con-tinental Europe (Behre 1989) requires alternativemigration routes for vegetation, and/or possible treerefugia, during previous stadials and different climategradients and conditions compared to today (e.g. Nils-son 1972: p. 225; Ukkonen et al. 2007).

Conclusions

� A new exposure at the Pilgrimstad site in centralSweden shows 44m thick subtill minerogenic andorganic sediments; these have been OSL-dated.

� The OSL ages from the lacustrine and fluvial sedi-ments range from 52 to 36ka, while the underlyingglacifluvial deposit is probably younger than�74ka.

� The OSL samples passed appropriate methodologi-cal checks and the ages are internally consistent. Wetherefore consider the results reliable.

� From sedimentological and chronological points ofview, we favour deposition during a single but vari-able event, rather than two separate interstadials.

� The OSL ages assign the Pilgrimstad Interstadialsite to MIS 3; the organic deposits could possiblycorrespond to one or more of Greenland inter-stadials 17–10.

� The central location of the site, together with resultsfrom other studies in Fennoscandia, indicates thatfor at least parts of MIS 3 the ice sheet was absent,

or at least restricted to the highest Scandinavianmountains.

Acknowledgements. – We thank Jan Lundqvist, Ann-Marie Rober-tsson and Barbara Wohlfarth at Stockholm University for valuablediscussions and for sharing experiences and data; Jan-Pieter Buylaertand the technical staff at the Nordic Laboratory for LuminescenceDating for helping with OSL measurements; Damian Steffen at theUniversity of Bern for introducing Alexanderson and Johnsen to theworld of de-convolution; and Dick Olsson and Claes in Pilgrimstadfor giving us access to the gravel pit and assisting with excavation. Thestudy was financed by a grant (no. 60-1356/2005) from the GeologicalSurvey of Sweden to Alexanderson, with additional fieldwork fundingfrom the Swedish Nuclear Fuel and Waste Management Company(SKB). Alexanderson was in charge of and carried out most of theOSL sampling, preparation and measurements, sedimentological andstratigraphical work as well as writing, with significant input at allstages by Johnsen. Johnsen analysed all the data, with some inputfrom Murray and Alexanderson. Murray contributed to discussionsregarding OSL preparation and measurement setup and interpreta-tion of results as well as provided laboratory access. The constructivecomments on themanuscript from reviewers PerMoller andOle Bennikeare appreciated.

References

Adrielsson, L. & Alexanderson, H. 2005: Interactions between theGreenland Ice Sheet and the Liverpool Land coastal ice cap duringthe last two glaciation cycles. Journal of Quaternary Science 20,269–283.

Alexanderson, H. & Murray, A. S. 2007: Was southern Sweden icefree at 19–25 ka, or were the post LGM glacifluvial sediments in-completely bleached? Quaternary Geochronology 2, 229–236.

Alexanderson, H., Johnsen, T., Wohlfarth, B., Naslund, J.-O. &Stroeven, A. 2008: Applying the optically stimulated luminescence(OSL) technique to date the Weichselian glacial history of southernSweden. Reports from the Department of Physical Geography andQuaternary Geology, Stockholm University 4, 1–33.

Arnold, N. S., van Andel, T. H. & Valen, V. 2002: Extent and dy-namics of the Scandinavian Ice Sheet during oxygen isotope stage 3(65,000–25,000 yr B.P.). Quaternary Research 57, 38–48.

Banerjee, D., Murray, A. S., Bøtter-Jensen, L. & Lang, A. 2001:Equivalent dose estimation using a single aliquot of polymineralfine grains. Radiation Measurements 33, 73–94.

Behre, K.-E. 1989: Biostratigraphy of the last glacial period in Eur-ope. Quaternary Science Reviews 8, 25–44.

Behre, K. E. & van der Plicht, J. 1992: Towards an absolute chronol-ogy for the last glacial period in Europe: Radiocarbon dates fromOerel, northern Germany.Vegetation History and Archaeobotany 1,111–117.

Bøtter-Jensen, L., Bulur, E., Duller, G. A. T. & Murray, A. S. 2000:Advances in luminescence instrument systems. Radiation Measure-ments 32, 523–528.

Choi, J. H., Duller, G. A. T. &Wintle, A. G. 2006: Analysis of quartzLM-OSL curves. Ancient TL 24, 9–20.

Frodin, G. 1954: De sista skedena av Centraljamtlands glaciala his-toria. Geographica 24, 1–251.

Fuchs, M. & Owen, L. A. 2008: Luminescence dating of glacial andassociated sediments: Review, recommendations and future direc-tions. Boreas 37, 636–659.

Garcıa Ambrosiani, K. 1990: Pleistocene Stratigraphy in Centraland Northern Sweden. Ph.D. thesis. Department of Quaternary Re-search, Report 16, 1–15. University of Stockholm.

Hansen, L., Funder, S., Murray, A. S. & Mejdahl, V. 1999: Lumi-nescence dating of the last Weichselian glacier advance in EastGreenland. Quaternary Geochronology (Quaternary Science Re-views) 18, 179–190.

Hattestrand, M. 2008: Vegetation and Climate During Weichselian IceFree Intervals in Northern Sweden. Ph.D. thesis. Dissertation from

Table 5. 14C ages from twigs.

Sample Stratigraphic unit 14C age years BP

LuS 6957 F 39 200�2000LuS 6958 E 440 000


the Department of Physical Geography and Quaternary Geology 15,1–35. Stockholm University.

Heijnis, H. 1992: Uranium/Thorium Dating of Late Pleistocene PeatDeposits in N.W. Europe. Thesis, Rijksuniversiteit Groningen, 149 pp.

Helmens, K. F., Bos, J. A. A., Engels, S., Van Meerbeeck, C. J.,Bohncke, S. J. P., Renssen, H., Heiri, O., Brooks, S. J., Seppa, H.,Birks, H. J. B. & Wohlfarth, B. 2007a: Present-day temperatures innorthern Scandinavia during the last glaciation. Geology 35,987–990.

Helmens, K. F., Johansson, P. W., Rasanen, M. E., Alexanderson, H.& Eskola, K. O. 2007b: Ice-free intervals at Sokli continuing intoMarine Isotope Stage 3 in the central area of the Scandinavian gla-ciations. Geological Society of Finland, Bulletin 79, 17–39.

Houmark-Nielsen, M. 2007: Extent and age of Middle andLate Pleistocene glaciations and periglacial episodes in southernJylland, Denmark. Bulletin of the Geological Society of Denmark 55,9–35.

Houmark-Nielsen, M. 2008: Testing OSL failures against a regionalWeichselian glaciation chronology from southern Scandinavia.Boreas 37, 660–677.

Houmark-Nielsen, M. & Kjær, K. H. 2003: Southwest Scandinavia,40–15 kyrBP: Palaeogeography and environmental change. Journalof Quaternary Science 18, 769–786.

Johnsen, S. J., Dahl-Jensen, D., Gundestrup, N., Steffensen, J. P.,Clausen, H. B., Miller, H., Masson-Delmotte, V., Sveinbjornsdot-tir, A. E. & White, J. 2001: Oxygen isotope and palaeotemperaturerecords from six Greenland ice-core stations: Camp Century, Dye-3, GRIP, GISP2, Renland and NorthGRIP. Journal of QuaternaryScience 16, 299–307.

Kjær, K. H., Lagerlund, E., Adrielsson, L., Thomas, P. J., Murray, A.& Sandgren, P. 2006: The first independent chronology forMiddle and LateWeichselian sediments from southern Sweden andthe island of Bornholm. Geologiska Foreningens; StockholmForhandlingar 128, 209–220.

Kulling, O. 1945: Om fynd av mammut vid Pilgrimstad i Jamtland.Geological Survey of Sweden C 473, 61 pp.

Kulling, O. 1967: Yttrande med anledning av J. Lundqvists foredrag‘Submorana sediment i Jamtland’. Geologiska Foreningens i Stock-holm Forhandlingar 89, 123–125.

Lagerback, R. 2007: Ventifacts – means to reconstruct glacial develop-ment and palaeoenvironment in northern and central Sweden. Geo-logiska Foreningens; Stockholm Forhandlingar 129, 315–324.

Larsen, E. & Mangerud, J. 1992: Subglacially Formed Clastic Dikes.Geological Survey of Sweden, SGU Ca 81, 163–170.

Linden, M., Moller, P. & Adrielsson, L. 2008: Ribbed moraineformed by subglacial folding, thrust stacking and lee-side cavityinfill. Boreas 37, 102–131.

Lundqvist, J. 1967: Submorana sediment i Jamtlands lan. GeologicalSurvey of Sweden C 618. 267 pp.

Lundqvist, J. 2002:Weichseltidens huvudfas. In Freden, C. (ed.): Bergoch jord, 124–135. Sveriges Nationalatlas, Vallingby.

Lundqvist, J. & Miller, U. 1992: Weichselian Stratigraphy and Glacia-tions in the Tasjo–Hoting Area, Central Sweden. Geological Survey ofSweden C 826, 35 pp.

Lunkka, J. P., Murray, A. & Korpela, K. 2008: Weichselian sedimentsuccession at Ruunaa, Finland, indicating a Mid-Weichselian ice-free interval in eastern Fennoscandia. Boreas 37, 234–244.

Mangerud, J. 2004: Ice-sheet limits on Norway and the Norwegiancontinental shelf. In Ehlers, J. & Gibbard, P. L. (eds.): QuaternaryGlaciations – Extent and Chronology, 271–294. Elsevier, Am-sterdam.

Murray, A. S. & Wintle, A. G. 2000: Luminescence dating of quartzusing an improved single-aliquot regenerative-dose protocol. Ra-diation Measurements 32, 57–73.

Murray, A. S. & Wintle, A. G. 2003: The single aliquot regenerativedose protocol: Potential for improvements in reliability. RadiationMeasurements 37, 377–381.

Murray, A. S., Marten, R., Johnson, A. & Martin, P. 1987: Analysisfor naturally occurring radionuclides at environmental concentra-tions by gamma spectrometry. Journal of Radioanalytical and Nu-clear Chemistry Articles 115, 263–288.

Nilsson, T. 1972: The Pleistocen. 508 pp. Esselte Studium, Stockholm.Olsen, L., Sveian, H. & Bergstrøm, B. 2001: Rapid adjustments of thewestern part of the Scandinavian Ice Sheet during theMid and LateWeichselian – a newmodel.Norsk Geologisk Tidsskrift – NorwegianJournal of Geology 81, 93–118.

Prescott, J. R. & Hutton, J. T. 1994: Cosmic ray contributions to doserates for luminescence and ESR dating: Large depths and long-term time variations. Radiation Measurements 23, 497–500.

Preusser, F., Andersen, B. G., Denton, G. H. & Schluchter, C. 2005:Luminescence chronology of Late Pleistocene glacial deposits inNorth Westland, New Zealand. Quaternary Science Reviews 24,2207–2227.

Robertsson, A.-M. 1986: The diatom flora of the till-covered sedi-ments at Pilgrimstad, Jamtland. Striae 24, 2123–2220.

Robertsson, A.-M. 1988a: Biostratigraphical Studies of Interglacialand Interstadial Deposits in Sweden. Ph.D. thesis. Department ofQuaternary Research, Report 10, 1–19. University of Stockholm.

Robertsson, A.-M. 1988b: A reinvestigation of the biostratigraphy ofthe till-covered sediments at Pilgrimstad and Oje, central Sweden.In Robertsson, A.-M. (ed.): Biostratigraphical Studies of Inter-glacial and Interstadial Deposits in Sweden, 1–35. University ofStockholm, Department of Quaternary Research, Stockholm.

Robertsson, A.-M. & Garcıa Ambrosiani, K. 1992: The Pleistocene inSweden – a Review of Research, 1960–1990. Geological Survey ofSweden Ca 81, 299–306.

Salonen, V.-P., Kaakinen, A., Kultti, S., Miettinen, A., Eskola, K. O.& Lunkka, J. P. 2008: Middle Weichselian glacial event in the cen-tral part of the Scandinavian Ice Sheet recorded in the Hitura pit,Ostrobothnia, Finland. Boreas 37, 38–54.

Spencer, J. Q. & Owen, L. A. 2004: Optically stimulated luminescencedating of Late Quaternary glaciogenic sediments in the upperHunza valley: Validating the timing of glaciation and assessingdating methods. Quaternary Science Reviews 23, 175–191.

Svendsen, J. I., Alexanderson, H., Astakhov, V. I., Demidov, I.,Dowdeswell, J. A., Funder, S., Gataullin, V., Henriksen,M., Hjort,C., Houmark-Nielsen, M., Hubberten, H. W., Ingolfsson, O.,Jakobsson, M., Kjær, K. H., Larsen, E., Lokrantz, H., Lunkka, J.P., Lysa, A., Mangerud, J., Matiouchkov, A., Murray, A., Moller,P., Niessen, F., Nikolskaya, O., Polyak, L., Saarnisto, M., Siegert,C., Siegert, M. J., Spielhagen, R. F. & Stein, R. 2004: Late Qua-ternary ice sheet history of northern Eurasia. Quaternary ScienceReviews 23, 1229–1271.

Thomas, P. J., Murray, A. S., Kjær, K. H., Funder, S. & Larsen, E.2006: Optically stimulated luminescence (OSL) dating of glacialsediments from Arctic Russia – depositional bleaching and metho-dological aspects. Boreas 35, 587–599.

Ukkonen, P., Arppe, L., Houmark-Nielsen, M., Kjær, K. H. & Karhu,J. A. 2007:MIS 3 mammoth remains from Sweden – implications forfaunal history, palaeoclimate and glaciation chronology.QuaternaryScience Reviews 26, 3081–3098.

Walker, M. J. C., Bjorck, S., Lowe, J. J., Cwynar, L. C., Johnsen, S.,Knudsen, K.-L. & Wohlfarth, B. & INTIMATE group 1999: Iso-topic ‘events’ in the GRIP ice core: A stratotype for the Late Pleis-tocene. Quaternary Science Reviews 18, 1143–1150.

Wohlfarth, B. 2009: Ice-free conditions in Fennoscandia duringMarine Oxygen Isotope Stage 3? Swedish Nuclear Fuel and WasteManagement Company (SKB) Report TR-09-12, 1–48.


Paper III

1

OSL ages in central Norway confirm a MIS 2 interstadial (25-20 ka) and a dynamic Scandinavian ice sheet Timothy F. Johnsen1,3, Lars Olsen2 and Andrew Murray3 1Corresponding author. Department of Physical Geography and Quaternary Geology, Stockholm University, 106 91 Sweden, [email protected] 2NGU, Geological Survey of Norway, N-7491 Trondheim, Norway, [email protected] 3Nordic Laboratory for Luminescence Dating, Aarhus University, Risø National Laboratory, Frederiksborgvej 399, Build. 201, P.O. 49, DK-4000 Roskilde, Denmark

Submitted to Quaternary Science Reviews, February, 2010 © Timothy F. Johnsen, ISBN: 978-91-7447-068-0, ISSN: 1653-7211

Abstract Recent work has suggested that the Scandinavian ice sheet was much more dynamic than previously be-lieved, and its western marine-based margin can provide an analogue to the rapid-paced fluctuations and deglaciation observed at the margins of the Antarctic and Greenland ice sheets.

In this study we used a complimentary dating technique, OSL (Optically Stimulated Luminescence dat-ing), to confirm the existence of the Trofors interstadial in central Norway; an ice-free period that existed from ~25 to 20 ka recorded at multiple sites throughout Norway (cf. Andøya interstadial) and that divides the Last Glacial Maximum (LGM) into two stadials. OSL signal component analysis was used to optimize data analysis, and internal (methodological) tests show the results to be of good quality. Both large and small aliquots gave consistent OSL ages (22.3 ±1.7 ka, n = 7) for sub-till glaciofluvial/fluvial sediments at the Langsmoen stratigraphic site, and an apparent old age (~100 ka) for a poorly bleached sample of glacio-lacustrine sediment at the nearby stratigraphically-related Flora site. Eight radiocarbon ages of sediment from the Flora site gave consistent ages (20.9 ±1.6 cal. ka BP) that overlap within 1� with OSL ages from the nearby Langsmoen site.

The similarity in age within and between these stratigraphically-related sites and using different geo-chronological techniques strongly suggests that this area was ice-free around ~21 or 22 ka. The existence of the Trofors interstadial along with other interstadials during the Middle and Late Weichselian (MIS 3 and MIS 2) indicates that not only the western margin, but the whole western part of the Scandinavian ice sheet, from the ice divide to the ice margin was very dynamic. These large changes in the ice margin and accom-panying drawdown of the ice surface would have affected the eastern part of the ice sheet as well. Keywords: Optically stimulated luminescence (OSL) dating, Scandinavian ice sheet, Trofors interstadial, MIS 2, LGM, Norway

Introduction

Stratigraphic and chronologic study of sub-till sediments from many sites throughout Norway has revealed that the Scandinavian ice sheet behaved in a much more dynamic manner than previously believed (Olsen 1997, Olsen et al. 2001a,b,c, 2002). Four interstadials during the Middle and Late Weichselian glaciation mark periods of major ice retreat, and sub-till sediments from these peri-ods have been identified and dated at many sites throughout Norway, indicating near-synchronous behaviour of the western margin of the ice sheet (Figs. 1 and 2). These interstadials are the Middle Weichselian pre-Hattfjelldal interstadial (~55 to 45 ka, cf. Bø and Austnes interstadials at the western coast, Andersen et al. 1981, 1983, Mangerud et al.

1981, 2003, 2010, Larsen et al. 1987), Hattfjelldal interstadial I (~39 to 34 ka, cf. Ålesund and Sandes interstadials at the western coast, Andersen et al. 1981, Mangerud et al. 1981, 2003, 2009), Hattfjelldal interstadial II (~32 to 29 ka, cf. two dates of bones representing the Hamnsund intersta-dial at the western coast, Valen et al. 1996), and Trofors interstadial (~25 to 20 ka, cf. Andøya in-terstadial at the northwestern coast, Vorren et al. 1988) with intervening stadials bracketing the Tro-fors interstadial termed LGM 1 and LGM 2 (Last Glacial Maximum; Fig. 2). The work by Olsen and colleagues represents a significant leap forward in our knowledge, and has important implications for our understanding of ice sheet dynamics and pa-leoglaciology, vegetation dynamics, and the rela-tionships between climate change, sea level

Timothy F. Johnsen, Lars Olsen and Andrew Murray

2

Fig. 1: Overview map with locations of the Hattfjelldal I, Hattfjelldal II and Trofors interstadials and their correlatives in different parts of Norway. Langsmoen “L” and Flora “F” stratigraphic sites indicated. Locations of other sites from Finland and Sweden, also mentioned in the paper are also indicated, and several of the sites where bones of mammoth “M” have been found and dated to Middle and early Late Weichselian age are also included (see the main text for references). Curved large and small arrows are for-mer ice streams, solid line is the LGM ice margin, and the dashed line is the Younger Dryas ice margin (compiled from various sources by Olsen et al. 2001a). Ruunaa is just off the east edge of map at N 63°26’ latitude. In cases where two or more of the Tro-fors interstadial sites are located within a few hundred metres to a few kms one dot in the map may represent more than one site. Thus, the sum of dots and triangles with dots is less than the actual number of Trofors interstadial sites (42).

OSL ages in central Norway confirm a MIS 2 interstadial (25-20 ka) and a dynamic Scandinavian ice sheet

3

change, physiography, and Earth rheology. Knowl-edge of very dynamic ice-margins in the past is especially pertinent given recent observations of ongoing rapid but episodic glacier acceleration and thinning from marine-terminating sectors of the Greenland and West Antarctica ice sheets (Shep-herd and Wingham 2007) and their possible im-pacts on our environment. The landforms and sediments of the former Scandinavian ice sheet are accessible for study, and its western marine-based margin can provide an analogue to the rapid-paced fluctuations and deglaciation observed at the mar-gins of contemporary polar ice sheets.

Despite the importance of the work by Olsen and colleagues it has not been fully embraced by the scientific community because it depends mostly (~60%) on results from AMS radiocarbon dating of glacial sediments with low organic carbon content, a method which has some inherent uncertainties (Olsen et al. 2001c) and consequently has been met by some scepticism (e.g., Mangerud 2004). An-other reason is that results from this work represent a significant departure from earlier views of a more stable ice sheet particularly during the Late Weich-selian (Denton and Hughes 1981, Mangerud 1991, 2004, Donner 1995, Ehlers 1996). Given the impli-cations and controversy of this work it is important to assess it further, and the optically stimulated luminescence (OSL) dating technique has devel-oped to a point that it can be confidently applied in

dating Quaternary sediments (Murray and Olley 2000, Murray and Wintle 2000, Duller 2004, Lian and Roberts 2006, Wintle 2008). Thus, to improve our understanding of the ice dynamics for the western portion of the Scandinavian ice sheet we have used the OSL dating technique to date sub-till sediments from central Norway as an independent dating control for comparison to radiocarbon dated sediments.

A dynamic ice sheet

The dynamics of the Scandinavian ice sheet have been explored by a number of workers (e.g., Sejrup et al. 1994, 2000, 2009, Arnold et al. 2002, Hou-mark-Nielsen and Kjær 2003, Mangerud et al. 2004), and the work by Olsen and collaborators (Olsen 1997, Olsen et al. 2001a,b,c, 2002) is par-ticularly relevant to this study. Based upon the regional Quaternary stratigraphy, using data from more than 50 sections from own studies and sev-eral others, more than 300 dates (of which ~120 were radiocarbon dates taken from own projects), fossil content and some paleomagnetic data, glacia-tion curves have been constructed for nine tran-sects from inland to shelf covering the latitudinal range of Norway (1800 km; Fig. 1). Comparison of these curves, including statistical analysis of the datings, indicates that the Middle to Late Weich-selian glaciation in Norway was punctuated by four major ice recessions (Fig. 2; details presented above). These interstadials correlate well with peaks in the ice-rafted debris record. This recon-struction of ice oscillations is based mainly (~60%) on AMS radiocarbon dates of glacial sediments with low organic carbon content (0.2-1.5%), which gave promising results with respect to accuracy and precision (Olsen et al. 2001c). As the age esti-mate of the oldest of these interstadials is close to the age-limit of the radiocarbon technique, we will for the remaining part of this paper only deal with the three younger interstadials. The rapid and rhythmic ice fluctuations, as reconstructed in this model, have been fairly synchronous in most parts of Norway. Given the regional synchroneity and rapidity of glacial fluctuations it is suggested that in addition to precipitation, the mountainous fjord and valley topography, glacial isostasy and relative sea level changes were probably more important for the size of the glacial fluctuations than were air

Fig. 2: Frequency distribution of uncalibrated radiocarbon dates (running mean ages ±l�) of organic-bearing sediments and shells from four ice-free intervals separated by glacialepisodes during the Middle to Late Weichselian. The sug-gested names, the number of dates (N) and their mean age (nin 14C ka) for each ice-free interval are indicated. The differ-ence in mean ages between these groups of dates is significantat a 99% confidence level (based on data presented in Olsen etal. 2001b). See text for corresponding calibrated ages. Figurefrom Olsen et al. 2002.


4

temperature changes (Olsen et al. 2001a; cf. Ar-nold et al 2002, Hubbard et al. 2009).

