Late Quaternary Ice Sheet History and Dynamics in Central and Southern Scandinavia Timothy F

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 Thesis in Quaternary Geology at Stockholm University, Sweden 2010

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

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 history. 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 technique. 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 comparative 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

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 deglaciation in the central Swedish mountains: implications for the timing of tree establishment. Manuscript.

5 Timothy F. Johnsen

Introduction millions of humans in coastal regions (IPCC 2007). These processes along with others related to global Ice sheets are a crucial component of the function- climate change remind us of how we as a species ing of the Earth system (Oerlemans and van der are linked to the activities of ice sheets that in turn Veen 1984). Their large size displaces vast areas of at least partially reflect our own activity, and that plants and animals (Robertsson 1994, Hewitt 2000) our fate is tied to how the Earth and its ice sheets and changes the albedo and climate of the Earth and climate will behave. Despite intensive scien- (Manabe and Broccoli 1985, Ruddiman 2003). tific efforts it remains difficult to accurately predict Tremendous quantities of water from the oceans the magnitude and rate of changes for the future, are stored on land as ice, causing global sea levels and this uncertainty is directly tied to our under- to lower over 100 metres, and their massive weight standing of the how the Earth system is operating depresses the surface of the Earth hundreds of and has operated. Answering important questions metres which leads to flooding of coastal areas and about the future climate and conditions on Earth the diversion of rivers (Lambeck and Chappell are inextricably linked to our understanding of how 2001). Immense quantities of meltwater can be the Earth has operated in the past. Thus, by under- discharged, disrupting ocean circulation and the standing the dynamics of past ice sheets, as in this climate system and resulting in sudden sea level doctoral research project, we will better predict rise (Fairbanks 1989). And, huge areas of the land- how modern ice sheets will respond to climate scape are altered and shaped by the erosional and change and affect society. depositional activity of ice sheets (Lundqvist During the Quaternary Period multiple glaci- 2002). In addition, our own past is closely linked to ations of varying spatial extent occurred in Scandi- that of ice sheets and ice dynamics. Modern hu- navia starting in the Early Quaternary (Mangerud mans evolved during the Late Quaternary, a period et al. 1996, Kleman et al. 2008) and with the first characterized by glaciations and rapid climatic and glaciation reaching the shelf-edge occurring ~1.1 environmental shifts resulting in great changes in Ma (Sejrup et al. 2000). As ice sheets are effective the distribution of organisms. The present genetic agents of erosion the best evidence for glaciations structure of populations, species and communities is from the most recent glaciation, the Weichselian, has been mainly formed by Quaternary ice ages spanning from ~117 to 11.7 ka, and the reconstruc- (Hewitt 2000). tion of glacial history prior to the Last Glacial In recent time dramatic changes in the margins Maximum (LGM) is in many cases difficult and of the Greenland and Antarctic ice sheets have ambiguous (Fig. 1 and 2). Numerous deposits and occurred including rapid but episodic glacier accel- landforms related to the last deglaciation dominate eration and thinning from their marine-terminating the landscape (Fredén 2002) while reconstruction sectors (Shepherd and Wingham 2007): e.g., the of earlier ice sheet activity mostly relies on discov- collapse of sections of the Larsen Ice Shelf in Ant- ery and study of terrestrial sediments or landforms arctica (Rott et al. 1996), and loss of about 100 Gt that have managed to survive being overrun by the yr-1 of mass from the the Greenland ice sheet Scandinavian ice sheet (SIS; e.g., Lagerbäck 1988, (Shepherd and Wingham 2007). The rate of mod- Robertsson and García Ambrosiani 1992, Kleman ern changes of ice sheets is occurring faster than and Stroeven 1997, Olsen et al. 2001b, Hättestrand many scientists anticipated, and have made it easier and Stroeven 2002, Lundqvist and Robertsson to imagine dynamic glacier activity for the past. As 2002, Heyman and Hättestrand 2006, Lokrantz and well, a major change in our understanding of the Sohlenius 2006), marine sediments (e.g., Sejrup et dynamics of ice sheets occurred when present and al. 1994, Baumann et al. 1995, Vorren and Laberg past ice streams were recognized for their impor- 1997), and ice sheet modelling (e.g., Holmlund and tance in the mass balance of ice sheets and dy- Fastook 1995, Kleman et al. 1997, Lambeck et al. namic behaviour (e.g., Bentley 1987, Stokes and 1998, Boulton et al. 2001, Siegert et al. 2001, Clark 2001, Alley et al. 2004). Decay of modern Charbit et al. 2002, Näslund et al. 2003). With the ice sheets along with global climate warming is set exception of the last deglaciation, the timing of ice to cause global sea levels to rise considerably dur- sheet advances and retreats is based mainly on ing this century and potentially displace tens of correlation to ‘global’ continuous records such as

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

western margin of the ice sheet was highly dynamic 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

Fig. 1: Chronostratigraphy of northwestern Europe (after (Lambeck et al. 1998) indicate that the ice sheet Mangerud 1991), also showing the oxygen isotope stages, was much thinner than earlier estimates (Denton stadials and interglacials. and Hughes 1981), while cosmogenic nuclide dating results from central Norway indicate that the ice sheet was not thin enough to expose large areas ice cores or marine sediment cores. This is partly of high elevation alpine land during the LGM in- due to the difficulty in dating glacial sediments, terval (Goehring et al. 2008) as proposed by earlier partly to the relative scarcity of datable deposits workers (Nesje and Dahl 1990). Archaeological older than the LGM, and the age limit (~50 ka) of evidence suggests that humans may have inhabited the popularly employed radiocarbon dating tech- Finland as far back as the Eemian interglacial nique. Altogether a general picture of the stadials (~120-130 ka; Schulz et al. 2002, cf. Pettitt and and interstadials of the Weichselian glaciation has Niskanen 2005), Sweden >40 ka (Lundqvist 1964, been produced although reconstructions of older Heimdahl 2006), and Arctic Russia ~35-40 ka stadials and interstadials are partly hypothetical (Pavlov et al. 2001). (Lundqvist 1992, Mangerud 2004; Fig. 2). A prob- Recent robust modelling of the nearby British- lem with correlation of more or less global records Irish ice sheet has produced a highly dynamic ice for climate change and inferred glacial response is sheet with numerous binge/purge, advance/retreat that information on the local variation in timing, (i.e., yo-yo) cycles dominated by ice streaming and regional leads or lags is missing. In order to fill (Hubbard et al. 2009). Phases of predominant ice this gap absolute dates for local and regional events streaming activity coincide with periods of maxi- must be acquired to build a more detailed glacial mum ice extent and are triggered by abrupt transi- history and a better understanding of how the (gla- tions from a cold to relatively warm climate, result- cial/climatic) system operates. ing in major iceberg/melt discharge events (Hub- Lately, several studies have shown that the Mid- bard et al. 2009). The fjord-dominated landscape of and Late Weichselian glacial history may be more Norway and its shelf are in no doubt a record of the complex than generally believed. For example, the dominance of ice streams in draining the SIS

7 Timothy F. Johnsen

Fig. 2: General history of the Scandinavian ice sheet over the Weichselian glaciation. All maps are shown with modern sea level. Black dots are study area locations (Fig. 3). The Younger Dryas outline is according to Andersen et al. (1995) and Mangerud (2004), ages after Rasmussen et al. (2006). The LGM 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 stage ages from Martinson et al. (1987). The pre-LGM outlines are hypothetical although using interpretations of known sites.

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

(Longva and Thorsnes 1997, Ottesen et al. 2005, sediments and address ice-free periods/interstadials Kessler et al. 2008, Kleman 2008; Fig. 3), and (Pilgrimstad and Langsmoen, Papers II and III), provide the setting for a potentially dynamic and while the third addresses the vertical rate of glacia- responsive ice sheet. Dynamic ice sheet behaviour tion during the Late Glacial (Mt. Åreskutan, Paper has been documented for Norway as well (Olsen et IV). Each site is the topic of one paper. All sites al. 2001a). Modelling of the SIS suggests that it is have complementary dating results either from the strongly sensitive to small-scale ice-sheet instabil- site or from nearby to allow comparison to multiple ity (Charbit et al. 2002). Perhaps the SIS (at least OSL or TCN dates. its western margin) behaved in an even more dynamic manner than documented so far, similar to Southern Sweden – Vimmerby moraine the British-Irish ice sheet. Vimmerby: The South Swedish Upland lies be- Therefore, the objective of this doctoral thesis tween the west and east coasts of southern Sweden work is to improve our understanding of the history and is 200-300 m asl. The crystalline bedrock and dynamics of the SIS for different areas and forms an undulating landscape with isolated insel- covering different time periods. This will be bergs, shallow valleys and locally deep-weathered achieved by employing two dating techniques, bedrock. The common occurrence of hummocky Optically Stimulated Luminescence dating (OSL) moraine in parts of the South Swedish Upland and Terrestrial Cosmogenic Nuclide (TCN) expo- indicates widespread stagnation (dead ice) instead sure dating. Thus, the main research questions of active retreat during the last deglaciation stemming from this objective include: (Björck and Möller 1987). The Vimmerby moraine is one of the few prominent features in the South Was the behaviour of the SIS characterized as Swedish Upland that is related to the former mar- stable or dynamic? For example, did its margin of the decaying SIS (Agrell et al. 1976, Malm- gins wax and wane often as modelled for the berg Persson et al. 2007). The Vimmerby moraine British-Irish ice sheet, or infrequently? site was selected for a number of reasons (Fig. 3). Was the ice sheet thick or thin during the Late The timing and pattern of deglaciation in the South Glacial? What was the vertical rate of deglacia- Swedish Upland is not well known (Lundqvist and tion? Wohlfarth 2001). There is an opportunity to evalu- Can OSL and TCN dating produce accurate re- ate TCN dating results by comparison to other sults and that are useful for improving our un- estimates for deglaciation from nearby (Lundqvist derstanding of the glacial history of the SIS? If and Wohlfarth 2001), and prior to attempting to so, when were areas deglaciated or when were date moraines in areas where there is even less they ice-free? dating control (e.g., Paper IV). By dating the Vimmerby moraine, well-known deglacial chro- By answering these questions, this knowledge nologies from dated moraine systems from the can be used to constrain ice sheet models and guide west coast and detailed varve chronology on the further research into refining the history and dy- east coast of Sweden can be better correlated namics of the SIS. across the South Swedish Upland (cf. Lundqvist and Wohlfarth 2001). As well, sandurs adjacent to Study areas the Vimmerby moraine have been dated thoroughly and give consistent OSL ages that are thousands of Two areas were selected for study (Fig. 2 and 3). years older (~5-10 ka) than the estimate for degla- The first area of study is in the Småland area of ciation of the Vimmerby moraine (Alexanderson southern Sweden at Vimmerby and Lannaskede and Murray 2007). (Paper I), which is located in the southern portion Central Sweden and Norway – Jämtland- of the former SIS. The second study area is the Trøndelag area Jämtland-Trøndelag area of central Sweden and Norway, which includes three study sites located in Pilgrimstad: The Pilgrimstad stratigraphic site is this former central area of the SIS and west of the an important site for paleoecological and paleogla- LGM ice-divide. Two of these sites contain sub-till ciological reconstruction (Kulling 1945, Frödin

9 Timothy F. Johnsen

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 indicated, 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.

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

1954, Lundqvist 1967, Robertsson 1988a,b, García The Langsmoen and Flora sites contain sub-till Ambrosiani 1990, Wohlfarth 2010, Paper II; Fig. sediments, are in central Norway (Trøndelag re- 3). The results from this critical central location, gion) ~110 km from the present coast, ~300 km just east of the Scandinavian mountains and west from the LGM ice sheet outline (Fig. 2), 4 km from of the former LGM ice-divide of the SIS, have one another, and lie within the Nea River valley implications for the SIS as a whole. It contains sub- (relief ~300 m). Sediments from the Flora site till organic and minerogenic sediments, including represent an ice-proximal, ice-dammed glacio- lower sediments of proglacial sub-aquatic sedi- lacustrine paleoenvironment, and sediments from ments, and upper sediments representing a transi- the nearby Langsmoen site represent a fluvial del- tion from glaciofluvial to fluvial to lacustrine taic ice-free paleoenvironment. Eight radiocarbon deposition (Lundqvist 1967), and possibly includ- dates from fine-grained sediment within the Flora ing some aeolian sediment (Robertsson 1988a,b). section give consistent ages of 20.9 ±1.6 cal. ka BP Paleoecological interpretation is of a cool, subarc- and indicate that these sediments were deposited tic-arctic environment based on several proxies, during what is traditionally thought of as the LGM mainly from the organic beds (Robertsson interval. The content of plant remains and other 1988a,b). A number of radiocarbon ages from organics is not sufficiently high in the Langsmoen organic-rich sediments have been at the limit of the sub-till sediments to allow testing by radiocarbon technique and so have been considered unreliable dating. Thus, the Langsmoen stratigraphic site was (cf. Wohlfarth 2010). Consequently the site has selected for study as stratigraphically-related sedi- been assigned an Early Weichselian age (MIS 5 ments from the nearby Flora section give consis- a/c; >74 ka), based on pollen stratigraphic correla- tent radiocarbon dates that fall within the Trofors tions with type sections in continental Europe. interstadial, and the sandy sediments of fluvial There is an opportunity to apply a complementary deltaic origin may be suitable for the OSL dating dating technique, OSL, on the sub-till sediments to technique. Comparison of OSL and radiocarbon determine the age of these sediments and ice-free dating results will allow evaluation of the existence conditions. And, detailed paleoenvironmental work of the Trofors interstadial in this area of Norway, has already been completed (Robertsson 1988a,b) and an assessment of the radiocarbon dating of and is ongoing (Wohlfarth et al. in prep.), so by bulk sediment of low organic content, a method accurately dating the sediments we can assign the commonly, although not exclusively, used in the paleoenvironmental inferences to the correct time work of Olsen et al. (2001b). period. Mt. Åreskutan: The county of Jämtland in cen- Langsmoen and Flora: Stratigraphic and geo- tral Sweden includes mountainous areas of moder- chronologic study of sub-till sediments from many ate relief (~800 m) and adjacent low-relief (~100 sites throughout Norway has revealed that the m) rolling hill landscape with numerous lakes and western margin of the SIS behaved in a much more peatlands. Within this area, Mt. Åreskutan is 1420 dynamic manner than previously believed (Olsen m asl and is largely isolated in its eastern position 1997, Olsen et al. 2001a,b, 2002). Four interstadi- from other mountains in the Scandinavian moun- als during the Middle and Late Weichselian glacia- tain chain. The modern tree-line is at about 950- tion mark periods of major ice retreat, and sub-till 1050 m in this area. Remains of three tree species sediments from these periods have been identified have been found at high-elevation alpine sites in and dated at many sites throughout Norway, indi- central Sweden, principally Mt. Åreskutan (Fig. 3), cating near-synchronous behaviour of the western that are hundreds of meters above modern tree-line, margin of the ice sheet. Of these interstadials, the and dating to a time (as old as ~17 cal. ka BP, Trofors interstadial, which separates the LGM into Kullman 2002) when it is commonly inferred that two stadials, represents the strongest evidence for a the sites were occupied by the SIS, and deglaciated dynamic ice sheet, and is the most difficult for the 10.3 to 10.0 cal. ka BP (Lundqvist 2002). The Quaternary community to accept. Thus, the Trofors existence of these tree remains, their elevation interstadial requires more study to verify its exis- above modern tree-line, their species, and age po- tence. tentially have tremendous implications for our understanding of the dynamics of the SIS, the pat-

11 Timothy F. Johnsen

tern and rate of migration of tree species, the loca- Sediments that were not adequately exposed to tion and role of refugia in re-establishing plants, light during transport may give maximum ages due and paleoclimatic conditions and variability. Con- to the inheritance of a luminescence signal from a sequently there has been debate on these data and previous burial event. Likewise, in some geo- what they represent (Birks et al. 2005, 2006, Kull- graphical areas the characteristics of the mineral man 2005, 2006), but there has not been an attempt (e.g., quartz) being measured may not be condu- to directly assess these data against a complemen- cive to the OSL technique (Preusser et al. 2009, tary dating technique, despite that the reliability of Alexanderson and Murray 2009, Steffen et al. the ages is central in any accurate paleoglaciologi- 2009), for example, not satisfying the assumption cal or paleoecological reconstruction. Thus, study that most of the luminescence signal is derived of the deglaciation of high-elevation sites in central from what is termed the ‘fast’ component, or the Sweden, by using the TCN dating method, can signal is too weak/dim. allow determination of the timing of deglaciation, The OSL technique typically uses sandy sedi- assessment of the thickness evolution of the ice ments; the same grain size available from sub-till sheet during the Late Glacial, and evaluation of sediment at the Pilgrimstad (Paper II) and radiocarbon results from high elevation tree re- Langsmoen (Paper III) sites. The sediments from mains. these sites are potentially glacial or at least glacially-related, which means that the grains may not have been adequately exposed to light (i.e., the Methods sediments were not fully bleached, or ‘zeroed’) prior to deposition (Fuchs and Owen 2008, The two dating methods used in this work (OSL, Thrasher et al. 2009), and/or the quartz minerals TCN) have the advantage of dating deposits that may not be sensitive enough (e.g., through repeated are directly related to the ice sheet, in addition to transport and burial episodes) to produce adequate dating ‘ice-free’ depositional events rather than the luminescence signals (Pietsch et al. 2008, Alex- invasion of plants or animals that radiocarbon is anderson and Murray 2009). Thus, adjustments to limited to. The age limit of these techniques greatly the OSL protocols and additional tests were made exceeds the radiocarbon technique, and potentially to accommodate these potential issues and to get includes the entire Weichselian Glacial or older. In the most reliable dates possible; explained below. these respects the OSL and TCN dating techniques A key methodological decision was the collection are complementary to the radiocarbon dating tech- and dating of multiple samples from a variety of nique; and, all of my sites have radiocarbon dating lithofacies and positions within the sediments at results or from nearby. each site, in order to evaluate the consistency of the Samples were collected by me in the field fol- results. lowing standard procedures. Study also included Defects and chemical impurities within miner- sedimentological analysis, and geomorphological als such as quartz or feldspar can act as traps for analysis using aerial photograph interpretation, free electrons (e.g. Aitken 1985, Preusser et al. GIS, and field mapping. 2009). These free electrons are produced from OSL dating ionising radiation from radioactive elements in sediment and from cosmic rays. Luminescence The optically stimulated luminescence (OSL) dat- results when some external stimulation (e.g., expo- ing technique (Papers II and III) determines the sure to light) ejects electrons from traps to emit time elapsed since buried sediment grains were last photons (light). The longer the mineral is exposed exposed to daylight. It is a reliable chronological to ionising radiation, the greater the store of elec- tool, reflected in its growing popularity in Quater- trons trapped within the mineral and the stronger nary studies (Murray and Wintle 2000, Murray and the luminescence signal, although the relative Olley 2002, Duller 2004, Lian and Roberts 2006, strength of this signal will vary aliquot-to-aliquot Wintle 2008). However, the OSL dating technique (an aliquot is a small portion of the sample placed may not always produce meaningful or good qual- on a small disc for measurements; Fig. 4). By ity results (e.g., Alexanderson and Murray 2009). comparing the natural luminescence of an aliquot

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

Fig. 4: (a) OSL samples were collected by hammering an opaque tube into sediments. (b) A Risø TL/OSL reader is open showing aliquots 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 second computer that controls the reader. (c) Analysis of repeated measurements of the luminescence (inset) for an aliquot using different 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. This process 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, and other factors.

to the relationship between known doses given and were wet-sieved below 250 μm and in some cases the luminescence signal produced (i.e., a ‘growth minerals magnetically separated using a Frantz curve’), the equivalent dose is determined (Fig. 4). magnetic separator. Chemical preparation to isolate The water content of the sediment over the entire quartz grains, including heavy liquid separation burial history will impede ionising radiation and so (2.62 g cm-3) in some cases to remove feldspars, must be considered in the final determination of the treatment with 10% HCl for 5-30 minutes to re-

sediment dose rate. The cosmic ray contribution for move carbonates, 10% H2O2 for 15-30 minutes to shallow-depth samples is also considered. Then the remove organics, 10% and 38% HF for 60-120 luminescence age is calculated as the equivalent minutes to remove everything but quartz and to dose divided by the sediment dose rate. etch the outer surface of the grains to remove the To not expose the samples to light, they were parts affected by alpha radiation, and 10% HCl taken in opaque plastic tubes and stored in black again for 40 minutes to remove any fluorides that bags until opened under darkroom conditions for may have formed in the previous step. This was initial preparation. This included separating the performed at the Nordic Laboratory for Lumines- sample into portions for OSL measurements, sedi- cence Dating in Risø, Denmark, where the OSL ment dose rate determination (gamma spectrome- measurements were also completed. try), and water content measurements. Samples

13 Timothy F. Johnsen

The samples were analysed using aliquots of nent analysis were used to select the peak and quartz (63 to 250 μm) on Risø TL/OSL-readers background portions of the luminescence signals (Fig. 4) equipped with calibrated 90Sr/90Y beta (i.e., ‘channels’) that produced the best results for radiation sources (dose rate 0.14-0.35 Gy s-1), blue dose recovery experiments, and that gave a domi- (470 ±30 nm; ~50 mW cm-2) and infrared (880 nm, nant fast component. Signal component analysis of ~100 mW cm-2) light sources, and detection was some natural signals also showed the fast compo- through 7 mm of U340 glass filter (Bøtter-Jensen nent to be dominant. The equivalent doses were et al. 2000). This system essentially automates the then calculated in Risø Luminescence Analyst numerous repeated measurements, heatings and software and in Microsoft Excel. To be accepted radiation doses given to multiple aliquots of each aliquots had to pass rejection criteria for the recy- sample (i.e., the protocol). Analyses employed cling ratio, recuperation, equivalent dose error, and post-IR blue SAR-protocols (single-aliquot regen- the signal had to be more than three sigma above erative-dose protocol, SAR; Murray and Wintle the background. Decay and growth curves also had 2000, 2003; Banerjee et al. 2001), adapted to suit to be regular in shape. Ages were calculated using the samples based on internal methodological tests the mean and median of the equivalent dose popu- such as dose recovery and preheat experiments. lation of accepted aliquots for each sample, as well The ‘post-IR blue’ portion of the protocol was as using the natural and saturated water contents. added to minimize the contribution from any po- As well, a sensitivity analysis was completed to tential feldspar minerals that passed the physical determine quantitatively which uncertainties have a and chemical treatments. A relatively high test- larger effect on the age. dose (~50 Gy) was necessary to get a statistically precise test signal since Swedish quartz is rela- TCN dating tively dim (Alexanderson and Murray 2009). 100 s The terrestrial cosmogenic nuclide (TCN) exposure of illumination at 280° between cycles improved dating technique can allow determination of the recuperation (response to zero dose), by emptying amount of time that a rock surface has been ex- traps prior to a new radiation dose being given posed to cosmic radiation (e.g., how long ago an (Murray and Wintle 2003). Sediment dose rates ice sheet deposited a boulder; Papers I and IV). were calculated from gamma spectrometry data Cosmic radiation causes the accumulation of 10Be (Murray et al. 1987) and included the cosmic ray in situ within quartz rock and the measurement of contribution (Prescott and Hutton 1994) for shal- the concentration of 10Be against the production low-depth samples. Natural and saturated water rate of 10Be for a given site of known latitude, ele- content was measured either using a portion of vation, topographic shielding, and sample thick- sediment from the sample tube or using pF-rings ness, provides determination of the exposure age (cylinder volumeters). (Lal 1991, Gosse and Phillips 2001). This tech- Blue-light stimulated luminescence signals have nique has proven useful in numerous studies of been shown to be made of a number of components deglacial history and landform preservation (e.g., (Bailey et al., 1997, Jain et al. 2003). If the initial Phillips et al. 1997, Licciardi et al. 2001, Balco et part of the luminescence signal is not dominated by al. 2002, Fabel et al. 2002, Clark et al. 2003, the fast component, inaccurate equivalent doses Rinterknecht et al. 2006). and ages may be produced (Choi et al. 2003, Tsu- The TCN dating technique is valuable to apply kamoto et al. 2003) because of the differing char- to the Vimmerby moraine (Paper I), and Mt. Åre- acteristics of different signal components (Singa- skutan and Mt. Snasahögarna (Paper IV) because rayer and Bailey 2003, Singarayer et al. 2004, glacially transported boulders are abundant in the Wintle and Murray 2006, Kitis et al. 2007). Simple area, and there is the opportunity to compare to component analyses of the continuous-wave OSL results already available that used other dating data from some aliquots was undertaken using techniques at or near these sites (i.e., altogether SigmaPlot 10.0, based on the parameters and for- radiocarbon, varve chronology and OSL). Similar mulas of Choi et al. (2006). This allowed quantifi- to my OSL methodology, a key decision was the cation of the fast, medium and slow components of collection and dating of multiple samples from the luminescence signal. The results of the compo- each site. To minimize the risk of processes modi-

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

fying the exposure history of the boulder, such as Samples for 10Be cosmogenic exposure dating were cosmogenic nuclide inheritance (e.g. Briner et al. collected from six glacially transported, rounded to 2001), boulder exhumation, erosion or moraine sub-rounded, weathering-resistant, granitic and deformation (Hallet and Putkonen 1994, Putkonen quartzitic boulders (0.9-2.3 m b-axis; Fig. 5) that and Swanson 2003, Zreda et al. 1994), samples were resting on top of stable and level surfaces or were from the tops of large, weathering-resistant on the broad crest of the Vimmerby moraine (Fig. boulders resting either on moraine crests (Papers I 3 and 5). The six ages were internally consistent, and IV) or bedrock (Paper IV). Boulder surfaces ranging from 14.9 ±1.5 to 12.4 ±1.3 ka with a and adjacent ground were carefully inspected for mean of 13.6 ±0.9 ka. Adjustments were made to indications of the rate and dominant mode of ero- these ages for the effects of surface erosion, snow sion. The height of quartz nodules and veins above burial and glacio-isostatic rebound, which causes adjacent softer rock within each boulder were used the mean age to increase by only ~6% to ~14.4 to estimate the rate of erosion. Altogether twelve ±0.9 ka. Thus, the southern margin of the SIS was samples were processed in the Glasgow Univer- at the Vimmerby moraine ~14 ka ago. sity-SUERC cosmogenic nuclide laboratory. It is not clear which of the moraines west of the Using the CRONUS-Earth 10Be-26Al exposure study area correlate with the Vimmerby moraine, age calculator (http://hess.ess.washington.edu), and even less clear for moraines from the southeast measured 10Be concentrations were converted to portion of the ice sheet (northern Poland and the surface exposure ages. The different 10Be produc- Baltic states). Nevertheless, the internal consis- tion rate scaling schemes were used to examine tency of the six cosmogenic ages and their com- variation in the calculated age for each sample. The patibility with previous radiocarbon ages and varve surface exposure ages were calculated using the chronology (Lundqvist and Wohlfarth 2001, ‘Lm’ scaling scheme which includes paleomag- Lundqvist 2002) indicate that the TCN (10Be) ex- netic corrections (Balco et al. 2008). Altogether posure dating technique works well for erratic ages were calculated by considering the sample boulders within this area and shows promise for thickness, latitude, weathering and erosion of the further TCN exposure studies in southern Sweden. rock surface during exposure, the changing eleva- Sandur sediments adjacent to the moraine were tion of the rock surface due to glacio-isostatic recently thoroughly dated with OSL and provide movement, and partial shielding of the rock surface consistent ages around the LGM (Alexanderson from cosmic rays by topography and seasonal and Murray 2007). While this was stated in the snow cover (Gosse and Phillips 2001). paper it was not discussed; please see the discussion below. My contribution to this work included boulder Presentation of papers selection, sampling, and sample crushing by John- sen and Alexanderson, age calculations by Fabel My contributions to this project are stated below at and Johnsen, interpretations by Johnsen and co- the end of each paper summary, and in the ac- authors, and writing mostly by Johnsen with input knowledgements section at the end of each of the from co-authors. four papers in the appendix. For details of the motivations and background Paper II: Pilgrimstad OSL for each study site refer to the study area section above or the individual papers in the appendix. Alexanderson, H., Johnsen, T., Murray, A.S. 2010. Re-dating the Pilgrimstad Interstadial Paper I: Vimmerby TCN with OSL: a warmer climate and a smaller ice sheet during the Swedish Middle Weichselian Johnsen, T.F., Alexanderson, H., Fabel, D., (MIS 3)? Boreas, 39: 367–376. Freeman, S.P.H.T. 2009. New 10Be cosmogenic ages from the Vimmerby moraine confirm the Pilgrimstad, an important stratigraphic site for timing of Scandinavian Ice Sheet deglaciation in Weichselian history, was re-excavated to expose southern Sweden. Geografiska Annaler: Series >4 m thick sub-till minerogenic and organic sedi- A, Physical Geography, 91: 113–120. ments. The architecture and lithofacies of the sediments were described in detail. Eight units

15 Timothy F. Johnsen

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. were identified representing paleoenvironmental quartz were analysed using a post-IR blue SAR- change from glaciofluvial to glaciolacustrine to protocol. Dose recovery tests were satisfactory at lacustrine and back to fluvial or glaciofluvial depo- 1.05 ±0.04 (n = 21) with deconvolution results sition; somewhat glaciotectonized and crosscut by indicating that over 90% of the luminescence sig- a clastic sandy dyke. Based on the sedimentologi- nal is derived from the fast component. The OSL cal observations of this new section at Pilgrimstad, ages are internally consistent lying in the range 52- a single interstadial rather than two is favoured 36 ka, except one from an underlying unit that is (Kulling 1967, Lundqvist 1967, Robertsson older; and compatible with existing radiocarbon 1988a,b). As well, the sand within brecciated ages, including two we measured with AMS. The lacustrine sediments is likely reworked from adja- mean of the OSL ages is 44 ±6 ka (n = 9). This cent glaciofluvial sand through glaciotectonic places the interstadial sediments in the Middle processes, rather than having an aeolian origin Weichselian (MIS 3) and possibly corresponds to (Robertsson 1988a,b). one or more of Greenland interstadials 17-10. The Ten samples from the variety of lithofacies OSL ages cannot be assigned to the Early Weich- were collected for OSL dating. Single aliquots of selian (MIS 5 a/c) as proposed by earlier workers

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

(Robertsson 1988a,b) for all reasonable adjust- niques strongly suggests that this area was ice-free ments to water content estimates and other parame- around ~21 or 22 ka. This supports findings from ters. These new ages indicate that, during the Mid- other sites throughout Norway that indicate ice-free dle Weichselian, the climate was relatively warm conditions ~25-20 ka, collectively termed the Tro- and the SIS was absent or restricted to the moun- fors interstadial, an interstadial that divides what is tains for at least part of MIS 3. This is supported by generally thought of as the LGM into a two part results from other recent studies completed in Fen- LGM (Olsen 1997, Olsen et al. 2001a,b, 2002). noscandia. The existence of the Trofors interstadial along with I provided significant input into all stages of other interstadials during the Mid- and Late this work including OSL sampling, preparation, Weichselian (MIS 3 and MIS 2) indicates that not sedimentological and stratigraphical work as well only the western margin, but the whole western as writing. I analysed all the data, with some input part of the SIS, from the ice divide to the ice mar- from co-authors. gin was highly dynamic (Fig. 6). These large changes in the ice margin and accompanying Paper III: Langsmoen OSL drawdown of the ice surface would have affected Johnsen, T.F., Olsen, L., Murray, A. Submitted. the eastern part of the ice sheet as well. OSL ages in central Norway confirm a MIS 2 For this work I was in charge of and performed interstadial (25-20 ka) and a dynamic Scandina- all aspects: conception, acquiring funding, OSL- vian ice sheet. Quaternary Science Reviews. sampling and measurements, data analysis, and writing. Olsen completed the original stratigraphic Four samples were collected from sub-till sandy and sedimentologic fieldwork years earlier, wrote sediments within the Langsmoen stratigraphic site, this portion of the paper, and recommended strati- as well as one from silty sand sediments in the graphic sites for us to visit. stratigraphically-related Flora site. The local paleoenvironment was such that Langsmoen sedi- Paper IV: Åreskutan TCN ments represent glaciofluvial/fluvial ice-distal Johnsen, T.F., Fabel, D., Stroeven, A. High- environment followed by ice damming and the elevation cosmogenic nuclide dating of the last deposition of ice-proximal lacustrine sediments at deglaciation in the central Swedish mountains: the Flora site. These sites were in-turn buried by implications for the timing of tree establish- LGM-2 till and glaciotectonized. ment. Manuscript. OSL dose recovery tests were good at 1.06 ±0.03, n = 18, and signal component analysis indi- TCN exposure ages of glacially transported boul- cated that over 90% of the luminescence signal was ders from the summit of Mt. Åreskutan (1420 m from the fast component. Both large and small asl) in central Sweden are consistent (adjusted aliquots were measured to see if there was a sig- mean age = 10.6 ±0.6 ka, n = 3) and similar to nificant age difference caused by incomplete lower elevation dates for deglaciation from the bleaching or other processes. OSL ages for all region. Thirty-five kilometres down-ice (1200 m Langsmoen samples and for both large and small asl; west) the highest elevation moraine in Sweden aliquots are consistent at 22.3 ±1.7 ka, n = 7. Study produced by the retreating SIS, the Snasahögarna of modern river sediments in the region indicates SIS moraine, gave consistent TCN ages (sampled that fluvial transported sediment can be bleached. at 1125-1149 m asl; adjusted mean age = 12.0 ±0.6 The sample from ice-proximal glaciolacustrine ka, n = 3). The ages from both sites were adjusted sediment at the Flora site gave an apparent old age for the effects of glacio-isostatic rebound and of ~100 ka likely reflecting incomplete bleaching shielding by snow cover, and the effect of erosion of the sediments prior to deposition. was considered negligible. The difference in TCN Eight radiocarbon ages of sediment from the ages between Mt. Åreskutan and Snasahögarna SIS Flora site gave consistent ages (20.9 ±1.6 cal. ka moraine probably reflects a geographical differ- BP) that overlap within 1 with OSL ages from the ence in the timing of deglaciation between sites as nearby Langsmoen site. The similarity in age shown in detailed ice margin reconstructions for within and between these stratigraphically-related the area (Borgström 1989), and/or possibly differ- sites and using different geochronological tech-

17 Timothy F. Johnsen ences in the actual historical snow depth between Discussion the sites. Together the ages from both these sites, but par- The results from the four studies (Papers I-IV) ticularly for Mt. Åreskutan, clash with unusually directly address the objective and research ques- old radiocarbon dates of the remains of three spe- tions. These research topics include the history and cies of tree from the summit area (as old as ~17 geochronology of the SIS, the behaviour of the ice cal. ka BP; Kullman 2002). We could not find a sheet (‘dynamic’ versus ‘stable’), the thickness and plausible hypothesis that accommodates both the vertical rate of deglaciation of the ice sheet, and the radiocarbon and TCN ages from this site. The un- application of two relatively new dating techniques usually old radiocarbon ages are rejected on the for understanding glacial history. As explained in basis of incompatibility with consistent TCN ages the introduction, there are strong motivations to for deglaciation, and incompatibility with pa- better understand the history and dynamics of for- leoecological and paleoglaciological reconstruc- mer ice sheets – scientific, environmental and so- tions. Nevertheless, the mere presence of tree re- cial. A part of this knowledge-seeking is to have mains, and of three different species, at this high good tools (e.g., dating techniques) that produce elevation well above (400-500 m) modern-tree line reliable results to help confidently decipher natural indicates that there was considerable variation in archives. Thus, I will first discuss the quality of the the climate during the Holocene. The problem lies TCN and OSL dating results before I discuss the in reliably dating tree remains from high elevation implications of my results for our understanding of in central Sweden. We suggest that contamination the history and dynamics of the SIS. from calcareous bedrock or neutron production from lightning may have caused the age bias. TCN dating quality Given these results it is strongly recommended that Both papers that use the TCN dating technique specimens for radiocarbon dating be thoroughly (Paper I and IV) demonstrate that the TCN dating tested to ascertain possible sources of contamina- technique can produce consistent results for the tion, and complementary dating techniques be timing of local deglaciation. In Paper I, ages for the employed before proposing radical changes to the Vimmerby moraine were consistent with a 900 paleoglaciological and paleoecological history. year standard deviation for six boulders with a High elevation areas deglaciated sometime be- mean of 13.6 ka – which is 7% standard deviation tween ~12.0 and 10.6 ka coinciding approximately of the mean. In other words, these results are of with the termination of the Younger Dryas cold high precision. These results are considered excel- interval (11.7 ka), giving a vertical rate of deglacia- lent for this dating technique and for the dating of tion ~50 m 100a-1. However, as this value is de- moraines (Fabel, pers. comm.). Studies of moraines rived from data over a large area, it is averaging elsewhere have included problems of nuclide in- estimates over space and time; and thus local verti- heritance which produces old ages, or problems of cal rates of deglaciation can be higher. The vertical boulder exhumation, erosion or moraine deforma- rate of deglaciation in the Mt Åreskutan area may tion (e.g., Fabel et al. 2006, Briner et al. 2001, have been as high as ~500 m 100a-1. Sometime Hallet and Putkonen 1994, Putkonen and Swanson after deglaciation Betula pubescens, Picea abies, 2003, Zreda et al. 1994) which produces young and Pinus sylvestris, grew at 1360 m asl near the ages. The six boulders selected from the Vimmerby summit area of Mt. Åreskutan. moraine do not appear to suffer from any of these I was in charge of and performed all aspects of processes as the TCN dates are highly consistent the project: conception, acquiring funding, TCN- with each other and with estimates for the deglacia- sampling, data analysis, and writing, except for tion of the Vimmerby moraine from the radiocar- sample measurements which were completed by bon dated varve chronology (Lundqvist and Wohl- Fabel. 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 radiocarbon dating. These results are highly promising and

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

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 modern sea level. Black dots are study area locations (Fig. 3). See text for explanation.

19 Timothy F. Johnsen

may mean that other moraine systems in southern age likely due to incomplete bleaching of the lumi- Sweden can be dated successfully (cf. Larsen et al. nescence signal. Thus this tenth age is a maximum 2010). In Paper IV, TCN samples were collected age for the deposition of overlying sediments. The from groups of three erratics from two mountain good correspondence of the nine ages indicates that sites. The first site was from the summit of Mt. incomplete bleaching of the sediments was not a Åreskutan, and the second was from a high eleva- problem. Put another way, if incomplete bleaching tion moraine of the SIS on Mt. Snasahögarna. were a significant process we would not expect Dates were exceptionally consistent at each site; sediments from a variety of lithofacies to have the i.e., very high precision. The percentage standard same degree of incomplete bleaching to result in deviation of the mean of the TCN dates for the consistent ages. In terms of the OSL dating tech- three erratics at each site is only 1%. Note that for nique these are considered consistent results of the Vimmerby moraine study, twice as many errat- adequate precision. A mean age of 44 ±6 ka (n = 9; ics were used in deriving this statistic. Neverthe- 14% standard deviation of the mean) for the Pil- less, the very high precision of the ages at each site grimstad sediments agrees with the bulk of previ- is considered outstanding. The difference in age ous age determinations from the site and sites in between the sites likely reflects the geographical northern and central Sweden, which fall between difference in the relative timing of deglaciation as ~60 and ~35 ka (Wohlfarth 2010). revealed in detailed mapping of the retreating SIS Fewer samples for OSL dating were collected margin (Borgström 1989). Similar to the Vim- from the Langsmoen site (n = 4; Paper III). How- merby moraine study, it appears that these samples ever, unlike for the Pilgrimstad site both large and do not suffer significantly from any processes that small aliquots were measured (for three of the could cause ages to appear too young or old, as the samples) to look for any differences in dose that TCN dates are highly consistent at each site and may reflect incomplete bleaching of the sediments with estimates for deglaciation from lower eleva- (Murray and Olley 2002). As well, a modern flu- tion sites in the area. However, as acknowledged in vial sediment sample was collected from the region the paper, the estimate of the snow depth covering to evaluate modern bleaching potential. Both large the boulders over the entire exposure history may and small aliquots gave consistent OSL ages (22.3 be important at these mountain sites and is an esti- ±1.7 ka, n = 7; standard deviation of the mean of mate that may even vary between the Mt. Åresku- 8%) for sub-till glaciofluvial/fluvial sediments at tan and Mt Snasahögarna sites. This snow depth the Langsmoen stratigraphic site, and an apparent estimate along with the inherent uncertainties of old age (~100 ka) for a poorly bleached sample of the TCN dating technique (Gosse and Phillips glaciolacustrine sediment at the nearby strati- 2001) mean that while the radiocarbon dates from graphically-related Flora site. The apparent old age the summit of Mt. Åreskutan can be rejected (Pa- for the Flora site sediment is expected for ice- per IV), the accuracy of the value for the vertical proximal glaciolacustrine sediments due to incom- rate of deglaciation is a rough estimate. plete bleaching (Alexanderson and Murray 2009). The consistency of the ages from Langsmoen OSL dating quality sediments indicate that they were likely completely Similar to results from the TCN dating technique, bleached prior to deposition, and are of good preci- both papers that use the OSL dating technique sion. Eight radiocarbon ages of bulk sediment in (Paper II and III) demonstrate that this technique the Flora section are fairly consistent (20.9 ±1.6 can produce consistent results. Nine out of ten cal. ka BP; standard deviation of the mean of also dates from the Pilgrimstad site gave fairly consis- 8%) and overlap within 1 with OSL ages from the tent ages (44 ±6 ka, n = 9; standard error of the nearby Langsmoen site. As sediments at these two mean of 14%), despite the ages being derived from sites are expected to have formed around the same variety of lithofacies representing different sedi- time based on stratigraphic study, and the ages mentary environments and transport histories. The overlap between both sites, both the OSL and ra- tenth date was from the lowest stratigraphic unit in diocarbon ages are considered to be accurate. This the excavation and made of coarse gravels that raises the credibility of using radiocarbon dating of were likely ice-proximal, giving an erroneous old bulk sediments; a technique that has been used to

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

study interstadial sites throughout Norway (Olsen highly important. Ages were not biased by varia- 1997, Olsen et al. 2001a,b, 2002). tion in the sediment dose rate as these did not cor- Incomplete bleaching results in an apparent age relate. The sediments from both sites had rather overestimation, and is fairly common in glacial dim signals, probably indicating that the sediments settings (Fuchs and Owen 2008, Alexanderson and experienced few transport and burial cycles; cycles Murray 2009). As sediments from both Langsmoen that would help ‘sensitize’ the quartz mineral and and Pilgrimstad were glacial or glacially-related, make it more capable of holding charge (Pietsch et incomplete bleaching was addressed, and in differ- al. 2008, Alexanderson and Murray 2009). This ent ways. The age of multiple samples from differ- resulted in lower signal-to-noise ratios, requiring ent lithofacies and over a vertical range of the more measurements of the equivalent dose as a sediments from both sites was arguably the best higher number of aliquots did not meet the rejec- approach to address incomplete bleaching. As tion criteria. stated earlier, it is not reasonable to expect sediments that represent different environments and Applying the TCN and OSL techniques transport histories to produce consistent ages if Within this work the TCN and OSL techniques incomplete bleaching is significant. As fewer sam- provided fruitful results. One important difference ples were available from the Langsmoen site, is that while the TCN dating technique was used to measurement of both large and small aliquots was determine the timing of local deglaciation, the undertaken to see if smaller populations of grains OSL dating technique was used to estimate the would give a significantly younger age due to in- timing of ice-free periods; parts of interstadials that complete bleaching (Murray and Olley 2002, likely spanned thousands of years. Like TCN dat- Duller 2008); ages were similar. As well, a sample ing technique, the OSL dating technique produces of modern fluvial sediment was measured and ages with typically large uncertainties. Thus, gen- shown to not have a significant residual signal. erally speaking, the radiocarbon technique is better Thus, incomplete bleaching is not an important to use when precision is important. However, at the phenomena for sediments at both these sites; if it Pilgrimstad site (Paper II) a number radiocarbon were then the ages would be considered maximum results were at or very near the limit of the radio- ages. carbon technique (~50 ka), while at the Flora site Specific tests and analyses were conducted to (Paper III) radiocarbon dates were from bulk sedi- ensure the OSL results were of good quality. ment samples and produced controversial ages. Sediments were bleached, given a known labora- Thus the reliability of the radiocarbon results at tory radiation dose, and then the equivalent dose both these sites was in question and benefitted was measured and compared to the given labora- from comparison to an alternative dating technique tory dose (i.e., a dose recovery experiment); results like OSL. Note that radiocarbon dates at Mt. Åre- were satisfactory (i.e., close to unity). Deconvolu- skutan were also questionable (Paper IV) and so tion of the luminescence signals was completed to warranted the use of an alternative although less see what portions of the signal produced fast, me- precise dating technique like TCN exposure dating. dium and slow components and to select integra- For glacial research, OSL dating will be mostly tion limits (‘channels’) that best isolate the fast limited to use in valley positions where sandy component and produce the best dose recovery sediments may be available. It clearly is useful in results; these channels were also used for equiva- the study of interstadial sediments and where the lent dose determinations. Values for the recycling lower precision is tolerable, or at least as a com- ratio and recuperation, and the shape of the lumi- plementary dating technique. Within Sweden, in- nescence signal and growth curves were all as- completely bleached sediments and low-sensitivity sessed and used to isolate aliquots that had good quartz are a common problem in OSL dating, and OSL characteristics (Fig. 4). A sensitivity analysis appears to be related to the degree of sediment was completed at each site to understand the rela- reworking and the source bedrock, with the Dala tive importance of different variables and their sandstone providing high-sensitivity quartz (Alex- uncertainties in the calculation of the age; and anderson and Murray 2009). In this doctoral re- revealed that water content estimates were not search, TCN dating was limited to surface expo-