Does the dynamic ice model described by Olsen et al. (2001a) allow for realistic ice retreat and advance rates? Ice retreats of 200-400 m per year and ice advances of the same rate would be much more than needed to fit with the age model for the described Middle to Late Weichselian glacier fluc-tuations. If Norway topographically had a tabular form with no deep fjords and valleys and no high mountains, then such rates for ice retreats and ad-vances would seem to be extremely high and quite unrealistic. However, Norway is highly dissected with deep fjords and valleys and with high moun-tains between. This means that the ice retreats from the coast and landwards occurred mainly through calving along the fjords, which generally is a very rapid process, and the ice advances are initiated from numerous ice growth areas between the fjords

and in the inland areas. The ice advances along the main fjords were supported by ice flowing from most tributary valleys and fjords. Consequently, in this model, the average rate of advance for the ice sheet moving from inland to coast may have been very rapid, possibly up to 200-400 m per year along fjords, although each contributing glacier may have advanced at much slower rates, less than 200 m per year. The west-east distance between the outer coastline and the Langsmoen and Flora sites is ~110-113 km, and with an ice retreat rate as high as 200-400 m per year the ice margin could have retreated from the coast to Langsmoen and Flora within less than 300-600 years. This rate of ice retreat, and a similar rate of ice advance, may be slowed down considerably, to half size or less, and the glacial system would still be very dynamic and still in agreement with the model described by Olsen et al. (2001a). The model may therefore be

Fig. 3: Location of sites where the OSL dating samples were taken (Øysand, Langsmoen, Flora), shown in a regional perspective. Trondheim and the adjacent fjord are in the northwestern corner. The Norwegian-Swedish international border that also forms the water divide is shown on right.


5

characterized as robust and not sensitive to small and moderate changes in ice retreat and ice ad-vance rates.

Methods

Stratigraphy and sedimentology

In this study we have focused on three stratigraphic sites, the Flora, Langsmoen and Drivvollen sec-tions, within the Selbu region of central Norway (Figs. 1, 3 and 4) approximately 120 km east-southeast of the city of Trondheim. The sections are located in the Nea River valley (relief ~300 m)

which drains into Selbu Lake which in turn drains into the Nid River which flows to the port city of Trondheim on Trondheim Fjord. The Flora section is located at 63° 6' 46"N, 11° 18' 29"E, the Langsmoen section is located approximately 4 km down-valley at 63° 8' 5"N, 11° 15' 29"E, and the Drivvollen section is located midway between Langsmoen and Flora sections. The watershed divide for the headwaters of this valley, and the Norway-Sweden border is located ~50 km up-valley from the sections.

The sections at the Flora site were studied for the first time in 1997 and at Langsmoen and Driv-vollen three years later. The sections of the Flora

Fig. 4: Location of the Flora, Langsmoen ad Drivvollen sites in the Nea River valley, with indication of the main regional ice flowdirection and the position of a damming ice barrier in the valley at some point in the history of deposition of sediments at Flora.


6

and Langsmoen sites were exposed by sliding and removal of surficial slope material due to heavy rain which followed a short time after removal of trees and other slope-stabilizing vegetation. The sections at Drivvollen were exposed by road exca-vations. Sediments were studied using standard observational techniques – sediment texture and structure, paleoflow direction indicators, bed con-tact relationships and lateral continuity and thick-nesses of units – and lithofacies were identified and interpreted based on their characteristics and asso-ciations, following sedimentologic principles (e.g., Middleton and Hampton 1976, Allen 1982).

Sampling, preparation and measurement

The OSL dating technique is used to estimate the time elapsed since buried sediment grains were last exposed to daylight and is a reliable chronological tool, reflected in its growing popularity in Quater-nary studies (Murray and Olley 2000, Murray and Wintle 2000, Duller 2004, Lian and Roberts 2006, Wintle 2008).

Five OSL samples were collected in August 2005 and June 2007 for the Langsmoen and Flora sections at the levels indicated in Figs. 5 and 6. The sites were overgrown with herbaceous and shrub vegetation since the original detailed stratigraphic work was completed but the stratigraphic units could nonetheless be identified securely. In addi-tion, a sample of young fluvial sediment was col-lected from the riverbank at the mouth of the Gaula River at Øysand (Fig. 3; 63° 20’ 27”N, 10° 13’ 44”E; ~13 km south-southwest of Trondheim; at a depth of 25 cm) to evaluate if the OSL signals from fluvial sediments in this region are adequately bleached (i.e., zeroed). These sediments are ex-pected to be less than 1 ka old and to have been transported at least 5 km from a landslide area upstream (Sveian et al. 2006).

The samples were collected using opaque plas-tic tubes and stored in an opaque box until opened under darkroom conditions at the Nordic Labora-tory for Luminescence Dating (2005) or Stockholm University (2007) where samples were apportioned and wet sieved. Final preparation, including heavy liquid separation (2.62 g/cm3) to remove feldspars, treatment with 10% HCl for 5-30 minutes, 10% H2O2 for 15-30 minutes, 38% HF for 60-120 min-utes and 10% HCl again for 40 minutes was per-formed at the Nordic Laboratory for Luminescence

Dating, where the OSL measurements were also completed.

The samples were analyzed using large (8 mm) and small (2 mm) aliquots of quartz, ranging in mean grain size from 83 to 170 �m, on Risø TL/OSL-readers equipped with calibrated 90Sr/90Y beta radiation sources (dose rate 0.089-0.32 Gy/s), blue (470±30 nm; ~50 mW/cm2) and infrared (880 nm, ~100 mW/cm2) light sources, and detection was through 7 mm of U340 glass filter (Bøtter-Jensen et al. 2000). Analyses employed post-IR blue SAR-protocols (Murray and Wintle 2000, 2003, Banerjee et al. 2001), adapted to suit the samples based on dose recovery and preheat ex-periments (preheat 260° for 10 s, cut-heat 220°). 100 s of illumination at 280° between cycles im-proved recuperation (response to zero dose). Some of the small aliquots were measured using a short-ened IR blue SAR-protocol to increase measure-ment productivity. The shortened protocol had three regenerative doses, selected to bracket the equivalent dose. The third dose was a recycled dose, and since recuperation rarely was greater than 5% using the full SAR-protocol, we excluded the zero dose measurements from the shortened protocol.

Dose rates were calculated from gamma spec-trometry data (Murray et al. 1987) and included the cosmic ray contribution (Prescott and Hutton 1994). Water content (weight percent) was meas-ured in the laboratory by weighing re-compacted sediment at natural, saturated and dry water con-tent. The historical water content was estimated to be 0.75 the saturated water content. This seems reasonable given that the sections are located on valley walls that are subject to significant ground-water flow during parts of the year. Also, sedi-ments in these sections have been observed at vari-ous times during spring and autumn to be quite wet, although their locations are well above the local groundwater level and the sediments appear to be quite dry during the summer season.

Samples for radiocarbon dating from Flora were taken from positions well inside the fine-grained units and therefore well away from potentially contaminated surfaces. Water flow through the sediments was not a major problem for the radio-carbon dates, because the water flow goes mainly through the adjacent coarse grained sediments and may therefore contaminate the boundary surfaces


7

Fig. 5: Generalised stratigraphic log from Flora, based on several sections which are partly separated, both in height and width, butlocated within a 200m by 300m area in the valley side (see Fig. 4). Shown are the location of eight radiocarbon samples and cali-brated ka age (black dots) and OSL sample 071301 (open circle ‘01’ at 252 m). Till fabric location indicated by circles with cross; solid arrow and value indicate the direction of the main cluster of data; dashed arrow is direction based on less than 25 clasts.


8

Fig. 6: Generalized stratigraphic log from the combined section based on several, small, overlapping sections in height and width at Langsmoen. Shown is the location of four OSL samples with the ka age using large and small aliquots (number inside circle is the last two digits of the sample number; see Table 3 and 4). Till fabric location indicated by circles with cross; solid arrow and value indicate the direction of the main cluster of data.


9

of fine-grained sediment units, but probably very little inside the fine-grained units. As well, there were no visible traces of oxidation, root penetra-tion, or fissures within the sampled sediment. Samples were air dried and stored in a cold room with a temperature of 0-4°C to avoid bacterial contamination (e.g., Wohlfarth et al. 1998). Sam-ples were measured at the Accelerator Mass Spec-trometry (AMS) facility at the University of Utrecht, which uses a 6 MV Van de Graaff tandem accelerator (van der Borg et al. 1997). Graphite targets were used for the AMS analysis. A standard alkali-acid-alkali treatment was used to remove contamination by humic acids and carbonate. After combustion of the organic carbon samples, the obtained CO2 gas was routinely separated from SO2 with KMnO4 (see Olsen et al. 2001c and refer-ences therein). The insoluble fraction was used to derive the sample age. At Langsmoen it was more difficult to find fine-grained units that were thick enough to avoid a potential contamination prob-lem, and therefore we have no radiocarbon dates from this site.

As we were calibrating radiocarbon dates (i.e., interstadial dates) from the literature that were well beyond the calibration curve IntCal04 (Reimer et al. 2004), we used Fairbanks et al. (2005) with the calibration curve Fairbanks0107 for all radiocarbon dates. Note that calibrations using IntCal04 versus Fairbanks0107 of radiocarbon dates from Flora section produced calibrated ages differing by less than one percent.

OSL data analysis

Simple component analyses of the continuous wave OSL data from selected aliquots was under-taken using SigmaPlot 10.0, based on the parame-ters and formulas of Choi et al. (2006). The results of the component analysis are discussed further below, but based on such information from dose recovery measurements we chose the 0-0.64 s and the 0.96-2.08 s portions of the signals for peak and background integration limits for all aliquots. The equivalent doses were then calculated in Risø Lu-minescence Analyst 3.24 (using exponential curve fitting) and in Microsoft Excel.

Aliquots were initially accepted when (1) the absolute value of the recycling ratios was within the range 0.7 to 1.3, and (2) when accounting for their uncertainty their value was within 15% of

unity. The first step was introduced to address aliquots with poor recycling ratios that would be accepted because they had unacceptably large un-certainties (~>30%). Aliquots also had to pass the following rejection criteria: recuperation <5%, equivalent dose error <50% and signal more than three sigma above the background. Decay and growth curves also had to be regular. More conser-vative rejection criteria values were used for the shortened protocol employed for some small ali-quots; they were scaled appropriately based upon the actual structure of the sequence used. Ages for all samples were calculated using the mean and median of the equivalent dose population of ac-cepted aliquots for each sample, as well as using the dry and saturated water contents.

A sensitivity analysis was also performed to de-termine quantitatively which uncertainties have the largest effect on the age. Using reasonable esti-mates of uncertainty for each parameter, ages were recalculated after adjustments were made to each of the parameters in turn. The estimated uncertain-ties (1 standard deviation) used for each parameter in these calculations were: depth below surface ±1 m, elevation ±50 m, grain size ±20%, water content (gamma and beta) ±20%, dose rate gamma ±5%, dose rate beta ±5%, internal dose rate ±30%, den-sity ±10%, cosmic ray contribution ±5%, and beta source calibration ±2%.

Results

The Flora, Langsmoen and Drivvollen stratigraphic sites

The Flora site The Flora site consists of several sections within a 200 m by 300 m area in the valley side (Fig. 4). The four main sections are situated 337-340 m a.s.l., 301-310 m a.s.l., 270-280 m a.s.l., and 252-265 m a.s.l. (Fig. 5 and Table 1). The lowest-lying of these was later extended to bedrock by excavat-ing mainly through a boulder-rich lower till and a coarse-grained diamicton, probably a till, with heavily weathered material at the base. The vertical intervals between the main sections have not been fully excavated and cleaned, but have been checked with the help of a spade and a 1 m-long drilling rod with a 1 inch-wide track for sediment sampling. This is done to make sure that no inter-


10

vening main sediment unit is missing in the strati-graphic record.

The ~25 m high lowermost section starts with a 1-4 m thick coarse-grained gravelly sandy diamic-ton, which is inferred to be a till, at the bedrock between ~240 and 245 m a.s.l. (Fig. 5; unit G). The bedrock is mica schist, with a glacially striated surface partly preserved. The striation indicates ice flow towards 280°. In other parts the bedrock is weathered and some of the weathered material has been reworked in the till. Overlying the lowermost till follows a 7-8 m-thick silty, sandy, compact, brownish-grey to greenish-grey till (unit F) with many cobbles and boulders. Clast fabrics in these two lower tills indicate preferred maxima between 280° and 305°, which correspond with striation on bedrock as well as on a big boulder in the contact zone between the tills. Overlying these tills follows unit E3 (Fig. 5), which is a 4.5 m-thick series of alternating layers of laminated sand and silt in a number of fining-upwards sediment successions. Thin irregular layers of diamict material or grav-elly sand, as well as some indications of glacial

deformation, occur also in unit E3. This unit is overlain by a brownish-grey till of, at least, 10 m thickness. It is separated in three parts, D1, D2 and D3, by intervening silt and sand horizons E1 and E2. The matrix of tills D1, D2 and D3 changes with a general fining-upwards trend from gravelly sandy in till D3 to sandy in till D2 and more silty in till D1. The compactness of these tills increases as the matrix changes upwards to a more fine-grained character. Overlying these tills follows again sev-eral metres (50-60 m) of till in beds (subunits B1, B2 and B3) alternating with laminated sand and silt (subunits C1, C2 and C3) of thicknesses from a few centimetres to one or several metres. Tills B1, B2 and B3 are all relatively compact and mainly silty and dark grey in colour, with a bluish-grey tone in most parts of tills B1 and B2.

The Flora site is completed on top by a 3 m-thick sequence of two coarse-grained, brownish-grey and medium to loose compacted tills (Fig. 5; unit A). Full fabric analyses have not been per-formed in the tills younger than D3, but the orien-tation of some 15-20 clasts and a few fold-axes

Table 1: Sedimentary diamict and other lithofacies used in Figures 5 and 7. Modified from Olsen et al. (1996). Code Facies description InterpretationDcs Clast supported diamicton, stratified, occurrence often in combination with S-

, G-, sDm-, gDm-, and Dms-facies.Ablation melt-out till and poorly sorted glacifluvial debris flow/ debris fall material.

Dmm Matrix supported, massive, normal silty-sandy composition. Variable clast fabric from random to strong parallel or occasionally transverse.

Till or debris flow deposits (and also some glacimarine sediments).

Dmm(s) Dmm with dense internal shearing, often with strong parallel clast fabric. Lodgement till.

Dmm(r) Dmm with evidence of resedimentation, often medium strong to strong parallel clast fabric.

Mainly lodgement and basal melt-out till.

Dms Matrix supported, stratified diamicton, with sand lenses and linings around clasts. Medium to relatively strong parallel clast fabric, transverse fabric may occur in some areas.

Mainly basal melt-out till, but often combined with intervening or alternating zones of lodgement till.

Dms(r) Dms with evidence of resedimentation, for instance with rafts of deformed sand/silt/clay laminae and abundant silt/clay stringers and rip-up clasts. Clasts fabric variable from random to fairly strong.

Mainly basal melt-out, but often combined with zones of lodgement till.

sDm Sand-enriched matrix supported diamicton, mainly massive (sDmm), but occasionally heterogeneous or stratified (sDms).

Tillised sediment/deformation till containing sand eroded from local substrate; and melt-out till.

gDm Gravel-enriched matrix supported diamicton, mainly massive, but occasionally heterogeneous or stratified.

Mainly melt-out till, but in some cases possibly also tillised sediment or deformation till with gravel eroded from local substrate.

cDm Cobble- and boulder-enriched matrix supported diamicton, mainly massive, but occasionally heterogeneous or stratified.

Mainly melt-out till, but in some cases possibly also tillised or deformed sediment reworked from local substrate.

FSi Fine siltSi SiltCSi Coarse siltFS Fine sandS/SSi Sand/ Sand, siltCS Coarse sandG/GSi/GS Gravel/ Gravel, silt/ Gravel, sandD Diamicton Till/ mudflow/ debris flow/ and debris fall material, mainly.Sil/Sild Silt, laminated/ Sil, deformed Fallout with or without traction, quit water/ later deformation. Si(r) Silt, resedimented Erosion and resedimentation, varied environments.Sl/Sm Sand, laminated/ Sand, massive Settling from suspension, shifting -/ increased sediment supply. Sld Sand, laminated, deformed Laminated sand, later deformed by glacial processes, mainly.Sp/Spd Sand, planar bedding/ Sp, deformed Fallout with or without traction, quit water/ later deformation.

Vertical accretion by settling of sediment from suspension in quiet water lakes (glaci-/lacustrine) and marine basins. This include FSi, Si, CSi, FS, and often also S/SSi facies.

Lateral accretion, sediments supplied by traction with fallout, glaci-/fluvial and marine coarse-grained sediments. This include S/SSi, CS, and G/GSi/GS facies.


11

from glacial deformation at three levels in the up-per 65 m of the Flora stratigraphy have helped to indicate approximate ice flow directions (stippled arrows in Fig. 5). All these are trending towards the NW.

Radiocarbon dating of bulk organics from the silty sediments was performed at eight levels from five of the silt and sand horizons between 309 and 252 m a.s.l. (Fig. 5). The mean and standard devia-tion of the ages from Flora are 20.9 ±1.6 cal. ka BP (Table 2). Also, based on microscopic and stereo-scopic examination of parallels to most of the dated bulk-organic sediment samples, terrestrial plant remains are the main organic component for these samples. OSL results for the Flora and Langsmoen sites are presented in detail in section OSL charac-teristics and ages.

The Langsmoen site A 15-17 m high section which was comprised by several 1-2 m wide and 2-3 m high sections that were overlapping in height and width is represent-ing the Langsmoen site (Fig. 6). The lowermost sedimentary unit here, unit D, is divided in four subunits D4, D3, D2 and D1, where D4 is mainly gravel and sand, and D2 is gravel. Subunits D3 and D1 are silt and sand with some silt-layers, respec-

tively. Unit D could alternatively be described as four fining-upwards sequences, with other bounda-ries for each sub-unit than illustrated in Fig. 6. The gravel and sand are deposited in layers sloping in the downstream direction towards the north. De-formation structures like sediment injections and folding are increasing in size and frequency up-wards in the upper three subunits D3-D2-D1, with only small deformation represented in the bounda-ries of subunit D3, whereas overturned layers of sand dipping steeply, almost vertically towards the northwest occur in subunit D1. The silt in subunit D3 contains foraminifers of e.g. Cibicides lobatu-lus and Elphidium excavatum (Olsen et al. 2002). All the C. lobatulus foraminifers indicate some corrosion or dissolution, possibly due to the effect of various environments during resedimentation from upstream, older strata (of probably Eemian age). The few other species show no sign of corro-sion and may belong to the last phase of deposition and, therefore, indicate marine influence in these sediments.

Overlying the waterlain sediments, which are thought to be fluvial deltaic in origin, follows two till beds with less than 1 metre of sandy slope ma-terial on top (Fig. 6). The lower till, unit C is rela-tively compact, with a sandy-silty matrix and some boulders, cobbles and pebbles, and a brownish-grey colour. The upper till (unit B) is more silty, darker grey to bluish-grey and very compact. Till fabrics indicate preferred clast orientation towards the NW and NNW for these tills, quite similar to the regional ice flow direction based on striation and drumlin orientation (Figs. 4 and 6).

The Drivvollen site Fresh road cuts along a local road at Drivvollen (Fig. 4; Olsen et al. 2002), midway between Langsmoen and Flora, revealed in the year 2000 a stratigraphy comprising both the till units recorded at Langsmoen and an underlying glacifluvial/ flu-vial/ fluvial marine gravelly and sandy sediment unit that may correlate with the Langsmoen sub-till deltaic sediments. At Drivvollen, these sediments reach up to 215-220 m a.s.l. (Figs. 4).

Table 2: Flora section AMS radiocarbon ages of sub-till waterlain sediments (from Olsen et al. 2001c)a.

Lab no. Elevation (m)

14C age (yr BP)

Calibrated age (yr BP)b

1 UtC 5977 309 17800 ±400 21100 ±10002 UtC 5978 309 15920 ±260 19100 ±4003 UtC 5979 305 17800 ±400 21100 ±10004 UtC 5981 273 16700 ±220 19800 ±6005 UtC 5982 272 15620 ±200 18800 ±4006 UtC 5984 260 19600 ±280 23400 ±8007 UtC 6042 255 19050 ±120 22600 ±2008 UtC 5985 253 18000 ±400 21400 ±1200

mean = 20900s tdev = 1600

Notea: All were measured using the insoluble fraction. Noteb: Calibrated using Fairbanks et al. 2005. Note that calibrations differ by less than 1% from OxCal 4.0 and the IntCal 04 calibration curve (Ramsey 1995, 2001, Reimer et al. 2004). Uncertainty shown is at 1�.


12

OSL characteristics and ages

The major sample properties, settings and results are listed in Tables 3 and 4, and Figs. 7 to 11. The quartz from Langsmoen and Flora sections is in-sensitive with photon counts less than 1000 for the first 0.08 s of stimulation (Fig. 7, insets), which leads to larger uncertainties (cf. Alexanderson and Murray in press). Careful selection of peak and background channels was made (0-0.64 s and 0.96-2.08 s portions, respectively) to best isolate the fast component (Fig. 8). For the eighteen accepted dose recovery experiments the average proportion of the net component signal to the total net signal used for dose calculation was: 95 ±6% (fast), 5 ±5% (medium), and 0.3 ±0.6% (slow). These results

indicate that the selected time intervals for peak and background integration resulted in a signal that was dominated by the fast component. Doses cal-culated with this peak and background selection gave good dose recovery (1.06 ±0.03, n = 18, un-certainty as the standard error; Fig. 9) and demon-strate that the SAR protocols used were able to accurately recover a known dose given before any heating. The signals from the Flora section were close to saturation, from the flattish part of the extended growth curve, while signals from the Langsmoen section were from the steep, near-linear, early part of the extended growth curve (Fig. 7). This has implications for the reliability of the final ages, see Discussion.

Sample Elev (m) Sampling depth (m)

Lithofacies sampled Saturated Selected a Mean grain s ize used

071302 210 8.0 Deformed, laminated sand 34.7 26.0 85081357 200 18.0 Diffusely graded sand with s ilty clay rip-ups 33.2 24.9 94081358 208 10.0 Deformed, laminated sand 31.0 23.3 122081373 210 8.0 Deformed, laminated sand with mass ive, matrix-supported diamicton 24.2 18.2 170

071301 252 35.0 Deformed, diffusely graded sandy s ilt 17.0 12.8 83

Notea: Selected water content es timated to be approximately 0.75 the saturated water content.

W ater content (weight %)

Table 3: Langsmoen OSL sample properties; listed in order of sample number. Flora section results at bottom (sample 071301). Note that the text sometimes refers only to the last two digits of the sample number.

Notea: The "Dry" age represents the age of the sample when it contains no water, whereas the "Saturated" age represents the age of the sample when it is saturated with water. Together these represent the maximum age range when considering the water content. Noteb: Two times the estimated standard error of skewness derived using 2 x SES = SQRT(6/N) (Tabachnick and Fidell, 1996). Note that none of the samples are skewed to a significant degree as the skew is less than two times the standard error of skewness - except for sample 071301 (see note c below). Notec: Sample 071301 from the Flora section had many aliquots that gave estimated doses far beyond the highest regenerative dose and so were not used in the calculations. Thus, the estimated dose and apparent ages are underestimated, and skewness is negatively biased.

Table 4: Langsmoen OSL sample results and ages. Flora section results at bottom (sample 071301).

Sample Aliquot size (mm)

n Dose rate (Gy/ka)

Estimated dose (Gy)

Recycling ratio

Mean Median Drya Saturated Skew 2 x SE of skewnessb

071302 8 22 1.81 ±0.07 37.5 ±2.3 1.09 ±0.10 20.7 ±1.6 19 16.1 ±1.3 22.1 ±1.7 0.35 1.04071302 2 66 1.81 ±0.07 38.6 ±1.8 1.02 ±0.10 21.3 ±1.4 21 16.6 ±1.2 22.8 ±1.4 0.03 0.60081357 8 26 1.51 ±0.07 33.9 ±1.3 0.95 ±0.14 22.5 ±1.4 22 17.6 ±1.2 24.1 ±1.5 0.48 0.96081357 2 49 1.51 ±0.07 35.3 ±1.7 1.06 ±0.17 23.4 ±1.6 24 18.3 ±1.4 25.1 ±1.7 -0.35 0.70081358 8 18 2.01 ±0.08 41.4 ±2.5 1.02 ±0.18 20.6 ±1.5 20 16.4 ±1.3 22.0 ±1.6 0.59 1.15081358 2 20 2.01 ±0.08 45.0 ±2.4 0.98 ±0.15 22.4 ±1.6 23 17.8 ±1.3 23.9 ±1.6 -0.74 1.10081373 8 37 2.01 ±0.09 51.0 ±2.0 1.02 ±0.12 25.4 ±1.6 25 21.2 ±1.4 26.7 ±1.6 0.64 0.81

071301c 8 41 1.62 ±0.08 164.8 ±6.6 1.03 ±0.07 101.5 ±6.6 106 88.4 ±6.1 105.7 ±6.8 -0.34 0.77

Age


13

For the Langsmoen section samples, equivalent doses range between 34 and 51 Gy and dose rates between 1.5 and 2.0 Gy/ka, while for the Flora

sample is 165 Gy and 1.6 Gy/ka respectively (Ta-ble 4). Recycling is exceptionally good with a mean and standard deviation for all samples of 1.02 ±0.04, n = 8. The age of the Flora section sample is 102 ka. The seven ages from large and small aliquots from the Langsmoen section range between 25.4 to 20.6 ka with a mean and standard deviation of 22.3 ±1.7 ka. The ages from the large and small aliquots for the same sample are similar and overlap at 1�. With the exception of the Flora section sample, the distributions of the equivalent doses used to calculate the age are not significantly skewed (Table 4, see footnote). The sensitivity analysis demonstrates that the calculated age is most sensitive to changes in the equivalent dose, beta source calibration and dose rate, followed by water content (Fig. 11).

Fig. 7: Single-aliquot regenerative-dose (SAR) OSL growthcurves for Flora section sample 071301, and Langsmoensection samples 071302 and 081358; first two samples are average of two aliquots with uncertainties as standard devia-tion. The range of estimated doses used to calculate the age ofeach sample is indicated by the width of the grey shading; themean equivalent dose is also shown (black triangle). Note that equivalent doses for 071301 (from the Flora section) have a large range and are from the flatter part of the growth curve.Repeated dose points (open squares) and a zero dose point(open circle) are also shown. The inset shows the natural decaycurve; the decay curve for the young sediment sample fromØysand (081360) is also shown in inset for 081358 (sedimentdose rate of 1.7 Gy/ka).

Fig. 8: Using data from the 18 dose recovery experiments, peak and background channels were chosen that provided the overall best recycling and dose recovery, and a dominance of the fast component after background subtraction. Example of signal component analysis from sample 081357 showing the natural signal deconvoluted into the fast, medium and slow components (left axis), and the natural and modeled (sum) signal (right axis). Also shown are the peak and background portions of the signal used (0-0.64 s and 0.96-2.08 s). Note that only the first 5 s of 40 s total stimulation are shown. Deconvo-lution of dose recovery data indicated that the channel selec-tion effectively isolated the fast component and the selection of peak and background channels provided the optimum recy-cling and dose recovery.


14

Discussion

Stratigraphy and paleoenvironment

The Flora site The interpretation of the Flora stratigraphy is that it represents a succession of repeated glacial ad-vances alternating with repeated phases of ice-damming with pulses of sand and silt flowing into lateral or subglacial glaciolacustrine basins. All the waterlain sediments (E and C subunits) are thought to represent the same sedimentary environment, a glaciolacustrine/ ice-dammed lake environment. For example, the 4.5 m-thick series of alternating layers of laminated sand and silt in a number of fining-upwards sediment successions that comprise unit E3 are thought to represent repeated melt-water-pulses into an ice-dammed lake environ-ment. Small amounts of plant remains in the sedi-ments indicate that at least some vegetation may have been established during the phases of ice retreat. The relative sea level during deposition of the Flora waterlain sediments is not known, except that it was probably lower than 252 m a.s.l. since

no marine fossils have been found in these sedi-ments.

The radiocarbon datings, with a mean and stan-dard deviation of the ages of 20.9 ±1.6 cal. ka BP (Table 2), indicate that all the glacial and glacio-lacustrine oscillation phases occurred during what is traditionally thought of as the LGM interval (Fig. 5 and Table 2). The results from this site along with other dated sub-till sediments in Nor-way define the Trofors interstadial, which is brack-eted by the LGM 1 and LGM 2 stadials (e.g., Olsen et al. 2001a, Figs. 2 and 10).

The Langsmoen site The interpretation of the Langsmoen stratigraphy is that it represents a phase of considerable ice retreat with fluvial deltaic sedimentation, possibly with some marine influence at the site, followed by phases of glacial advances during the last glacia-tion. The coarse-grained lower till and the fine-grained upper till here are texturally similar and with clasts of similar lithology and are therefore probably correlative to the coarse-grained, brownish-grey lower till and the fine-grained, blu-ish-grey tills, respectively, at Flora. However, the waterlain, sub-till sediments at these two localities are different, with ice-free conditions at Langsmoen and ice proximal and ice-damming conditions represented at Flora. These different environments may have occurred during the same interval, and even possibly with ice temporarily acted as a damming barrier between the localities (Figs. 4 and 12). The content of plant remains and other organics is, however, not sufficiently high in the Langsmoen sub-till sediments to allow testing of this hypothesis by radiocarbon dating. The rela-tive sea level during deposition of the Langsmoen waterlain sediments was higher than 210 m a.s.l., which corresponds with the elevation of the eroded top of the sub-till deltaic foreset beds at Langsmoen. The relative sea level was probably at least 220 m a.s.l., in agreement with the elevation of the top of similar and possibly correlative fluvial beds, also with marine influence at the neighbour-ing Drivvollen site (Fig. 4, Drivvollen; Olsen et al. 2002). This is ~10% higher than the late-/postglacial marine limit in this area.

Fig. 9: Histogram of dose recovery tests, excluding aliquots that did not pass rejection criteria. As the mean is close to unity, this indicates that the samples can be used for OSL-dating with our analytical protocols.


15

Ice flow directions During initial ice growth and final decay of the last ice sheet, the ice flow has followed the main trends of the valleys and fjords. In the region east of Trondheim (Fig. 3) this resulted in a dominating ice flow direction during these phases towards the west and northwest. During maximum phases, the ice sheet was thicker and the ice flow was less influenced by the underlying topography, and was trending towards the northwest and north-northwest in this region.

Westerly to northwesterly trending ice flow di-rections are inferred from striation on bedrock and boulders at Flora and till fabrics in the lower tills and deformation in the sub-till sediments at both Flora and Langsmoen. These ice flow directions are in best agreement with phases during ice growth or final decay of the ice sheet. Till fabrics indicate that the fine-grained, dark bluish-grey upper tills at both sites may be a result of a more northwest to north-northwesterly trending ice flow,

which is in agreement with the main regional ice flow direction (Figs. 4 and 6).

The general glacial stratigraphy model of most valleys in Norway is characterized, not only with more of local materials, including more weathered material in the lower tills, but also often with more coarse-grained material. The general increase of fine-grained material in younger tills is thought to derive from more long-transported material which has been crushed and abraded during a longer transport history. During deglaciation the youngest tills may again include more coarse-grained mate-rials, which may derive from ice-margin oscilla-tions and a changed pattern of erosion, transporta-tion and deposition. The total Flora glacial strati-graphy shows this trend, and therefore seems to be in concert with the general model. Therefore, it seems likely that the lower tills have been depos-ited during ice growth, whereas the younger tills derive from later phases. Of course there may be exceptions from the general model in some loca-

Fig. 10: Graphical summary of OSL ages for all Langsmoen section samples, including large and small aliquots; corresponding recycling ratios below. Age calculated from the population of accepted aliquots using the mean, median, no water content, andsaturated water content. Error bars represent one standard error. The mean of the ages is 22.3 ±1.7 ka (uncertainty as standard devia-tion which ranges from 24.0 to 20.6 ka). The ages are reasonably consistent and correspond very well to the Trofors interstadial, ~25 to 20 ka (Olsen et al. 2001a). Stadials (grey) and interstadials (white) indicated on the right.


16

tions, but to really test this hypothesis carefully at each site requires much more detailed work.

OSL and ages

Of primary importance in employing OSL dating in glacial and glacially-related sediments is to de-termine whether the sediments were adequately exposed to light prior to deposition; otherwise the ages will be maximum ages (Murray and Olley 2000, Fuchs and Owen 2008, Alexanderson and Murray in press). The good correspondence in age between samples, between large and small aliquots, between mean and median values, and the insig-nificant skewness strongly indicate that the Langsmoen sampled sediment was adequately exposed to light (i.e., “well-bleached”, or “zeroed”) prior to deposition (Table 4, Fig. 10). This also justifies the use of the mean age and not for exam-ple a minimum age model for the small-aliquot data (Galbraith et al. 1999). Besides, it is very unlikely that all the samples suffer from incom-

plete bleaching, as it is not reasonable to expect sediments from different portions of the section to suffer the same degree of incomplete bleaching to produce similar ages. Also, if there were bleaching problems we would expect measurements on small aliquots to produce positively skewed dose distri-butions or at the very least a significantly younger age (Fuchs and Owen 2008, and references therein); neither is the case (Table 4), and the ages from large and small aliquots from the same given sample overlap within 1�. In addition the very low OSL signals from young fluvial sediments at Øy-sand indicate that fluvial sediments in this region can be properly bleached (see sample 081360 in bottom inset of Fig. 7).

Based on that sample 081373 contained some diamict sediment, which may have been derived from subaqueous debris flow or till tectonized into the fluvial sediments, there is a risk for incomplete bleached grains to be incorporated. This sample is also slightly older than the others (Table 4, Fig.

Fig. 11: Sensitivity analysis showing by how much the OSL age will change by percent change in a given parameter for sample (071)302, large aliquots (age = 20.7 ka). Steeper slopes indicate that the age is more sensitive to changes in the given parameter. Noteworthy is that the age is most sensitive to changes in the equivalent dose, beta source calibration and dose rate, followed by water content. The x-axis range of a parameter’s line is its 2� uncertainty. Quite similar results were acquired for the other samples.


17

10). As we measured only large aliquots for this sample, the dose distribution unfortunately cannot be used to evaluate skewness-effects of incomplete bleaching (as there is a large averaging-out effect for large aliquots; Duller 2008). Nevertheless, on the whole we conclude that the Langsmoen sam-ples were well-bleached before deposition.

In contrast, the Flora section sample gave an er-roneously old age based on comparison to the eight radiocarbon dates (Fig. 5). This is anticipated as it is not expected that glaciolacustrine sediments would be adequately bleached prior to deposition (Alexanderson and Murray in press), especially if the ice-front was nearby or if sediments were de-rived from a submerged ice-front. Unlike the well-bleached Langsmoen sediments, the equivalent dose population for the Flora sediment is very broad and from the flattish, high dose, portion of the growth curve (Figs. 8 and 10, Table 4). Note that many aliquots gave equivalent doses well beyond the highest regenerative dose as well, and were not included in the determination of the equivalent dose; thus the dose population would be even wider than portrayed in Figure 8. Even the

lowest equivalent dose for this sample gives an apparent age of about 45 ka, which is well beyond the 21 ka (n = 8) radiocarbon age of the Flora sediments. Thus it is very likely that the glacio-lacustrine sediment from Flora is poorly-bleached and therefore we reject its OSL age (Wallinga 2002).

The sediment dose rates for the various samples (1.8 ±0.2 Gy/ka) were quite consistent and suggest that there is not a significant problem with the sediment dose rates measured (Table 4). Finally, the water content history of the sampled sediment can only be estimated. As stated earlier, our best estimate is that over their lifetime the sediments experienced on average ~75% their saturation. Despite this limitation the sensitivity analysis indi-cates that even with a 20% change of the percent water content values, ages change by less than ~3% (Fig. 11).

The radiocarbon dates from the Flora section are considered to be reliable as (1) sampled sedi-ments were derived from the interior of laminae or beds that were adjacent to more hydraulically con-ductive sediments, thus minimizing possible

Fig. 12: Long-valley profile cartoon showing the relationship between stratigraphic sites and theice dam at Hoggkjølen (see Fig. 4).


18

groundwater flow and contamination through the sampled sediment, (2) the ages are consistent over the 54 meter range of the section, which is unlikely to occur if groundwater contamination was an important process, (3) the ages fit well with similar dates from other sites throughout Norway that belong to the Trofors interstadial, and finally (4) the ages are in agreement with the OSL ages from sub-till sediments from the nearby Langsmoen section. Terrestrial plant remains, which appear to be the main organic constituent of the dated mate-rial, are considered to be reliable for radiocarbon dating, and are supported by the agreement be-tween the radiocarbon and OSL ages.

Implications for ice sheet history and dynamics

The seven OSL ages of the Langsmoen section sediments of 22.3 ±1.7 ka correspond very well to the Trofors interstadial of ~25.2 to 20.1 ka which is based on 40 radiocarbon ages including eight from the Flora section of 20.9 ±1.6 cal. ka BP (Fig. 10, Table 2; Olsen et al. 2001b). Thus, the OSL ages from the Langsmoen section support the results of the AMS radiocarbon dating method on glacial sediments with low organic content (Olsen et al. 2001c), and confirm the occurrence of the Trofors interstadial in central Norway which also has been identified elsewhere in Norway (Fig. 1). The simi-larity in age within and between these stratigraphi-cally-related sites and using different geochro-nological techniques strongly suggests that this area was ice-free around ~21 or 22 ka.

The ages of the Langsmoen and Flora section sediments fall within the early to middle portion of the Trofors interstadial (Fig. 10). We imagine a paleoenvironment whereby the Langsmoen gla-ciofluvial sediments (at 200 to 210 m a.s.l.) were deposited in ice-distal conditions with local sea level at least 210 m a.s.l., possibly ~215-220 m a.s.l. if the Langsmoen sediments correlate with the nearby Drivvollen sediments as mentioned before. Further advance of the ice sheet leading into LGM 2 resulted in ice damming down-valley (Figs. 4 and 12) and deposition of till alternating with glacio-lacustrine sediments in ice-proximal conditions at the Flora site with the highest lake level up to 310 m a.s.l. Further advance of the ice sheet resulted in the moderate glaciotectonizing of these sediments and their burial beneath lodgement till. The climate

during the Andøya – Trofors interstadial was probably cold to boreal (Vorren et al. 1988, Alm 1993, Olsen et al. 2001b) with arctic to subarctic conditions where the mollusc Mya truncata, which is a temperate Atlantic water indicator, reached northwards at least to the southern coast of Norway (Olsen and Bergstrøm 2007).

The geographical extensive distribution of Tro-fors interstadial sediments, as well as that for the Hattfjelldal I and Hattfjelldal II interstadials (Fig. 3), indicates that causes behind these interstadials were regional in nature. In addition, the very dy-namic ice sheet fluctuations exceed rates of change that could be brought on by air temperatures changes alone. Thus, in addition to precipitation, the mountainous fjord and valley topography, gla-cial isostasy and relative sea level changes were probably more important for the size of the glacial fluctuations than were air temperature changes. This statement is simply based on that the fjord and valley topography, considerable glacial isostatic depression and high relative Middle to Late Weichselian sea levels are all documented and obvious important factors for ice sheet margin fluctuations (e.g., Olsen and Grøsfjeld 1999), whereas air temperature changes may be balanced with precipitation so that the glacier fluctuations are reduced, as seen from many modern glaciers.

Would the ice sheet fluctuations recorded along the western margin of the Scandinavian ice sheet also occur in other areas of the ice sheet? Argua-bly, ice streams, especially widespread in Norway, played a very important role in the dynamics of the Scandinavian ice sheet (Fig. 1), especially as they would be sensitive to relatively small changes in relative sea level and would be very effective at removing mass from more interior areas of the ice sheet and encourage retreat from both lowlands and high elevation (Shepherd and Wingham 2007, Hubbard et al. 2009); this is reflected in the re-gional response of the ice sheet (Olsen et al. 2001a). We suspect that the south-western areas of the ice sheet in Denmark and Sweden may also have responded in a similar fashion to the ice mar-gin in Norway (Hillefors 1969, 1974, Lundqvist 1992, 2004) because like Norway the south-western area of the ice sheet was adjacent to the influencing-effect of the sea and was host to ice streams in the North Sea and Baltic Sea (Fig. 1;


19

Houmark-Nielsen and Kjær 2003 and references therein).

In Denmark, the Kattegat Till and Mid Danish Till may correspond to the similarly dated LGM 1 and LGM 2 tills in Norway; and the Vennebjerg-Sejerø-Klintholm interstadial, the Lønstrup-Uranienborg-Gärdslöv interstadial and the Ribjerg-Fegge-Spøttrup-Glumslöv-Kobbelgård interstadial in Denmark and SW Sweden correspond to the Ålesund-Sandnes-Hattfjelldal I interstadial, the Hamnsund-Hattfjelldal II interstadial and the Andøya-Trofors interstadial in Norway, respec-tively (Olsen et al. 2001b; Houmark-Nielsen and Kjær 2003). While one could assume that higher elevation areas within southern Sweden may have been influenced less by the draw of ice streaming, thoroughly dated sediments in southern Sweden (Småland, part of the south Swedish highland) give OSL ages with a standard deviation range of the mean from 23.7 to 19.1 ka (Alexanderson and Murray 2007, in press), and correspond very well to the Trofors interstadial (Fig. 1).

The marine records include megascale linea-tions that indicate ice-stream activity (fast-moving wet-based ice constrained by slow-moving ice) localized to many channels, troughs and elongated basins crossing the continental shelf off Norway (e.g., Ottesen et al. 2005). Also included are large trough mouth fans comprised of numerous, stacked sediment wedges from debris flows produced at the ice margin during maximum ice extent (e.g., Vor-ren and Laberg 1997). These features, and the chronology based on dates of marine fossils in sediments that subsequently have been overlain by debris flow sediments, in several stacked marine-debris flow sediment successions, indicate a very unstable ice margin during the MIS 3 - MIS 2 in-terval (e.g., King et al. 1996, Nygård et al. 2005, Sejrup et al. 2009).

We appreciate that it may be difficult for some to accept that there was an interstadial during what has been traditionally thought of as the Last Glacial Maximum in Scandinavia (now a two part LGM separated by the Trofors interstadial; Olsen et al. 2001a). Perhaps it is easier to accept by examining the dynamics of the ice sheet over a longer time scale. Large ice retreats occurred after moderate to large ice advances during the Early and Middle Weichselian (Mangerud 2004). These retreats and advances stretched over vast areas of low relief

topography. For example, the Middle Weichselian interstadial was longer lived than the Trofors inter-stadial and consequently is more likely to have been felt in the more central areas of the ice sheet, and sites at Pilgrimstad in central Sweden, Sokli, Ruunaa, and Hitura in Finland, several sites in Skåne, southernmost Sweden, as well as numerous AMS-radiocarbon dated mammoth finds suggest that the ice sheet was either very small or absent during MIS 3 (Figs. 1 and 2, Arnold et al. 2002, Svendsen et al. 2004, Kjær et al. 2006, Ukkonen et al. 2007, Helmens et al. 2007a, 2007b, Alexander-son et al. 2008, in press, Lunkka et al. 2008, Sa-lonen et al 2008, Wohlfarth 2010). In contrast to the Early and Middle Weichselian interstadials, the Trofors interstadial was of smaller geographical extent, shorter-lived, and occurred in the most responsive portion of the ice sheet (Figs. 1 and 2). Nevertheless, the Trofors interstadial represents acknowledgement that the Scandinavian ice sheet was indeed dynamic and that we should be open to the possibility that the existing ice sheets on Earth could act dramatically as well.

Recent robust modelling of the nearby British-Irish ice sheet has produced a very dynamic ice sheet with numerous binge/purge, advance/retreat (i.e., yo-yo) cycles dominated by ice streaming (Hubbard et al. 2009); behaviour similarly docu-mented for Norway. The modelled behaviour dem-onstrates alternating periods of relatively cold-based ice, with a high aspect ratio associated with net growth, and wet-based ice with a lower aspect ratio characterized by streaming. Phases of pre-dominant streaming activity coincide with periods of maximum ice extent and are triggered by abrupt transitions from a cold to relatively warm climate, resulting in major iceberg/melt discharge events (Hubbard et al. 2009). The fjord-dominated land-scape of Norway and its shelf are in no doubt a record of the dominance of ice streams in draining the Scandinavian ice sheet (Longva and Thorsnes 1997, Ottesen et al. 2005). Perhaps the Scandina-vian ice sheet (at least its western margin) behaved in an even more dynamic manner than documented so far, similar to the British-Irish ice sheet. Argua-bly the modelling has exceeded the practicalities of the field science as dynamic ice sheet conditions mean that it is difficult to find sediments from earlier ice-free periods that have survived subse-quent, and multiple, glacial erosion events. Fortu-


20

nately in Scandinavia there are many sub-till sites that have already been discovered and that can be studied further to verify the timing and location of ice-free periods and to improve our understanding of the dynamic nature of ice sheets (e.g., Roberts-son and Ambrosiani 1992, Olsen et al. 2001b, Lundqvist and Robertsson 2002, Lokrantz and Sohlenius 2006).

Conclusion

OSL datings of Langsmoen section sub-till sedi-ments from central Norway, using large (8 mm) and small (2 mm) aliquots, give consistent ages of 22.3 ±1.7 ka (n = 7). These ages are consistent with radiocarbon datings of sub-till sediments from the nearby Flora section, 21.0 ±1.6 ka (n = 8). The similarity in age within and between these strati-graphically-related sites and using different geo-chronological techniques strongly suggests that this area was ice-free around ~21 or 22 ka. These posi-tive results show further promise in employing the OSL dating technique for dating sub-till ice-distal glaciofluvial/fluvial, as well as other sediments in Norway.

We imagine a local paleoenvironment whereby glaciofluvial/fluvial sediments from the Langsmoen site were deposited in ice-distal condi-tions, in a narrow fjord with relative sea level at 210 m a.s.l. or more, perhaps up to ~220 m a.s.l. Further advance of the ice sheet leading into LGM 2 resulted in ice damming down-valley from Flora, ice-margin oscillations and deposition of glacio-lacustrine sediments in ice-proximal conditions, alternating with till deposition at the Flora sections with the highest lake level up to at least 310 m a.s.l. These sites were in-turn buried by LGM 2 till and glaciotectonized.

Sediments from Langsmoen and Flora sections indicate ice-free conditions that are consistent with other similarly-aged sediments in similar strati-graphic settings at many other sites in Norway that collectively define the Trofors interstadial (~25-20 ka; 42 sites and 67 dates); a regional ice-free pe-riod during MIS 2 that is bracketed by LGM 1 and LGM 2 stadials. The existence of the Trofors inter-stadial along with other interstadials during the Middle and Late Weichselian (MIS 3 and MIS 2) indicates that not only the western margin, but the whole western part of the ice sheet, from the ice

divide to the ice margin, and likely including areas in Denmark and southern Sweden, was very dy-namic. These large changes in the ice margin and accompanying drawdown of the ice surface would have affected the eastern part of the ice sheet as well. These conditions have important implications for our understanding of ice sheet dynamics and paleoglaciology, vegetation dynamics, and the relationships between climate change, sea level change, physiography, and Earth rheology. The dynamic nature of the SIS allows us to better ac-cept and understand the rapid-paced fluctuations and deglaciation observed at the margins of con-temporary polar ice sheets, and imagine what is possible in the future.