21 Timothy F. Johnsen

sure dating of glacially transported boulders as I geography of alpine blockfields (Nesje and Dahl was most interested in the deglaciation history of 1990). Extrapolation of this upper ice sheet limit to the ice sheet. However, TCN dating can also be central Sweden would have meant that numerous used in paleoglaciology to measure the pattern and mountain tops, including Mt. Åreskutan, were ice- depth of glacial erosion, and dating the burial of free during the LGM. This notion seemed like a sediments (Fabel and Harbor 1999). distinct possibility when considering the results from the Langsmoen site in central Norway for ice- Late Quaternary history and dynamics of free conditions sometime from 25-20 ka (Paper the SIS III). However, the TCN dating results from the Results from the four papers have improved our summit of Mt. Åreskutan (Paper IV) do not support understanding of the history, geochronology, and this hypothesis. Recent TCN study in central and dynamics of the SIS. TCN dating results from the south Norway indicates that the apparent trend in Vimmerby moraine (Paper I) were in agreement the lower limit of blockfields mapped by Nesje et with earlier estimates of the timing of deglaciation al. (1988) appears to represent an englacial thermal and thus confirm the age of this feature and the limit of the SIS, rather than the upper limit of SIS former position of the SIS in southern Sweden. during the LGM (Goehring et al. 2008; cf. Ballan- New mapping of the Vimmerby moraine com- tyne and Hall 2008). A site near the Swedish- pleted by the Swedish Geological Survey (Malm- Norwegian border ~195 km south of Mt. Åreskutan berg Persson et al. 2007) has required some altera- indicates that the LGM ice sheet was above 1460 tion in the pattern of deglaciation for this time m asl and rapid deglaciation commenced approxi- period (Fig. 5). These results suggest that prior to mately coinciding with the termination of the the formation of the Vimmerby moraine (~14 ka), Younger Dryas cold interval (i.e., similar to results the mean ice retreat rate in eastern portion of from Paper IV; Goehring et al. 2008). southern Sweden was rather high at >150 m a-1 Papers II and III present evidence of ice-free pe- (Lundqvist and Wohlfarth 2001). The local retreat riods during times when it has been assumed the rate in the Vimmerby area was very approximately sites were covered by the SIS (Mangerud 2004; 70 m a-1, while north of the moraine a concentra- Fig. 2). For the Pilgrimstad site, it was inferred that tion of varve data indicates the retreat rate was 135 the Pilgrimstad interstadial sediments belonged to m a-1 (Wohlfarth et al. 1998). the MIS 5a/c based mostly upon pollen stratigra- TCN dating results from Mt. Snasahögarna and phy (Robertsson 1988a,b), while OSL dates and the summit of Mt. Åreskutan (Paper IV) contribute the bulk of radiocarbon dates indicate ice-free to the three-dimensional understanding of the ice conditions during MIS 3 (Paper II, Wohlfarth sheet history. Results were in agreement with local 2010). The extent of the MIS 3 ice sheet has not estimates for deglaciation while controversial ra- been agreed upon and several different ice-sheet diocarbon ages of high elevation tree mega-fossils scenarios exist (Wohlfarth 2010 and references (Kullman 2002) were rejected. The TCN results therein). The central location of Pilgrimstad east of indicate that the summit of Mt. Åreskutan (1420 m the Scandinavian mountain range and west of the asl) was covered by the SIS during the Late Gla- LGM ice-divide has implications for the ice sheet cial. High elevation areas may have deglaciated as a whole. The findings from the Pilgrimstad site around the termination of the Younger Dryas inter- mean the SIS was very small and at least restricted val (~11.7 ka) and shortly before adjacent valley to the Scandinavian mountain range during at least bottoms. The vertical rate of deglaciation in the Mt a part of MIS 3; smaller in extent than proposed for Åreskutan area may have been as high as ~500 m the possibly correlative Ålesund interstadial and 100a-1. Thus, the SIS thinned rapidly (cf. Krabill et perhaps similar in extent to the older Brørup or al. 2000). Clay-varve studies in valleys indicate Odderade interstadials (Fig. 2), or similar to the that the margin of the SIS retreated horizontally minimum MIS 3 outline according to Arnold et al. 350 m a-1 (mean; Lundqvist 1973). Earlier studies (2002; Fig. 6). The Pilgrimstad interstadial may suggested that the ice sheet during the LGM may correlate with the pre-Hattfjelldal interstadial of not have covered large areas of mountainous ter- Olsen et al. (2001a; cf. Bø and Austnes interstadi- rain in central and south Norway delineated by the als at the western coast, Andersen et al. 1981,

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

1983, Larsen et al. 1987, Mangerud et al. 2003, of the ice sheet and the formation of the Vimmerby 2010) which occurred ~55-45 ka, and perhaps is a moraine. However, the moraine is found in inter- more reasonable correlation than with the younger fluves and may be cross-cut by the sandurs, mean- Ålesund interstadial (~35 ka). A very recent review ing that the sandurs may be younger. Alternatively, of published and unpublished TL/OSL, 14C and U- no or only small moraines were deposited in the series dates for Sweden, and that includes OSL valleys. Possibly the sandur and Vimmerby mo- results from Paper II, suggests that central and raine formed during the Trofors interstadial, and northern Sweden were ice-free during the early and then during readvance and retreat of the ice sheet middle part of MIS 3 and that southern Sweden during LGM-2 the dated boulders were deposited remained ice-free until ~25 cal. ka BP (Wohlfarth on top of the moraine. A similar formation for 2010). Greenland ice core data indicates that the relict moraines in the Swedish mountains has been MIS 3 stage had a mild interstadial climate marked suggested (Fabel et al. 2006). Alternatively, the by numerous rapid switches between brief cold and OSL ages may be incorrect. Further research is longer warm times (i.e., Dansgaard/Oeschger needed to solve this mystery. events; GRIP 1993). Together, the limited ice sheet extent during the The confirmation of the Trofors interstadial in Pilgrimstad interstadial and the existence of the central Norway (Paper III) indicates that the west- Trofors interstadial require the SIS to behave in a ern marine- and ice streaming- influenced portion more dynamic manner than previously thought. of the ice sheet experienced a two-part LGM Field science is limited by the discovery and accu- (LGM-1 and LGM-2) separated by this interstadial rate dating of interstadial sediments that are un- that occurred sometime between 25 and 20 ka. If common. Thus, ice sheet models play a potential results from other interstadial sites throughout role in filling the gaps in our field-based knowl- Norway are considered reliable (Olsen et al. edge of ice sheet history and dynamics. At the 2001a,b, 2002), the Trofors interstadial is a re- same time those models must be constrained by gional phenomenon (Fig. 3; cf. Andøya interstadial field results for the geography and timing of degla- at the northwest coast, Vorren et al. 1988). The ciation and ice-free intervals. Recent robust model- LGM ice sheet had the largest extent of all stadials ling of the nearby British-Irish ice sheet has pro- during the Weichselian glacial period (Mangerud duced a highly dynamic ice sheet with numerous 2004), yet the ice sheet margins behaved dynami- advance/retreat cycles dominated by ice streaming cally during this stadial. The ice sheet margin in (Hubbard et al. 2009). Arguably modelling has central Norway was at least 110 km inland during exceeded the practicalities of field science as dy- the Trofors interstadial but later grew in extent to namic ice sheet conditions mean that it is difficult reach the Norwegian shelf (Olsen et al. 2001a) and to find sediments from earlier ice-free periods that thickened to cover mountain tops in central Swe- have survived subsequent, and multiple, glacial den (Paper IV). This indicates quite dynamic con- erosion events. It is also difficult to have the reso- ditions. This required average ice margin retreat lution to distinguish between successive events in and advance rates of less than 200-400 m per year the field. Thus, relying on field data alone could (Paper III). lead to a biased less-dynamic view of the behav- The Trofors interstadial may have extended into iour of the SIS or any paleo-ice sheet. Even the the South Swedish Highland as indicated by con- recent dynamic activity of the Greenland and Ant- sistent OSL ages (~19-25 ka; Alexanderson and arctic ice sheets (e.g., ice margins and ice streams) Murray 2007). However, these ages are of sandur has surprised many scientists and reshaped our sediments that are adjacent to the younger-dated view of what may be possible for ice sheets both Vimmerby moraine (~14 ka; Paper I). A hypothesis past and future. Both field data and modelling have to explain the results from both dating results could their limitations, but our knowledge will be best include the following events: (1) formation of san- advanced through employing both and in an inte- dur system during the Trofors interstadial, (2) cold- grated manner. based ice conditions and readvance over the sandurs without modification of the sandurs or deposition of the erratics on top of the sandurs, (3) retreat

23 Timothy F. Johnsen

A refined SIS history hypothetical to resolve on a map, but its margins may have approached those of the MIS 4 stadial Incorporation of results from newer studies can (Lambeck et al. 2010); age from (Mangerud et al. result in a more detailed picture of the history and 2010). The Ålesund/Sandes/Hattfjelldal-1 intersta- dynamics of the ice sheet. Compiling such results dial outline and age is according to Olsen (2001a; into a coherent picture is difficult given large gaps cf. Andersen et al. 1981, Mangerud et al. 1981, in our spatial and temporal knowledge, the large 2003, 2010). The ice extent during the last part of scale of study (northern Europe), the regionally this interstadial is assumed to have been small. For asynchronous behaviour of the ice sheet, and com- example, the ice thickness inferred from glacioi- peting ideas on this history. Nevertheless, such sostatic depression in the west and northwest dur- reconstructions have been proposed previously ing this period was probably only 50-70 % of that (e.g., Denton and Hughes 1981, Lundqvist 1992, of the YD interval (Olsen 2010). The Rogne stadial Holmlund and Fastook 1995, Kleman et al. 1997, outline is according to Olsen et al. (2001a), with Lambeck et al. 1998, 2010, Boulton et al. 2001, name from Mangerud et al. (1981) and age from Arnold et al. 2002, Mangerud 1991, 2004, Svend- Mangerud et al. (2010). During this stadial the ice sen et al. 2004; Fig. 2), and I attempt to present a sheet may have extended into Denmark to corre- slightly more refined reconstruction of the Weich- spond with the Klintholm advance (Houmark- selian Glacial from MIS 4 to the Younger Dryas Nielsen et al. 2005, Houmark-Nielsen 2010). The drawing from the literature, and motivated from Hamnsund/Hattfjelldal-2 interstadial outline and my own research (Fig. 6). The ages of the stadials ages are according to Olsen et al. (2001a; cf. Valen and interstadials are approximate, and the outlines et al. 1996). The LGM-1 outline for the western of the ice sheet are partly hypothetical and based margin is according to Olsen et al. (2001a), and the on limited field data. southern margin from Houmark-Nielsen and Kjær Given the strong results in favour of the Trofors (2003), age from Olsen et al. (2001a). The Trofors interstadial (Paper III), ice sheet reconstructions for interstadial outline is according to Olsen et al. this interstadial and neighbouring stadials are in- (2001a), ages from Olsen et al. (2001a) and Paper corporated (LGM-1 and LGM-2; Olsen 1997, Ol- III. The LGM-2 outline is according to Mangerud sen et al. 2001a). As this is one of most difficult (2004) and Vorren and Mangerud (2007), age from interstadials to accept, the other older and less Olsen et al. (2001a). Note that at 19 ka the ice controversial interstadials proposed by Olsen 1997 sheet had started to retreat in the south and in and Olsen et al. (2001a) and other workers (see Denmark in the southwest (e.g., Houmark-Nielsen references below) are included as well. Altogether and Kjær 2003; Ehlers et al. 2004) so this age es- these comprise, from oldest to youngest, the Kar- timate is too young for this portion of the ice sheet. møy/MIS 4 maximum extent stadial, Bø/Austnes/ However, the ice margin reached its last maximum Pre-Hattfjelldal interstadial, Skjonghelleren stadial, position at the mouth of the Norwegian Channel as Ålesund/Sandnes/Hattfjelldal-1 interstadial, Rogne late as ~19 ka according to, e.g., Nygaard et al. stadial, Hamnsund/Hattfjelldal-2 interstadial, (2007), which clearly indicate the asynchronous LGM-1 stadial, Trofors interstadial, LGM-2 sta- and highly dynamic behaviour of the ice sheet. dial, and the Younger Dryas cold interval. The Finally, the Younger Dryas outline is according to following is a summary of the information com- Andersen et al. (1995) and Mangerud (2004), ages piled to produce Figure 6. The Karmøy/MIS 4 after Rasmussen et al. (2006). Maximum extent stadial outline is according to It is not clear which of the stadials the Ristinge Mangerud (2004), age from Lambeck et al. (2010). advance (Houmark-Nielsen 2010) would correlate The Bø/Austnes/Pre-Hattfjelldal interstadial out- to; possibly the early part of the Karmøy/MIS 4 line is quite hypothetical but if ice-free conditions Maximum extent stadial? Note that the Bø and at Pilgrimstad (Paper II) correlate with this inter- Austnes may not correlate directly with each other stadial, then the ice sheet would be restricted to the as the Bø interstadial may be slightly older than the Scandinavian mountain range perhaps similar to Austnes interstadial (Mangerud et al. 2010). Fur- the MIS 3 minimum ice sheet outline that proposed thermore, the pre-Hattfjelldal interstadial correlates by Arnold et al. (2002); age from Olsen et al. most likely with the Austnes interstadial, but the (2001a). The Skjonghelleren stadial outline is too

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

pre-Hattfjelldal interstadial is less accurately dated Weichselian (cf. Fig. 2). Thus, a dynamic ice sheet so it may well include or overlap with the Bø inter- is revealed. Further study of interstadial sediments, stadial as well (Olsen et al. 2001a). The very recent relict moraines (e.g., Fabel et al. 2006), and state- review of absolute ages of Mid-Weichselian inter- of-the-art modelling of the SIS constrained by field stadial deposits for Sweden suggests that northern data will help fill gaps in our knowledge and may and central Sweden was ice-free from ~60 to 35 potentially reveal an even more dynamic and com- cal. ka BP, and southern Sweden from ~40 to 25 plex ice sheet history (cf. Hubbard et al. 2009). cal. ka BP (Wohlfarth 2010). If this is true, depending on the real age of events, it may require a Recommendations for future research smaller ice sheet for the Rogne stadial and possibly The study of the history and dynamics of any ice the Skjonghelleren stadial (Fig. 6). sheet is a very broad research area and so I limit Are the rates of ice sheet change reasonable in a my recommendations to those that are most related glaciological sense? Without ice sheet modelling, to this doctoral research. only a back-of-the-envelope evaluation can be As emphasized, field study of interstadial sites made by assuming that it could take approximately must continue. Many interstadial sites have already 5 ka for a LGM-sized ice sheet to retreat to the been discovered in Scandinavia (e.g., Robertsson mountains with more rapid retreat in the Baltic, and García Ambrosiani 1992, Olsen et al. 2001b, and perhaps 10 ka for the ice sheet to grow to Lundqvist and Robertsson 2002, Lokrantz and LGM-size from the Scandinavian mountains. Ice Sohlenius 2006) and await further study using growth is precipitation-limited and so much slower complementary dating techniques like OSL. Many than deglaciation (Näslund and Wohlfarth 2008). relict moraines await more detailed study as well. The smaller an ice sheet is during a given intersta- However, reconstructions of the ice sheet must be dial, a disproportionally large amount of time is complemented by state-of-the-art ice sheet models needed to re-grow the ice sheet for the following that rely less on global continuous data, and more stadial. The timing and extent of the various stadi- on regional field data. These models will help fill als and interstadials are approximate, and large the still rather large gaps in our understanding of uncertainties exist. Nevertheless, during the MIS 2 the SIS history. the growth phase from the Trofors interstadial into I wholeheartedly recommend taking multiple LGM-2 may be problematic for the east and south samples for dating from study sites as this appears portions of the ice sheet if southern Sweden and to be the best way to address potential issues with the Baltic were deglaciated during the Trofors any dating technique. Where possible, using more interstadial. The minimum extent of the Trofors than one dating technique would be fruitful as well; interstadial can not be too small or too long in even so-called ‘negative’ results can contribute to duration as more time would be needed to build the new understanding and developments in geochro- ice sheet thickness for the margin to advance to the nological techniques. If I only got positive results, I LGM-2 position. This is not limited for the west would have half the skill and knowledge I now marine-proximal portion of the ice sheet (Paper possess. Perhaps completing a regional reconnais- IV). During the MIS 3 the advance into the Rogne sance-level study would be reasonable to discover stadial may again be problematic for the east and sites where the technique is ‘working’ prior to south portions of the ice sheet, and if the Klintholm doing intensive study (cf. Alexanderson and stadial in Denmark correlates with this stadial. As Murray 2009). I also strongly recommend that potentially about 10 ka spans between the students or researchers work in the dating lab with Bø/Austnes/Pre-Hattfjelldal interstadial and the their samples especially for OSL dating, and that following Skjonghelleren stadial, a large ice sheet such arrangements are made early in a research could have existed during this stadial (cf. Lambeck project. For OSL samples consider setting up a et al. 2010). preparation laboratory at your home institution Despite uncertainties in the actual outlines of where samples can be partitioned and wet-sieved in the ice sheets or the accuracy of the ages, the most dim light. This will likely expedite the processing striking feature of Figure 6 is that there are many of submitted samples. When using TCN dating for stadials and interstadials over the Mid- and Late study of deglaciation I recommend multiple sam-

25 Timothy F. Johnsen

ples of glacial erratics from each site especially in appropriate material and adequate resources are areas where cold-based ice or minimal glacial ero- available. sion had occurred, as nuclide inheritance may be a Field data on deglaciation is mostly from the problem. outer margins of ice sheets to generate map-views At the very least consult with experienced users of deglaciation. A significant contribution to de- of the dating technique about field procedures and glaciation studies would be to examine the three- interesting approaches, collect more samples than dimensional (thickness evolution) of ice sheets, as I thought are needed, and collect and submit samples have done in Paper IV and like studies by Landvik early in a research program. For instance it never et al. (2003), Paus et al. (2006) and Goehring et al. occurred to me that it would be valuable to OSL- (2008). This important data would provide impor- date till at Langsmoen, but after more discussions tant vertical and subsequent volumetric constraints with other workers I realized it would have been on ice sheet evolution. The most valuable data worthy. Networking and reaching out to other would be from high elevation sites of the interior workers around you and around the world will of ice sheets, but may also provide the greatest enlarge your experience and provide creative ideas, challenge to acquire meaningful TCN results due knowledge and opportunity. to frost-weathering and possible nuclide inheri- Building directly on the doctoral results from tance processes for high elevation erratic boulders. the four study sites I recommend the following. As a compromise, mountainous sites closer to the The consistency of TCN ages from the Vimmerby margin of the ice sheet may provide more fruitful moraine, and OSL ages from adjacent sandur thou- results. sands of years older is as yet unexplained (Alex- anderson and Murray 2007, 2009). Finding similarly-aged sediments elsewhere in the South Swed- Conclusions ish Highland would be a further evaluation of the Study using the TCN and OSL dating techniques OSL results and improve our understanding of the from four sites located in southern Sweden and geographical extent of the Trofors interstadial. central Scandinavia has improved our understand- Future research efforts should focus on more de- ing of the history and dynamics of the SIS. TCN tailed mapping of the various moraine systems in dating of the Vimmerby moraine has confirmed southern Sweden, and employ an integrated dating earlier estimates for the formation of this important approach (e.g. radiocarbon dating of the first estab- landform in the South Swedish Upland around 14 lishment of vegetation, 10Be dating of erratic boul- ka. The Pilgrimstad interstadial in central Sweden ders, and OSL dating of glaciofluvial and other belongs to MIS 3 rather than to MIS 5a/c and gla- sediments). Current work is being conducted by ciers were restricted to the Scandinavian mountain Per Möller of Lund University, and Larsen et al. range during at least part of MIS 3. The existence (2010). I suspect that LIDAR aerial surveys will of the Trofors interstadial was confirmed for cen- revolutionize geomorphic mapping and the pattern tral Norway, an ice-free period that occurred ~25- of deglaciation. 20 ka and separates the LGM into two parts. Fur- Many interstadial sites throughout Scandinavia ther study is needed to confirm whether southern could potentially benefit from OSL dating. Part of Sweden was ice-free during this interstadial. High the challenge in formerly glaciated landscapes is elevation tree remains of Late Glacial age in cen- finding sediments that do not have too dim quartz tral Sweden have been rejected in favour of consis- or that do not suffer from incomplete bleaching. tent TCN dating results. High elevation areas de- For this reason I recommend a reconnaissance- glaciated coinciding approximately with the termi- level study be conducted prior to intensive study at nation of the Younger Dryas interval. Altogether selected sites where preferably there is some age- these results indicate dynamic behaviour for the control. The age of multiple samples from different SIS. Promising results for TCN and OSL dating lithofacies and over a vertical range of the sedi- were achieved for these sites and future research ments is arguably the best approach to addressing will benefit from application of these techniques to incomplete bleaching, and an approach I recom- decipher the history and dynamics of the SIS. mend for probably any stratigraphic study where