Acknowledgements

We wish to thank Helena Alexanderson of Stock-holm University/Norwegian University of Life Sciences and Jan Lundqvist of Stockholm Univer-sity for valuable input into the writing; Jan-Pieter Buylaert and the technical staff at the Nordic Labo-ratory for Luminescence Dating for helping with OSL-measurements; Damian Steffen at University of Bern for introducing Johnsen to the deconvolu-tion technique; Irene Lundquist at NGU for techni-cal assistance with maps and several drawings; and, field assistants Marie Koitsalu and Jakob Heyman. This project was supported with funding from the Swedish Society for Anthropology and Geography, Carl Mannerfelts fund, and Ahlmanns fund.

Johnsen was in charge of and performed all as-pects of the project: conception, acquiring funding, OSL-sampling and measurements, data analysis, and writing. Olsen suggested sites for OSL-sampling, guided in the field, provided important input to the writing, and completed the original stratigraphic and sedimentologic fieldwork years earlier. Murray gave valuable contributions to discussions regarding OSL-preparation and meas-urement setup, and interpretation of results as well as provided laboratory access.

References Alexanderson, H., Johnsen, T.F., Murray, A.S., in press. Re-

dating the Pilgrimstad interstadial with OSL: a warmer cli-


21

mate and a smaller ice sheet during the Swedish Middle Weichselian (MIS 3)? Boreas.

Alexanderson, H., Murray, A.S., in press. Problems and poten-tial of OSL dating Weichselian and Holocene sediments in Sweden. Quaternary Science Reviews.

Alexanderson, H., Eskola, K.O, Helmens, K.F., 2008. Optical dating of a Late Quaternary sediment sequence from Sokli, northern Finland. Geochronometria 32, 51-59.

Allen, J. R., 1982. Sedimentary structures: their character and physical basis, volumes I and II. Developments in Sedimen-tology 30A and 30B, 593 and 663 p.

Andersen, B.G., Nydal, R.,Wangen, O.P., Østmo, S.R., 1981. Weichselian before 15,000 years B.P. at Jaeren-Karmøy in southwestern Norway. Boreas 10, 297–314.

Andersen, B.G., Sejrup, H.P., Kirkhus, Ø., 1983. Eemian and Weichselian deposits at Bø on Karmøy, S.W. Norway, a preliminary report. Norges Geologisk Undersøkelse 380, 189–201.

Arnold, N.S., van Andel, T.H., Valen, V., 2002. Extent and Dynamics of the Scandinavian Ice Sheet during Oxygen Iso-tope Stage 3 (65,000-25,000 yr B.P.) Quaternary Research 57, 38-48.

Banerjee, D., Murray, A. S., Bøtter-Jensen, L., Lang, A., 2001. Equivalent dose estimation using a single aliquot of po-lymineral fine grains. Radiation Measurements 33, 73-94.

Bøtter-Jensen, L., Bulur, E., Duller, G.A.T., Murray, A.S., 2000. Advances in luminescence instrument systems. Radia-tion Measurements 32, 523-528.

Choi, J.H., Duller, G.A.T., Wintle, A.G., 2006. Analysis of quartz LM-OSL curves. Ancient TL 24, 9-20.

Denton, G.H., Hughes, T.H., 1981. The Last Great Ice Sheets. Wiley-Interscience, New York, 484 p.

Donner, J., 1995. The Quaternary History of Scandinavia. Cambridge University Press, 206 p.

Duller, G.A.T., 2004. Luminescence dating of Quaternary sediments: recent advances. Journal of Quaternary Science 19, 183-192.

Duller, G.A.T., 2008. Single-grain optical dating of Quater-nary sediments: why aliquot size matters in luminescence dating. Boreas 37, 589-612.

Ehlers J., 1996. Quaternary and Glacial Geology. John Wiley & Sons, Chichester.

Fairbanks, R.G., Mortlock, R.A., Chiu, T.-C., Cao, L., Kaplan, A., Guilderson, T.P., Fairbanks, T.W., Bloom, A.L., 2005. Marine radiocarbon calibration curve spanning 0 to 50,000 years B.P. based on paired 230Th/234U/238U and 14C dates on pristine corals. Quaternary Science Reviews, 24, 1781-1796.

Fuchs, M., Owen, L.A., 2008. Luminescence dating of glacial and associated sediments: review, recommendations and future directions. Boreas 37, 636-659.

Galbraith, R.F., Roberts, R.G., Laslett, G.M., Yoshida, H., Olley, J.M., 1999. Optical dating of single and multiple grains of quartz from Jinmium rock shelter, northern Austra-lia. Part I: experimental design and statistical models. Ar-chaeometry 41, 339-364.

Helmens, K.F., Bos, J.A.A., Engels, S., van Meerbeeck, C.J., Bohncke, S.J.P., Renssen, H., Heiri, O., Brooks, S.J., Seppä, H., Birks, H.J.B., Wohlfarth, B., 2007a. Present-day tem-

peratures in northern Scandinavia during the last glaciation. Geology 35, 987-990.

Helmens, K.F., Johansson, P.W., Räsänen, M.E., Alexander-son, H., Eskola, K.O., 2007b. Ice-free intervals at Sokli con-tinuing into Marine Isotope Stage 3 in the central area of the Scandinavian glaciations. Geological Society of Finland Bulletin 79, 17-39.

Hillefors, Å., 1969. Västsveriges glaciala historia och mor-fologi. Naturgeografiska studier. Meddelanden från Lunds Universitets geografiska institution, Avhandlingar 60. 319 p.

Hillefors, Å., 1974. The stratigraphy and genesis of the Döse-backa and Ellesbo drumlins. A contribution to the knowl-edge of the Weichsel-glacial history in Western Sweden. Geologiska Föreningens i Stockholm Förhandlingar 96, 355-374.

Houmark-Nielsen, M. and Kjær, K.H, 2003. Southwest Scan-dinavia, 40–15 ka BP: palaeogeography and environmental change. Journal of Quaternary Science 18, 769-786.

Hubbard, A., Bradwell, T., Golledge, N., Hall, A., Patton, H., Sugden, D., Cooper, R., Stoker, M., 2009. Dynamic cycles, ice streams and their impact on the extent, chronology and deglaciation of the British–Irish ice sheet. Quaternary Sci-ence Reviews, 28, 758-776.

King, E. L., Sejrup, H. P., Haflidason, H., Elverhøi, A., Aarseth, I., 1996. Quaternary seismic stratigraphy of the North Sea Fan: glacially-fed gravity flow aprons, hemipelagic sediments, and large submarine slides. Marine Geology 130: 293-315.

Kjær, K.H., Lagerlund, E., Adrielsson, L., Thomas, P., J., Murray, A., Sandgren, P., 2006. The first independent chro-nology for Middle and Late Weichselian sediments from southern Sweden and the island of Bornholm. GFF 128, 209-220.

Larsen, E., Gulliksen, S., Lauritzen, S.-E., Lie, R., Løvlie, R., Mangerud, J., 1987. Cave stratigraphy in western Norway; multiple Weichselian glaciations and interstadial vertebrate fauna. Boreas 16, 267-292.

Lian, O.B., Roberts, R.G., 2006. Dating the Quaternary: pro-gress in luminescence dating of sediments. Quaternary Sci-ence Reviews 25, 2449–2468.

Lokrantz, H., Sohlenius, G., 2006. Ice marginal fluctuations during the Weichselian glaciation in Fennoscandia, a litera-ture review. SKB technical report no. TR-06-36. Swedish Nuclear Fuel and Waste Management Co., Stockholm.

Longva, O., Thorsnes, T., 1997. Skagerak in the past and the present - an integrated study of geology, chemistry, hydro-graphy and microfossil ecology. Norges geologiske un-dersøkelse, Special Publication, 8, 1-100.

Lundqvist, J., 1992. Glacial stratigraphy of Sweden. Geologi-cal Survey of Finland, Special Paper 15, 43-59.

Lundqvist, J., 2004. Glacial history of Sweden. In: Ehlers, J., Gibbard, P. L.: Quaternary Glaciations – Extent and Chro-nology. Elsevier B. V., p. 401-412.

Lundqvist, J., and Robertsson, A.-M., 2002. Istider och mel-lanistider. In: Fredén, C. (Ed.) Sveriges Nationalatlas: Berg och jord. 3rd ed. pp. 120-124.

Lunkka, J.P., Murray, A., Korpela, K, 2008. Weichselian sediment succession at Ruunaa, Finland, indicating a Mid-Weichselian ice-free interval in eastern Fennoscandia. Bo-reas 37, 234-244.


22

Mangerud, J., 1991. The Scandinavian Ice Sheet through the last interglacial/glacial cycle. Paläoklimaforschung, 1, 307-330.

Mangerud, J., 2004. Ice-sheet limits on Norway and the Nor-wegian continental shelf. In: Ehlers, J. and Gibbard, P. L. (Eds.), Quaternary glaciations - extent and chronology. El-sevier, Amsterdam, pp. 271-294.

Mangerud, J., Gulliksen, S., Larsen, E., Longva, O., Miller, G., Sejrup, H.P., Sønstergaard, E., 1981. A middle Weich-selian ice-free period in western Norway: The Ålesund In-terstadial. Boreas 10, 447–462.

Mangerud, J., Løvlie, R., Gulliksen, S., Hufthammer, A.-K., Larsen, E., Valen, V., 2003. Paleomagnetic correlations be-tween Scandinavian Ice-Sheet fluctuations and Greenland Dansgaard-Oeschger Events, 45,000-25,000 yr B.P. Quater-nary Research 59, 213-222.

Mangerud, J., Gulliksen, S., Larsen, E., 2010. 14C-dated fluc-tuations of the western flank of the Scandinavian Ice Sheet 45–25 kyr BP compared with Bølling–Younger Dryas fluc-tuations and Dansgaard–Oeschger events in Greenland. Bo-reas, 39: 328-342.

Middleton, G. V., Hampton, M. A., 1976. Subaqueous sedi-ment transport and deposition by sediment gravity flows. In: Stanley, D. J., Swift, D. J. P. (Eds.), Marine Sediment Transport and Environmental Management. John Wiley & Sons, New York, pp. 197-218.

Murray, A.S., Olley, J.M., 2000. Precision and accuracy in the optically stimulated luminescence dating of sedimentary quartz: a status review. Geochronometria 21, 1-16.

Murray, A.S. and Wintle, A.G., 2000. Luminescence dating of quartz using an improved single-aliquot regenerative-dose protocol. Radiation Measurements 32, 57-73.

Murray, A.S., Wintle, A.G., 2003. The single aliquot regenera-tive dose protocol: potential for improvements in reliability. Radiation Measurements 37, 377-381.

Murray, A.S., Marten, R., Johnson, A., Martin, P., 1987. Analysis for naturally occurring radionuclides at environ-mental concentrations by gamma spectrometry. Journal of Radioanalytical and Nuclear Chemistry Articles 115, 263-288.

Nygård, A., Sejrup, H.P., Haflidason, H., Bryn, P., 2005. The glacial North Sea Fan, southern Norwegian Margin: Archi-tecture and evolution from the upper continental slope to the deep-sea basin. Marine and Petroleum Geology, 22, 71-84.

Olsen, L., 1997. Rapid shifts in glacial extension characterise a new conceptual model for glacial variations during the Mid and Late Weichselian in Norway. Norges Geologiske Un-dersøgelse Bulletin 433, 54-55.

Olsen, L., Bergstrøm, B. 2007. Glacier variations during the LGM interval in the Karmøy-Jæren district, SW Norway. Norwegian Geologial Society, Abstracts and Proceedings, No.1, 73-74.

Olsen, L., Grøsfjeld, K., 1999. High Middle and Late Weich-selian relative sea levels in Norway – with implications for the glacial isostasy and the ice-retreat rates. Norges geolo-giske undersøkelse Bulletin 435, 43-51.

Olsen, L., Mejdahl, V., Selvik, S.F., 1996. Middle and Late Pleistocene stratigraphy, chronology and glacial history in Finnmark, North Norway. Norges geologiske undersøkelse Bulletin 429, 1-111.

Olsen, L., Sveian, H. and Bergstrøm, B., 2001a. Rapid adjust-ments of the western part of the Scandinavian ice sheet dur-ing the Mid- and Late Weichselian – a new model. Norges Geologiske Tidsskrift 81, 93-118.

Olsen, L., Sveian, H., Bergstrøm, B., Selvik, S.E, Lauritzen, S.-E., Stokland, O., Gmsfjeld, K., 2001b. Methods and stratigraphies used to reconstruct Mid- and Late Weich-selian palaeoenvironmental and palaeoclimatic changes in Norway. Norges geologiske undersøkelse Bulletin 438, 21-46.

Olsen, L., van der Borg, K., Bergstrøm, B., Sveian, H., Lauritzen, S.-E., Hansen, G., 2001c. AMS radiocarbon dat-ing of glacigenic sediments with low organic carbon con-tent-an important tool for reconstructing the history of gla-cial variations in Norway. Norges Geologiske Tidsskrift 81, 59-92.

Olsen, L., Sveian, H., van der Borg, K., Bergstrøm, B., Broekmans, M., 2002. Rapid and rhythmic ice sheet fluctua-tions in western Scandinavia 15-40 Kya – a review. Polar Research 21, 235-242.

Ottesen, D., Rise, L., Knies, J. Olsen, L., Henriksen, S. 2005. The Vestfjorden-Trænadjupet palaeo-ice stream drainage system, mid-Norwegian continental shelf. Marine Geology 218, 175- 189.

Prescott, J.R., Hutton, J.T., 1994. Cosmic ray contributions to dose rates for luminescence and ESR dating: large depths and long-term time variations. Radiation Measurements 23, 497-500.

Ramsey, C., 1995. Radiocarbon calibration and analysis of stratigraphy: The OxCal program, Radiocarbon 37, 425-430

Ramsey, C., 2001. Development of the radiocarbon calibration program OxCal, Radiocarbon 43, 355-363

Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Bertrand, C.J.H., Blackwell, P.G., Buck, C.E., Burr, G.S., Cutler, K.B., Damon, P.E., Edwards, R.L., Fairbanks, R.G., Friedrich, M., Guilderson, T.P., Hogg, A.G., Hughen, K.A., Kromer, B., McCormac, G., Manning, S., Bronk Ramsey, C., Reimer, R.W., Remmele, S., Southon, J.R., Stuiver, M., Talamo, S., Taylor, F.W., van der Plicht, J., Weyhenmeyer, C.E., 2004. IntCal04 terrestrial radiocarbon age calibration, 0-26 cal kyr BP. Radiocarbon, 46, 1029-1058.

Robertsson, A.-M., García Ambrosiani, K., 1992. The Pleisto-cene in Sweden – a review of research, 1960-1990. Geo-logical Survey of Sweden, SGU Ca 81, 299-306.

Salonen, V.-P., Kaakinen, A., Kultti, S., Miettinen, A., Eskola, K. O., Lunkka, J. P, 2008. Middle Weichselian glacial event in the central part of the Scandinavian Ice Sheet recorded in the Hitura pit, Ostrobothnia, Finland. Boreas 37, 38-54.

Sejrup, H.P., Haflidason, H., Aarseth, I., King, E., Forsberg, C.F., Long, D., Rokoengen, K., 1994. Late Weichselian gla-ciation history of the northern North Sea. Boreas 23, 1-13.

Sejrup, H.P., Larsen, E., Landvik, J., King, E.L., Haflidason, H., Nesje, A., 2000. Quaternary glaciations in southern Fen-noscandia: evidence from southwestern Norway and the northern North Sea region. Quaternary Science Reviews 19, 667-685.

Sejrup, H. P., Nygård, A., Hall, A.M., Haflidason, H., 2009. Middle and Late Weichselian (Devensian) glaciation history of south-western Norway, North Sea and eastern UK. Qua-ternary Science Reviews 28, 370-380.


23

Shepherd, A., Wingham, D., 2007. Recent sea-level contribu-tions of the Antarctic and Greenland Ice Sheets. Science 315, 1529–1532.

Sveian, H., Hansen, L., Solberg, I.-L., Rokoengen, K., 2006. Veibygging avdekker skredhistorien (“Road-construction reveals the clay slide history”). GeoPublishing AS, Trond-heim, GEO 3-2006, 20-25.

Svendsen, J. I., Alexanderson, H., Astakhov, V. I., Demidov, I., Dowdeswell, J. A., Funder, S., Gataullin, V., Henriksen, M., Hjort, C., Houmark-Nielsen, M., Hubberten, H. W., Ingólfsson, Ó., Jakobsson, M., Kjær, K. H., Larsen, E., Lok-rantz, H., Lunkka, J. P., Lyså, A., Mangerud, J., Matiouch-kov, A., Murray, A., Möller, P., Niessen, F., Nikolskaya, O., Polyak, L., Saarnisto, M., Siegert, C., Siegert, M. J., Spiel-hagen, R. F., Stein, R, 2004. Late Quaternary ice sheet his-tory of northern Eurasia. Quaternary Science Reviews 23, 1229-1271.

Ukkonen, P., Arppe, L., Houmark-Nielsen, M., Kjær, K. H., Karhu, J. A, 2007. MIS 3 mammoth remains from Sweden - implications for faunal history, palaeoclimate and glaciation chronology. Quaternary Science Reviews 26, 3081-3098.

Valen, V., Mangerud, J., Larsen, E., Hufthammer, A.K., 1996. Sedimentology and stratigraphy in the cave Hamnsundhel-leren, western Norway. Journal of Quaternary Science, 11, 185-201.

Van der Borg, K., Alderliesten. C., de Jong, A.F.M., van den Brink, A., de Haas. A.P., Kersemaekers, H.J.H., Raaymakers, J.E.M.J., 1997. Precision and Mass Fractiona-tion in 14C analysis with AMS. Nuclear Instruments and Methods B123, 97-101.

Vorren, T.O., Laberg, J.S., 1997. Trough Mouth Fans – Pa-laeoclimate and ice-sheet Monitors. Quaternary Science Re-views 16, 865-882.

Vorren, T. O., Vorren, K.-D., Alm, T., Gulliksen, S., Løvlie, R., 1988. The last deglaciation (20,000 to 11,000 B.P.) on Andøya, northern Norway. Boreas 17, 41-77.

Wallinga, J., 2002. On the detection of OSL age overestima-tion using single-aliquot techniques. Geochronometria 21, 17-26.

Wintle, A.G., 2008. Luminescence dating: where it has been and where it is going. Boreas, 37, 471–482.

Wohlfarth, B. 2010. Ice-free conditions in Sweden during Marine Oxygen Isotope Stage 3? Boreas 39, 377-398.

Wohlfarth, B., Skog, G., Possnert, G., Holmquist, B., 1998. Pitfalls in the AMS radiocarbon-dating of terrestrial macro-fossils. Journal of Quaternary Science 13, 137-145.

Paper IV

1

High-elevation cosmogenic nuclide dating of the last deglaciation in the central Swedish mountains: implications for the timing of tree establishment Timothy F. Johnsen1, Derek Fabel2, Arjen P. Stroeven1 1Department of Physical Geography and Quaternary Geology, Stockholm University, SE-10691 Stockholm, Sweden, [email protected]; [email protected] 2Department of Geographical and Earth Sciences, East Quadrangle, Main Building, University of Glasgow, Glasgow, G12 8QQ, UK, [email protected]

Manuscript © Timothy F. Johnsen, ISBN: 978-91-7447-068-0, ISSN: 1653-7211

Abstract We use cosmogenic exposure ages to determine the timing of deglaciation of the Scandinavian ice sheet

(SIS) at summit elevation in the central Swedish mountains. Mean exposure ages for boulders on the sum-mit of Mt. Åreskutan (10.6 ±0.6 ka, n = 3, 1420 m a.s.l.) and from the highest-elevation moraine related to SIS deglaciation in Sweden (12.0 ±0.6 ka, n = 3, 1135 m a.s.l.) are consistent with previous lower-elevation radiocarbon age estimates for the timing of deglaciation. Summit areas in this region deglaciated ~12.0-10.6 ka, coinciding approximately with the termination of the Younger Dryas cold interval (11.7 ka). Un-usually old radiocarbon ages of tree remains previously studied from the summit-area of Mt. Åreskutan are rejected on the basis of incompatibility with consistent TCN ages for deglaciation, and incompatibility with established paleoecological and paleoglaciological reconstructions. Analysis of the new exposure ages against radiocarbon ages from lower elevation indicates that the SIS decayed rapidly during final deglacia-tion. Keywords: cosmogenic exposure dating, radiocarbon dating, moraine, deglaciation, nunatak

Introduction

Tree remains of three species, Betula pubescens, Picea abies, and Pinus sylvestris, have been found at high-elevation alpine sites in central Sweden that are 400-500 m above the modern tree-line. More-over, these remains are as old as 16.7 ±0.2 cal. ka BP (Kullman 2000, 2001, 2002a), at times when it is commonly perceived that the sites were covered by the Scandinavian ice sheet (SIS; Kleman et al. 1997, Lundqvist 2002). Widespread low-elevation radiocarbon evidence indicates, instead, that degla-ciation in the area occurred around 10.3 to 10.0 cal. ka BP (De Geer 1940, Borell and Offerberg 1955, Lambeck et al. 1998, Lundqvist 1969, 2002). The incongruent dates of the tree remains, therefore, potentially have tremendous implications for our understanding of the dynamics of the SIS, the pat-tern and rate of migration of tree species, the loca-tion and role of refugia in re-establishing plants, paleoclimatic conditions, and nunatak microcli-mate variability. Although the reliability of age constraints is central in obtaining accurate pa-leoglaciological or paleoecological reconstructions

(Birks et al. 2005, 2006, Kullman 2005, 2006) there has been no attempt to directly assess these conflicting radiocarbon data against an independ-ent dating technique. Thus, the objective of our study was to determine independently the timing of deglaciation and to discuss their implications for the potential of tree establishment at high elevation in central Sweden and for the vertical rate of de-glaciation. We evaluate the radiocarbon ages of the wood remains against new results from terrestrial cosmogenic nuclide (TCN) exposure dating of glacial erratics at high elevation and discuss their divergence in terms of paleoecological and pa-leoglaciological implications.

Study area

The study area in the county of Jämtland, central Sweden, includes mountainous areas of moderate relief (~800 m) and adjacent low-relief (~100 m) rolling hill landscape with numerous lakes and peatlands (Fig. 1). Mountains are generally concen-trated in the southern half of the study area except

Timothy F. Johnsen, Derek Fabel, and Arjen P. Stroeven

2

for Mt. Åreskutan which occurs largely isolated in the northeast (Å in Fig. 1). The modern tree-line is at about 950-1050 m above sea level (a.s.l.) in this area.

The study area deglaciated 10.3 to 10.0 cal. ka BP (De Geer 1940, Borell and Offerberg 1955, Sveian 1997, Lambeck et al. 1998, Lundqvist 2002). Above the Late Glacial highest coastline and within the study area, studies on moderate elevation lake sediments estimate deglaciation 10.4 ±0.1 cal. ka BP (Bergman et al. 2005). Thus, there is general agreement for the timing of deglaciation between these studies. However, some of the high elevation tree remains are considerably older (Kullman 2002a, 2004b, Kullman and Kjällgren 2006). While traditionally evidence has been inter-preted in terms of a thick ice sheet in the study area at that time (e.g., Lundqvist 1969, 2002, Denton and Hughes 1981, Kleman et al., 1997), there re-

mains a possibility that the ice sheet was relatively thin during the Late Glacial (Lambeck et al. 1998) leading to early deglaciation of high elevation areas.

Previous studies have shown the SIS to be more dynamic than previously believed during the Mid-dle and Late Weichselian (Arnold et al. 2002, Ol-sen et al. 2002, Kjær et al. 2006, Helmens et al. 2007a,b, Ukkonen et al. 2007, Lunkka et al. 2008; Salonen et al. 2008, Alexanderson et al. 2010) including deglaciation of inland portions of neighbouring Norway during the global last glacial maximum (LGM) (Johnsen et al. submitted). The tree mega-fossils (stems, cones and roots) of Betula pubescens, Picea abies, and Pinus sylvestris dated on Mt. Åreskutan are anomalous in age, oldest at 16.7 ±0.2, 13.0 ±0.1, and 13.6 ±0.1 cal. ka BP, respectively (Kullman 2002a, Table 1), and imply that high elevation areas around Mt Åreskutan deglaciated thousands of years before adjacent valleys. Altogether Kullman (2002a, 2005) dated sixteen well-preserved specimens found on the ground surface in the forefield of a receding 'per-ennial' snow-patch at 1360 m a.s.l. which is only 60 m below the summit and 400-500 m above the modern tree-limits. The surrounding landscape on the summit of Mt Åreskutan is dominated by ex-posed and glacially sculpted bedrock. Numerous ponds exist that are sometimes bordered by shal-low peat or mineral rich soils. Overall the impres-sion is that this mountain-top location is hostile for tree growth even within the modern climate. Within the counties of Jämtland and Dalarna only four other unusually old tree remains have been found (Table 1, Fig. 1, see inset map). Thus, the geographical distribution of unusually old tree remains is especially concentrated to the summit area of Mt. Åreskutan.

Methods

The TCN exposure dating technique can allow determination of the amount of time that a rock surface has been exposed to cosmic radiation (e.g., how long ago an ice sheet deposited an erratic boulder). Cosmic radiation causes the accumula-tion of 10Be in situ within quartz rock and the measurement of the concentration of 10Be against the production rate of 10Be for a given site of known latitude, elevation, topographic shielding,

Fig. 1: Regional map of the study area in the county ofJämtland showing location of radiocarbon dates relevant todeglaciation (cf. numbering system in Table 1); radiocarbondates from the county of Dalarna (Table 1), approximately 80km south of the study area, are not shown. TCN sample sitesfor Snasahögarna SIS moraine area at “S” (see Fig. 3) and forMt. Åreskutan summit erratics at “Å” (arrow). Dashed curvesare early ice-free areas in Sweden (Borgström 1989); note Mt.Åreskutan deglaciated later. Ice movement during the lateglacial maximum was towards the west (Lundqvist 1969).Inset map shows location of study area just east of the Swed-ish-Norwegian border in central Sweden, the counties ofJämtland and Dalarna (dark grey), and the LGM and YoungerDryas ice margins (compiled from various sources by Olsen etal. 2001).

High-elevation cosmogenic nuclide dating of the last deglaciation in the central Swedish mountains

3

No.

inAn

alyz

ed m

ater

ial

Latit

ude

Long

itude

Elev

evat

ion

14C

yr B

PC

alib

rate

d ag

eaSa

mpl

e nu

mbe

rR

efer

ence

Fig.

1no

rthea

st(m

asl

)(1

� er

ror)

(1�

mid

-poi

nt)

Jäm

tland

pro

vinc

e - m

ega-

foss

il w

ood

Mt.

Lills

ylen

13B

etul

a pu

besc

ens,

woo

d63

.027

12.2

0114

9580

30 ±

60

8900

±10

0Be

ta-1

2231

3Ku

llman

200

4aM

t. Li

llsyl

en13

Bet

ula

pube

scen

s, w

ood

63.0

2712

.201

1495

8150

±60

91

00 ±

100

Beta

-172

310

Kullm

an 2

004a

Mt.

Lills

ylen

13B

etul

a pu

besc

ens,

woo

d63

.027

12.2

0114

9584

30 ±

70

9500

±10

0Be

ta-1

7231

2Ku

llman

200

4aM

t. Li

llsyl

en13

Bet

ula

pube

scen

s, w

ood

63.0

2712

.201

1495

8710

±60

96

00 ±

100

Beta

-172

311

Kullm

an 2

004a

Mt.

Åres

kuta

n1

Pic

ea,

cone

63.4

3313

.100

1360

8640

±60

96

00 ±

100

Beta

-160

729

Kullm

an 2

002a

Mt.

Åres

kuta

n1

Bet

ula

pube

scen

s, s

tem

63.4

3313

.100

1360

8850

±70

10

000

±200

Beta

-152

673

Kullm

an 2

002a

Mt.

Åres

kuta

n1

Pin

us,

stem

63.4

3313

.100

1360

9700

±90

11

000

±200

Beta

-121

829

Kullm

an 2

002a

Mt.

Åres

kuta

n1

Bet

ula

pube

scen

s, ro

ot63

.433

13.1

0013

6010

200

±90

11

900

±200

Beta

-121

828

Kullm

an 2

002a

Mt.

Åres

kuta

n1

Pic

ea,

stem

63.4

3313

.100

1360

1025

0 ±9

0

1200

0 ±2

00Be

ta-1

2183

0Ku

llman

200

2aM

t. År

esku

tan

1B

etul

a pu

besc

ens

, ste

m63

.433

13.1

0013

6010

380

±60

12

300

±100

Beta

-152

671

Kullm

an 2

002a

Mt.

Åres

kuta

n1

Pin

us,

stem

63.4

3313

.100

1360

1070

0 ±6

0

1280

0 ±1

00Be

ta-1

3367

3Ku

llman

200

2aM

t. År

esku

tan

1P

icea

, st

em63

.433

13.1

0013

6011

020

±90

13

000

±100

Beta

-121

826

Kullm

an 2

002a

Mt.

Åres

kuta

n1

Pin

us,

stem

c63

.433

13.1

0013

6011

440

±100

1330

0 ±1

00Be

ta-1

2182

7bKu

llman

200

5M

t. År

esku

tan

1P

inus

, st

emc

63.4

3313

.100

1360

1172

0 ±9

0

1360

0 ±1

00Be

ta-1

2079

9bKu

llman

200

2aM

t. År

esku

tan

1B

etul

a pu

besc

ens,

ste

m63

.433

13.1

0013

6012

870

±70

15

200

±100

Beta

-158

301

Kullm

an 2

002a

Mt.

Åres

kuta

n1

Bet

ula

pube

scen

s, s

tem

63.4

3313

.100

1360

1402

0 ±8

0

1670

0 ±2

00Be

ta-1

3367

2Ku

llman

200

2aM

t. G

etry

ggen

6P

inus

, w

ood

63.1

8112

.326

1250

9310

±70

10

500

±100

Beta

-184

490

Kullm

an a

nd K

jällg

ren

2006

Mt.

Lills

nase

n4

Pin

us,

stem

63.2

1712

.317

1180

9840

±90

11

300

±100

Beta

-127

895

Kullm

an a

nd K

jällg

ren

2000

Mt.

Hel

agfjä

llet

17P

inus

, w

ood

62.9

1712

.483

1150

1116

0 ±8

0

1300

0 ±1

00Be

ta-1

2789

4Ku

llman

and

Kjä

llgre

n 20

00M

t. St

or-U

lvåf

jälle

t9

Pin

us,

stem

63.1

5012

.283

1030

8420

±80

94

00 ±

100

Beta

-108

782

Kullm

an a

nd K

jällg

ren

2000

Mt.

Get

rygg

en7

Bet

ula

pube

scen

s, w

ood

63.1

7812

.349

1000

8680

±10

097

00 ±

200

St-1

2980

Kullm

an 1

995

Klop

panå

stug

an24

Pin

us,

woo

d62

.819

12.1

4299

081

60 ±

190

9100

±30

0St

-103

3Lu

ndqv

ist 1

969

Ham

mar

en15

Pin

us,

woo

d62

.961

12.3

2094

085

50 ±

120

9600

±10

0St

-102

8Lu

ndqv

ist 1

969

Ham

mar

en14

Pin

us,

woo

d62

.995

12.3

4892

583

85 ±

105

9400

±10

0St

-109

9Lu

ndqv

ist 1

969

Vålå

stug

an12

Pin

us,

woo

d63

.039

12.7

6988

080

30 ±

90

8900

±20

0St

-106

4Lu

ndqv

ist 1

969

Ham

mar

en16

Pin

us,

woo

d62

.937

12.3

6487

081

50 ±

50

9100

±10

0St

-628

0Lu

ndqv

ist 1

969

Mt.

Enkä

len

11P

inus

, w

ood

63.0

9612

.395

855

9050

±10

010

200

±200

St-1

2729

Kullm

an 1

995

Mt.

Lillu

lvåf

jälle

t10

Pin

us,

woo

d63

.142

12.3

5483

581

90 ±

120

9200

±20

0Be

ta-5

7608

Kullm

an 1

995

O B

unne

rån

5P

inus

, w

ood

63.2

0312

.540

810

8295

±11

093

00 ±

100

St-1

579

Lund

qvis

t 196

9M

t. G

etry

ggen

8A

lnus

, w

ood,

bar

k63

.167

12.3

6774

081

00 ±

50

9100

±10

0Be

ta-1

0634

8Ku

llman

199

8M

t. G

etry

ggen

8P

inus

, w

ood

63.1

6712

.367

740

8220

±70

92

00 ±

100

Beta

-914

97Ku

llman

199

8M

t. G

etry

ggen

8C

oryl

us,

nut

63.1

6712

.367

740

8270

±50

93

00 ±

100

Beta

-108

756

Kullm

an 1

998

Mt.

Get

rygg

en8

Bet

ula,

woo

d63

.167

12.3

6774

082

70 ±

50

9300

±10

0Be

ta-1

0634

7Ku

llman

199

8M

t. G

etry

ggen

8C

oryl

us,

nut

63.1

6712

.367

740

8300

±50

93

00 ±

100

Beta

-108

756

Kullm

an 1

998

Mt.

Get

rygg

en8

Aln

us,

cone

63.1

6712

.367

740

8330

±50

94

00 ±

100

Beta

-108

554

Kullm

an 1

998

Mt.

Get

rygg

en7

Pin

us,

woo

d63

.178

12.3

4972

084

60 ±

120

9400

±10

0Be

ta-5

7645

Kullm

an 1

995

Stor

kluk

en3

Pin

us,

woo

d63

.265

11.9

8470

086

00 ±

30

9500

±30

St

-819

Lund

qvis

t 196

9Kl

ocka

Bog

2P

inus

, w

ood

63.3

0812

.483

526

8390

±10

094

00 ±

100

LuA-

5134

Berg

man

et a

l. 20

04

(by

coun

ty a

nd ty

pe)

Loca

tion

Tabl

e 1:

Sum

mar

y of

radi

ocar

bon

date

s rel

ated

to d

egla

ciat

ion

in th

e co

unty

of J

ämtla

nd p

lus d

ates

of m

ega-

foss

il w

ood

from

the

coun

ty o

f Dal

arna

. Lis

ted

in

orde

r of c

ount

y, sa

mpl

e ty

pe, e

leva

tion,

and

age

. Con

tinue

d on

nex

t pag

e.


4

No.

inAn

alyz

ed m

ater

ial

Latit

ude

Long

itude

Elev

evat

ion

14C

yr B

PC

alib

rate

d ag

eaSa

mpl

e nu

mbe

rR

efer

ence

Fig.

1no

rthea

st(m

asl

)(1

� er

ror)

(1�

mid

-poi

nt)

Jäm

tland

pro

vinc

e - l

ake

sedi

men

tLa

ke S

tent

järn

20S

alix

, bu

d sc

ales

63.1

0012

.242

987

9165

±12

010

400

±100

LuA-

5477

Berg

man

et a

l. 20

05La

ke S

tent

järn

20E

mpe

trum

, see

ds a

nd le

af p

arts

63.1

0012

.242

987

9250

±50

10

400

±100

Poz-

2750

Berg

man

et a

l. 20

05La

ke S

påim

e19

Sal

ix,

fruits

, sca

les,

leav

es a

nd tw

igs

63.1

1712

.317

887

9315

±16

010

500

±200

Ua-

1638

8H

amm

arlu

nd e

t al.

2004

Lake

Ulls

jön

18D

ryas

oct

opet

ala,

see

ds a

nd le

af p

arts

63.4

5213

.010

734

9225

±75

10

400

±100

LuS-

6447

this

stu

dy

Jäm

tland

pro

vinc

e - p

eat

Stor

ulvå

n23

bulk

pea

t63

.167

12.3

5075

088

35 ±

115

9900

±20

0U

a-14

621

Sege

rströ

m a

nd v

on S

tedi

ngk

2003

Stor

ulvå

n23

bulk

pea

t63

.167

12.3

5075

089

30 ±

145

1000

0 ±2

00U

a-14

920

Sege

rströ

m a

nd v

on S

tedi

ngk

2003

Mt.

Get

rygg

en8

Que

rcus

, le

af63

.167

12.3

6774

080

30 ±

50

8900

±10

0Be

ta-1

0875

2Ku

llman

199

8M

t. G

etry

ggen

8A

lnus

, le

af63

.167

12.3

6774

080

60 ±

50

8900

±10

0Be

ta-1

0875

3Ku

llman

199

8M

t. G

etry

ggen

8U

lmus

, le

af63

.167

12.3

6774

085

10 ±

60

9500

±60

Be

ta-1

0875

8Ku

llman

199

8År

ebjö

rnen

21m

oss

and

Eric

acea

e le

aves

63.3

8313

.167

700

8750

±65

98

00 ±

100

Ua-

2175

8Ek

200

4St

orsn

asen

I22

bulk

pea

t63

.233

12.4

1768

085

30 ±

95

9500

±10

0U

a-13

310

Sege

rströ

m a

nd v

on S

tedi

ngk

2003

Stor

snas

en II

22bu

lk p

eat

63.2

3312

.417

680

8855

±10

510

000

±200

Ua-

1331

5Se

gers

tröm

and

von

Ste

ding

k 20

03Kl

ocka

Bog

2M

oss

stem

s an

d le

aves

63.3

0812

.483

526

8055

±95

89

00 ±

200

LuA-

5135

Berg

man

et a

l. 20

04

Dal

arna

pro

vinc

e - m

ega-

foss

il w

ood

Mt.

Stor

vätte

shåg

naP

inus

, w

ood

62.1

1712

.450

1180

8500

±60

95

00 ±

60

Beta

-172

317

Kullm

an 2

004b

Mt.

Stor

vätte

shåg

naP

inus

, w

ood

62.1

1712

.450

1180

9070

±70

10

300

±100

Beta

-172

305

Kullm

an 2

004b

Mt.

Stor

vätte

shåg

naP

inus

, w

ood

62.1

1712

.450

1180

9230

±50

10

400

±100

Beta

-178

795

Kullm

an 2

004b

Mt.

Nip

fjälle

tP

inus

, w

ood

61.9

8312

.850

1160

8050

±70

89

00 ±

100

Beta

-169

410

Kullm

an 2

004b

Mt.

Städ

jan

Pin

us,

woo

d61

.917

12.8

8311

0010

500

±60

12

500

±100

Beta

-158

305

Kullm

an 2

004b

Mt.

Städ

jan

Pin

us,

woo

d61

.917

12.8

8311

0010

500

±60

12

500

±100

Beta

-158

305

Kullm

an 2

004b

Mt.

Stor

vätte

shåg

naP

inus

, w

ood

62.1

1712

.450

1070

8380

±50

94

00 ±

100

Beta

-158

314

Kullm

an 2

004b

Mt.

Städ

jan

Pin

us,

woo

d61

.917

12.8

8310

5578

90 ±

60

8800

±20

0Be

ta-1

7879

7Ku

llman

200

4bM

t. St

orvä

ttesh

ågna

Pin

us,

woo

d62

.117

12.4

5010

3580

50 ±

70

8900

±10

0Be

ta-1

6941

1Ku

llman

200

4bM

t. St

ädja

nP

inus

, w

ood

61.9

1712

.883

1035

8190

±60

91

00 ±

100

Beta

-178

794

Kullm

an 2

004b

Mt.

Stor

vätte

shåg

naP

inus

, w

ood

62.1

1712

.450

940

8040

±60

89

00 ±

100

Beta

-172

316

Kullm

an 2

004b

Mt.

Stor

vätte

shåg

naP

inus

, w

ood

62.1

1712

.450

940

8040

±60

89

00 ±

100

Beta

-172

316

Kullm

an 2

004b

Mt.

Barfr

edhå

gna

Larix

, tw

ig a

nd c

one

62.0

6712

.417

915

8160

±70

91

00 ±

100

Beta

-178

796

Kullm

an 2

004b

Mt.

Stor

vätte

shåg

naB

etul

a, w

ood

62.1

1712

.450

915

8360

±60

94

00 ±

100

Beta

-178

799

Kullm

an 2

004b

Mt.

Stor

vätte

shåg

naQ

uerc

us,

nuts

hell

62.1

1712

.450

910

8560

±40

95

00 ±

40

Beta

-158

310

Kullm

an 2

004b

Mt.

Stor

vätte

shåg

naC

oryl

us,

acor

n62

.117

12.4

5091

086

70 ±

40

9600

±10

0Be

ta-1

5830

9Ku

llman

200

4bM

t. Ba

rfred

hågn

aP

icea

, la

rge

cone

62.0

6712

.417

850

8490

±70

95

00 ±

70

Beta

-108

767

Kullm

an 2

001

b Sam

ples

Bet

a-12

0799

and

Bet

a-12

1827

from

Mt.

Åres

kuta

n ar

e fro

m th

e sa

me

piec

e of

woo

d w

ith th

e fir

st fr

om th

e yo

unge

r par

t of t

he tr

unk

whi

le th

e se

cond

was

from

its

cent

er.

a Cal

ibra

ted

age

is th

e m

idpo

int o

f the

1�

calib

rate

d ag

e ra

nge

with

the

unce

rtain

ty a

s ha

lf th

is ra

nge.

Cal

ibra

ted

usin

g O

xCal

4.0

and

the

IntC

al 0

4 ca

libra

tion

curv

e (R

eim

er e

t al.

2004

).

Loca

tion

(by

coun

ty a

nd ty

pe)

Tabl

e 1:

(con

tinue

d).


5

and sample thickness, provides determination of the exposure age (Lal 1991, Gosse and Phillips 2001). This technique has proven useful in numer-ous studies of deglacial histories and landform preservation (e.g., Phillips et al. 1997, Licciardi et al. 2001, Balco et al. 2002, Fabel et al. 2002, in review, Stroeven et al., 2002, Clark et al. 2003, Rinterknecht et al. 2006). Unlike the radiocarbon dating technique that dates events following degla-ciation, TCN exposure dating can yield the direct age of deglaciation.

The data presented in this paper are part of a large TCN dating campaign for central Sweden. In this paper we present data from erratic boulders from the mountains Åreskutan and Snasahögarna (Å and S in Fig. 1). These data are most relevant to addressing the high elevation radiocarbon ages and for assessing the vertical rate of deglaciation in central Sweden. We collected quartz rich samples from glacially transported boulders from the sum-mit of Mt. Åreskutan (at 1415-1420 m a.s.l.) and from a newly discovered moraine on Mt. Snasa-högarna (1125-1149 m a.s.l.), which, incidentally, is the highest elevation moraine related to the SIS discovered so far in Sweden (Heyman 2004, Hey-man and Hättestrand 2006). The Snasahögarna SIS moraine stretches to an elevation of 1190 m a.s.l. To minimize the risk of having ages biased by processes such as cosmogenic nuclide inheritance (e.g., Briner et al. 2001), boulder exhumation, surface erosion or moraine degradation (Hallet and Putkonen 1994, Zreda et al. 1994, Putkonen and Swanson 2003), we sampled the tops of large (>1.0 m b-axis), weathering-resistant (granitic and quartzitic) boulders, and in sets of three. Boulder surfaces and adjacent ground were carefully in-spected for indications of the amount of differential erosion of erosion. The height of quartz nodules and veins above adjacent softer rock within each boulder were used to estimate the amount of sur-face erosion. In total six boulder samples were processed in the Glasgow University-SUERC cos-mogenic nuclide laboratory.

Measurements and calculations

All samples were processed for 10Be from quartz following procedures based on methods modified from Kohl and Nishiizumi (1992) and Child et al. (2000). Approximately 20 g of pure quartz was separated from each sample, purified, spiked with

c. 0.25 mg 9Be carrier, dissolved, separated by ion chromatography, selectively precipitated as hy-droxides, and oxidized. Accelerated Mass Spec-trometry (AMS) measurements were completed at the SUERC AMS Facility. Measured 10Be/9Be ratios were corrected by full chemistry procedural blanks with 10Be/9Be of <3 x 10-15. Independent measurements of AMS samples were combined as weighted means with the larger of the total statisti-cal error or mean standard error. We calculated the analytical uncertainty by assuming that the uncer-tainties in AMS measurement and Be carrier are normal and independent, adding them in quadra-ture in the usual fashion (e.g., Bevington and Rob-inson 1992). The resulting analytical uncertainties range from 3 to 6% (Table 2). All 10Be concentra-tions were converted to exposure ages by using a production rate linked to a calibration data set us-ing a 10Be half-life of 1.5 Ma.

Using the CRONUS-Earth exposure age calcu-lator version 2.2 (http://hess.ess.washington.edu), measured 10Be concentrations were converted to surface exposure ages, assuming no prior exposure and no erosion since deposition. The results for the different 10Be production rate scaling schemes used by the online calculator yielded ages that vary by about 5% for each sample. The surface exposure ages were calculated using the ‘Lm’ scaling scheme which includes paleomagnetic corrections (Balco et al. 2008; Table 2).

Fig. 2: Land uplift curve of study area adjusted for relative sea level changes. Ten percent confidence interval shown. See text for explanation.


6

10Be age adjustments

The accuracy of the exposure age calculations can be affected by various physical factors: (1) moraine degradation (Putkonen and Swanson 2003), (2) cosmogenic nuclide inheritance (e.g., Briner et al. 2001), (3) the weathering and erosion of the rock surface during exposure, (4) the partial shielding of the rock surface from cosmic rays by topography, seasonal snow cover, and vegetation, or (5) the changing elevation of the rock surface due to gla-cio-isostatic movement (Gosse and Phillips 2001). Apart from inheritance, all these factors cause 10Be ages to appear too young and without adjusting for them, 10Be ages will be minimum ages. On the

other hand, if inheritance is dominating 10Be ages will give maximum ages.

During the exposure history of a site, the air pressure will change due to the combined affect of glacio-isostatic rebound and eustatic change; these factors cause changes in the local air pressure over time that in turn alters the cosmogenic nuclide production rate and apparent exposure age. Thus, we estimated the land uplift corrected for sea level changes and used this information to adjust the apparent TCN exposure age for each site (cf. Stone 2000, Johnsen et al. 2009). The employed land uplift curve was generated for the study area by interpolating between shoreline displacement curves located to the west and east (using Dahl and Nesje 1996, and shoreline displacement curves of

Fig. 3: Map of recessional ice sheet moraines and TCN sample locations in the Snasahögarna area (cf. Fig. 1, location 22). Inferred ice flow direction is also shown. Letters refer to same locations in Fig. 4.


7

Lidén 1938, Mörner 1980, Kjemperud 1981, and Sveian and Olsen 1984) for ages less than 9 cal. ka BP, and referring to regional glacio-isostatic mod-elling results of Lambeck et al. (1998) for ages from 9 to 12 ka (Fig. 2).