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

Acknowledgements och dynamik (hur isen ändras med tiden) på fyra platser i Sverige och Norge som också The summary chapter of the thesis was conceived representerar olika tidpunkter under den senaste and written by myself with feedback from Helena istiden. Jag har använt mig av två ganska nya Alexanderson, Jan Lundqvist, Lars Olsen, Mona metoder för åldersbestämning för att få hållpunkter Henriksen, and Stefan Wastegård. för nedisningshistorien: optiskt stimulerad Many people helped and inspired me over the luminiscensdatering (OSL) och kosmogen course of my doctoral studies. Foremost, I am very exponeringsdatering. thankful to my supervisor Helena Alexanderson, På flera platser i Skandinavien finns avlagringar and co-supervisors Arjen Stroeven and Jan från så kallade interstadialer, faser av den senaste Lundqvist – just simply excellent people to work istiden då det var mer eller mindre isfritt. with. Collaborations with Helena, Arjen, Jan, Lars Pilgrimstad i Jämtland, Sverige och Langsmoen i Olsen, Derek Fabel and Andrew Murray were en- Trøndelag, Norge är två sådana platser men vilka joyable and very fruitful. I was inspired by discus- tidpunkter representerar de? OSL-datering av nio sions with the above people and many others in- prover från avsättningarna i Pilgrimstad tyder på cluding: Jonas Bergman, Barbara Wohlfarth, Har- att det var isfritt där för ca 50 000 - 38 000 år ald Sveian, Hilary Birks, Ann-Marie Robertsson, sedan. Detta är betydligt senare än vad man Terri Lacourse, Robert Lagerbäck, Ingmar tidigare antagit. Utifrån analyser av pollen i Borgström, Martina Hättestrand, Clas Hättestrand, Pilgrimstad-avsättningarna har man trott att just Johan Kleman, Stefan Wastegård, Jan Risberg, den här isfria fasen ägde rum i början av den Sven Karlsson, Jakob Heyman, Jan-Pieter senaste istiden, för ca 90 000 - 75 000 år sedan. Buylaert, Dimitri Vandenberghe, Damian Steffen, Mina resultat placerar istället den här händelsen i and others. Daniel Veres, Jakob Heyman, Marie mitten av Weichsel-istiden och det betyder att den Koitsalu, Helena Alexanderson and Jonas Bergman skandinaviska inlandsisen då måste ha varit mycket assisted in the fieldwork and provided enjoyable liten och bara funnits i fjällen, eller kanske varit company. helt försvunnen. Klimatet var antagligen också I am grateful to the many organizations that fi- förhållandevis varmt. Resultat från andra nancially supported my research activities: Swed- undersökningar både i Skandinavien och i Finland ish Society for Anthropology, Gerard De Geer stödjer mina slutsatser. fund, Carl Mannerfelt fund, Royal Swedish Acad- Mina OSL-dateringar av avlagringarna i emy of Sciences, Lars Hiertas remembrance fund, Langsmoen bekräftar att delar av Norge var isfria Swedish Tourist Association, and the Geological för ca 22 000 år sedan, under den så kallade Survey of Sweden. Trofors-interstadialen. Detta är under en tid, det Last but not least, I am once again amazed by senaste nedisningsmaximat, då den skandinaviska the tremendous support and patience of my partner inlandsisen traditionellt ansetts ha varit som störst Tina who accompanied me on this journey. I love under de senaste dryga hundratusen åren. you so much. Exponeringsdatering av flyttblock ger My apologies if I neglected to acknowledge you tidpunkten för när den senaste inlandsisen smälte here. bort från ett område. Sex flyttblock från Vimmerbymoränen i Småland, södra Sverige ger samstämmiga resultat och visar att isen försvann Summary in Swedish därifrån för ca 14 000 år sedan. Detta stämmer bra med tidigare uppskattningar av tidpunkten för Den skandinaviska inlandsisen har ansetts vara isavsmältningen som baserats på kolfjortondatering stor, tjock och ganska stabil under stora delar av och lervarvskronologi. den senaste istiden (Weichselistiden, 117 000 - Tre flyttblock från toppen av Åreskutan i 11 700 år sedan). Undersökningar som gjorts de Jämtland ger likadana exponeringsåldrar och senaste åren ger däremot en helt ny bild av daterar isens avsmältning där till ca 11 000 år inlandsisen: en mer aktiv is som snabbt växlat i sedan, medan isen försvann lite tidigare längre storlek. I min doktorsavhandling har jag särskilt västerut på Snasahögarna, enligt studerat den skandinaviska inlandsisens historia

27 Timothy F. Johnsen

exponeringsdatering av tre flyttblock därifrån. Balco, G., Stone, J.O.H., Porter, S.C., Caffee, M.W. 2002. Cosmogenic nuclide ages for New England coastal mo- Detta är i överensstämmelse med tidigare resultat raines, Martha’s Vineyard and Cape Cod, Massachusetts, och visar att den skandinaviska inlandsisen snabbt USA. Quaternary Science Reviews, 21: 2127-2135. smälte bort och sjönk ihop under den senaste Balco, G., Stone, J.O.H., Lifton, N.A., Dunai, T.J. 2008. A isavsmältningen. Mina resultat innebär också att complete and easily accessible means of calculating surface exposure ages or erosion rates from 10Be and 26Al measure- man inte kan acceptera de ovanligt gamla ments. Quaternary Geochronology, 3: 174-195. kolfjortondateringar av trädrester från Åreskutan Ballantyne, C.K., Hall, A.M. 2008. The altitude of the last ice som gjorts tidigare. Träd på Åreskutan för upp till sheet in Caithness and east Sutherland, Northern Scotland. 17 000 år sedan går inte ihop med mina dateringar Scottish Journal of Geology, 44: 169-181. av isavsmältningen och med rekonstruktioner av is- Banerjee, D., Murray, A.S., Bøtter-Jensen, L., Lang, A. 2001. och vegetationsutbredning. Equivalent dose estimation using a single aliquot of polymineral fine grains. Radiation Measurements, 33: 73-94. Sammantaget visar min forskning, som utförts i olika områden, för olika tidsperioder och med olika Baumann, K.H., Lackschewitz, K.S., Mangerud, J., Spielha- gen, R.F., Wolf-welling, T.C.W., Henrich, R., Kassens, H. metoder, att den skandinaviska inlandsisen var 1995. Reflection of Scandinavian ice sheet fluctuations in mycket dynamisk och känslig för Norwegian Sea sediments during the past 150,000 years. miljöförändringar. Detta leder till en omvärdering Quaternary Research, 43: 185-197. av vår uppfattning av hur forna och nutida istäcken Bentley, C.R. 1987. Antarctic ice streams: a review. Journal of Geophysical Research, 92: 8843-8858. reagerar, och av vår förståelse för sambanden mellan istäcken och olika miljöförändringar. 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. References 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- Agrell, H., Friberg, N., Oppgården, R. 1976. The Vimmerby ther comment. Journal of Biogeography, 33: 376-378. line – an ice-margin zone in north-eastern Småland. Svensk Geografisk Årsbok, 52: 71-91. Björck, S., Möller, P. 1987. Late Weichselian Environmental History in Southeastern Sweden during the Deglaciation of Aitken, M.J. 1985. Thermoluminescence dating. Academic, the Scandinavian Ice Sheet. Quaternary Research, 28: 1-37. London. Bøe, A.G, Murray, A., Dahl, S.O. 2007. Resetting of sedi- Alexanderson, H., Murray, A. 2007. Was southern Sweden ice ments mobilised by the LGM ice-sheet in southern Norway. free at 19-25 ka, or were the post LGM glacifluvial sedi- Quaternary Geochronology, 2: 222-228. ments incompletely bleached? Quaternary Geochronology, 2: 229-236. Borgström, I. 1989. Terrängformerna och den glaciala utvecklingen i södra fjällen. Meddelande från Naturgeografiska Alexanderson, H., Murray. A.S. 2009. Problems and potential institutionen vid Stockholms universitet, Nr A 234, 133 p. of OSL dating Weichselian and Holocene sediments in (Ph.D. thesis). Sweden. Quaternary Science Reviews, in press. Bøtter-Jensen, L., Bulur, E., Duller, G.A.T., Murray, A.S. Alley, R.B., Anandakrishnan, S., Dupont, T.K., Parizek, B.R. 2000. Advances in luminescence instrument systems. Radia- 2004. Ice streams--fast, and faster? Comptes Rendus Phy- tion Measurements, 32: 523-528. sique, 5: 723-734. Boulton G.S., Dongelmans P., Punkari M., Broadgate M. Andersen, B.G., Nydal, R., Wangen, O.P., Østmo, S.R. 1981. 2001. Palaeoglaciology of an ice sheet through a glacial cy- Weichsel before 15,000 years B.P. at Jæren – Karmøy in cle: the European ice sheet through the Weichselian. Qua- southwestern Norway. Boreas, 10: 297-314. ternary Science Reviews, 20: 591-625. Andersen, B.G., Sejrup, H.P., Kirkhus, L. 1983. Eemian and Briner, J.P., Swanson, T.W., Caffee, M. 2001. Late Pleisto- Weichselian deposits at Bø on Karmøy, SW Norway: A pre- cene cosmogenic Cl-36 glacial chronology of the south- liminary report. Norges geologiske undersøkelse, 380: 189- western Ahklun Mountains, Alaska. Quaternary Research, 201. 56: 148-154. Andersen, B.G., Lundqvist, J., Saarnisto, M. 1995. The Charbit, S., Ritz, C., Ramstein, G. 2002. Simulations of Younger Dryas margin of the Scandinavian ice sheet—an Northern Hemisphere ice-sheet retreat: sensitivity to physi- introduction. Quaternary International, 28: 145-146. cal mechanisms involved during the Last Deglaciation. Qua- Arnold, N.S., van Andel, T.H., Valen, V. 2002. Extent and ternary Science Reviews, 21: 243-265. Dynamics of the Scandinavian Ice Sheet during Oxygen Iso- Choi, J.H., Murray, A.S., Cheong, C.S., Hong, D.G., Chang, tope Stage 3 (65,000-25,000 yr B.P.). Quaternary Research, H.W. 2003. The resolution of stratigraphic inconsistency in 57: 38-48. the luminescence ages of marine terrace sediments from Ko- Bailey, R.M., Smith, B.W., Rhodes, E.J. 1997. Partial bleach- rea. Quaternary Science Reviews, 22: 1201-1206. ing and the decay form characteristics of quartz OSL. Radia- Choi, J.H., Duller, G.A.T., Wintle, A.G. 2006. Analysis of tion Measurements, 27: 123-136. quartz LM-OSL curves. Ancient TL, 24: 9-20.

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

Clark, P.U., Brook, E.J., Raisbeck, G.M., Yiou, F., Clark, J. Hättestrand, C., Stroeven, A. 2002. A preglacial landscape in 2003. Cosmogenic Be-10 ages of the Saglek Moraines, the centre of Fennoscandian glaciation: geomorphological Torngat Mountains. Labrador Geology, 31: 617-620. evidence of minimal Quaternary glacial erosion. Geomor- phology, 44: 127–143. Clark, P.U., Dyke, A.S., Shakun, J.D., Carlson, A.E., Clark, J., Wohlfarth, B., Mitrovica, J.X., Hostetler, S.W., McCabe, Heimdahl, J. 2006. Spår av en mellanpaleolitisk befolkning i A.M. 2009. The Last Glacial Maximum. Science 325: 710- Sverige? Förnyad granskning av Mårtensöfynden 1938 och 714. 1964. (Traces of a Middle Palaeolithic population in Swe- den? A renewed assessment of the finds from Mårtensön in Denton, G.H., Hughes, T.J. (eds.). 1981. The Last Great Ice 1938 and 1964.) Fornvännen 101. Stockholm. Sheets. Wiley, New York, NY. Helmens, K.F., Bos, J.A.A., Engels, S., van Meerbeeck, C.J., Duller, G.A.T. 2004. Luminescence dating of Quaternary Bohncke, S.J.P., Renssen, H., Heiri, O., Brooks, S.J., Seppa, sediments: recent advances. Journal of Quaternary Science, H., Birks, H.J.B., Wohlfarth, B. 2007a. Present-day tem- 19: 183-192. peratures in northern Scandinavia during the last glaciation. Duller, G.A.T. 2008. Single-grain optical dating of Quaternary Geology, 35: 987-990. sediments: why aliquot size matters in luminescence dating. Helmens, K.F., Johansson, P.W., Räsänen, M.E., Alexander- Boreas, 37: 589-612. son, H., Eskola, K.O. 2007b. Ice-free intervals at Sokli con- Ehlers, J., Eissmann, L., Lippstreu, L., Stephan, H.J., Wansa, tinuing into Marine Isotope Stage 3 in the central area of the S. 2004. Pleistocene glaciations of North Germany. In: Scandinavian glaciations. Geological Society of Finland Ehlers, J., Gibbard, P.L. (eds.), Quaternary Glaciations – Bulletin, 79: 17-39. Extent and chronology, 135-146. Elsevier, Amsterdam. Helmut, R., Skvarca, P., Nagler, T. 1996. Rapid Collapse of Fabel, D., Harbor, J. 1999. The use of in-situ produced cos- Northern Larsen Ice Shelf, Antarctica. Science, 271: 788- mogenic radionuclides in glaciology and glacial geomor- 792. phology. Annals of Glaciology, 28: 103-110. Hewitt, G. 2000. The genetic legacy of the Quaternary ice Fabel, D., Stroeven, A.P., Harbor, J., Kleman, J., Elmore, D., ages. Nature, 405: 907-913. Fink, D. 2002. Landscape preservation under Fennoscandian Heyman, J., Hättestrand, C. 2006. Morphology, distribution ice sheets determined from in situ produced 10Be and 26Al. and formation of relict marginal moraines in the Swedish Earth and Planetary Science Letters, 201: 397-406. mountains. Geografiska Annaler, 88: 253-265. Fabel, D., Fink, D., Fredin, O., Harbor, J., Land, M., Stroeven, Holmlund, P., Fastook, J. 1995. A time dependent glaciologi- A.P. 2006. Exposure ages from relict lateral moraines over- cal model of the Weichselian ice sheet. Quaternary Interna- ridden by the Fennoscandian ice sheet. Quaternary Re- tional, 27: 53-58. search, 65: 136-146. Houmark-Nielsen, M., Krüger, J., Kjær, K. 2005. De seneste Fairbanks, R.G. 1989. A 17 000-year glacio-eustatic sea-level 150.000 år i Danmark. Geovidden - Geologi og Geografi, 2: record: influence of glacial melting rates on the younger 1-20. dryas event and deep ocean circulation. Nature, 342: 637- 642. Houmark-Nielsen, M. 2010. Extent, age and dynamics of Marine Isotope Stage 3 glaciations in the southwestern Bal- Fredén, C. (ed.). 2002. Sveriges Nationalatlas: Berg och jord. tic Basin. Boreas, 39: 343-359. 3rd ed. 208 p. Houmark-Nielsen, M., Kjaer. K.H. 2003. Southwest Scandi- Frödin, G. 1954. De sista skedena av Centraljämtlands glaciala navia, 40-15 kyr BP: palaeogeography and environmental historia. Geographica, 24. change. Journal of Quaternary Science, 18: 769-786. Fuchs, M., Owen, L.A. 2008. Luminescence dating of glacial Hubbard, A., Bradwell, T., Golledge, N., Hall, A., Patton, H., and associated sediments: review, recommendations and Sugden, D., Cooper, R., Stoker, M. 2009. Dynamic cycles, future directions. Boreas, 37: 636-659. ice streams and their impact on the extent, chronology and García Ambrosiani, K. 1990. Pleistocene stratigraphy in cen- deglaciation of the British–Irish ice sheet. Quaternary Sci- tral and northern Sweden. University of Stockholm, De- ence Reviews, 28: 758-776. partment of Quaternary Research, Report 14. IPCC. 2007. Climate Change 2007: Mitigation of Climate Goehring, B.M., Brook, E. J., Linge, H., Raisbeck, G. M., Change. Contribution of Working Group III to the Fourth Yiou, F. 2008. Beryllium-10 exposure ages of erratic boul- Assessment Report of the Intergovernmental Panel on Cli- ders in southern Norway and implications for the history of mate Change. B. Metz, O.R. Davidson, P.R. Bosch, R. the Fennoscandian Ice Sheet. Quaternary Science Reviews, Dave, L.A. Meyer (eds.). Cambridge University Press, 27: 320-336. Cambridge, United Kingdom and New York, NY, USA., 851 p. Gosse, J.C., Phillips, F.M. 2001. Terrestrial in situ cosmogenic nuclides: theory and application. Quaternary Science Re- Jain, M., Murray, A.S., Bøtter-Jensen, L. 2003. Characterisa- views, 20: 1475-1560. tion of blue-light stimulated luminescence components in different quartz samples: implications for dose measure- GRIP (Greenland Ice-Core Project Members). 1993. Climate ments. Radiation Measurements, 37: 441-449. instability during the last interglacial period recorded in the GRIP ice core. Nature, 364, 203-207. Kessler, M.A., Anderson, R.S., Briner, J.P. 2008. Fjord inser- tion into continental margins driven by topographic steering Hallet, B., Putkonen, J. 1994. Surface dating of dynamic of ice. Nature Geoscience, 1: 365-369. landforms: young boulders on aging moraines. Science, 265: 937-940. Kjær, K.H., Lagerlund, E., Adrielsson, L., Thomas, P. J., Murray, A., Sandgren, P. 2006. The first independent chro-

29 Timothy F. Johnsen

nology for Middle and Late Weichselian sediments from Larsen, N.K, Linge, H., Hilkansson, L, Fabel, D. 2010. Rapid southern Sweden and the island of Bornholm. GFF, 128: disintegration of the marine-based parts of the Scandinavian 209-220. Ice sheet during the last deglaciation. 29th Nordic Geologi- cal Winter Meeting, Oslo, January, 2010. Abstract, p. 108. Kitis, G., Polymeris, G.S., Kiyak, N.G. 2007. Component- resolved thermal stability and recuperation study of the LM- Lian, O.B, Roberts, R.G. 2006. Dating the Quaternary: pro- OSL curves of four sedimentary quartz samples. Radiation gress in luminescence dating of sediments. Quaternary Sci- Measurements, 42: 1273-1279. ence Reviews, 25: 2449-2468. Kleman, J. 2008. Geomorphology: Where glaciers cut deep. Licciardi, J.M., Clark, P.U., Brook, E.J., Pierce, K.L., Kurz, Nature Geoscience, 1: 343-344. M.D., Elmore, D., Sharma, P. 2001. Cosmogenic He-3 and Be-10 chronologies of the late Pinedale northern Kleman, J., Stroeven, A.P. 1997. Preglacial surface remnants Yellowstone ice cap, Montana, USA. Geology, 29: 1095- and Quaternary glacial regimes in northwestern Sweden. 1098. Geomorphology 19: 35-54. Lokrantz, H., Sohlenius, G. 2006. Ice marginal fluctuations Kleman J., Hättestrand, C., Borgström I., Stroeven, A. 1997. during the Weichselian glaciation in Fennoscandia, a Fennoscandian paleoglaciology reconstructed using a glacial literature review. SKB technical report no. TR-06-36. geological inversion model. Journal of Glaciology, 43: 283- Swedish Nuclear Fuel and Waste Management Co., 299. Stockholm. Kleman, J., Stroeven, A.P., Lundqvist, J. 2008. Patterns of Longva, O., Thorsnes, T. 1997. Skagerak in the past and the Quaternary ice sheet erosion and deposition in Fennoscandia present - an integrated study of geology, chemistry, and a theoretical framework for explanation. hydrography and microfossil ecology. Norges geologiske Geomorphology, 97: 73-90. undersøkelse, Special Publication, 8: 1-100. Krabill, W., Abdalati, W., Frederick, E., Manizade, S., Martin, Lundqvist, G. 1964. Interglaciala avlagringar i Sverige. C., Sonntag, J., Swift, R., Thomas, R., Wright, W., Yungel, Sveriges geologiska undersökning C 600. 60 p. J. 2000. Greenland Ice Sheet: High-Elevation Balance and Peripheral Thinning. Science, 289: 428-430. Lundqvist, J. 1967. Submoräna sediment i Jämtlands län. Sveriges Geologiska Undersökning C 618. 267 p. Kulling, O. 1945. Om fynd av mammut vid Pilgrimstad i Jämtland. Sveriges Geologiska Undersökning C 473. 61 p. Lundqvist, J. 1973. Isavsmältningens förlopp i Jämtlands län. Sveriges Geologiska Undersökning C 681. 187 p. Kulling, O. 1967. Yttrande med anledning av J. Lundqvists föredrag ‘Submoräna sediment i Jämtland’. Geologiska Lundqvist, J. 1992. Glacial stratigraphy in Sweden. Geological Föreningens i Stockholm Förhandlingar, 89: 123-125. Survey of Finland, Special paper, 15: 43-59. Kullman, L. 2002. Boreal tree taxa in the central Scandes Lundqvist, J. 2002. Weichsel-istidens huvudfas. In: Fredén, C. during the Late-Glacial: implications for Late-Quaternary (ed.), Berg och jord. Sveriges Nationalatlas. p. 124-135. forest history. Journal of Biogeography, 29: 1117-1124. Lundqvist, J., Robertsson, A.-M. 2002. Istider och mellanis- Kullman, L. 2005. Reply - On the presence of late-glacial trees tider. In: Fredén, C. (ed.), Sveriges Nationalatlas: Berg och in the Scandes. Journal of Biogeography, 32: 1499-1500. jord. 3rd ed. p. 120-124. Kullman, L. 2006. Late-glacial trees from arctic coast to alpine Lundqvist, J., Wohlfarth, B. 2001. Timing and east–west tundra: response to Birks et al. 2005 and 2006. Journal of correlation of south Swedish ice marginal lines during the Biogeography, 33: 376-378. Late Weichselian. Quaternary Science Reviews, 20: 1127- 1148. Lagerbäck, R. 1988. The Veiki moraines in northern Sweden – widespread evidence of an Early Weichselian deglaciation. Lunkka, J.P., Murray, A., Korpela, K. 2008. Weichselian Boreas, 17: 469-486. sediment succession at Ruunaa, Finland, indicating a Mid- Weichselian ice-free interval in eastern Fennoscandia. Bo- Lal, D. 1991. Cosmic-ray labeling of erosion surfaces: in situ reas, 37: 234-244. nuclide production rates and erosion models. Earth and Planetary Science Letters, 104: 424-439. Malmberg Persson, K., Persson, M., Lindén A.G. 2007. Is- randstråket Vimmerbymoränen mellan Knivshult och Van- Lambeck, K., Purcell, A., Zhao, J., Svensson, N.-O. 2010. The stad i nordöstra Småland. Sveriges Geologiska Undersökn- Scandinavian Ice Sheet: from MIS 4 to the end of the Last ing Rapport 2007, 7. Glacial Maximum. Boreas, 39: 410-435. Manabe, S., Broccoli, A.J. 1985. The influence of continental Lambeck, K., Chappell, J. 2001. Sea level change through the ice sheets on the climate of an Ice Age. Journal of Geo- last glacial cycle. Science, 292: 679-686. physical Research, 90: 2167-2190. Lambeck, K., Smither, C., Johnston. P. 1998. Sea-level Mangerud, J. 1991. The last interglacial/glacial cycle in change, glacial rebound and mantle viscosity for northern Northern Europe. In: Shane, L.C.K., Cushing, E.J. (eds.), Europe. Geophysical Journal International, 134: 102-144. Quaternary Landscapes. University of Minnesota Press, Landvik, J.Y., Brook, E.J., Gualtieri, L., Raisbeck, G., Minneapolis, p. 38-75. Salvigsen, O., Yiou, F. 2003. Northwest Svalbard during the Mangerud, J. 2004. Ice sheet limits on Norway and the Nor- last glaciation: Ice-free areas existed. Geology, 31: 905-908. wegian continental shelf. In: Ehlers, J., Gibbard, P. (eds.), Larsen, E., Gulliksen, S., Lauritzen, S.-E., Lie, R., Løvlie, R., Quaternary Glaciations—Extent and Chronology. Vol. 1. Mangerud, J. 1987. Cave stratigraphy in western Norway: Europe. Elsevier, Amsterdam. p. 271-294. Multiple Weichselian glaciations and interstadial vertebrate fauna. Boreas, 16: 267-292.