Results

Winter fieldwork helped reveal a system of mo-raines that were produced during decay of the SIS within a broad u-shaped valley, named In-golvskalet, that cuts into the west shoulder of Mt. Snasahögarna (Fig. 1, location 22, and Fig. 3). Using aerial photographs coupled with field obser-

vations, and GIS, detailed mapping of these mo-raines was completed (Fig. 3, 4, and 5). The geo-graphical pattern of the moraines, including the location of the highest moraines at only 1 km from a saddle, indicate that they were produced from the decay of the SIS rather than from a retreating local alpine glacier (Fig. 1); as concluded by Lundqvist (1969, 1973) and for similar moraines east of the Mt. Snasahögarna (Borgström 1979). Three boul-ders from the crest of the highest moraine gave consistent ages (10.1 ±0.6 ka, n = 3; Table 2, Fig. 3). The three ages overlap within 1� when consid-ering the measurement errors alone.

Erratic boulders from the summit of Mt. Åre-skutan also produced exposure ages that are consis-tent with each other (9.0 ±0.5 ka, n = 3; Table 2, Fig. 1, location 1, Fig. 6). The three ages overlap within 1� when considering the measurement er-rors alone. Close inspection of the boulder surfaces and adjacent ground surface indicated little signs of erosion. Boulders were either sitting on bare bed-rock or shallow (<20 cm) minerogenic soils likely of weathered till origin. Close examination of the lithologies and angularity of clasts on the ground surface against each sampled boulder revealed that spallation was not an important process. Measure-ment of the height of quartz veins and nodules against softer parent rock divided by the TCN ex-posure age indicates a reasonable weathering rate of about 1 mm ka-1. Such an erosion rate would

Fig. 4: Photograph looking south up into the west side of the Ingolvskalet valley - see Fig. 3. Snow-free moraine crests are indicated by black arrows. White letters correspond to lettered locations in Fig. 3.

Fig. 5: Photograph of moraine extending up the eastern, snow-covered slope of upper-Ingolvskalet valley. This is the highestelevation moraine related to the Scandinavian Ice Sheet yetidentified in Sweden (1190 m a.s.l.). Note person for scale.Inset photograph shows sample TJ-15 and arrow indicates location in photograph.


8

produce less than a 1% age increase and so is ig-nored in our calculations. For interest, a large 3 mm ka-1 erosion rate would increase ages less than only 2.5% (~275-305 years). Similar observations to Mt. Åreskutan were made at and adjacent to each boulder on the Snasahögarna moraine, indi-cating that an erosion rate of 1 mm ka-1 is also reasonable here.

Further adjustments to the apparent TCN expo-sure ages must be made when considering the ef-fect of land uplift and snow shielding during the exposure history of the boulders. By considering the land uplift history of the sites (Fig. 2), the TCN ages increase by 4.7% and 6.4% for Mt. Åreskutan and the Snasahögarna moraine, respectively. The difference in these adjustments is caused by their differences in apparent age and elevation. Recalcu-

Fig. 6: Photographs of Mt. Åreskutan; located at point 1 in Fig. 1. (A) Overview photograph taken from valley at 400 m a.s.l. to northwest indicating the location of the mega-fossil tree remains at 60 m below the summit and 400-500 m above modern tree-line (Kullman 2002a), and the location of the three TCN samples at the summit. (B) TCN sample TJ-1 from boulder partially resting on bedrock east of communications tower. (C) Example of glacially-moulded bedrock next to gondola station that was also found up to the summit; indicating warm-based ice sheet conditions.


9

lating the TCN ages using conservative 10% bounds on the land uplift curve (Fig. 2) results in adjustments that vary by only 0.6%. Thus the accu-racy of the land uplift curve insignificantly affects the calculated age.

Based on precipitation data from the region (Sweden Meteorological and Hydrological Insti-tute), field observation, and local knowledge, we estimate the snow thickness over the tops of boul-ders to be 3 m and persist for four months each year. This snow thickness and of medium density of 0.2 g cm-3 (cf. Lundberg et al. 2006) causes a 12% increase in the TCN ages (Gosse and Phillips 2001). Note that doubling this thickness to 6 m per four months a year less than doubles the increase in the TCN ages to 22%. Thus, when adjusting for the combined affects of land uplift and snow burial, the apparent mean TCN exposure age for Mt. Åre-skutan increases by 17.6% to 10.6 ±0.6 ka, and for the Snasahögarna moraine increases by 19.4% to 12.0 ±0.6 ka.

Discussion

TCN ages

The three consistent TCN exposure ages of erratic boulders from the summit of Mt. Åreskutan indi-

cate that the mountain top deglaciated about 10.6 ±0.6 ka when erosion rates, glacio-isostatic re-bound, and snow shielding effects are taken into consideration (Table 2). Exhumation processes that could cause ages to appear younger than their true age are disregarded because the erratic boulders were resting on bedrock or shallow soils. If nuclide inheritance was an important process then the ages would appear older than their true depositional age; which would be unreasonable to expect as at least the valley bottoms deglaciated around the same time or slightly later (Sveian 1997, Lundqvist 2002). Thus, the TCN exposure ages from the summit of Mt. Åreskutan are considered reliable.

We collected a lake core from Lake Ullsjön (5 km northwest of Mt. Åreskutan, at 734 m a.s.l.; Table 1, Fig 1, location 18). Within the first organ-ics deposited above glacial clay we found and dated Dryas octopetala seeds and leaf parts to 10.4 ±0.1 cal. ka BP. This species is commonly associ-ated with the tundra paleoenvironment following deglaciation in Scandinavia (e.g., Bergman et al. 2005). Thus, this date is considered a reliable esti-mate of the timing of local deglaciation and is in good agreement with other palaeoecological evi-dence in the area (e.g., Bergman et al. 2005) and TCN adjusted exposure ages.

Table 2: Terrestrial cosmogenic nuclide (10Be) exposure data for central Sweden boulders (county of Jämtland) from the summit of Mt. Åreskutan and the Snasahögarna SIS moraine.

Lab ID Elevation Lat Long Sample Shielding Thicknessa [10Be]b Exposure agec Adjusted aged

(m asl) (°N) (°E) lithology factor correction (104 atom/g) (kyr) (kyr)

Mt. Åreskutan summit erraticsTJ-1 1415 63.431 13.095 granite 0.999 0.960 16.30 ±0.74 9.2 ±0.9 (0.4) 10.8 ±1.0TJ-2 1414 63.431 13.094 quartz band 1.000 0.976 16.15 ±1.05 8.9 ±1.0 (0.6) 10.5 ±1.1TJ-3 1420 63.431 13.094 quartzite 1.000 0.960 15.96 ±0.77 8.9 ±0.9 (0.4) 10.5 ±1.0

Mt. Snasahögarna SIS moraineTJ-14 1125 63.225 12.309 quartz band 0.999 0.976 14.41 ±0.50 10.1 ±0.9 (0.3) 12.1 ±1.1TJ-15 1149 63.225 12.314 quartz band 1.000 0.976 14.80 ±0.50 10.2 ±0.9 (0.3) 12.2 ±1.1TJ-20 1130 63.225 12.307 quartz band 0.998 0.976 14.26 ±0.43 10.0 ±0.9 (0.3) 11.9 ±1.1

d Adjusting for the effects of snow cover (~12% for both sites) and crustal rebound (6.4% for Snasahögarna, and 4.7% for Mt. Åreskutan). See text for explanation.

a Calculated using a rock density of 2.65 g/cm3, an effective attenuation length for production by neutron spallation of 160 g/cm2, and erosion rate of 1 mm/kyr.b Measured at SUERC-AMS relative to NIST SRM with a nominal value of 10Be/9Be = 3.06 x 10-11 (Middleton et al. 1993). Uncertainties propagated at ±1� level including all known sources of analytical error.c Exposure ages calculated using the CRONUS-Earth 10Be-26Al exposure age calculator version 2.2 (http://hess.ess.washington.edu) assuming no prior exposure and no erosion during exposure. The quoted values are for the ‘Lm’ scaling scheme which includes palaeomagnetic corrections (Balco et al. 2008). Uncertainties are ±1� (68% confidence) including 10Be measurement uncertainties and a 10Be production rate uncertainty of 9%, to allow comparison with ages obtained with other methods. Values in parentheses are uncertainties based on measurement errors alone, for sample-to-sample comparisons.


10

The TCN adjusted exposure ages from the Sna-sahögarna moraine are older than those from Mt. Åreskutan (adjusted mean age of 12.0 ±0.6 ka, n = 3; Table 2). This age difference probably reflects a geographical pattern of deglaciation whereby the SIS margin was at the Snasahögarna moraine at 1150 m a.s.l., while Mt. Åreskutan, 280 m higher but also 35 km east (up-ice) from Mt. Snasa-högarna, still remained ice-covered. This is consis-tent with detailed ice margin reconstructions for deglaciation of the area (cf. Borgström 1989, maps A and D). In addition, differences in average snow thickness over the Holocene may have been impor-tant between the sites in which case we may have underestimated it at Mt. Åreskutan or overesti-mated it at Snasahögarna. For example, we would have to more than double the snow thickness from our current estimate at Mt. Åreskutan to adjust the ages to be close to the ages at Snasahögarna. Be-cause all boulders from the Snasahögarna moraine were from the crest, this minimized the chance that exhumation may have been an important process that could cause TCN ages to appear too young. Nevertheless, the TCN ages from the Snasahögarna moraine and Mt. Åreskutan are quite similar as they overlap within one sigma of each other (Table 2).

Radiocarbon ages and comparison to TCN ages

Figure 7 and Table 1 show a summary of radiocar-bon dates related to deglaciation in county of Jämtland plus dates of mega-fossil wood from the southern neighbouring county of Dalarna. The majority of the dates fall within the 8.5 to 10.5 cal. ka BP range, while most of the older dates are from Mt. Åreskutan (Fig. 7). Ideally, we might expect the TCN and radiocarbon samples to date the start of ecological succession, whereby the TCN ages represent immediate deglaciation and a barren landscape, and where pioneering species (from lake cores) would perhaps be 200 years younger (Bergman et al. 2005), followed by tree and peat species. Agreeing with this succession, dates from lake cores of pioneering flora representing a tundra paleoenvironment are consistent at 10.4 to 10.5 cal. ka BP (n = 4) which overlap within uncertainty with the TCN ages from Mt. Åreskutan (see Table 2), although central values are slightly younger. Peat dates are in turn slightly younger than the dates from lake sediment. Thus, the pattern of the histogram of radiocarbon dates from the region reflects the stages of ecological succession with the exception of some old dates from tree remains, principally from Mt. Åreskutan.

A summary of ages that relate to deglaciation and from a range of elevations from the study area can provide a general picture of the vertical rate of deglaciation (i.e., a vertical deglaciation curve). Using radiocarbon ages alone would indicate that high elevation areas deglaciated considerably ear-lier than lower elevation areas as indicated by de-glaciation curve 1 in Figure 8. Note that the portion of the curve older than 11 ka is the minimum-age deglaciation curve since it is based only on tree remains. When plotting the TCN ages from high elevation against the radiocarbon ages, there is a clear conflict for those ages from Mt. Åreskutan (Fig. 8). How is it that the three ages for glacial erratics on the mountain summit (adjusted mean age of 10.6 ±0.6 ka) are so much younger than radiocarbon ages nearby on the same mountain? The following three hypotheses attempt to accom-modate both the old radiocarbon ages from high elevation and the TCN ages, however, all three hypotheses are ultimately rejected.

Hypothesis 1: Erosion and/or snow burial led to much younger apparent ages than the true deposi-

Fig. 7: Count versus age (cal. ka BP) of calibrated radiocarbon dates related to deglaciation in the counties of Jämtland andDalarna (See Table 1). The graph shows the contribution ofeach type of date to the total per given 500 year interval. Datetypes are: mega-fossil wood from Dalarna, mega-fossil woodfrom Mt. Åreskutan, all other mega-fossil wood fromJämtland, lake sediment, and peat. Also shown are TCN dat-ings from Mt. Åreskutan and Snasahögarna SIS moraine; errorbars are 1�.


11

tional age of the erratics. If this hypothesis is true, the exposure age of the erratics should be at least as old as the oldest tree megafossils on Mt. Åresku-tan, ~17 cal. ka BP. When considering erosion only, an unrealistic erosion rate of 50 mm ka-1 would be needed make 17 ka erratic boulders on Mt. Åreskutan have an apparent exposure age of 10.6 ka. This extremely high rate of erosion would mean 85 cm of rock loss from the top surface of each sampled boulder. Furthermore, if the boulders experienced a high amount of erosion it is highly unlikely that this amount would be similar for all three boulders to get similar ages. As described above, careful inspection of boulder surfaces and adjacent material indicated that erosion rates were low (Fig. 6). Similarly, as snow has a low density and poorly shields cosmic rays, unrealistic snow thicknesses, in excess of 10 m for a third of a year (Gosse and Phillips 2001), would be required to

cause 17 ka boulders to be perceived as 10.6 ka. Therefore hypothesis 1 is rejected.

Hypothesis 2: There was a small glacier or dead-ice remnant on the mountain summit shield-ing the glacier erratics while trees were growing nearby. Ice flow indicators (striae, and moulded and plucked bedrock) at and surrounding the sum-mit of Mt. Åreskutan indicate only westward pa-leo-ice flow despite variations in topography (ground slope and aspect). The summit area is also covered in glacial erratics. If the summit area was host to an alpine glacier, there would be ice flow indicators that would reflect a thin glacier with flow that was topographically-influenced. In addi-tion, it is unlikely that there would be many erratics under a former alpine glacier due to glacier flow to the margins of the glacier. Thus, abundance of erratics in the summit area and ice flow indicators reveal that glacier flow was largely uninfluenced by the underlying topography (i.e., produced by an

Fig. 8: Vertical deglaciation curves (elevation versus age) for the study area using radiocarbon dates related to deglaciation of mega-fossil wood from the counties of Jämtland (including Mt. Åreskutan) and Dalarna, dates from lake sediment and peat, and TCN adjusted exposure ages. Error bars are 1�. Curve 1 is based on the acceptance of high-elevation radiocarbon ages, principally from Mt. Åreskutan. Curve 2 is based on TCN ages and rejection of old radiocarbon ages from high elevation (~0.5 m year-1). However, as the value for ice sheet thinning of 0.5 m year-1 is derived from data over a large area, it is averaging estimates over space andtime; and local vertical rates of deglaciation can therefore be higher. Curve 3 is the approximate local deglaciation curve for the Mt. Åreskutan area, using central datapoints on TCN adjusted exposure ages and radiocarbon dates (data points 6, 18, 19, and 20; Fig. 1,Table 1), providing a local vertical rate of deglaciation of ~5 m year-1. Also shown are the end of the Younger Dryas cold interval at 11.7 ka BP, and the deglaciation of the east and west edges of the study area at ~10.3 and 10.0 cal. ka BP, respectively. Note that the small numbers indicate the number of identical or near-identical data points.


12

ice sheet) and that there was not an alpine glacier on the summit area of Mt. Åreskutan. Moreover, to grow an alpine glacier on the summit of the moun-tain would require a cooler climate than today which would mean that the tree-line would be even lower than today; the tree megafossils were found 400-500 m above modern tree-line and only 60 m below the summit (Fig. 6). One could invoke a dead-ice remnant on the mountain covering the sites of the boulders, but not the site where trees were living, for a duration of approximately seven thousands years after deglaciation at 17 ka. Be-cause dead-ice is dynamically inert, it wouldn’t have left any traces of ice flow such as striae, moulding or bedrock plucking, or move erratic boulders around. However, it would require a rather cold climate to allow for a slow melt-back over thousands of years. Whereas this may have been the case for during the Younger Dryas, a slow melt-back climate before the Younger Dryas would potentially have been hostile to tree growth. As well, the similarly-aged Snasahögarna moraine supports the TCN ages from Mt. Åreskutan (Fig. 8). Thus, glacier erratics from the summit of Mt. Åreskutan were not shielded by an alpine glacier or dead-ice remnant to give apparent young ages. Therefore, hypothesis 2 is rejected.

Hypothesis 3: Trees were growing at this site and then were overrun but not excavated by the SIS; then erratics were deposited on the mountain summit during the final deglaciation. Glacially moulded, plucked and striated bedrock found from the summit of the mountain to its base indicate that the ice sheet was warm-based in this area at some point during glacial history. Thus the ice sheet that moulded, plucked and striated bedrock would cer-tainly also have excavated soils, peat and wood from its base. The almost uninterrupted range of radiocarbon ages of tree remains from Mt. Åresku-tan from 9.6 to 16.7 cal. ka BP indicates that there would not be an opportunity for the ice sheet to readvance over the mountain summit and deposit erratic boulders (Table 1) at 10.6 ±0.6 ka, even if it did so without erosion of the substrate. Therefore, hypothesis 3 is rejected.

The rejection of these three hypotheses implies that it is unlikely that a sequence of events or proc-esses could lead to both the radiocarbon ages and TCN ages from the summit area of Mt. Åreskutan being compatible with each other. Based on TCN

measurements and other deglaciation evidence cited above, it appears objectively much more reasonable to favour the TCN apparent exposure ages as representing deglaciation of the summit area of Mt. Åreskutan over the old radiocarbon ages because: (1) the TCN ages are consistent with each other and with well-established ages for de-glaciation from middle and lower elevation sites in the study area, (2) the TCN ages for the Snasa-högarna moraine are consistent and overlap within 1� with TCN ages from Mt. Åreskutan (Table 2), and (3) the radiocarbon ages are incompatible with other paleoecological and paleoglaciological evi-dence for Scandinavia (Birks et al. 2005, 2006; see below).

Paleoglaciological and paleoecological considerations

While, as argued above, the strongest evidence to oppose the old radiocarbon dates from Mt. Åresku-tan are incompatible TCN exposure ages, a number of other arguments have been presented in the literature that address paleoecological aspects of high-elevation Late Glacial trees (Kullman 2002a, 2005, 2006, Birks et al. 2005, 2006). In the further analysis, it is assumed that Mt. Åreskutan repre-sented a nunatak environment if trees were to have grown there since 17 ka. This is because many lines of evidence show that the SIS at 17 ka ex-tended beyond the west coast of Norway (cf. Kle-man et al. 1997, Sveian 1997, Lundqvist 2002, Svendsen et al. 2004) and to southern Sweden (Lundqvist and Wohlfarth 2001, Sandgren and Snowball 2001; Fig. 1, see inset map).

A nunatak environment would seem to be in-hospitable to tree growth, as the summer tempera-ture would most likely be below the thermal re-quirement of even the hardiest species, Betula pubescens, which currently requires ~9 oC mean July temperature in the west Norwegian mountains (Odland 1996). It appears even more strenuous to argue in favour of germination or growth condi-tions during the Younger Dryas (~12.8 to 11.7 ka, Muscheler et al. 2008) at 1360 m a.s.l. in the Swed-ish mountains, in light of documented rapid and sustained fall in temperatures during this cold in-terval (Coope et al. 1998, Isarin and Bohncke 1999, Brooks and Birks 2000), yet four of the ra-diocarbon dates from Mt. Åreskutan fall within this interval (Table 1). In addition, soil was skeletal or


13

non-existent on the glacially scraped Mt. Åresku-tan landscape. This would certainly not satisfy the ecological requirements of Picea abies which re-quires acid soils with adequate nutrients (Nikolov and Helmisaari 1992). Although, Kullman (2002a) has reported the discovery of Picea abies saplings at Mt. Åreskutan (1385 m a.s.l.; ~415 m above modern Picea abies tree-line) where the standard-level mean temperature for June to August is ~5°C, which is a lower thermal level for sustained tree growth and reproduction than generally understood (Kullman 2002b). Whatever the actual thermal requirements for tree growth, at sometime after deglaciation Betula pubescens, Picea abies, and Pinus sylvestris grew at 1360 m a.s.l. near the summit area of Mt. Åreskutan.

If Betula, Picea and Pinus occupied nunataks or other near-ice-margin areas, it is surprising that they did not spread out from there after deglacia-tion of lower-elevation terrain (cf. Segerström and von Stedingk 2003, Birks et al. 2005, Hammarlund et al. 2004). As revealed through lake sediments, the first flora in the area at lower elevation at 10.5 cal. ka BP was a pioneer, mountain-type, assem-blage. Picea spread in the area thousands of years later, in fact only 3500 years ago (Lundqvist 1969, Giesecke and Bennett 2004, Bergman et al. 2005), probably from the west rather than from high ele-vation areas in central Sweden (Persson 1975; see Giesecke and Bennett 2004).

Kullman (2002a) proposed that boreal trees sur-vived the glaciation along the southwest coast of Norway and migrated eastward early in the Late Glacial to early deglaciated parts of the central Swedish mountains. However, numerous macro-fossil studies show that during the Late Glacial tree-Betula was absent, or occurred so localized that it has remained undetected, in southwest Nor-way (Birks et al. 2005 and references therein). Similarly, there is no macrofossil or fossil stomata evidence for the presence of Picea and Pinus to indicate that they survived the glaciation in south-west Norway or that they were present in the Late Glacial (Giesecke and Bennett 2004, Birks et al. 2005). Hence, the refuge area for the trees thought to have grown on Mt Åreskutan appears to have been unpopulated with them for the time period of the Late Glacial. We refer the reader to Kullman (2002a, 2005, 2006) and Birks et al. (2005, 2006) for further detailed discussion on paleoecological

aspects of Late Glacial high elevation trees in cen-tral Scandinavia.

Causes of radiocarbon age bias

Based on new TCN apparent exposure age results and paleoecological arguments, the old radiocarbon ages from high elevation sites in central Sweden, and particularly from Mt. Åreskutan, would seem to be unreliable. Re-investigations of the dated tree samples were not possible (Kullman, pers. com. 2006) and we therefore examine four potential explanations as to the causes of the radiocarbon age bias. Firstly, continuous decay processes probably occurred on exposed wood remains as protective perennial snow banks would not always have existed, especially during the Holocene cli-matic optimum. Such decay processes would pre-dictably lead to erroneously young ages (e.g., Baker et al. 1987). However, the sixteen dated specimens were reported to be well-preserved and included bark for some specimens despite sitting on the ground surface for thousands of years (Kullman 2002a). The inner and outer portions of a stem from Mt. Åreskutan (lab numbers Beta-121827 and Beta-120799, Kullman 2002a, 2005; Table 1) gave a similar age (11,440 ±100 and 11,720 ±90 14C ka BP, respectively) indicating that if contamination did not affect the inner portion of the trunk as much as the outer portion, then it may not be an important process for this sample.

Secondly, another possible cause of older radio-carbon ages is lightning (Libby and Lukens 1973, Harkness and Burleigh 1974, Bowman 1990, Paiva 2009). Lightning is more common at high elevation and generates neutrons which may alter the proper-ties of the wood. In this respect, it is noteworthy that the old radiocarbon ages (Fig. 8) occur at high elevation where there is a higher frequency of lightning.

Thirdly, contamination processes may lead to radiocarbon ages being too old. It is more likely that contamination of ancient samples by younger carbon would have a much greater affect than con-tamination of younger specimens by older carbon (Bowman 1990). In this case, the dated wood would be ancient. Ancient trees may therefore derive from stratigraphic units in the up-ice region that date beyond the limit of radiocarbon dating (>50 14C ka BP; Lundqvist 1967) and that contain pollen and wood fragments of the same tree species


14

as those found on Mt. Åreskutan. However, the well-preserved physical condition of the Mt. Åre-skutan samples (Kullman 2002a) potentially pre-cludes this possibility.

Fourthly, contamination processes may lead to radiocarbon ages being too young. Although less effective (Bowman 1990), there also is a possibility for old/dead carbonate in the catchments where tree remains have been found. Lundqvist (1969, figure 17) shows the occurrence of limestone out-cropping 4-6 km east (up-ice) and southeast of Mt. Åreskutan (Lundegårdh et al. 1984). We speculate that the ice sheet could have incorporated lime-stone erosional products in till on Mt Åreskutan and that the old/dead carbonate could be utilised by the plants, or absorbed after their death through groundwater (Bowman 1990), leading to some erroneously older radiocarbon dates from Mt Åre-skutan.