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

Mangerud, J., Jansen, E., Landvik, J. 1996. Late Cenozoic and Late Weichselian in Norway. Norges Geologiske Un- history of the Scandinavian and Barents Sea ice sheets. dersøgelse Bulletin, 433: 54-55. Global and Planetary Change, 12: 11-26. Olsen, L. 2010. A buried late MIS 3 shoreline in northern Mangerud, J., Gulliksen, S., Larsen, E., Longva, O., Miller, Norway – implications for ice extent and volume. Norges G.H., Sejrup, H.-P., Sønstegaard, E. 1981. A Middle Weich- geologiske undersøkelse. Bulletin 450: 1-14. selian ice-free period in Western Norway: The Ålesund In- Olsen, L., Sveian, H., Bergstrøm, B., Selvik, S.F., Lauritzen, terstadial. Boreas, 10: 447-462. S.-E., Stokland, Ø, Grøsfjeld, K. 2001a. Methods and strati- Mangerud, J., Løvlie, R., Gulliksen, S., Hufthammer, A.-K., graphies used to reconstruct Mid- and Late Weichselian pa- Larsen, E., Valen, V. 2003. Paleomagnetic correlations be- laeoenvironmental and palaeoclimatic changes in Norway. tween Scandinavian Ice-Sheet fluctuations and Greenland Norges geologiske undersøkelse Bulletin, 438: 21-46. Dansgaard–Oeschger Events, 45,000–25,000 yrs B.P. Qua- Olsen, L., Van der Borg, K., Bergstrøm, B., Sveian, H., ternary Research, 59: 213-222. Lauritzen, S.-E., Hansen, G. 2001b. AMS radiocarbon dat- Mangerud, J., Gulliksen, S., Larsen, E. 2010. 14C-dated fluc- ing of glacigenic sediments with low organic carbon content tuations of the western flank of the Scandinavian Ice Sheet – an important tool for reconstructing the history of glacial 45–25 kyr BP compared with Bølling–Younger Dryas fluc- variations in Norway. Norwegian Journal of Geology, 81: tuations and Dansgaard–Oeschger events in Greenland. Bo- 59-92. reas, 39: 328-342. Olsen, L., Sveian, H., Van der Borg, K., Bergstrøm, B., Martinson, D.G., Pisias, N.G., Hays, J.D., Imbrie, J., Moore Broekmans, M. 2002. Rapid and rhythmic ice sheet fluctua- Jr., T.C., Shackleton, N.J. 1987. Age dating and the orbital tions in western Scandinavia 15-40 kya – a review. Polar theory of the ice ages: development of a high-resolution 0 to Research, 21: 235-242. 300,000-yr chronostratigraphy. Quaternary Research, 27: 1- Ottesen, D., Rise, L., Knies, J. Olsen, L., Henriksen, S. 2005. 29. The Vestfjorden-Trænadjupet palaeo-ice stream drainage Murray, A.S, Olley, J.M. 2002. Precision and accuracy in the system, mid-Norwegian continental shelf. Marine Geology, optically stimulated luminescence dating of sedimentary 218: 175- 189. quartz: a status review. Geochronometria, 21: 1-16. Paus, A., Velle, G., Larsen, J., Nesje, A., Lie, Ø. 2006. Late- Murray, A.S., Wintle, A.G. 2000. Luminescence dating of glacial nunataks in central Scandinavia: Biostratigraphical quartz using an improved single-aliquot regenerative-dose evidence for ice thickness from Lake Flåfattjønn, Tynset, protocol. Radiation Measurements, 32: 57-73. Norway. Quaternary Science Reviews, 25: 1228-1246. Murray, A.S., Wintle, A.G. 2003. The single aliquot regenera- Pavlov, P., Svendsen, J.I., Indrelid, S. 2001. Human presence tive dose protocol: potential for improvements in reliability. in the European Arctic nearly 40,000 years ago. Nature, Radiation Measurements, 37: 377-381. 413: 64-67. Murray, A.S., Marten, R., Johnson, A., Martin, P. 1987. Pettitt P., Niskanen, M. 2005. Neanderthals in Susiluola cave, Analysis for naturally occurring radionuclides at environ- Finland, during the last interglacial period? Fennoscandia mental concentrations by gamma spectrometry. Journal of archaeological, XXII: 79-87. Radioanalytical and Nuclear Chemistry Articles, 115: 263- Phillips, F.M., Zreda, M.G., Evenson, E.B., Hall, R.D., 288. Chadwick, O.A., Sharma, P. 1997. Cosmogenic Cl-36 and Näslund, J.-O., Wohlfarth, B. (eds.). 2008. Fennoscandian Be-10 ages of Quaternary glacial and fluvial deposits of the paleo-environment and ice sheet dynamics during Marine Wind River Range, Wyoming. Geological Society of Amer- Isotope Stage (MIS) 3. Swedish Nuclear Fuel and Waste ica Bulletin, 109: 1453-1463. Management Co, Report R-08-79, p. 1-52. Pietsch, T.J., Olley, J.M., Nanson, G.C. 2008. Fluvial transport Näslund J.-O., Rodhe L., Fastook J.L., Holmlund, P. 2003. as a natural luminescence sensitiser of quartz. Quaternary New ways of studying ice sheet flow directions and glacial Geochronology, 3, 365-376. erosion by computer modelling – examples from Fenno- Prescott, J.R., Hutton, J.T. 1994. Cosmic ray contributions to scandia. Quaternary Science Reviews, 22: 245-258. dose rates for luminescence and ESR dating: large depths Nesje, A., Dahl, S.O. 1990. Autochthonous block fields in and long-term time variations. Radiation Measurements 23, southern Norway: implications for the geometry, thickness, 497-500. and isostatic loading of the Late Weichselian ice sheet. Preusser, F., Chithambo, M.L., Götte, T., Martini, M., Ram- Journal of Quaternary Science, 5: 225-234. seyer, K., Sendezera, E.J., Susino, G.J., Wintle, A.G. 2009. Nesje, A., Dahl, S.O., Anda, E., Rye, N. 1988. Block fields in Quartz as a natural luminescence dosimeter. Earth-Science southern Norway: significance for the Late Weichselian ice Reviews, 97: 196-226. sheet. Norsk Geologisk Tidsskrift, 68: 149-169. Putkonen, J., Swanson, T. 2003. Accuracy of cosmogenic ages Nygård, A., Sejrup, H.P., Haflidason, H., Lekens, W.A.H., for moraines. Quaternary Research, 59: 255-261. Clark, C.D., Bigg, G.R. 2007. Extreme sediment and ice Rasmussen, S.O., Andersen, K.K., Svensson, A.M., Steffen- discharge from marine based ice streams; new evidence sen, J.P., Vinther, B.M., Clausen, H.B., Siggaard-Andersen, from the North Sea. Geology, 35: 395-398. M.L., Johnsen, S.J., Larsen, L.B., Dahl-Jensen, D., Bigler, Oerlemans, J., van der Veen, C.J. 1984. Ice sheets and climate. M., Rothlisberger, R., Fischer, H., Goto-Azuma, K., Hans- Reidel, Dordrecht. 217 p. son, M.E., Ruth, U. 2006. A new Greenland ice core chronology for the last glacial termination. Journal of Geophysi- Olsen, L. 1997. Rapid shifts in glacial extension characterise a cal Research-Atmospheres, 111: 1-16. new conceptual model for glacial variations during the Mid

31 Timothy F. Johnsen

Rinterknecht, V.R., Clark, P.U., Raisbeck, G.M. , Yiou, F., Stokes, C. R, Clark, C.D. 2001. Palaeo-ice streams. Quater- Bitinas, A., Brook, E., Marks, J.L., Zelcs, V., Lunkka, J.-P., nary Science Reviews, 20: 1437-1457. Pavlovskaya, I.E., Piotrowski, J.A., Raukas, A. 2006. The Svendsen, J.I, Alexanderson, H., Astakhov, V.I., Demidov, I., last deglaciation of southeastern sector of the Scandinavian Dowdeswell, J.A., Funder, S., Gataullin, V., Henriksen, M., Ice Sheet. Science, 311: 1449-1452. Hjort, C., Houmark-Nielsen, M., Hubberten, H.W., Robertsson, A.-M. 1988a. Biostratigraphical studies of inter- Ingólfsson, Ó., Jakobsson, M., Kjær, K.H., Larsen, E., Lok- glacial and interstadial deposits in Sweden. Ph.D. thesis, rantz, H., Lunkka, J.-P., Lyså, A., Mangerud, J., Matiouch- Department of Quaternary Research Report 10, University kov, A., Murray, A., Möller, P., Niessen, F., Nikolskaya, O., of Stockholm. Polyak, L., Saarnisto, M., Siegert, C., Siegert, M. J., Spiel- hagen, R.F., Stein, R. 2004. Late Quaternary ice sheet his- Robertsson, A.-M. 1988b. A reinvestigation of the biostrati- tory of northern Eurasia. Quaternary Science Reviews, 23: graphy of the till-covered sediments at Pilgrimstad and Öje, 1229–1271. central Sweden. In: Robertsson, A.-M. (ed.), Biostrati- graphical studies of interglacial and interstadial deposits in Thrasher, I.M., Mauz, B., Chiverrell, R.C., Lang, A. 2009. Sweden. p. 1-53. University of Stockholm, Department of Luminescence dating of glaciofluvial deposits: A review. Quaternary Research, Stockholm. Earth-Science Reviews, 97: 145-158. Robertsson, A.-M. 1994. Vegetational history. In: Fredén, C. Tsukamoto, S., Rink, W.J., Watanuki, T. 2003. OSL of tephric (ed.), Geology, National Atlas of Sweden. Almqvist & Wik- loess and volcanic quartz in Japan and an alternative proce- sell International, Stockholm, 208 p. dure for estimating De from a fast OSL component. Radia- tion Measurements, 37: 459-465. Robertsson, A.-M., García Ambrosiani, K. 1992. The Pleisto- cene in Sweden – a review of research, 1960–1990. Sveriges Ukkonen, P., Arppe, L., Houmark-Nielsen, M., Kjær, K.H., Geologiska Undersökning Ca 81, 299-306. Karhu, J.A. 2007. MIS 3 mammoth remains from Sweden - implications for faunal history, palaeoclimate and glaciation Rott, H., Skvarca, P., Nagler, T., 1996. Rapid Collapse of chronology. Quaternary Science Reviews, 26: 3081-3098. Northern Larsen Ice Shelf, Antarctica. Science, 271: 788- 792. Valen, V., Larsen, E., Mangerud, J., Hufthammer, A.K. 1996. Sedimentology and stratigraphy in the cave Hamnsundhel- Ruddiman, W.F. 2003. Orbital insolation, ice volume, and leren, western Norway. Journal of Quaternary Science, 11: greenhouse gases. Quaternary Science Reviews, 22: 1597- 185-201. 1629. Vorren, T.O., Laberg, J.S. 1997. Trough Mouth Fans – Pa- Salonen, V.-P., Kaakinen, A., Kultti, S., Miettinen, A., Eskola, laeoclimate and ice-sheet Monitors. Quaternary Science Re- K.O., Lunkka, J.P. 2008. Middle Weichselian glacial event views, 16: 865-882. in the central part of the Scandinavian Ice Sheet recorded in the Hitura pit, Ostrobothnia, Finland. Boreas, 37: 38-54. Vorren T.O., Mangerud J. 2007. Istider kommer og går. In: Ramberg I.B., Bryhni I., Nøttvedt A. (eds.), Landet blir til. Schulz, H.-P., Eriksson, B., Hirvas, H., Huhta, P., Jungner, H., 2nd ed. Norsk Geologisk Forening, Trondheim. p. 478-531. Purhonen, P., Ukkonen, P., Rankama, T. 2002. Excavations at Susiluola Cave. Suomen Museo, 109: 5-45. Vorren, T.O., Vorren, K.-D., Alm, T., Gulliksen, S., Løvlie, R. 1988. The last deglaciation (20,000 to 11,000 B.P.) on Sejrup, H.P., Haflidason, H., Aarseth, I., King, E., Forsberg, Andøya, northern Norway. Boreas, 17: 41-77. C.F., Long, D., Rokoengen, K. 1994. Late Weichselian glaciation history of the northern North Sea. Boreas, 23: 1-13. Wintle, A.G. 2008. Luminescence dating: where it has been and where it is going. Boreas, 37: 471-482. Sejrup, H.P., Larsen, E., Landvik, J., King, E.L., Haflidason, H., Nesje, A. 2000. Quaternary glaciations in southern Fen- Wintle, A.G., Murray, A.S. 2006. A review of quartz optically noscandia: evidence from southwestern Norway and the stimulated luminescence characteristics and their relevance northern North Sea region. Quaternary Science Reviews, 19: in single-aliquot regeneration dating protocols. Radiation 667-685. Measurements, 41: 369-391. Shepherd, A., Wingham, D. 2007. Recent Sea-Level Contribu- Wohlfarth, B. 2010. Ice-free conditions in Sweden during tions of the Antarctic and Greenland Ice Sheets. Science, Marine Oxygen Isotope Stage 3? Boreas, 39: 377-398. 315,: 1529-1532. Wohlfarth, B., Björck, S., Possnert, G., Holmquist, B. 1998. A Siegert, M.J., Dowdeswell, J.A., Hald, M., Svendsen, J.-I. 800-year long, radiocarbon-dated varve chronology from 2001. Modelling the Eurasian Ice Sheet through a full south-eastern Sweden. Boreas, 27: 243-257. (Weichselian) glacial cycle. Global and Planetary Change, Zreda, M., Phillips, F.M., Elmore, D. 1994. Cosmogenic 36Cl 31: 367-385. accumulation in unstable landforms 2. Simulations and Singarayer, J.S., Bailey, R.M. 2003. Further investigations of measurements on eroding surfaces. Water Resources Re- the quartz optically stimulated luminescence components search, 30: 3127-3136. 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.

32 Paper I

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

BY TIMOTHY F. JOHNSEN1, HELENA ALEXANDERSON1,2, DEREK FABEL3 AND STEWART P.H.T. FREEMAN4

1Department of Physical Geography and Quaternary Geology, Stockholm University, Sweden 2 Department of Plant and Environmental Sciences, Norwegian University of Life Sciences, Ås, Norway 3Department of Geographical and Earth Sciences, University of Glasgow, UK 4SUERC-AMS, Scottish Universities Environmental Research Centre, East Kilbride, UK

Johnsen, T.F., Alexanderson, H., Fabel, D. and Freeman, former ice sheets. The deglaciation pattern of the 10 S.P.H.T., 2009: New Be cosmogenic ages from the Vimmerby Scandinavian Ice Sheet is generally well recon- moraine confirm the timing of Scandinavian Ice Sheet deglacia tion in southern Sweden. Geogr. Ann. 91 A (2): 113 120 structed, but the absolute timing and the dynamics of the retreating ice margin are less well known ABSTRACT. The overall pattern of deglaciation of (Lundqvist and Wohlfarth 2001). Several recent the southern part of the Scandinavian Ice Sheet has studies have shown that the Late Weichselian gla- been considered established, although details of cial and deglacial history may be more complicat- the chronology and ice sheet dynamics are less well ed than generally believed. For example, the west- known. Even less is known for the south Swedish Upland because the area was deglaciated mostly by ern margin of the ice sheet was very dynamic with stagnation. Within this area lies the conspicuous multiple ice-free periods during the last 40 ka in- Vimmerby moraine, for which we have used the ter- cluding around the Last Glacial Maximum restrial cosmogenic nuclide (10Be) exposure dating (LGM) (Olsen et al. 2001a,b, 2002); large mo- technique to derive the exposure age of six glacially raine systems from the southeast portion of the ice transported boulders. The six 10Be cosmogenic ages are internally consistent, ranging from 14.9 ± sheet may be younger than previous estimates 1.5 to 12.4 ± 1.3 ka with a mean of 13.6 ± 0.9 ka. Ad- (Rinterknecht et al. 2006); there were possibly justing for the effects of surface erosion, snow bur- ice-free conditions around the LGM in southern ial and glacio-isostatic rebound causes the mean Sweden (Alexanderson and Murray 2007) and age to increase only by c. 6% to c. 14.4 ± 0.9 ka. The 10 southern Norway (Bøe et al. 2007); and tree Be derived age for the Vimmerby moraine is in mega-fossils dated from high elevation areas in agreement with previous estimates for the timing of deglaciation based on radiocarbon dating and central Sweden suggest ice-free conditions as ear- varve chronology. This result shows promise for ly as 17 ka cal. ka BP (Kullman 2002). further terrestrial cosmogenic nuclide exposure In southern Sweden, the deglaciation history is studies in southern Sweden. best known in the coastal areas where the Late Weichselian ice margin actively retreated and can 10 Key words: terrestrial cosmogenic nuclide ( Be) exposure dating, be traced by conspicuous moraines (west coast) or deglaciation, Scandinavian Ice Sheet, Vimmerby moraine, Swe den, south Swedish Upland by patterns of varved clay deposition (east coast). Above the highest Late Weichselian coastline is the south Swedish Upland, an area with a poor degla- Introduction cial chronology. The correlation between the west Accurate reconstructions of past ice sheets are and east coasts, across the south Swedish Upland, needed to better understand their contributions to is problematic due to a lack of continuous geo- changes in climate, sea level, and solid Earth geological and geomorphological evidence and be- physics. Ice sheet models play a central role in this cause the chronologies are based on different tech- effort but are too frequently poorly constrained by niques. ﬁeld data, especially for the interior areas of The Vimmerby moraine (Agrell et al. 1976), is

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 deglaciation pattern. (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 30 1730/2006)

© The authors 2009 114 Journal compilation © 2009 Swedish Society for Anthropology and Geography NEW COSMOGENIC AGES FROM THE VIMMERBY MORAINE a discontinuous ice-marginal zone, at least 100- still-stand and/or readvance in the general ice-mar- km-long, lying in the eastern part of the upland gin retreat during the last deglaciation (e.g. Malm- (Fig. 1); it is a distinctive feature since ice-marginal berg Persson 2001). According to the most current features reflecting active ice are very scarce in this accepted deglaciation model for southern Sweden, part of the upland. New mapping of the moraine which uses a combination of calibrated radiocar- shows that it strikes across the tentative ice-mar- bon ages, partly radiocarbon-dated clay-varve ginal lines (isochrones) of the standard deglacia- chronology, and geomorphology (Lundqvist and tion model, and this indicates that the deglaciation Wohlfarth 2001), this still-stand and/or readvance pattern emerging from this new mapping is differ- happened between 14.0 and 13.3 cal. ka BP, roughly ent from the deglaciation pattern previously as- corresponding to Greenland Interstadial 1 (Björck sumed (Lundqvist and Wohlfarth 2001; Lundqvist et al. 1998). A maximum age for deglaciation in the 2002). Thus, dating of this feature would efficiently study area is provided by a well dated site 200 km fill a gap in our knowledge of the deglacial history south, where AMS radiocarbon dating of leaf frag- in the south Swedish Upland and would be useful ments and twigs gives ages for early vegetation es- to compare to other deglaciation chronologies from tablishment and a minimum age of deglaciation the region. there around 15.1–14.4 cal. ka BP (Davies et al. We have used and present results of one of the 2004). A minimum deglaciation age for the study first applications of the terrestrial cosmogenic nu- area, c. 10.0 cal. ka BP (Lindén 1999; calibrated by clide (10Be) exposure dating technique in southern us using OxCal v4.0.5 and IntCal04 atmospheric Sweden with the aim of improving our understand- curve; Ramsey 2007; Reimer et al. 2004), is given ing of the chronological history of the decay of the by a radiocarbon date of early organic sedimenta- Scandinavian Ice Sheet, and for comparison to ra- tion in a kettle hole within the moraine. diocarbon and varve chronology. Methodology Study area and previous research Sampling The south Swedish Upland (57–58˚N, 13–16˚E) is The accumulation of in situ produced terrestrial the highest area in southernmost Sweden and is sit- cosmogenic 10Be in quartz exposed to cosmic radi- uated 200–300 m above present sea level (Fig. 1). ation provides a means of determining the amount The crystalline bedrock forms an undulating land- of time the rock has been at or near the ground sur- scape with isolated inselbergs, valleys and deep- face (Lal 1991; Gosse and Phillips 2001). This weathered bedrock. The study area in the eastern technique has proven useful in numerous studies of part of the upland (Fig. 1a) is situated above the deglacial histories and landform preservation (e.g. Late Weichselian highest shoreline and the surface Balco et al. 2002; Clark et al. 2003; Fabel et al. cover is dominated by till (cover moraine, drum- 2002; Licciardi et al. 2001; Phillips et al. 1997; lins, hummocky moraine), glaciofluvial deposits Rinterknecht et al. 2006). Unlike the radiocarbon (valley fills, deltas) and peat (Fig. 1b). The common dating technique that gives the age of events fol- occurrence of hummocky moraine in parts of the lowing deglaciation (e.g. migration and establish- south Swedish Upland indicates widespread stag- ment of vegetation followed by deposition and nation (dead ice) instead of active retreat (e.g. preservation of plant remains in a basin), the 10Be- Björck and Möller 1987); stagnation discourages dating technique can give a direct age of deglacia- the formation of end moraines. The Vimmerby mo- tion. raine (Agrell et al. 1976; Lindén 1984; Malmberg In this study, we collected quartz-rich samples Persson 2001; Persson 2001; Malmberg Persson et from glacially transported granitic and quartzitic al. 2007) is thus an exception in the area. It consists boulders on features belonging to the Vimmerby of small end moraines and partly till-covered ice- moraine: an end moraine at Vimmerby and a partly marginal glaciofluvial deposits, and separates an till-covered marginal delta at Lannaskede (Figs 1b, ice-proximal landscape with thicker till cover from 2; Table 1). To minimize the risk of processes mod- one with thin and discontinuous till. Distally, san- ifying the exposure history of the boulder, such as durs fill the river valleys down to the highest coast- cosmogenic nuclide inheritance (e.g. Briner et al. line (115–130 m a.s.l.; Malmberg Persson et al. 2001), boulder exhumation, erosion or moraine de- 2007). formation (Hallet and Putkonen 1994; Putkonen The Vimmerby moraine is believed to reflect a and Swanson 2003; Zreda et al. 1994), we sampled