We concur with Birks et al. (2005) that the ra-diocarbon dated high elevation wood samples should be assessed carefully for possible sources of contamination and have their tree-rings analysed against the Scandinavian Holocene tree-ring se-quence. In particular we suggest for these samples that cellulose be chemically extracted from both the inner and outer portions of all stems, and dated.

Implications of new deglaciation ages

With the rejection of radiocarbon dates from high elevation locations in central Sweden and using new high-elevation TCN apparent exposure age data, the timing of deglaciation at high elevation and the vertical rate of deglaciation must be re-examined. These values can only be estimated because of the uncertainties associated with the TCN ages along with the difference in mean age between the Mt. Åreskutan and Snasahögarna mo-raine sites. Nevertheless, high-elevation areas de-glaciated sometime after ~12.0-10.6 ka, coinciding approximately with the termination of the Younger Dryas cold interval (11.7 ka, Muscheler et al. 2008; Fig. 8, curve 2). During the termination of the Younger Dryas, the margin of the SIS was ~85 km and ~60 km west of Mt. Åreskutan and Mt. Snasa-högarna, respectively (Reite 1994, Sveian 1997). The mean vertical rate of deglaciation for the study area was ~0.5 m year-1. However, as this value is derived from data over a large area, it is averaging estimates over space and time; and thus local verti-

cal rates of deglaciation can be higher. The ap-proximate vertical rate of deglaciation for the Mt. Åreskutan area, using central values for TCN ad-justed exposure ages and radiocarbon dates (data points 6, 18, 19, and 20; Fig. 1, and Fig. 8, curve 3), may have been as high as ~5 m year-1. There is no evidence apart from the old radiocarbon dates for a local vertical rate of deglaciation as low as 0.007 m year-1 (Fig. 8, curve 1). This means that high elevation areas within central Sweden re-mained ice covered and could not be colonised by trees as early as 17 ka. Interestingly, TCN exposure age dating work ~195 km south of our study area at Elgåhogna close to the Swedish-Norwegian border indicates that during the LGM the ice sheet surface was above 1460 m a.s.l., and that rapid deglacia-tion commenced about 12 ka (Goehring et al. 2008). This corresponds well with estimates for deglaciation from our study area.

Conclusion

Three terrestrial cosmogenic nuclide (TCN) expo-sure ages from the summit of Mt. Åreskutan (1360 m a.s.l.) in central Sweden are consistent with each other (adjusted mean age of 10.6 ±0.6 ka) and are similar to lower-elevation dates for deglaciation from the region. However, the TCN ages are in-compatible with reported old radiocarbon dates from wood of three tree species from the summit area (as old as 16.7 ±0.2 cal. ka BP). We cannot find a plausible hypothesis that accommodates both the radiocarbon and TCN ages from this site.

Three TCN samples were collected from the highest elevation SIS moraine in Sweden. Located 35 km down-ice from Mt Åreskutan, the Snasa-högarna SIS moraine samples yielded consistent TCN ages (sampled at 1125-1149 m a.s.l.; adjusted mean age of 12.0 ±0.6 ka). The difference in TCN ages between Mt. Åreskutan and the Snasahögarna SIS moraine probably reflects a geographical dif-ference in the timing of deglaciation between sites and possibly a difference in the actual historical snow depths.

We reject the reported old radiocarbon ages on the basis of (1) incompatibility with consistent TCN ages and (2) incompatibility with pa-leoecological and paleoglaciological evidence for deglaciation (Birks et al. 2005). Nevertheless, the mere presence of tree remains of three different


15

species at this high elevation, and well-above (400-500 m) the modern-tree line, still allows for some speculation with regards to their climatic interpre-tation. The problem lies in reliably dating these tree remains. Assuming these are younger tree remains, consistent with TCN age results, we sug-gest that contamination from calcareous bedrock or neutron production from lightning may have caused the age bias. We also strongly recommend that specimens for radiocarbon dating be thor-oughly tested to ascertain possible sources of con-tamination and that complementary dating tech-niques be employed before proposing radical changes to the established paleoglaciological and paleoecological history.

Mt. Åreskutan within central Sweden did not deglaciate as early as ~17 ka. High-elevation areas in this region deglaciated ~12.0-10.6 ka, coinciding approximately with the termination of the Younger Dryas cold interval (11.7 ka). The vertical rate of deglaciation may have been as high as ~5 m year-1 but was more typically 0.5 m year-1. We propose that sometime after deglaciation Betula pubescens, Picea abies, and Pinus sylvestris grew at 1360 m a.s.l. near the summit area of Mt. Åreskutan.

Acknowledgements

We thank Maria Miguens-Rodriguez and Henriette Linge for chemistry and preparation of AMS tar-gets. Jan Lundqvist, Hilary Birks, and Terri La-course provided fruitful discussions, and Daniel Veres, Jakob Heyman, Marie Koitsalu, Helena Alexanderson and Jonas Bergman assisted in the fieldwork. Helena Alexanderson, Jan Lundqvist, and Hilary Birks, helped improve the manuscript. Funding was provided by the Swedish Society for Anthropology, the Gerard De Geer fund, and the Carl Mannerfelt fund. Johnsen was in charge of and performed all aspects of the project: concep-tion, acquiring funding, TCN-sampling, data analy-sis, and writing, except for sample measurements which were completed by Fabel. Fabel and Stro-even provided feedback on the science and writing.

References Alexanderson, H., Johnsen, T. F., Murray, A.S., 2010. Re-

dating the Pilgrimstad Interstadial with OSL: a warmer cli-

mate and a smaller ice sheet during the Swedish Middle Weichselian (MIS 3)? Boreas 39, 367-376.

Arnold, N. S., van Andel, T. H., Valen, V., 2002. Extent and dynamics of the Scandinavian Ice Sheet during oxygen iso-tope stage 3 (65,000–25,000 yr B.P.). Quaternary Research 57, 38–48.

Baker, P. E., Harkness, D. D., Roobol, M. J., 1987. Discussion on differences in radiocarbon ages reported from St. Kitts, West Indies. Journal of the Geological Society 144, 205-206.

Balco, G., Stone, J.O.H., Porter, S.C., Caffee, M.W., 2002. Cosmogenic nuclide ages for New England coastal mo-raines, Martha’s Vineyard and Cape Cod, Massachusetts, USA. Quaternary Science Reviews 21, 2127-2135.

Balco, G., Stone, J.O.H., Lifton, N.A., Dunai, T.J., 2008. A complete and easily accessible means of calculating surface exposure ages or erosion rates from 10Be and 26Al meas-urements. Quaternary Geochronology 3, 174-195.

Bergman, J., Hammarlund, D., Hannon, G., Barnekow, L., Wohlfarth, B., 2005. Deglacial vegetation succession and Holocene tree-limit dynamics in the Scandes Mountains, west-central Sweden: stratigraphic data compared to mega-fossil evidence. Review of Palaeobotany and Palynology 134, 129-151.

Bergman, J., Wastegård, S., Hammarlund, D., Wohlfarth, B., Roberts, S.J., 2004. Holocene tephra horizons at Klocka Bog, west-central Sweden: aspects of reproducibility in su-barctic peat deposits. Journal of Quaternary Science 19, 241-249.

Bevington and Robinson, 1992. Data Reduction and Error Analysis for the Physical Sciences. 2nd edition. New York, McGraw-Hill.

Birks, H. H., Larsen, E., Birks, H. J. B., 2005. Did tree-Betula, Pinus and Picea survive the last glaciation along the west coast of Norway? A review of the evidence, in light of Kullman (2002). Journal of Biogeography 32, 1461-1471.

Birks, H. H., Larsen, E., Birks, H. J. B., 2006. On the presence of late-glacial trees in western Norway and the Scandes: a further comment. Journal of Biogeography 33, 376-378.

Borell, R., Offerberg, J., 1955. Geokronologiska undersökningar inom Indalsälvens dalgång mellan Bergeforsen och Ragunda. SGU, Ca 31.

Borgström, I., 1979. De Geer moraines in a Swedish mountain area?, Geografiska Annaler 61, 35-42.

Borgström, I., 1989. Terrängformerna och den glaciala utvecklingen i södra fjällen. Meddelande från Naturgeografiska institutionen vid Stockholms universitet, Nr A 234, 133 p. (PhD thesis).

Bowman, S., 1990. Radiocarbon dating. University of Califor-nia Press, Berkeley. 64 p.

Briner, J.P., Swanson, T.W., Caffee, M., 2001. Late Pleisto-cene cosmogenic Cl-36 glacial chronology of the south-western Ahklun Mountains, Alaska. Quaternary Research 56, 148-154.

Brooks, S.J., Birks, H.J.B., 2000. Chironomid-inferred late-glacial and early-Holocene mean July air temperatures for Kråkenes Lake, western Norway. Journal of Paleolimnology 23, 77-89.


16

Child, D., Elliott, G., Mifsud, C., Smith, A.M., Fink, D., 2000. Sample processing for earth science studies at ANTARES. Nuclear Instruments and Methods in Physics Research Sec-tion B, Beam Interactions with Materials and Atoms 172, 856-860.

Clark, P.U., Brook, E.J., Raisbeck, G.M., Yiou, F., Clark, J., 2003. Cosmogenic Be-10 ages of the Saglek Moraines, Torngat Mountains. Labrador Geology 31, 617-620.

Coope, G.R., Lemdahl, G., Lowe, J. J., Walkling, A., 1998. Temperature gradients in northern Europe during the last glacial–Holocene transition (14–9 14C kyr BP) interpreted from coleopteran assemblages. Journal of Quaternary Sci-ence 13, 419-433.

Dahl, S. O., Nesje, A., 1996. A new approach to calculating Holocene winter precipitation by combining glacier equilib-rium-line altitudes and pine-tree limits: a case study from Hardangerjøkulan, central Norway. The Holocene 6, 381-398.

De Geer, G., 1940. Geochronologia Suecica Principles. Kungl. Svenska Vetenskapsakademiens Handlingar III : 18 : 6, 367 pp.

Denton, G.H., Hughes, T.J. (eds.)., 1981. The Last Great Ice Sheets. Wiley, New York, NY.

Ek, L.-G., 2004. The establishment of Norway spruce (Picea abies (L.) Karst.) on two mountains in the Åre area - a fol-low-up of the macrofossil finds on Mount Åreskutan. Institutionen för skoglig vegetationsekologi, Sveriges lantbruksuniversitet, Umeå, Sweden. Examensarbeten nr 1 (Masters thesis)

Fabel, D., Stroeven, A. P., Harbor, J., Kleman, J., Elmore, D., Fink, D., 2002. Landscape preservation under Fennoscan-dian ice sheets determined from in situ produced 10Be and 26Al. Earth and Planetary Science Letters 201, 397-406.

Fabel, D., A.P. Stroeven, K.I.T. Dahlgren, J. Harbor, D. Fink, M.W. Caffee, in review. Enhanced ice sheet deglaciation chronologies using cosmogenic nuclide dating of eskers, glaciofluvial deltas, meltwater channels and lineations. Qua-ternary Research.

Giesecke, T., Bennett, K.D., 2004. The Holocene spread of Picea abies (L.) Karst. in Fennoscandia and adjacent areas. Journal of Biogeography 31, 1523-1548.

Goehring, B.M., Brook, E.J., Linge, H., Raisbeck, G.M., Yiou, F., 2008. Beryllium-10 exposure ages of erratic boulders in southern Norway and implications for the history of the Fennoscandian Ice Sheet. Quaternary Science Reviews 27, 320-336.

Gosse, J. C., Phillips, F. M., 2001. Terrestrial in situ cos-mogenic nuclides: theory and application. Quaternary Sci-ence Reviews 20, 1475-1560.

Hallet, B., Putkonen, J., 1994. Surface dating of dynamic landforms: young boulders on aging moraines. Science 265, 937-940.

Hammarlund, D., Velle, G., Wolfe, B. B., Edwards, T. W.D, Barnekow, L. , Bergman, J., Holmgren, S., Lamme, S., Snowball, I., Wohlfarth, B., Possnert, G., 2004. Palaeolim-nological and sedimentary responses to Holocene forest re-treat in the Scandes Mountains, west-central Sweden. The Holocene 14, 862-876.

Harkness, D.D., Burleigh, R., 1974. Possible carbon-14 en-richment in high altitude wood. Archaeometry 16, 121-127.

Helmens, K. F., Bos, J. A. A., Engels, S., Van Meerbeeck, C. J., Bohncke, S.J.P., Renssen, H., Heiri, O., Brooks, S. J., Seppä , H., Birks, H. J. B., Wohlfarth, B., 2007a. Present-day temperatures in northern Scandinavia during the last glaciation. Geology 35, 987-990.

Helmens, K.F., Johansson, P.W., Räsänen, M. E., Alexander-son, H., Eskola, K.O., 2007b. Ice-free intervals at Sokli con-tinuing into Marine Isotope Stage 3 in the central area of the Scandinavian glaciations. Geological Society of Finland, Bulletin 79, 17-39.

Heyman, J., 2004. Pre-recent End Moraines in the Swedish Mountains. Examensarbete i naturgeografi Stockholms universitet, Examensarbete N-54.

Heyman, J., Hättestrand, C., 2006. Morphology, distribution and formation of relict marginal moraines in the Swedish mountains. Geografiska Annaler 88, 253-265.

Isarin, R.F.B., Bohncke, S.J.P., 1999. Mean July Temperatures during the Younger Dryas in Northwestern and Central Europe as Inferred from Climate Indicator Plant Species. Quaternary Research 51, 158-173.

Johnsen, T.F., Alexanderson, H., Fabel, D., Freeman, S.P.H.T., 2009. New 10Be cosmogenic ages from the Vim-merby moraine confirm the timing of Scandinavian Ice Sheet deglaciation in southern Sweden. Geografiska An-naler 91, 113-120.

Johnsen, T.F., Olsen, L., Murray, A.S. submitted. OSL ages in central Norway confirm a MIS 2 interstadial (25-20 ka) and a dynamic Scandinavian ice sheet. Quaternary Science Re-views.

Kjær, K.H., Lagerlund, E., Adrielsson, L., Thomas, P.J., Murray, A., Sandgren, P., 2006. The first independent chro-nology for Middle and Late Weichselian sediments from southern Sweden and the island of Bornholm. Geologiska Föreningens; Stockholm Förhandlingar 128, 209-220.

Kjemperud, A., 1981. A shoreline displacement investigation from Frosta in Trondheimsfjorden, Nord-Trønde1ag, Norway. Norsk Geologisk Tidsskrift 61, 1-15.

Kleman, J., Hättestrand, C., Borgström, I., Stroeven, A., 1997. Fennoscandian paleoglaciology reconstructed using a glacial geological inversion model. Journal of Glaciology 43, 283-299.

Kohl, C.P., Nishiizumi, K., 1992. Chemical isolation of quartz for measurement of in situ-produced cosmogenic nuclides, Geochim. Cosmochim. Acta 56, 3586-3587.

Kullman, L., 1995. Holocene tree-limit and climate history from the Scandes Mountains, Sweden. Ecology 76, 2490-2502.

Kullman, L., 1998. Non-analogous tree flora in the Scandes Mountains, Sweden, during the early Holocene-macrofossil evidence of rapid geographic spread and response to palaeo-climate. Boreas 27, 153-161.

Kullman, L., 2000. The Geoecological histroy of Picea abies in northern Sweden and adjacent parts of Norway. A con-trarian hypothesis of postglacial tree immigration patterns. Geoöko, 21, 141-172.

Kullman, L., 2001. Immigration of Picea abies into North-Central Sweden. New evidence of regional expansion and tree-limit evolution. Nordic journal of Botany 21, 39-54.


17

Kullman, L., 2002a. Boreal tree taxa in the central Scandes during the Late-Glacial: implications for Late-Quaternary forest history. Journal of Biogeography 29, 1117-1124.

Kullman, L., 2002b. Rapid recent range-margin rise of tree and shrub species in the Swedish Scandes. Journal of Ecol-ogy 90, 68-77.

Kullman, L., 2004a. Early Holocene appearance of mountain birch (Betula pubescens ssp. tortuosa) at unprecedented high elevations in the Swedish Scandes: megafossil evidence ex-posed by recent snow and ice recession. Arctic, Antarctic, and Alpine Research 36, 172-180.

Kullman, L., 2004b. Tree-limit landscape evolution at the southern fringe of the Swedish Scandes (Dalarna province) - Holocene and 20th century perspectives. Fennia 182, 73-94.

Kullman, L., 2005. Reply - On the presence of late-glacial trees in the Scandes. Journal of Biogeography 32, 1499-1500.

Kullman, L., 2006. Late-glacial trees from arctic coast to alpine tundra: response to Birks et al. 2005 and 2006. Jour-nal of Biogeography 33, 376-378.

Kullman, L., Kjällgren, L., 2000. A coherent postglacial tree-limit chronology (Pinus sylvestris L.) for the Swedish Scan-des: aspects of paleoclimate and "recent warming," based on megafossil evidence. Arctic, Antarctic, and Alpine Research 32, 419-428.

Kullman, L., Kjällgren, L., 2006. Holocene pine tree-line evolution in the Swedish Scandes: Recent tree-line rise and climate change in a long-term perspective. Boreas 35, 159-168.

Lal, D., 1991. Cosmic-ray labeling of erosion surfaces: in situ nuclide production rates and erosion models. Earth and Planetary Science Letters 104, 424-439.

Lambeck, K., Smither, C., Johnston. P., 1998. Sea-level change, glacial rebound and mantle viscosity for northern Europe. Geophysical Journal International 134, 102-144.

Libby, L.M., H.R. Lukens, 1973, Production of Radiocarbon in Tree Rings by Lightning Bolts, J. Geophys. Res. 78, 5902-5903.

Licciardi, J.M., Clark, P.U., Brook, E.J., Pierce, K.L., Kurz, M.D., Elmore, D., Sharma, P., 2001. Cosmogenic He-3 and Be-10 chronologies of the late Pinedale northern Yellow-stone ice cap, Montana, USA. Geology, 29, 1095-1098.

Lidén, R., 1938. Den senkvartara strandförskjutningens förlopp och kronologi i Angermanland. Geologiska Foreningens i Stockholm Forhandlingar 60, 397-404.

Lundberg, A., Richardson-Näslund, C., Andersson, C., 2006. Snow density variations: consequences for ground-penetrating radar. Hydrological Processes 20, 1483-1495.

Lundegårdh, P.H., Gorbatschev, R., Kornfält, K.-A., Strömberg, A., Karis, L., Zachrisson, E., Sjöstrand, T., Skoglund, R., 1984. Karta över berggrunden i Jämtlands län (utom förutvarande Fjällsjö k n). Bedrock geology of Jämtland County. Sveriges Geologiska Undersökning Ca 53. Map.

Lundqvist, J., 1967. Submoräna sediment i Jämtlands län. Sveriges Geologiska Undersökning C 618. 267 p.

Lundqvist, J., 1969. Beskrivning till jordartskarta över Jämtlands län. Sveriges Geologiska Undersökning (Ser. Ca.) 45, 418 p.

Lundqvist, J., 1973. Isavsmältningens förlopp i Jämtlands län. Sveriges Geologiska Undersökning (Ser. C.) 681, 187 p.

Lundqvist, J., 2002. Istider och mellanistider. In: Fredén, C. (Ed.), Berg och jord, Sveriges Nationalatlas. Stockholm, p. 124-135.

Lundqvist J., Wohlfarth, B., 2001. Timing and east-west correlation of south Swedish ice marginal lines during the Late Weichselian. Quaternary Science Reviews 20, 1127-1148

Lunkka, J.P., Murray, A., Korpela, K., 2008. Weichselian sediment succession at Ruunaa, Finland, indicating a Mid-Weichselian ice-free interval in eastern Fennoscandia. Bo-reas 37, 234-244.

Middleton, R., Brown L., Dezfouly-Arjomandy, B., Klein, J., 1993. On 10Be standards and the half-life of 10Be. Nuclear Instruments and Methods in Physics Research B, 82, 399-403.

Mörner, N.-A., 1980. The Fennoscandian uplift: Geological data and their geodynamical implications. In: Mörner, N.-A. (ed.), Earth rheology, isostasy and eustasy. Chichester: John Wiley, 251-284.

Muscheler, R., Kromer, B., Bjorck, S., Svensson, A., Frie-drich, M., Kaiser, K.F., Southon, J., 2008. Tree rings and ice cores reveal 14C calibration uncertainties during the Younger Dryas. Nature Geoscience 1, 263-267.

Nikolov, N., Helmisaari, H., 1992. Silvics of the circumpolar boreal forest trees. In: A systems analysis of the global bo-real forest. Shugart, H., Leemans, R., Bonan, G.B., Eds. Cambridge University Press, Cambridge. 13-84 p.

Odland, A., 1996. Differences in the vertical distribution pattern of Betula pubescens in Norway and its ecological significance. Paläoklimaforschung, 20, 43-59.

Olsen, L., Sveian, H., Bergstrøm, B., Selvik, S.F., Lauritzen, S.-E., Stokland, Ø, Grøsfjeld, K., 2001. Methods and strati-graphies used to reconstruct Mid- and Late Weichselian pa-laeoenvironmental and palaeoclimatic changes in Norway. Norges geologiske undersøkelse Bulletin 438, 21-46.

Olsen, L., Sveian, H., van der Borg, K., Bergstrøm, B., Broekmans, M., 2002. Rapid and rhythmic ice sheet fluctua-tions in western Scandinavia 15-40 Kya – a review. Polar Research 21, 235-242.

Paiva, G.S., 2009. Terrestrial gamma-ray flashes caused by neutron bursts above thunderclouds. Journal of Applied Physics 105, 083301.

Persson, C., 1975. Speculations on the immigration of spruce into Sweden. Geologiska föreningens förhandlingar.

Phillips, F.M., Zreda, M.G., Evenson, E.B., Hall, R.D., Chadwick, O.A., Sharma, P., 1997. Cosmogenic Cl-36 and Be-10 ages of Quaternary glacial and fluvial deposits of the Wind River Range, Wyoming. Geological Society of Amer-ica Bulletin 109, 1453-1463.

Putkonen, J., Swanson, T., 2003. Accuracy of cosmogenic ages for moraines. Quaternary Research 59, 255-261.

Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Bertrand, C., Blackwell, P.G., Buck, C.E., Burr, G., Cutler, K.B., Damon, P.E., Edwards, R.L., Fairbanks, R.G., Friedrich, M., Guilderson, T.P., Hughen, K.A., Kromer, B., Mc-Cormac, F.G., Manning, S., Bronk Ramsey, C., Reimer, R.W., Remmele, S., Southon, J.R., Stuiver, M., Talamo, S., Taylor, F.W., van der Plicht, J., Weyhenmeyer, C.E., 2004.


18

INTCAL04 terrestrial radiocarbon age calibration, 0–26 cal kyr BP. Radiocarbon 46, 1029-1058.

Reite, A.J., 1994. Weichselian and Holocene geology of Sør-Trøndelag and adjacent parts of Nord-Trøndelag county, Central Norway. Norges geologiske undersøkelse Bulletin 426, 1-30.

Rinterknecht, V.R., Clark, P.U., Raisbeck, G.M. , Yiou, F., Bitinas, A., Brook, E., Marks, J.L., Zelcs, V., Lunkka, J.-P., Pavlovskaya, I.E., Piotrowski, J.A., Raukas, A., 2006. The last deglaciation of southeastern sector of the Scandinavian Ice Sheet. Science 311, 1449-1452.