Fig. 2. Sampled boulders at the till covered ice marginal delta at Lannaskede and at an end moraine close to Vimmerby. Both sites are part of the Vimmerby moraine the tops of large (0.9–2.3 m b-axis) rounded to sub- Measurements and calculations rounded, weathering-resistant (granitic and quartz- All samples were processed for 10Be from quartz itic) boulders. Boulders were resting on top of sta- following procedures based on methods modiﬁed ble and level surfaces or on the broad crest of the from Kohl and Nishiizumi (1992) and Child et al. moraine. In total six boulders were processed in the (2000). Approximately 20 g of pure quartz was Glasgow University–SUERC cosmogenic nuclide separated from each sample, puriﬁed, spiked with laboratory. c. 0.25 mg 9Be carrier, dissolved, separated by ion

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

Altitude Shielding Thickness* [10Be]† Exposure Sample 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). Uncertainties propagated 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) as suming no prior exposure and no erosion during exposure. The quoted values are for the ‘Lm’ scaling scheme which includes palaeo magnetic corrections (Balco et al. 2008). Uncertainties are ±1σ (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.

chromatography, selectively precipitated as hy- variation in the initial conditions of the population droxides, and oxidized. AMS measurements were of boulders delivered to the site, for example, cos- carried out at the SUERC AMS Facility. Measured mogenic nuclide inheritance (e.g. Briner et al. 10Be/9Be ratios were corrected by full chemistry 2001); (3) the weathering and erosion of the rock procedural blanks with 10Be/9Be of <3 × 10 15. In- surface during exposure that removes in situ pro- dependent measurements of AMS samples were duced terrestrial cosmogenic nuclides from the combined as weighted means with the larger of the sampled surface; (4) the partial shielding of the total statistical error or mean standard error. We cal- rock surface from cosmic rays by seasonal snow culated the analytical uncertainty by assuming that cover; (5) the partial shielding of the rock surface the uncertainties in AMS measurement and Be car- from cosmic rays by vegetation or by burial under rier are normal and independent, adding them in water; or (6) the changing elevation of the rock sur- quadrate in the usual fashion (e.g. Bevington and face due to glacio-isostatic movement (Gosse and Robinson, 1992). The resulting analytical uncer- Phillips 2001). Apart from inheritance, all these tainties range from 5 to 8% (Table 1). All 10Be con- factors cause 10Be ages to appear too young and centrations were converted to exposure ages by us- without adjusting for them 10Be ages will be mini- ing a production rate linked to a calibration data set mum ages. On the other hand, if inheritance is dom- using a 10Be half-life of 1.5 Ma. inating, 10Be ages will give maximum ages. Measured 10Be concentrations were converted Glacio-isostatic rebound changes the elevation to surface exposure ages using the CRONUS-Earth of the sample site during exposure, which in the 10Be–26Al exposure age calculator version 2 (http:/ case of uplift will cause air pressure to decrease /hess.ess.washington.edu), assuming no prior ex- over time. This means that the cosmic ray flux will posure and no erosion during exposure. The results vary through time. Usually this can be corrected for for the different 10Be production rate scaling by using a nearby relative sea-level curve and inte- schemes used by the online calculator yielded ages grating the changes in 10Be production related to that vary by less than 2%. The quoted surface ex- relative sea-level changes over time. However, the posure ages (Table 1) are for the ‘Lm’ scaling nearest sea-level curves are from the east coast of scheme which includes palaeomagnetic correc- Sweden where before 9.5 ka cal. ka BP water levels tions (Balco et al. 2008). were influenced by the isolated Baltic Ice Lake and Ancylus Lake. As air pressure changes were likely more influenced by changes in sea level than local Physical factors influencing 10Be ages lake levels, the portion of the water level curve prior Various physical factors can affect the accuracy of to 9.5 ka cal. ka BP does not accurately reflect the the exposure age calculations such as: (1) moraine changes in air pressure. Fortunately, detailed mod- degradation (Putkonen and Swanson, 2003); (2) elling of shoreline displacements in south-central

Sweden and the evolution of the Baltic Sea since six boulder samples of c. 14.4 ± 0.9 ka (an increase the LGM has been completed (Lambeck et al. of c. 6% from the unadjusted exposure age). 1998). As part of this work water-level curves were produced showing both the inclusion and exclusion of ice and land damming. In other words, the water- Comparisons with other data from the area level curve that excludes the effect of damming is The mean apparent cosmogenic exposure age of the effectively the sea-level curve that should represent six boulders of 13.6 ± 0.9 ka is within the previously changes in air pressure. Therefore we used the estimated deglaciation age range for the area of nearest modelled water level curve from Oskars- 14.0–13.3 cal. ka BP (Lundqvist and Wohlfarth 2001 hamn, c. 55 km south of Vimmerby. The highest and references therein) and within uncertainties to coastline using the Oskarshamn water-level curve the 10Be age adjusted for the effects of erosion, is 130 m and is similar to the highest coastline for snow and glacio-isostatic rebound of 14.4 ± 0.9 ka. the Vimmerby moraine at 115–130 m (Malmberg Thus, given the similarity between the results of the Persson et al. 2007). Since the highest coastline el- different approaches used (radiocarbon dating, evations are similar between these two sites we did varve chronology, and now 10Be dating), we are not apply a scaling factor. confident that the deglacial age for the Vimmerby moraine is c. 14 ka and possibly slightly older if we consider the adjusted mean 10Be age. This result Results and discussion also demonstrates that the 10Be dating approach Glacially transported boulder 10Be ages works well in this area for boulders from moraines The six apparent exposure ages range from 14.9 ± and till surfaces, and may be a useful technique for 1.5 to 12.4 ± 1.3 ka with a mean of 13.6 ± 0.9 ka other areas in southern Sweden. It is not yet under- (with uncertainty at 1σ standard deviation of the stood why optically stimulated luminescence ages ages; Table 1). of glaciofluvial sediments associated with the Vim- Boulder exhumation and cosmogenic nuclide merby moraine are many thousands of years older inheritance do not appear to be significant issues (Alexanderson and Murray 2007, submitted). since our ages were consistent for boulders from Confident correlation of the Vimmerby moraine stable and level surfaces and from the broad crest to moraines in the west half of southern Sweden re- of the moraine. There were also no indications in mains problematic because (1) there is not a large the field of significant erosion or weathering of the difference in the ages of moraines as the overall rate boulder surfaces. If we assume a reasonable boul- of deglaciation was relatively fast compared to the der-surface constant erosion rate of 1 mm/ka, the resolution of the dating methods, and (2) at Lake mean apparent exposure age increases by c. 1% Vättern located in central southern Sweden, ice- (Table 1). A medium density (0.3 g/cm3) snow cov- marginal positions for different time periods were in er of 0.3 m thickness on top of the boulders for four similar positions (Lundqvist and Wohlfarth 2001). months a year would similarly increase the mean Thus, the Vimmerby moraine may correlate with the exposure age by only 1% (Wastenson 1995; Gosse Trollhättan moraine, or with the Berghem or Levene and Phillips 2001). The effect of shielding by the moraines (Fig. 1a; Lundqvist and Wohlfarth 2001). burial from water is not considered as both sites are Tracing the ice margin east of the Vimmerby mo- above the Late Weichselian highest coastline and raine is more complicated because (1) ice positions above any local lakes. Also, the shielding effect within the Baltic Sea basin are not well known, and from vegetation is less than 1% and so is not con- (2) the rate of deglaciation in the southeast portion of sidered (Plug et al. 2007). the ice sheet (northern Poland and the Baltic states) The integrated production rate from isostatic re- was also relatively fast compared to the resolution of bound accounts for a c. 4% increase in the exposure the dating methods (Rinterknecht et al. 2006). age within the study area. Another factor that affects air pressure, in addition to elevation changes, is the presence of the nearby ice sheet (Staiger et al. Conclusion 2007); however, since the southern margin of the 10Be ages for the Vimmerby ice-marginal zone of Scandinavian ice sheet retreated rapidly this effect 13.6 ± 0.8 (14.4 ± 0.9 adjusted) ka are in agreement is unlikely to have persisted for long enough to af- with previous estimates for the timing of deglacia- fect the calculated exposure ages. Altogether, these tion based on radiocarbon dating and varve chro- effects give an adjusted mean exposure age for the nology. Thus, the southern margin of the Scandi-

© The authors 2009 118 Journal compilation © 2009 Swedish Society for Anthropology and Geography NEW COSMOGENIC AGES FROM THE VIMMERBY MORAINE navian Ice Sheet was at the Vimmerby moraine c. Derek Fabel, Department of Geographical and 14 ka ago. It is not clear which of the moraines west Earth Sciences, East Quadrangle, Main Building, of the study area correlate with the Vimmerby mo- University of Glasgow, Glasgow, G12 8QQ, UK raine, and even less clear for moraines from the E-mail: [email protected] southeast portion of the ice sheet (northern Poland and the Baltic states). Nevertheless, the internal Stewart P.H.T. Freeman, SUERC-AMS, Scottish consistency of the six 10Be ages and their compat- Universities Environmental Research Centre, East ibility with previous radiocarbon ages and varve Kilbride, G75 0QF, UK chronology indicate that the terrestrial cosmogenic nuclide (10Be) dating technique works well for erratic boulders within this area. This result shows References promise for further terrestrial cosmogenic nuclide Agrell, H., Friberg, N. and Oppgården, R., 1976: The Vimmerby exposure studies in southern Sweden. Future re- line an ice margin zone in north eastern Småland. Svensk Geografisk Årsbok, 52: 71 91. search efforts should focus on more detailed map- Alexanderson, H. and Murray, A., 2007: Was southern Sweden ping of the various moraine systems in southern ice free at 19 25 ka, or were the post LGM glacifluvial sedi Sweden, and employ an integrated dating approach ments incompletely bleached? Quaternary Geochronology, (e.g. radiocarbon dating of the first establishment 2: 229 236. 10 Alexanderson, H. and Murray, A., submitted: Why doesn’t OSL of vegetation, Be dating of erratic boulders, and work well for deglacial sediments in Sweden? Quaternary optically stimulated luminescence dating of gla- Geochronology ciofluvial and other sediments). 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. Qua ternary Science Reviews, 21: 2127 2135. Acknowledgements Balco, G., Stone, J.O.H., Lifton, N.A. and Dunai, T.J., 2008: A We thank Maria Miguens-Rodriguez and Henriette complete and easily accessible means of calculating surface 10 26 Linge for chemistry and preparation of AMS tar- exposure ages or erosion rates from Be and Al measure ments. Quaternary Geochronology, 3: 174 195. gets. Jan Lundqvist and Arjen Stroeven provided Bevington, P. R. and Robinson, D. K., 1992: Data reduction and fruitful discussions, and Jakob Heyman assisted in error analysis for the physical sciences. McGraw Hill, New the fieldwork. Two anonymous reviewers and Hil- York. 328 p. dred Crill helped improve the manuscript. Funding Björck, S. and Möller, P., 1987: Late Weichselian Environmental History in Southeastern Sweden during the Deglaciation of was provided by a grant from the Geological Sur- the Scandinavian Ice Sheet. Quaternary Research, 28: 1 37. vey of Sweden (no. 60–1356/2005). Björck, S., Walker, M.J.C., Cwynar, L., Johnsen, S., Knudsen, The contributions by the co-authors included: K.L., Lowe, J.J., Wohlfarth, B. and INTIMATE members, project conception by Alexanderson, boulder se- 1998: An event stratigraphy for the last termination in the North Atlantic region based on the Greenland ice core record: lection and sampling by Johnsen and Alexander- a proposal by the INTIMATE group. Journal of Quaternary son, samples crushed by Alexanderson and Science, 13: 283 292. Johnsen and chemically processed by Fabel, AMS Bøe, A.G., Murray, A.S. and Dahl, S.O., 2007: Resetting of sedi measurements by Freeman, age calculations by ments mobilised by the LGM ice sheet in southern Norway. Quaternary Geochronology, 2: 222 228. doi: 10.1016/j.qua Fabel and Johnsen, interpretations by Johnsen, geo.2006.05.031 Fabel and Alexanderson, and writing mostly by Briner, J.P., Swanson, T.W. and Caffee, M., 2001: Late Pleis Johnsen with input from Alexanderson and Fabel. tocene cosmogenic Cl 36 glacial chronology of the south western Ahklun Mountains, Alaska. Quaternary Research, 56: 148 154. Timothy F. Johnsen, Department of Physical Geo- Child, D., Elliott, G., Mifsud, C., Smith, A.M. and Fink, D., 2000: graphy and Quaternary Geology, Stockholm Uni- Sample processing for earth science studies at ANTARES. versity, SE-10691 Stockholm, Sweden Nuclear Instruments and Methods in Physics Research Sec E-mail: [email protected] tion B, Beam Interactions with Materials and Atoms, 172: 856 860. Clark, P.U., Brook, E.J., Raisbeck, G.M., Yiou, F. and Clark, J., Helena Alexanderson, Department of Physical 2003: Cosmogenic Be 10 ages of the Saglek Moraines, Torn Geography and Quaternary Geology, Stockholm gat Mountains. Labrador Geology, 31: 617 620. University, SE-10691 Stockholm, Sweden and De- Davies, S.M., Wohlfarth, B., Wastegård, S., Andersson, M., Block ley, S. and Possnert, G., 2004: Were there two Borrobol Te partment of Plant and Environmental Sciences, phras during the early Lateglacial period: implications for te Norwegian University of Life Sciences, phrochronology? Quaternary Science Reviews, 23: 581 589. P.O. Box 5003, N-1432 Ås, Norway Fabel, D., Stroeven, A.P., Harbor, J., Kleman, J., Elmore, D. and E-mail: [email protected] Fink, D., 2002: Landscape preservation under Fennoscandian

ice sheets determined from in situ produced 10Be and 26Al. S. E. and Hansen, G., 2001b: AMS radiocarbon dating of gla Earth and Planetary Science Letters, 201: 397 406. cigenic sediments with low organic carbon content an im Gosse, J.C. and Phillips, F.M., 2001: Terrestrial in situ cos portant tool for reconstructing the history of glacial variations mogenic nuclides: theory and application. Quaternary Sci in Norway. Norwegian Journal of Geology, 81: 59 92. ence Reviews, 20: 1475 1560. Olsen, L., Sveian, H., Van der Borg, K., Bergstrøm, B. and Broek Hallet, B. and Putkonen, J., 1994: Surface dating of dynamic land mans, M., 2002: Rapid and rhythmic ice sheet fluctuations in forms: young boulders on aging moraines. Science, 265: 937 western Scandinavia 15 40 kya a review. Polar Research, 940. 21: 235 242. Kohl, C.P. and Nishiizumi, K., 1992: Chemical isolation of quartz Persson, M., 2001: Beskrivning till jordartskartan 6F Vetlanda for measurement of in situ produced cosmogenic nuclides, SV. (Description of the Quaternary map). Sveriges Geologis Geochimica Cosmochimica ACTA, 56: 3586 3587. ka Undersökning Ae 147. Kullman, L., 2002: Boreal tree taxa in the central Scandes during Phillips, F.M., Zreda, M.G., Evenson, E.B., Hall, R.D., Chadwick, the Late Glacial: implications for Late Quaternary forest his O.A. and Sharma, P., 1997: Cosmogenic Cl 36 and Be 10 tory. Journal of Biogeography, 29: 1117 1124. ages of Quaternary glacial and fluvial deposits of the Wind Lal, D., 1991: Cosmic ray labeling of erosion surfaces: in situ nu River Range, Wyoming. Geological Society of America Bul clide production rates and erosion models. Earth and Plane letin, 109: 1453 1463. tary Science Letters, 104: 424 439. Plug, L.J., Gosse, J.C., McIntosh, J.J. and Bigley, R., 2007: At Lambeck, K., Smither, C. and Johnston, P., 1998: Sea level tenuation of cosmic ray flux in temperate forest. Journal of change, glacial rebound and mantle viscosity for northern Eu Geophysical Research, 112, F02022. doi: 10.1029/ rope. Geophysical Journal International, 134: 102 144. 2006JF000668 Licciardi, J.M., Clark, P.U., Brook, E.J., Pierce, K.L., Kurz, M.D., Putkonen, J. and Swanson, T., 2003: Accuracy of cosmogenic Elmore, D. and Sharma, P., 2001: Cosmogenic He 3 and Be ages for moraines. Quaternary Research, 59: 255 261. 10 chronologies of the late Pinedale northern Yellowstone ice Ramsey, C.B., 2007: Deposition models for chronological cap, Montana, USA. Geology, 29: 1095 1098. records. Quaternary Science Reviews, 27: 42 60. Lindén, A.G., 1984: Some ice marginal deposits in the east cen Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., tral part of the south Swedish Upland. Sveriges Geologiska Bertrand, C., Blackwell, P.G., Buck, C.E., Burr, G., Cutler, Undersökning, C 805. K.B., Damon, P.E., Edwards, R.L., Fairbanks, R.G., Frie Lindén, A.G., 1999: Översiktlig dokumentation av Lannaskede drich, M., Guilderson, T.P., Hughen, K.A., Kromer, B., Mc platån. Sveriges Geologiska Undersökning JRAP 99001. 26 p. Cormac, F.G., Manning, S., Bronk Ramsey, C., Reimer, R.W., Lundqvist, J., 2002: Weichsel istidens huvudfas. In: Fredén, C. Remmele, S., Southon, J.R., Stuiver, M., Talamo, S., Taylor, (ed.): Berg och jord. Sveriges Nationalatlas. 124 135. F.W., van der Plicht, J. and Weyhenmeyer, C.E., 2004: Lundqvist J. and Wohlfarth, B., 2001: Timing and east west cor INTCAL04 terrestrial radiocarbon age calibration, 0 26 cal relation of south Swedish ice marginal lines during the Late kyr BP. Radiocarbon, 46: 1029 1058. Weichselian. Quaternary Science Reviews, 20: 1127 1148. Rinterknecht, V. R., Clark, P. U., Raisbeck, G. M., Yiou, F., Biti Malmberg Persson, K., 2001: Beskrivning till jordartskartan 6E nas, A., Brook, E., Marks, J. L., Zelcs, V., Lunkka, J. P., Pav Nässjö SO. Description to the Quaternary map. Sveriges Geo lovskaya, I. E., Piotrowski, J. A. and Raukas, A., 2006: The logiska Undersökning Ae 145 (in Swedish). last deglaciation of southeastern sector of the Scandinavian Malmberg Persson, K., Persson, M. and Lindén, A.G., 2007: Is Ice Sheet. Science, 311: 1449 1452. randstråket Vimmerbymoränen mellan Knivshult och Vans Staiger, J. Gosse, J., Toracinta, R., Oglesby, B., Fastook, J. and tad i nordöstra Småland. Sveriges Geologiska Undersökning. Johnson, J.V., 2007: Atmospheric scaling of cosmogenic nu Rapport 2007, 7. clide production: Climate effect. Journal of Geophysical Re Middleton, R., Brown L., Dezfouly Arjomandy, B., and Klein, J., search, 112, B02205. doi:10.1029/2005JB003811 1993: On 10Be standards and the half life of 10Be. Nuclear In Wastenson, L. (ed.) 1995: National Atlas of Sweden: Climate, struments and Methods in Physics Research B, 82: 399 403. Lakes and Rivers. SNA. 176 pp. Olsen, L., Sveian, H., Bergstrøm, B., Selvik, S.F., Lauritzen, S. E., Zreda, M., Phillips, F.M. and Elmore, D., 1994: Cosmogenic Stokland, Ø. and Grøsfjeld, K., 2001a: Methods and strati 36Cl accumulation in unstable landforms 2. Simulations and graphies used to reconstruct Mid and Late Weichselian measurements on eroding surfaces. Water Resources Re palaeoenvironmental and palaeoclimatic changes in Norway. search, 30: 3127 3136. Norges geologiske undersøkelse Bulletin, 438: 21 46. Olsen, L., Van der Borg, K., Bergstrøm, B., Sveian, H., Lauritzen, Manuscript received Aug. 2008, revised and accepted Febr. 2009.

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

HELENA ALEXANDERSON, TIMOTHY JOHNSEN AND ANDREW S. MURRAY

Alexanderson, H., Johnsen, T. & Murray, A. S. 2010 (April): Re-dating the Pilgrimstad Interstadial with OSL: a BOREAS warmer 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 last glacial 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 pollen stratigraphic correlations with type sections in continental Europe, but the few absolute dating attempts so far have given uncertain results. We re-excavated the site and collected 10 samples for optically stimulated luminescence (OSL) dating from mineral- and organic-rich sediments within the new Pilgrimstad section. Single aliquots of quartz were analysed using a post-IR blue single aliquot regenerative-dose (SAR) protocol. Dose recovery tests were satisfactory and OSL ages are internally consistent. All, except one from an underlying unit that is older, lie in the range 52–36 ka, which places the interstadial sediments in the Middle Weichselian (MIS 3); this is compatible with existing radiocarbon ages, including two measured with accelerator mass spectrometry (AMS). The mean of the OSL ages is 446ka(n = 9). The OSL ages cannot be assigned to the Early Weichselian for all reasonable adjustments to water content estimates and other parameters. The new ages suggest that climate was relatively mild and that the Scandinavian Ice Sheet was absent or restricted to the mountains for at least parts of MIS 3. These results 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 A˚s, 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; received 10th February 2009, accepted 6th October 2009.

Pilgrimstad in central Sweden (Fig. 1) is an important The Pilgrimstad site site for reconstructing environmental changes during the Weichselian in Scandinavia, since it is situated close Kulling (1945) recognized three separate series of sand to the former ice divide of the Scandinavian Ice Sheet and gravel within the subtill sediments at Pilgrimstad. and contains subtill organic and minerogenic sediments The lowermost series was interpreted as deposited in a (Fig. 2). The site has been investigated and described by proglacial sub-aquatic environment, while the upper a number of authors over the past 70 years (mainly two series represent a transition from glacifluvial to flu- Kulling 1945; Frodin¨ 1954; Lundqvist 1967; Robertsson vial to lacustrine deposition and contain fine-grained 1988a, b; Garcı´ a Ambrosiani 1990), but the absolute minerogenic and organic material (Lundqvist 1967). A chronology of the site is still poorly known. In this well-sorted sandy bed within the lacustrine sediments study, we present and evaluate results of optically sti- has been interpreted as an aeolian deposit (Robertsson mulated luminescence (OSL) dating that place the Pil- 1988a, b). The sediments were exposed at the surface for grimstad Interstadial in marine isotope stage (MIS) 3. some time before the most recent ice advance. Detailed OSL has been used successfully to date Weichselian descriptions of the stratigraphic units and a review of deposits in, for example, Russia (Svendsen et al. 2004; interpretations are available in Lundqvist (1967). Thomas et al. 2006), Greenland (Hansen et al. 1999; The palaeoecological interpretation as a cool, sub- Adrielsson & Alexanderson 2005), the Himalayas (Spen- arctic–arctic environment is based on several proxies, cer & Owen 2004) and New Zealand (Preusser et al. mainly from the organic beds. According to pollen and 2005), but in Scandinavia results have been of varying coleoptera, open herb–shrub vegetation was followed quality (Kjær et al. 2006; Alexanderson & Murray 2007; by a forest border setting during a climatic optimum Lagerback¨ 2007; Houmark-Nielsen 2008). Because of (Robertsson 1988a, b). This warm interval was this, and because our results are controversial with respect succeeded by a colder climatic phase with periglacial to previous age determinations of the site, in this article conditions and a subsequent phase of herb–shrub we focus on the methodology and reliability of the OSL vegetation (Robertsson 1988b). Both diatoms and ages, while the palaeoglaciological and palaeoecological insects record deposition in a nutrient-rich lake (Lund- implications of the dates will be discussed elsewhere. 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 368 Helena Alexanderson et al. BOREAS

0° 10°E 20°E 30°E 40°E 70°N

Sokli

65°N

Hitura Ruunaa

Pilgrimstad FINLAND

60°N

NORWAY

SWEDEN RUSSIA ESTONIA

LATVIA Fig. 1. Location map of Pilgrimstad. Mam- 55°N DENMARK Skåne moth locations from Ukkonen et al. (2007), Last 55°N Glacial Maximum (LGM) margin from Svend- sen et al. (2004). Other sites that also indicate ice- LITHUANIA free conditions during MIS 3 are shown: Sokli (50 ka; Helmens et al. 2007a, b), Ruunna 10°E 20°E 0100 200 400 Kilometers (50–25 ka; Lunkka et al. 2008), Hitura (deglaciation 62–55 ka; Salonen et al. 2008) and several 0-600 m a.sl. >600 m a.s.l. mammoth 38-32 ka LGM margin sites in Skane˚ (39–24 ka; Kjær et al. 2006).

The site has been assigned an Early Weichselian (MIS 5 a/c) age based on a correlation of the palaeoenvironmental reconstruction with type sections in continental Europe and represents the type section for the local Pilgrimstad Interstadial (Kulling 1967; cf. also Jamtland¨ Interstadial; Lundqvist 1967; Lundqvist & Miller 1992). Robertsson (1988a, b), for example, correlated the organic-rich beds with one or possibly two Early Weichselian interstadials (Brørup and/or Odder- ade) depending on how the climatic cooling in the middle of the section is interpreted – as a phase within one interstadial or as separating two different interstadials. So far, there have been few absolute dating attempts and these have given uncertain results. According to a recent evaluation of all radiocarbon dates from Pil- Fig. 2. Photograph of the upper part of the new section at Pilgrim- grimstad, 10 samples that are acceptable from a quality stad (see Fig. 3 for comparisons). D–H represent the lithological units. perspective range between 59 and 46 cal. ka BP in age (Wohlfarth 2009) (Fig. 3). Moreover, a single thermoluminescence measurement resulted in 60 ka for the indicates a cool, arctic–subarctic climate, consistent sandy unit within the organic beds (Garcı´ a Ambrosiani with the ﬁnds of mammoth (Mammuthus primigenius), 1990) and one U/Th measurement provided an age of reindeer (Rangifer tarandus) and elk (Alces alces) 384 ka (Heijnis in Robertsson & Garcı´ a Ambrosiani (Lundqvist 1967). 1992). Recently, Ukkonen et al. (2007) also re-dated a BOREAS A warmer climate and a smaller ice sheet during the Swedish MIS 3? 369

H A Pilgrimstad section B Sediment description C Chronology G OSL C Correlated ages 318 x 319 (ka) (ka BP) (cal. ka BP) x SE x 344 NW Unit H. Yellow sand, well- 38±3 SiSl 45±4 sorted, homogeneous. Wedge- Sm NE 36±3 60 ka TL like structure. Contains rip-ups 52±4 x 345 Sm of gyttja. 49±4 Unit G. Laminated and massive 320 39±3 F x sand, deformed in lower part. 346 x SGy Yellowish to rusty colour. 43±5 Unit F. Brown sandy gyttja with 48±8 39.2±2 ~46 ~47 occasional pebbles <15 cm. ~48 SGy 347 x Sm/Sl ~47 H ~50 ~52~55 348 x ~52 ~59 Unit E. Sandy gyttja, smaller ~52 >40 brecciated pieces than in D. E Unit D. Grey sandy silty gyttja. SiGyb Compact, brecciated. D 34-29 Unit C. Organic-rich sand. GySb x C 46±3 321 Compact, brecciated. B SSil Unit B. Varved clay and silt. CSil

CoGmm Unit A. Sandy gravel with 74±5 A lenses of sand and silt.

m Sm(ng)

1 x 322 SiSm SiSm

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

Fig. 3. A. Sketch of the new section at Pilgrimstad, lithology and position of OSL and 14C samples. B. Brief sediment description. C. Results of OSL and 14C age determination. mammoth molar from Pilgrimstad to 34–29 cal. ka BP. by blast stone (J. Lundqvist, pers. comm. 2007). Thus, we The exact stratigraphic position of the mammoth re- adjusted the sample depths by adding 1 m, corresponding mains is not clear, but based on Kulling’s (1945) and to the general till thickness in the area. The position of Robertsson’s (1988a) stratigraphic descriptions, it the investigated section is 62157.50N, 15101.10E and its seems to derive from the lower part of the organic beds elevation 300 m a.s.l., as measured by a Garmin GPS analysed by Robertsson (1988a). Vista C and checked against topographic maps. These ages are all younger than the age estimate based The beds in the excavated section were correlated to on pollen stratigraphy; a possible correlation of Pilgrim- the stratigraphies presented by Kulling (1945) and Ro- stad with the Moershoofd Interstadial (46–44 ka BP) bertsson (1988b). Robertsson’s original section was si- (Behre & van der Plicht 1992) has therefore been pro- tuated adjacent to and at right angles to our new posed but considered less likely (Robertsson 1988a). section; her stratigraphy is therefore best comparable to the new stratigraphy.

Setting Methods The study site is located near Pilgrimstad in Jamtland,¨ central Sweden (Fig. 1) and is situated in an abandoned Sampling, preparation and measurement gravel pit located at the edge of a valley ﬂoor. Most of the previously described sections have been mined or cov- Fieldwork was conducted in September 2006 and in July ered by colluvium or ﬁll. A small hill in the southern part 2007. Five OSL samples were collected during each of the pit was the focus of our excavation (Fig. 2). The campaign at the levels marked in Fig. 3 (see also Table 1). top (including the till cover) had been removed pre- The samples were taken in opaque plastic tubes and viously, and for a short time the section had been covered stored in black bags until opened under darkroom 370 Helena Alexanderson et al. BOREAS

Table 1. Pilgrimstad OSL sample properties and settings. Listed in order of sample number. For lithofacies codes and stratigraphic unit numbers, see Fig. 3. Ã Sample Sampling depth (m) Stratigraphic unit Lithofacies code Water content (weight %) Organic content (%) w1 w2 w3 w

061344 1.2 G Sm 30 29.5 9.1 28 2.2 061345 1.8 H SiSm 30 29.4 5.2 27 3.9 061346 2.4 F SGy 150 41.9 16.9 72 9.1 061347 2.9 H Sm/Sl 40 35.7 8.5 34 2.0 061348 3.1 H Sm/Sl 35 32.8 8.0 31 1.9 081318 0.5 H SiSm 35 32.8 5.5 31 1.4 081319 0.5 G SiSl 40 38.0 4.9 35 1.4 081320 2.1 H Sm 35 34.6 3.8 32 1.2 081321 4.1 C GySb 35 34.2 9.0 32 1.1 081322 6.3 A Sm(ng) 35 32.7 6.2 31 2.7 Ã Depth used in calculation was sampling depth plus 1 m; see text for explanation. conditions at Stockholm University (2006) and the Nor- Data analysis wegian University of Life Sciences (2007), where the in- Simple component analysis of the continuous wave itial preparation was completed. Final preparation, OSL data from some aliquots was undertaken using including heavy liquid separation (2.62 g/cm3) to remove SigmaPlot 10.0 based on the parameters and formulas feldspars (081318–22 only), treatment with 10% HCl for of Choi et al. (2006). The results of the component 5–30 min, 10% H O for 15–30 min, 38% HF for 2 2 analysis are discussed further below (Fig. 4), but based 60–120 min and 10% HCl again for 40 min, was done at on such information from dose recovery measurements the Nordic Laboratory for Luminescence Dating, where (Fig. 5) we chose channels 1–2 (first 0.16 s) and channels the OSL measurements were also undertaken. 4–6 (0.32–0.56 s) as peak and background integration The samples were analysed using large aliquots of limits for all aliquots. The equivalent doses were then quartz (180–250 mm) on Risø TL/OSL readers equip- calculated in Risø Luminescence Analyst 3.24 (ex- ped with calibrated 90Sr/90Y beta radiation sources ponential curve fitting) and in Microsoft Excel. To be (dose rate 0.14–0.35 Gy/s), blue (47030 nm; 50 mW/ accepted, aliquots had to pass the following rejection cm2) and infrared (880 nm, 100 mW/cm2) light sour- criteria: recycling ratio within 20% of unity, recupera- ces, and detection was through 7 mm of U340 glass tion o5%, equivalent dose error o50% and signal filter (Bøtter-Jensen et al. 2000). Analyses employed more than 3s above the background. Decay and post-IR blue SAR protocols (Murray & Wintle 2000, growth curves also had to be regular. Ages were calcu- 2003; Banerjee et al. 2001) adapted to suit the samples lated using the mean and median of the equivalent dose based on dose recovery and preheat experiments (first population of accepted aliquots for each sample, as well batch: preheat 2601C for 10 s, cut-heat 2201C; second as using the natural and saturated water contents. batch 2401/2001C). A relatively high test-dose (50 Gy) We also did a sensitivity analysis to determine was necessary to obtain a statistically precise test signal, quantitatively which uncertainties have the largest and 100 s of illumination at 2801C between cycles im- effect on age. Ages were recalculated after adjustments proved recuperation (response to zero dose). were made to each of the parameters in turn, using We calculated the dose rates from gamma spectro- reasonable estimates of uncertainty for each parameter. metry data (Murray et al. 1987) (Table 2) and included The estimated uncertainties (1 SD) used in these cal- the cosmic ray contribution (Prescott & Hutton 1994). culations were: depth below surface 1 m, elevation Natural and saturated water content was measured 50 m, grain size 10%, water content (gamma with pF rings (cylinder volumeters) (Table 1). To ac- and beta) 10%, dose rate gamma 5%, dose rate count for water content changes through time due beta 5%, internal dose rate 30%, density 10%, mainly to compaction, especially for the organic-rich cosmic ray contribution 5% and beta source cali- sediments, we applied a simple three-stage model to all bration 2%. our samples (Table 3). The mean water content ðwÞ since time of deposition was then calculated as:

¼ þ þ ð Þ w w1 t1 w2 t2 w3 t3 1 Results Two samples were collected for radiocarbon dating. Sedimentology and stratigraphy Small twigs (unknown species) were picked out and dated by AMS 14C at the Lund University Radio- We distinguished eight units in the new section at Pil- carbon Dating Laboratory. grimstad (shown and brieﬂy described in Figs 2 and 3). BOREAS A warmer climate and a smaller ice sheet during the Swedish MIS 3? 371

Table 2. Summary of radionuclide concentrations measured with high-resolution gamma spectrometry on the Pilgrimstad OSL samples. Beta and 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 396 42.10.7 35.40.7 64212 2.180.04 1.250.04 061345 345 42.80.6 34.70.5 6689 2.140.04 1.150.04 061346 17810 100.81.2 450.9 67513 3.190.07 1.840.09 061347 335 49.10.7 370.5 71510 2.380.04 1.370.04 061348 283 48.50.5 370.4 7167 2.360.03 1.360.04 081318 408 40.20.8 34.30.8 66315 2.220.05 1.240.04 081319 276 38.50.7 33.20.7 67012 2.170.04 1.210.03 081320 295 31.90.5 30.70.5 68012 2.150.04 1.140.03 081321 4110 100.31.3 35.10.9 64915 2.500.07 1.650.09 081322 427 74.40.9 30.70.7 61411 2.260.05 1.390.06

Table 3. Water-content modelling to calculate average water content back to fluvial or glacifluvial (unit G), in line with pre- since the time of deposition, accounting for compaction and environ- vious interpretations. The sandy unit H might represent mental changes. For values, see Table 1. the aeolian sand of Robertsson (1988a, b). However, Assumptions Three-stage hydrological the cross-cutting relationship with the other units in- evolution dicates formation after the deposition of units F and G, and suggests a possible glacitectonic rather than aeolian The sediments have never been 1. Lake/fluvial stage before the drier than at the time of ice advance (sediments loose and genesis. We therefore interpret unit H as a clastic dyke sampling, and the natural water saturated). formed subglacially during the Late Weichselian ice content is a minimum water (a) Water content w1 is the advance (cf. Larsen & Mangerud 1992; Linde´ n et al. content. (The samples were saturated value rounded up to the 2008). The sand is thus likely reworked sand from unit taken within a large gravel pit in nearest 5 or 0 for minerogenic G, and the gyttja clasts are rip-ups from the surround- which the groundwater table samples and an assumed value of has been lowered.) 150% for the organic sample ing beds. This has implications for the interpretation of (061346), derived from young the organic beds as belonging to one or two events, and The saturated water content is samples with similar organic for the ages of unit H, as discussed below. the maximum water content for content. the sediments in their present (b) Duration is 30% of the time state. (It is limited by porosity.) since deposition (t1 = 0.3). OSL characteristics and ages The porosity of minerogenic 2. Ice cover (sediments compacted sediments did not change and saturated). The major sample properties, settings and results are lis- significantly with compaction. (a) Water content w2 is the tedinTables1–3andareshowninFigs3and7.The saturated value. Most of the compaction took (b) Duration is 60% of the time quartz from Pilgrimstad is insensitive, which necessitated place early in the sediments’ since deposition (t2 = 0.6). careful selection of peak and background channels to history due to continued best isolate the fast component (Fig. 4). For the 21 ac- sedimentation and, later, 3. After deglaciation (sediments cepted dose recovery experiments (out of 27), the average pressure from the ice cover. compacted and drier). proportions of the net component signal to the total net (a) Water content w3 is the natural value. signal used for dose calculation were 928% (fast), (b) Duration is 10% of the time 87% (medium) and 0.20.3% (slow). When these since deposition (t3 = 0.1). channels were selected for peak and background integration, the resulting signal was dominated by the fast component. Doses calculated with this channel selection gave good dose recovery (1.050.04, n = 21) (Fig. 5) The sediments have been tectonized, as indicated by and demonstrated that the SAR protocols used were able brecciation and deformation structures. Within the accurately to recover a known dose administered before limited exposure available to us, the sediments never- any heating. The signals were not close to saturation theless seem to have a pancake stratigraphy, with the (Fig. 6). exception of unit H, which cuts units E, F and G. Equivalent doses range between 89 and 145 Gy and Unit A in the new section correlates with the ‘lime- dose rates between 2.5 and 3.1 Gy/ka (Table 4). The stone-rich pebbly gravel series’ of Kulling (1945), while nine upper samples are 52–36 ka, while the lowermost is units B–H correspond to his ‘silt-stratified sand series’. older (74 ka) (Table 4 and Fig. 7). OSL ages from unit We interpret these sediments as showing an environ- H average 447ka (n = 5), from unit G 413ka mental succession from glacifluvial (unit A) to glacila- (n = 2) and single ages from unit F and C are 488ka custrine (unit B) to lacustrine (units C, D, E, F) and and 463 ka, respectively. The sensitivity analysis 372 Helena Alexanderson et al. BOREAS

A 100% 4000 slow 90% 3500 80% Sample 061347

background 3000 4 70% Mean = 1.05 60% 2500 Std err = 0.04 50% 2000 n = 21 40%

1500 Bulk signal natural signal

% of bulk signal 30% 1000 20% modeled signal Frequency 2 medium 10% 500 fast 0% 0 012345 Stimulation time (sec)

B Growth curve using standard data 2.5 0 Sample 061347

0-0.1 2.0 1-1.1

0.2-0.3

0.4-0.5

0.6-0.7

0.8-0.9

1.2-1.3

1.4-1.5

1.6-1.7

1.8-1.9 1.5 Ratio of measured to given dose Fig. 5. Histogram of dose recovery tests, excluding aliquots that did Lx/Tx 1.0 Recycling = 1.11 not pass rejection criteria. A mean close to unity indicates that the Dose recovery = 0.92 samples can be used for OSL-dating with our analytical protocols. 0.5 given dose 0.0 New radiocarbon ages from twigs are 39.22 and 0 200 400 600 800 1000 1200 440 14C ka BP (Table 5). Dose (s)

C Growth curve using derived fast component data 2.5 Sample 061347 Discussion 2.0 OSL and sediments 1.5 From a sedimentological perspective, we can identify

Lx/Tx two potential problems for OSL-dating at this site: (1) 1.0 Recycling = 1.13 Dose recovery = 0.90 the glacifluvial deposit (unit A) may be incompletely 0.5 bleached and (2) we cannot properly estimate the ori- given dose ginal mean water content of the relatively organic-rich 0.0 unit F. 0 200 400 600 800 1000 1200 Dose (s) Incomplete bleaching results in an apparent age overestimation, and is fairly common in glacial settings Fig. 4. Data derived from the 27 dose recovery experiments allowed (e.g. Fuchs & Owen 2008). As we have used only large selecting peak and background channels that provided the overall best recycling and dose recovery and a dominance of the fast aliquots for measurements, it is difficult to infer any- component after background subtraction. A. Signal component thing about incomplete bleaching from the dose dis- analysis of sample 061347 showing the natural signal which was tribution of sample 081322, which is from unit A. In de-convoluted into fast, medium and slow components (left axis), and the natural and modelled (sum) signal (right axis). Also shown combination with the stratigraphic position of unit A are the peak and background portions of the signal used (0–0.16 s (lowest), we thus consider the OSL age from unit A as and 0.32–0.56 s). Note that only the first 5 s of the total length of the providing a maximum age for the overlying organic stimulation (40 s) are shown. B. The upper growth curve uses the standard data (bulk signal) with the selected channels for the beds. The good correspondence of ages (n = 9) from all same aliquot. C. The lower growth curve is produced by using the the other beds (derived from various depositional set- derived (de-convoluted) fast component signal data only. The tings) suggests that there are no problems with in- insignificant difference between the growth curves indicates that complete bleaching for those. the channel selection effectively isolated the fast component and that the selection of peak and background channels provided the Organic-rich deposits tend to be fairly loose at the optimum recycling and dose recovery. time of deposition and may have very high water content, i.e. up to several hundred percent. With time they demonstrates that the calculated age is most sensitive will become compacted due to sediment loading and the to changes in the equivalent dose, beta source calibra- overriding ice sheet; the water content will change sig- tion and dose rate, followed by water content (Fig. 8). nificantly, so that what is measured today is not BOREAS A warmer climate and a smaller ice sheet during the Swedish MIS 3? 373 representative of the mean water content since time of Based on the stratigraphic information and on the deposition (cf. Alexanderson et al. 2008). An under- OSL ages, we consider the OSL ages from units B to H estimation of the mean water content results in OSL ages as representing one event, while unit A is older. The that appear too young, and vice versa. In our case, the OSL ages from units B–H are internally consistent, i.e. OSL ages from the surrounding minerogenic beds can all lie in the range 52–36 ka, with a mean of 446 ka; provide some constraints, since these sediments do not this places the interstadial sediments in the Middle suffer to the same degree from water-content variations. Weichselian (MIS 3; 58–24 ka). Taken at face value, the The sandy unit H is the youngest, since it cross-cuts 74 ka age from unit A gives the timing of the preceding units E, F and G. The OSL ages from unit H should deglaciation, but we cannot rule out the possibility that thus be minimum ages for the organic deposits. How- the sediment suffers from some incomplete bleaching, ever, as mentioned above, re-interpretation of the unit and that the true deposition age is younger. as a clastic dyke with remobilized material implies that the sand in unit H could be contemporaneous with unit G and that the OSL ages do not represent the timing of What is required to make these OSL ages older the formation of the dyke. It also implies that the se- (MIS 5)? paration into units E and F is secondary. To capture uncertainties in the ages conservatively, Fig. 7 and Table 4 show the OSL ages calculated under a 3 variety of assumptions. Even when considering the Sample 061345 broadest age range for each sample, all upper nine ages fall within the Middle Weichselian (MIS 3) and no reasonable adjustments in assumptions can collectively