Salonen, V.-P., Kaakinen, A., Kultti, S., Miettinen, A., Eskola, K.O., Lunkka, J.P., 2008. Middle Weichselian glacial event in the central part of the Scandinavian Ice Sheet recorded in the Hitura pit, Ostrobothnia, Finland. Boreas 37, 38-54.

Sandgren, P., Snowball, I.F., 2001. The Late Weichselian sea level history of the Kullen Peninsula in northwest Skåne, southern Sweden. Boreas 30, 115-130.

Segerström, U., von Stedingk, H., 2003. Early-Holocene spruce, Picea abies (L.) Karst., in west central Sweden as revealed by pollen analysis. The Holocene 13, 897.

Stone, J.O., 2000. Air pressure and cosmogenic isotope pro-duction. Journal of Geophysical Research 105, 23753-23759.

Stroeven, A.P., Fabel, D., Hättestrand, C., Harbor, J., 2002. A relict landscape in the centre of Fennoscandian glaciation: Cosmogenic radionuclide evidence of tors preserved through multiple glacial cycles. Geomorphology 44, 145-154.

Sveian, H., 1997. Ice-marginal deposits and deglaciation chronology in Nord-Trøndelag and Fosen, central Norway. Norges geologiske undersøkelse 433, 52-53.

Sveian, H., Olsen, L., 1984. A shoreline-displacement curve from Veralsøra, Nord-Trøndelag, central Norway. Boreas 8, 241-46.

Svendsen, J.I, Alexanderson, H., Astakhov, V.I., Demidov, I., Dowdeswell, J.A., Funder, S., Gataullin, V., Henriksen, M., Hjort, C., Houmark-Nielsen, M., Hubberten, H.W., Ingólfsson, Ó., Jakobsson, M., Kjær, K.H., Larsen, E., Lok-rantz, H., Lunkka, J.-P., Lyså, A., Mangerud, J., Matiouch-kov, A., Murray, A., Möller, P., Niessen, F., Nikolskaya, O., Polyak, L., Saarnisto, M., Siegert, C., Siegert, M. J., Spiel-hagen, R.F., Stein, R., 2004. Late Quaternary ice sheet his-tory of northern Eurasia. Quaternary Science Reviews 23, 1229-1271.

Ukkonen, P., Arppe, L., Houmark-Nielsen, M., Kjær, K.H., Karhu, J.A., 2007. MIS 3 mammoth remains from Sweden – implications for faunal history, palaeoclimate and glaciation chronology. Quaternary Science Reviews 26, 3081-3098.

Zreda, M., Phillips, F.M., Elmore, D., 1994. Cosmogenic 36Cl accumulation in unstable landforms 2. Simulations and measurements on eroding surfaces. Water Resources Re-search 30, 3127-3136.

afppboq^qflkp fk nr^qbok^ov dblildv NVVRJOMMR

qÜÉ aÉé~êíãÉåí çÑ nì~íÉêå~êó oÉëÉ~êÅÜI píçÅâÜçäã råáîÉêëáíóaçÅíçê çÑ mÜáäçëçéÜó aáëëÉêí~íáçåë NVVRJOMMN

nì~íÉêå~êá~I pÉêK ^

t~ëíÉÖ™êÇI píÉÑ~åI NVVRW i~íÉ tÉáÅÜëÉäá~å J b~êäó eçäçÅÉåÉ ã~êáåÉ ëíê~íáÖê~éÜó áå ëçìíÜïÉëíÉêås®êãä~åÇ ~åÇ åçêíÜïÉëíÉêå a~äëä~åÇI pt pïÉÇÉåK aáëëÉêí~íáçå kçK NK

_êìååÄÉêÖI i~êëI NVVRW `ä~óJî~êîÉ ÅÜêçåçäçÖó ~åÇ ÇÉÖä~Åá~íáçå ÇìêáåÖ íÜÉ vçìåÖÉê aêó~ë ~åÇmêÉÄçêÉ~ä áå íÜÉ É~ëíÉããçëí é~êí çÑ íÜÉ jáÇÇäÉ pïÉÇáëÜ fÅÉ j~êÖáå~ä wçåÉK aáëëÉêí~íáçå kçK OK

pçÜäÉåáìëI dìëí~îI NVVSW qÜÉ Üáëíçêó çÑ íÜÉ _~äíáÅ éêçéÉê ëáåÅÉ íÜÉ i~íÉ tÉáÅÜëÉäá~å ÇÉÖä~Åá~íáçå ~ëêÉÅçêÇÉÇ áå ëÉÇáãÉåíëK aáëëÉêí~íáçå kçK PK

jçâÜí~êá c~êÇI ^ãáêI NVVUW o~éáÇ ÇÉÖä~Åá~íáçå ~åÇ áÅÉ ëÜÉÉí êÉíêÉ~í áå íÜÉ píçÅâÜçäã ~êÉ~I pïÉÇÉå J~ ëÉÇáãÉåíçäçÖáÅ~ä éÉêëéÉÅíáîÉK aáëëÉêí~íáçå kçK QK

tÉëíã~åI mÉêI NVVUW p~äáåáíó ~åÇ íêçéÜáÅ ÅÜ~åÖÉë áå íÜÉ åçêíÜJïÉëíÉêå _~äíáÅ pÉ~ ÇìêáåÖ íÜÉ ä~ëíURMM óÉ~êë ~ë áåÇáÅ~íÉÇ Äó ãáÅêçÑçëëáäë ~åÇ ÅÜÉãáÅ~ä é~ê~ãÉíÉêë áå ëÉÇáãÉåíëK aáëëÉêí~íáçåkç RK

_à�êÅâI gçå~ëI NVVVW bîÉåí ëíê~íáÖê~éÜó Ñçê íÜÉ i~ëí dä~Åá~äJeçäçÅÉåÉ íê~åëáíáçå áå É~ëíÉêå ãáÇÇäÉpïÉÇÉåK aáëëÉêí~íáçå kçK SK

hêáëíá~å pÅÜçåáåÖI OMMNK j~êáåÉ ÅçåÇáíáçåë áå ãáÇÇäÉ pïÉÇÉå ÇìêáåÖ íÜÉ ä~íÉ tÉáÅÜëÉäá~å ~åÇ É~êäóeçäçÅÉåÉ ~ë áåÑÉêêÉÇ Ñêçã Ñçê~ãáåáÑÉê~I lëíê~ÅçÇ~ ~åÇ ëí~ÄäÉ áëçíçéÉëK aáëëÉêí~íáçå kçK UK

i~áãÇçí~ h~äåáå~I OMMNK jáÇÇäÉ ~åÇ i~íÉ mäÉáëíçÅÉåÉ ÉåîáêçåãÉåí~ä ÅÜ~åÖÉë êÉÅçêÇÉÇ áå íÜÉ i~íîá~åé~êí çÑ íÜÉ _~äíáÅ pÉ~ Ä~ëáåK aáëëÉêí~íáçå kçK VK

^åå~ eÉÇÉåëíê�ãI OMMNK b~êäó eçäçÅÉåÉ ëÜçêÉ Çáëéä~ÅÉãÉåí áå É~ëíÉêå pîÉ~ä~åÇI pïÉÇÉåI Ä~ëÉÇ çåÇá~íçã ëíê~íáÖê~éÜóI ê~ÇáçÅ~êÄçå ÅÜêçåçäçÖó ~åÇ ÖÉçÅÜÉãáÅ~ä é~ê~ãÉíÉêëK aáëëÉêí~íáçå kçK NMK

qááí e~åÖI OMMNK mêçÖä~Åá~ä ëÉÇáãÉåí~êó ÉåîáêçåãÉåíI î~êîÉ ÅÜêçåçäçÖó ~åÇ ä~íÉ tÉáÅÜëÉäá~åÇÉîÉäçéãÉåí çÑ íÜÉ i~âÉ mÉáéëáI É~ëíÉêå bëíçåá~K aáëëÉêí~íáçå kçK NNK

qÜÉ aÉé~êíãÉåí çÑ mÜóëáÅ~ä dÉçÖê~éÜó ~åÇ nì~íÉêå~êó dÉçäçÖóI píçÅâÜçäã råáîÉêëáíóaçÅíçê çÑ mÜáäçëçéÜó aáëëÉêí~íáçåë OMMOJOMMR

qÜÉëáë áå nì~íÉêå~êó dÉçäçÖó

dêÉÖÉê iáåÇÉÄÉêÖI OMMOK qÜÉ pïÉÇáëÜ î~êîÉÇ Åä~óë êÉîáëáíÉÇW péÉÅíê~äJ ~åÇ áã~ÖÉ ~å~äóëáë çÑ ÇáÑÑÉêÉåííóéÉë çÑ î~êîÉ ëÉêáÉë Ñêçã íÜÉ _~äíáÅ _~ëáåK aáëëÉêí~íáçå kçK NK c~âìäíÉíëçééçåÉåíW mêçÑK _à�êåj~äãÖêÉå K

o~íÜå~ëáêá mêÉã~íÜáä~âÉI OMMPW i~íÉ nì~íÉêå~êó é~ä~ÉçÉÅçäçÖáÅ~ä ÉîÉåí ëíê~íáÖê~éÜó áå íÜÉ eçêíçåmä~áåëI ÅÉåíê~ä pêá i~åâ~ J ïáíÜ ÅçåíêáÄìíáçåë íç íÜÉ êÉÅÉåí éçääÉå Ñäçê~K aáëëÉêí~íáçå kçK OKc~âìäíÉíëçééçåÉåíW mêçÑK cê~åÅçáëÉ d~ëëÉK

^åÖÉäáÅ~ cÉìêÇÉ~åI OMMQW m~ä~ÉçÉåîáêçåãÉåí áå åçêíÜJïÉëíÉêå oçã~åá~ ÇìêáåÖ íÜÉ ä~ëí NRIMMM óÉ~êëKaáëëÉêí~íáçå kçK PK c~âìäíÉíëçééçåÉåíW mêçÑK h~íÜÉêáåÉ gK táääáëK

^åÇÉêë _çêÖã~âI OMMRW qÜÉ Åçäçìê çÑ Åäáã~íÉW ÅÜ~åÖÉë áå éÉ~í ÇÉÅçãéçëáíáçå ~ë ~ éêçñó Ñçê Åäáã~íÉÅÜ~åÖÉK aáëëÉêí~íáçå kçK QK c~âìäíÉíëçééçåÉåíW aêK _~ë î~å dÉÉäK

gÉåë eÉáãÇ~ÜäI OMMRW rêÄ~åáëÉÇ å~íìêÉ áå íÜÉ é~ëí Ó ëáíÉ Ñçêã~íáçå ~åÇ ÉåîáêçåãÉåí~ä ÇÉîÉäçéãÉåíáå íïç pïÉÇáëÜ íçïåëI ^a NOMMJNUMMK aáëëÉêí~íáçå kçK RK c~âìäíÉíëçééçåÉåíW aêK g~åÉ páÇ~ääK

afppboq^qflkp colj

qeb abmôqjbkq lc mevpf`î dbldo^mev ^ka nr^qbokôv dblildvOMMSÓ

e™â~å dêìÇÇI OMMSK qêÉÉ êáåÖë ~ë ëÉåëáíáîÉ éêçñáÉë çÑ é~ëí Åäáã~íÉ ÅÜ~åÖÉK aáëëÉêí~íáçå kçK NKc~âìäíÉíëçééçåÉåíW mêçÑK _êá~å iìÅâã~åK

räÑ gçåëÉääI OMMSK pìäÑìê áå éçä~ê áÅÉ ~åÇ ëåçïK fåíÉêéêÉí~íáçåë çÑ é~ëí ~íãçëéÜÉêÉ ~åÇ Åäáã~íÉíÜêçìÖÜ Öä~Åá~ä ~êÅÜáîÉëK aáëëÉêí~íáçå kçK OK c~âìäíÉíëçééçåÉåíW aêK j~êâ `ìêê~åK

e~åå~ pK pìåÇèîáëíI OMMTK péÉäÉçíÜÉãë ~ë ÉåîáêçåãÉåí~ä êÉÅçêÇÉêë J ^ ëíìÇó çÑ eçäçÅÉåÉëéÉäÉçíÜÉãë ~åÇ íÜÉáê ÖêçïíÜ ÉåîáêçåãÉåíë áå pïÉÇÉåK aáëëÉêí~íáçå kçK PK c~âìäíÉíëçééçåÉåíWmêçÑK cê~åâ jÅaÉêãçííK

m~íêáâ häáåíÉåÄÉêÖI OMMTK jçêÉ ï~íÉêI äÉëë Öê~ëë\ ^å ~ëëÉëëãÉåí çÑ êÉëçìêÅÉ ÇÉÖê~Ç~íáçå ~åÇëí~âÉÜçäÇÉêëÛ éÉêÅÉéíáçåë çÑ ÉåîáêçåãÉåí~ä ÅÜ~åÖÉ áå lãÄìÖ~ Öê~ëëä~åÇI åçêíÜÉêå k~ãáÄá~KaáëëÉêí~íáçå kçK QK c~âìäíÉíëçééçåÉåíW mêçÑK píÉáå _áÉK

j~êá~ oóåÉêI OMMTK m~ëí ÉåîáêçåãÉåí~ä ~åÇ Åäáã~íÉ ÅÜ~åÖÉë áå åçêíÜÉêå q~åò~åá~K sÉÖÉí~íáçå ~åÇä~âÉ äÉîÉä î~êá~Äáäáíó áå bãé~â~~á `ê~íÉêK aáëëÉêí~íáçå kçK RK c~âìäíÉíëçééçåÉåíW mêçÑK eÉåêói~ãÄK

a~åáÉä pK sÉêÉëI OMMTK qÉêêÉëíêá~ä êÉëéçåëÉ íç a~åëÖ~~êÇJlÉëÅÜÖÉê ÅóÅäÉë ~åÇ eÉáåêáÅÜ ÉîÉåíëW íÜÉä~ÅìëíêáåÉ êÉÅçêÇ çÑ iÉë bÅÜÉíëI ëçìíÜJÉ~ëíÉêå cê~åÅÉK aáëëÉêí~íáçå kçK SK c~âìäíÉíëçééçåÉåíWmêçÑK gçÜå gK içïÉK

vçëÜáÜáêç pÜáÄìçI OMMTK jçÇÉääáåÖ ï~íÉê ~åÇ ëçäìíÉ Ñäçïë ~í ä~åÇJëÉ~ ~åÇ ä~åÇJ~íãçëéÜÉêÉáåíÉêÑ~ÅÉë ìåÇÉê Ç~í~ äáãáí~íáçåëK aáëëÉêí~íáçå kçK TK c~âìäíÉíëçééçåÉåíW aêK `äáÑÑçêÇ sçëëK

dÉëëÉëëÉ aÉëëáÉI OMMTK cçêÉëí aÉÅäáåÉ áå pçìíÜ `Éåíê~ä bíÜáçéá~W bñíÉåíI Üáëíçêó ~åÇ éêçÅÉëëKaáëëÉêí~íáçå kçK UK c~âìäíÉíëçééçåÉåíW mêçÑK j~íë táÇÖêÉåK

eÉêå•å aÉ ^åÖÉäáëI OMMTK m~ä~ÉçJáÅÉ ëíêÉ~ãë áå íÜÉ åçêíÜJÉ~ëíÉêå i~ìêÉåíáÇÉ fÅÉ pÜÉÉíK aáëëÉêí~íáçåkçK VK c~âìäíÉíëçééçåÉåíW aê `çäã þ `çÑ~áÖÜK

^ã¨äáÉ a~êê~ÅèI OMMTK içåÖJíÉêã ÇÉîÉäçéãÉåíI ãçÇÉäáåÖ ~åÇ ã~å~ÖÉãÉåí çÑ åìíêáÉåí äç~ÇáåÖ íçáåä~åÇ ~åÇ Åç~ëí~ä ï~íÉêëK aáëëÉêí~íáçå kçK NMK c~âìäíÉíëçééçåÉåíW mêçÑK ^åÇêÉ~ oáå~äÇçK

bäáå kçêëíê�ãI OMMUK i~íÉ nì~íÉêå~êó Åäáã~íÉ ~åÇ ÉåîáêçåãÉåí~ä ÅÜ~åÖÉ áå íÜÉ ëìããÉê ê~áåÑ~ääêÉÖáçå çÑ pçìíÜ ^ÑêáÅ~ J ^ ëíìÇó ìëáåÖ íêÉÉë ~åÇ ïÉíä~åÇ éÉ~í ÅçêÉë ~ë å~íìê~ä ~êÅÜáîÉëKaáëëÉêí~íáçå kçK NNK c~âìäíÉíëçééçåÉåíW mêçÑK jáÅÜ~Éä jÉ~ÇçïëK

cêÉÇêáâ e~ååÉêòI OMMUK j~âáåÖ ï~íÉê áåÑçêã~íáçå êÉäÉî~åí çå äçÅ~ä íç ÖäçÄ~ä ëÅ~äÉ Ó íÜÉ êçäÉ çÑfåÑçêã~íáçå póëíÉãë Ñçê fåíÉÖê~íÉÇ t~íÉê j~å~ÖÉãÉåíK aáëëÉêí~íáçå kçK NOKc~âìäíÉíëçééçåÉåíW mêçÑK aÉååáë iÉííÉåã~áÉêK

j~ííá~ë ÇÉ tçìäI OMMUK oÉëéçåëÉ çÑ Öä~ÅáÉêë íç Åäáã~íÉ ÅÜ~åÖÉK aáëëÉêí~íáçå kçK NPKc~âìäíÉíëçééçåÉåíW mêçÑK oçÖÉê _ê~áíÜï~áíÉK

_ê~ÇäÉó tK dççÇÑÉääçïI OMMUK oÉäáÅí åçåJÖä~Åá~ä ëìêÑ~ÅÉë ~åÇ ~ìíçÅÜíÜçåçìë ÄäçÅâÑáÉäÇë áå íÜÉåçêíÜÉêå pïÉÇáëÜ ãçìåí~áåëK aáëëÉêí~íáçå kçK NQK c~âìäíÉíëçééçåÉåíW aêK ^Çêá~å e~ääK

j~êíáå~ e®ííÉëíê~åÇI OMMUK sÉÖÉí~íáçå ~åÇ Åäáã~íÉ ÇìêáåÖ tÉáÅÜëÉäá~å áÅÉ ÑêÉÉ áåíÉêî~äë áå åçêíÜÉêåpïÉÇÉå Ó áåíÉêéêÉí~íáçåë Ñêçã Ñçëëáä ~åÇ ãçÇÉêå éçääÉå êÉÅçêÇëK aáëëÉêí~íáçå kçK NRKc~âìäíÉíëçééçåÉåíW mêçÑK açå~íÉää~ j~ÖêáK

iáåÇ~ ^ãéÉäI OMMUK a~åëÖ~~êÇJlÉëÅÜÖÉê ÅóÅäÉë ~åÇ eÉáåêáÅÜ ÉîÉåíë áå ïÉëíÉêå bìêçéÉ Ó ^ Çá~íçãéÉêëéÉÅíáîÉK aáëëÉêí~íáçå kçK NSK c~âìäíÉíëçééçåÉåíW mêçÑK pÜÉêáäóå cêáíòK

dìää lääáI OMMUK t~íÉêÄçêåÉ ëÉÇáãÉåí ~åÇ éçääìí~åí íê~åëéçêí áåíç ä~âÉë ~åÇ ~ÅÅìãìä~íáçå áå ä~âÉëÉÇáãÉåíëK aáëëíÉêí~íáçå kçKNTK c~âìäíÉíëçééçåÉåíW mêçÑK fåÖã~ê oÉåÄÉêÖK

`Üêáëíáå~ bK gçåëëçåI OMMVK eçäçÅÉåÉ Åäáã~íÉ ~åÇ ~íãçëéÜÉêáÅ ÅáêÅìä~íáçå ÅÜ~åÖÉë áå åçêíÜÉêåcÉååçëÅ~åÇá~ J áåíÉêéêÉí~íáçåë Ñêçã ä~ÅìëíêáåÉ çñóÖÉå áëçíçéÉ êÉÅçêÇëK aáëëÉêí~íáçå kçK NUKc~âìäíÉíëçééçåÉåíW aêK mÜáäáé _~êâÉêK

h~êáå bÄÉêíI OMMVK `ÉåçòçáÅ ä~åÇëÅ~éÉ Éîçäìíáçå áå åçêíÜÉêå pïÉÇÉåK dÉçãçêéÜçäçÖáÅ~äáåíÉêéêÉí~íáçå ïáíÜáå ~ dfpJÑê~ãÉïçêâK aáëëÉêí~íáçå kçK NVK c~âìäíÉíëçééçåÉåíW mêçÑK m~ìä_áëÜçéK

pçÑá~ ^åÇÉêëëçåI OMNMK i~íÉ eçäçÅÉåÉ ÜìãáÇáíó î~êá~Äáäáíó áå ÅÉåíê~ä pïÉÇÉåK aáëëÉêí~íáçå kçK OMKc~âìäíÉíëçééçåÉåíW mêçÑK cê~åâ `Ü~ãÄÉêëK

g~âçÄ eÉóã~åI OMNMK m~ä~ÉçÖä~ÅáçäçÖó çÑ íÜÉ åçêíÜÉ~ëíÉêå qáÄÉí~å mä~íÉ~ìK aáëëÉêí~íáçå kçK ONKc~âìäíÉíëçééçåÉåíW mêçÑK cê~åâ iÉÜãâìÜäK

qáãçíÜó gçÜåëÉåI OMNMK i~íÉ nì~íÉêå~êó áÅÉ ëÜÉÉí Üáëíçêó ~åÇ Çóå~ãáÅë áå ÅÉåíê~ä ~åÇ ëçìíÜÉêåpÅ~åÇáå~îá~K aáëëÉêí~íáçå kçK OOK c~âìäíÉíëçééçåÉåíW mêçÑK g~ãÉë qK qÉääÉêK

Part of the key to predicting the future behaviour of the Earth is linked to our understanding of how ice sheets have operated in the past. Recent work suggests an emerging new paradigm for the Scan-dinavian ice sheet (SIS); one of a dynamically fluctuating ice sheet. This doctoral research project explicitly examines the history and dynamics of the SIS at four sites within Sweden and Norway, and provides results covering different time periods of glacial history. Two relatively new dating techniques are used to constrain the ice sheet history.

Dating of sub-till sediments in central Sweden and central Norway indicate ice-free conditions during times when it was previously inferred the sites were occupied by the SIS. Consistent exposure ages of boulders from the Vimmerby moraine in southern Sweden indicate that the southern margin of the SIS was at the Vimmerby moraine ~14 kyr ago. In central Sweden, consistent exposure ages for boulders at high elevation agree with previous estimates for the timing of deglaciation around 10 ka ago, and indicate rapid thinning of the SIS during deglaciation.

Altogether this research conducted in different areas, covering dif-ferent time periods, and using comparative geochronological met-hods demonstrates that the SIS was highly dynamic and sensitive to environmental change.

Department of Physical Geography and Quaternary Geology

ISBN 978-91-7447-068-0ISSN 1653-7211

I was born and raised on Vancouver Island on the west coast of Canada surrounded by beautiful mountains and coastline, where I developed a deep curiosity and passion for understanding the workings of nature. I completed a Bachelor of Science degree with distinction in Geography 1998 at the University of Victoria, Canada. Then I completed a Masters of Science degree in Geography 2004 at Simon Fraser University, Canada, for which I was awarded the Canadian Association of Geographers Starkey-Robinson Award 2005. I began a PhD in 2004 in Stockholm, Sweden investigating the dynamics of the Scandinavian ice sheet, and eating brown cheese with waffles under the midnight sun.

late quaternary ice sheet history and dynamics in central ...311283/fulltext01.pdf · late...

Documents