SL bring their ages to the Early Weichselian (MIS 5 a/c, 2 474 ka). As the sensitivity analysis showed the calculated age to depend most on changes in the equivalent Stimulation time (s) 01020 3040 dose and various factors inﬂuencing the dose rate (Fig. 3000 8), we tested what those factors would have to be to 2500 obtain 90 ka ages from these samples. To get early 1 2000 Weichselian ages from the current data we would need

Sensitivity corrected O 1500 to double the equivalent doses or half the dose rates; the 1000 latter for example by using water contents 4100% by 500 weight, or a combination of these. We consider it un-

Photon counts / 0.16s 0 likely that any of these parameters is in error to this 0 100 200 300 400 500 600 degree. Regenerative dose (Gy) Fig. 6. Single-aliquot regenerative-dose (SAR) OSL growth curve for a single aliquot of sample 061345. The equivalent dose (ED) (open Comparison with other chronologies and records diamond) is obtained by interpolating the corrected natural signal (open triangle) on the growth curve. Repeated dose points (open A mean age of 446ka(n = 9) for the Pilgrimstad se- squares) and a zero dose point (open circle) are also shown. (Inset) diments agrees with the bulk of previous age determi- The natural decay curve. nations from the site, which fall between 60 and 45 ka

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 43537364435 11612 18 2.720.10 1.140.24 061345 45444374464 11810 19 2.620.10 1.060.26 061346 48849325397 13423 18 2.820.09 1.060.31 061347 52450424534 14510 16 2.770.10 1.080.25 061348 49449403504 13910 22 2.830.10 1.030.20 081318 38336303383 1017 28 2.690.10 1.170.09 081319 39334293403977 31 2.510.09 1.070.12 081320 36335272363897 32 2.520.09 1.050.17 081321 46344373473 1436 30 3.100.13 1.100.12 081322 74572584755 2029 30 2.740.11 1.070.12

1Age calculated with water content at time of sampling. 2Age calculated with saturated water content. 374 Helena Alexanderson et al. BOREAS

A 0 10 20 30 40 50 Fig. 7. A. Graphical summary of ages for all Agr (ka) samples. B. Corresponding recycling ratios. 60 OSL ages are shown in stratigraphic order with 70 youngest ages to the left and oldest to the right. 80 H–A refers to the stratigraphic units in Fig. 3. 90 Age calculated from the population of accepted 318,H 345,H 320,H 347,H 348,H 344,G 319,G 346,F 321,C 322,A aliquots using the mean and median, and for Sample and stratigraphic unit natural water content at sampling (dry) and saturated water content (range shown by grey B square). Error bars represent 1 SD. The SD of 1.6 mean of ages excludes the older, strati- 1.4 graphically lower sample (081)322. Note that 1.2 the water content for organic-rich sample 1.0 (061)346 has been adjusted to compensate for 0.8 compression. Marine isotope stages (MIS) in-

Recycling ratio 0.6 dicated on the right.

Fig. 8. Sensitivity analysis showing by how 40 much the age will change (percentage change) in a given parameter for sample (061)344 Change in age (ka) (age = 43 ka). Steeper slopes indicate that the age is more sensitive to changes in the given parameter. The age is most sensitive to changes 35 in the equivalent dose, beta source calibration and dose rate, followed by water content; very large changes in these parameters are needed to make the age consistent with MIS 5. The range 30 of x-axis parameter values corresponds to our –100 –80 –60 –40 –20 0 20 40 60 80 100 estimate of 2 SD in that parameter. Similar re- Change in parameter (%) sults were obtained for the other samples.

(Fig. 3) (Wohlfarth 2009). These radiocarbon dates are Greenland interstadials 12–10 (47–41 ka), but con- considered to be of acceptable quality, although close sidering the uncertainties and the actual spread in ages, to the methodological limit (Wohlfarth 2009). The interstadials 17–10 are also possible options (cf. Walker mean OSL age also agrees with the finite of our two et al. 1999). AMS 14C ages (39.22 14C ka BP; Table 5) and is not This is in agreement with Wohlfarth (2009), who contradicted by the non-finite one. tentatively places the Pilgrimstad site within Greenland Oxygen isotope curves from the Greenland Ice Sheet interstadials 17–12. Hattestrand¨ (2008) proposes a si- provide estimates of temperature variations during milar age shift of the Tarend¨ o¨ II Interstadial in north- the Weichselian in the North Atlantic region (e.g. ern Sweden, i.e. from the Early Weichselian (MIS 5 a) Johnsen et al. 2001). If the sediments at Pilgrimstad are to the Middle Weichselian (MIS 3). Our OSL data, of MIS 3 age, as the OSL and other ages suggest, from a critical location in the former central area of the we would expect a correlation with one of the warmest ice sheet, thus support the documented ice-free condi- and/or longest interstadials during this time – with tions reported elsewhere in Fennoscandia during MIS 3 duration long enough to allow for the subarctic–arctic (e.g. Olsen et al. 2001; Arnold et al. 2002; Kjær et al. flora and fauna to become established. Taking the 2006; Helmens et al. 2007a, b; Ukkonen et al. 2007; OSL ages at face value leads to a correlation with Lunkka et al. 2008; Salonen et al. 2008) (see also Fig. 1). BOREAS A warmer climate and a smaller ice sheet during the Swedish MIS 3? 375

Table 5. 14C ages from twigs. or at least restricted to the highest Scandinavian mountains. Sample Stratigraphic unit 14C age years BP

LuS 6957 F 39 200 2000 Acknowledgements. – We thank Jan Lundqvist, Ann-Marie Rober- 4 LuS 6958 E 40 000 tsson and Barbara Wohlfarth at Stockholm University for valuable discussions and for sharing experiences and data; Jan-Pieter Buylaert and the technical staff at the Nordic Laboratory for Luminescence Dating for helping with OSL measurements; Damian Steffen at the Implications for Weichselian history University of Bern for introducing Alexanderson and Johnsen to the world of de-convolution; and Dick Olsson and Claes in Pilgrimstad Moving the age of the Pilgrimstad Interstadial from for giving us access to the gravel pit and assisting with excavation. The study was financed by a grant (no. 60-1356/2005) from the Geological Early to Middle Weichselian (MIS 5 a/c to MIS 3) has Survey of Sweden to Alexanderson, with additional fieldwork funding implications for the understanding of glacial history from the Swedish Nuclear Fuel and Waste Management Company and environmental change in Scandinavia during the (SKB). Alexanderson was in charge of and carried out most of the Weichselian. An ice-free Pilgrimstad at 50–40 ka re- OSL sampling, preparation and measurements, sedimentological and stratigraphical work as well as writing, with significant input at all quires that the Scandinavian Ice Sheet at that time was stages by Johnsen. Johnsen analysed all the data, with some input restricted, possibly limited to the highest mountains (cf. from Murray and Alexanderson. Murray contributed to discussions Arnold et al. 2002), i.e. much smaller than previously regarding OSL preparation and measurement setup and interpretation of results as well as provided laboratory access. The constructive believed (Lundqvist 2002; Mangerud 2004; Houmark- comments on the manuscript from reviewers Per Moller¨ and Ole Bennike Nielsen 2007) (Fig. 1). It also implies that the ice sheet are appreciated. must have expanded rapidly thereafter to reach southeastern Denmark just prior to 30 ka (the Klintholm advance; Houmark-Nielsen & Kjær 2003; Ukkonen et al. References 2007). Although the vegetation reconstruction for Pilgrim- Adrielsson, L. & Alexanderson, H. 2005: Interactions between the stad (Robertsson 1988b) still largely holds true, Greenland Ice Sheet and the Liverpool Land coastal ice cap during recent studies indicate that forest may not have been as the last two glaciation cycles. Journal of Quaternary Science 20, 269–283. close as suggested (Helmens et al. 2007a). Nevertheless, Alexanderson, H. & Murray, A. S. 2007: Was southern Sweden ice a reconciliation of the Pilgrimstad record with con- free at 19–25 ka, or were the post LGM glacifluvial sediments in- temporaneous records of tundra in northern con- completely bleached? Quaternary Geochronology 2, 229–236. Alexanderson, H., Johnsen, T., Wohlfarth, B., Naslund,¨ J.-O. & tinental Europe (Behre 1989) requires alternative Stroeven, A. 2008: Applying the optically stimulated luminescence migration routes for vegetation, and/or possible tree (OSL) technique to date the Weichselian glacial history of southern refugia, during previous stadials and different climate Sweden. Reports from the Department of Physical Geography and gradients and conditions compared to today (e.g. Nils- Quaternary Geology, Stockholm University 4, 1–33. Arnold, N. S., van Andel, T. H. & Valen, V. 2002: Extent and dy- son 1972: p. 225; Ukkonen et al. 2007). 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: Conclusions Equivalent dose estimation using a single aliquot of polymineral fine grains. Radiation Measurements 33, 73–94. A new exposure at the Pilgrimstad site in central Behre, K.-E. 1989: Biostratigraphy of the last glacial period in Eur- Sweden shows 44 m thick subtill minerogenic and ope. Quaternary Science Reviews 8, 25–44. Behre, K. E. & van der Plicht, J. 1992: Towards an absolute chronol- organic sediments; these have been OSL-dated. ogy for the last glacial period in Europe: Radiocarbon dates from The OSL ages from the lacustrine and fluvial sedi- Oerel, northern Germany. Vegetation History and Archaeobotany 1, ments range from 52 to 36 ka, while the underlying 111–117. Bøtter-Jensen, L., Bulur, E., Duller, G. A. T. & Murray, A. S. 2000: glacifluvial deposit is probably younger than 74 ka. Advances in luminescence instrument systems. Radiation Measure- The OSL samples passed appropriate methodologi- ments 32, 523–528. cal checks and the ages are internally consistent. We Choi, J. H., Duller, G. A. T. & Wintle, A. G. 2006: Analysis of quartz LM-OSL curves. Ancient TL 24, 9–20. therefore consider the results reliable. Frodin,¨ G. 1954: De sista skedena av Centraljamtlands¨ glaciala his- From sedimentological and chronological points of toria. Geographica 24, 1–251. view, we favour deposition during a single but vari- Fuchs, M. & Owen, L. A. 2008: Luminescence dating of glacial and associated sediments: Review, recommendations and future direc- able event, rather than two separate interstadials. tions. Boreas 37, 636–659. The OSL ages assign the Pilgrimstad Interstadial Garcı´ a Ambrosiani, K. 1990: Pleistocene Stratigraphy in Central site to MIS 3; the organic deposits could possibly and Northern Sweden. Ph.D. thesis. Department of Quaternary Re- search, Report 16, 1–15. University of Stockholm. correspond to one or more of Greenland inter- Hansen, L., Funder, S., Murray, A. S. & Mejdahl, V. 1999: Lumi- stadials 17–10. nescence dating of the last Weichselian glacier advance in East The central location of the site, together with results Greenland. Quaternary Geochronology (Quaternary Science Re- views) 18, 179–190. from other studies in Fennoscandia, indicates that Hattestrand,¨ M. 2008: Vegetation and Climate During Weichselian Ice for at least parts of MIS 3 the ice sheet was absent, Free Intervals in Northern Sweden. Ph.D. thesis. Dissertation from 376 Helena Alexanderson et al. BOREAS

the Department of Physical Geography and Quaternary Geology 15, Murray, A. S. & Wintle, A. G. 2003: The single aliquot regenerative 1–35. Stockholm University. dose protocol: Potential for improvements in reliability. Radiation Heijnis, H. 1992: Uranium/Thorium Dating of Late Pleistocene Peat Measurements 37, 377–381. Deposits in N.W. Europe. Thesis, Rijksuniversiteit Groningen, 149 pp. Murray, A. S., Marten, R., Johnson, A. & Martin, P. 1987: Analysis Helmens, K. F., Bos, J. A. A., Engels, S., Van Meerbeeck, C. J., for naturally occurring radionuclides at environmental concentra- Bohncke, S. J. P., Renssen, H., Heiri, O., Brooks, S. J., Seppa,¨ H., tions by gamma spectrometry. Journal of Radioanalytical and Nu- Birks, H. J. B. & Wohlfarth, B. 2007a: Present-day temperatures in clear Chemistry Articles 115, 263–288. northern Scandinavia during the last glaciation. Geology 35, Nilsson, T. 1972: The Pleistocen. 508 pp. Esselte Studium, Stockholm. 987–990. Olsen, L., Sveian, H. & Bergstrøm, B. 2001: Rapid adjustments of the Helmens, K. F., Johansson, P. W., Ras¨ anen,¨ M. E., Alexanderson, H. western part of the Scandinavian Ice Sheet during the Mid and Late & Eskola, K. O. 2007b: Ice-free intervals at Sokli continuing into Weichselian – a new model. Norsk Geologisk Tidsskrift – Norwegian Marine Isotope Stage 3 in the central area of the Scandinavian gla- Journal of Geology 81, 93–118. ciations. Geological Society of Finland, Bulletin 79, 17–39. Prescott, J. R. & Hutton, J. T. 1994: Cosmic ray contributions to dose Houmark-Nielsen, M. 2007: Extent and age of Middle and rates for luminescence and ESR dating: Large depths and long- Late Pleistocene glaciations and periglacial episodes in southern term time variations. Radiation Measurements 23, 497–500. Jylland, Denmark. Bulletin of the Geological Society of Denmark 55, Preusser, F., Andersen, B. G., Denton, G. H. & Schluchter,¨ C. 2005: 9–35. Luminescence chronology of Late Pleistocene glacial deposits in Houmark-Nielsen, M. 2008: Testing OSL failures against a regional North Westland, New Zealand. Quaternary Science Reviews 24, Weichselian glaciation chronology from southern Scandinavia. 2207–2227. Boreas 37, 660–677. Robertsson, A.-M. 1986: The diatom flora of the till-covered sedi- Houmark-Nielsen, M. & Kjær, K. H. 2003: Southwest Scandinavia, ments at Pilgrimstad, Jamtland.¨ Striae 24, 2123–2220. 40–15 kyr BP: Palaeogeography and environmental change. Journal Robertsson, A.-M. 1988a: Biostratigraphical Studies of Interglacial of Quaternary Science 18, 769–786. and Interstadial Deposits in Sweden. Ph.D. thesis. Department of Johnsen, S. J., Dahl-Jensen, D., Gundestrup, N., Steffensen, J. P., Quaternary Research, Report 10, 1–19. University of Stockholm. Clausen, H. B., Miller, H., Masson-Delmotte, V., Sveinbjornsdo¨ ´ t- Robertsson, A.-M. 1988b: A reinvestigation of the biostratigraphy of tir, A´ . E. & White, J. 2001: Oxygen isotope and palaeotemperature the till-covered sediments at Pilgrimstad and O¨ je, central Sweden. records from six Greenland ice-core stations: Camp Century, Dye- In Robertsson, A.-M. (ed.): Biostratigraphical Studies of Inter- 3, GRIP, GISP2, Renland and NorthGRIP. Journal of Quaternary glacial and Interstadial Deposits in Sweden, 1–35. University of Science 16, 299–307. Stockholm, Department of Quaternary Research, Stockholm. Kjær, K. H., Lagerlund, E., Adrielsson, L., Thomas, P. J., Murray, A. Robertsson, A.-M. & Garcı´ a Ambrosiani, K. 1992: The Pleistocene in & Sandgren, P. 2006: The first independent chronology for Sweden – a Review of Research, 1960–1990. Geological Survey of Middle and Late Weichselian sediments from southern Sweden and Sweden Ca 81, 299–306. the island of Bornholm. Geologiska Foreningens;¨ Stockholm Salonen, V.-P., Kaakinen, A., Kultti, S., Miettinen, A., Eskola, K. O. Forhandlingar¨ 128, 209–220. & Lunkka, J. P. 2008: Middle Weichselian glacial event in the cen- Kulling, O. 1945: Om fynd av mammut vid Pilgrimstad i Jamtland¨ . tral part of the Scandinavian Ice Sheet recorded in the Hitura pit, Geological Survey of Sweden C 473, 61 pp. Ostrobothnia, Finland. Boreas 37, 38–54. Kulling, O. 1967: Yttrande med anledning av J. Lundqvists foredrag¨ Spencer, J. Q. & Owen, L. A. 2004: Optically stimulated luminescence ‘Submorana¨ sediment i Jamtland’.¨ Geologiska Foreningens¨ i Stock- dating of Late Quaternary glaciogenic sediments in the upper holm Forhandlingar¨ 89, 123–125. Hunza valley: Validating the timing of glaciation and assessing Lagerback,¨ R. 2007: Ventifacts – means to reconstruct glacial develop- dating methods. Quaternary Science Reviews 23, 175–191. ment and palaeoenvironment in northern and central Sweden. Geo- Svendsen, J. I., Alexanderson, H., Astakhov, V. I., Demidov, I., logiska Foreningens;¨ Stockholm Forhandlingar¨ 129, 315–324. Dowdeswell, J. A., Funder, S., Gataullin, V., Henriksen, M., Hjort, Larsen, E. & Mangerud, J. 1992: Subglacially Formed Clastic Dikes. C., Houmark-Nielsen, M., Hubberten, H. W., Ingo´ lfsson, O´ ., Geological Survey of Sweden, SGU Ca 81, 163–170. Jakobsson, M., Kjær, K. H., Larsen, E., Lokrantz, H., Lunkka, J. Linde´ n, M., Moller,¨ P. & Adrielsson, L. 2008: Ribbed moraine P., Lysa,˚ A., Mangerud, J., Matiouchkov, A., Murray, A., Moller,¨ formed by subglacial folding, thrust stacking and lee-side cavity P., Niessen, F., Nikolskaya, O., Polyak, L., Saarnisto, M., Siegert, infill. Boreas 37, 102–131. C., Siegert, M. J., Spielhagen, R. F. & Stein, R. 2004: Late Qua- Lundqvist, J. 1967: Submorana¨ sediment i Jamtlands¨ lan¨ . Geological ternary ice sheet history of northern Eurasia. Quaternary Science Survey of Sweden C 618. 267 pp. Reviews 23, 1229–1271. Lundqvist, J. 2002: Weichseltidens huvudfas. In Frede´ n, C. (ed.): Berg Thomas, P. J., Murray, A. S., Kjær, K. H., Funder, S. & Larsen, E. och jord, 124–135. Sveriges Nationalatlas, Vallingby.¨ 2006: Optically stimulated luminescence (OSL) dating of glacial Lundqvist, J. & Miller, U. 1992: Weichselian Stratigraphy and Glacia- sediments from Arctic Russia – depositional bleaching and metho- tions in the Tasj˚ o–Hoting¨ Area, Central Sweden. Geological Survey of dological aspects. Boreas 35, 587–599. Sweden C 826, 35 pp. Ukkonen, P., Arppe, L., Houmark-Nielsen, M., Kjær, K. H. & Karhu, Lunkka, J. P., Murray, A. & Korpela, K. 2008: Weichselian sediment J. A. 2007: MIS 3 mammoth remains from Sweden – implications for succession at Ruunaa, Finland, indicating a Mid-Weichselian ice- faunal history, palaeoclimate and glaciation chronology. Quaternary free interval in eastern Fennoscandia. Boreas 37, 234–244. Science Reviews 26, 3081–3098. Mangerud, J. 2004: Ice-sheet limits on Norway and the Norwegian Walker, M. J. C., Bjorck,¨ S., Lowe, J. J., Cwynar, L. C., Johnsen, S., continental shelf. In Ehlers, J. & Gibbard, P. L. (eds.): Quaternary Knudsen, K.-L. & Wohlfarth, B. & INTIMATE group 1999: Iso- Glaciations – Extent and Chronology, 271–294. Elsevier, Am- topic ‘events’ in the GRIP ice core: A stratotype for the Late Pleis- sterdam. tocene. Quaternary Science Reviews 18, 1143–1150. Murray, A. S. & Wintle, A. G. 2000: Luminescence dating of quartz Wohlfarth, B. 2009: Ice-free conditions in Fennoscandia during using an improved single-aliquot regenerative-dose protocol. Ra- Marine Oxygen Isotope Stage 3? Swedish Nuclear Fuel and Waste diation Measurements 32, 57–73. Management Company (SKB) Report TR-09-12, 1–48. Paper III

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 believed, 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 dating), 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 glaciolacustrine 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 geochronological 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 accompanying 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 1981, 2003, 2010, Larsen et al. 1987), Hattfjelldal interstadial I (~39 to 34 ka, cf. Ålesund and Sandes Stratigraphic and chronologic study of sub-till interstadials at the western coast, Andersen et al. sediments from many sites throughout Norway has 1981, Mangerud et al. 1981, 2003, 2009), revealed that the Scandinavian ice sheet behaved in Hattfjelldal interstadial II (~32 to 29 ka, cf. two a much more dynamic manner than previously dates of bones representing the Hamnsund intersta- believed (Olsen 1997, Olsen et al. 2001a,b,c, dial at the western coast, Valen et al. 1996), and 2002). Four interstadials during the Middle and Trofors interstadial (~25 to 20 ka, cf. Andøya in- Late Weichselian glaciation mark periods of major terstadial at the northwestern coast, Vorren et al. ice retreat, and sub-till sediments from these peri- 1988) with intervening stadials bracketing the Tro- ods have been identified and dated at many sites fors interstadial termed LGM 1 and LGM 2 (Last throughout Norway, indicating near-synchronous Glacial Maximum; Fig. 2). The work by Olsen and behaviour of the western margin of the ice sheet colleagues represents a significant leap forward in (Figs. 1 and 2). These interstadials are the Middle our knowledge, and has important implications for Weichselian pre-Hattfjelldal interstadial (~55 to 45 our understanding of ice sheet dynamics and pa- ka, cf. Bø and Austnes interstadials at the western leoglaciology, vegetation dynamics, and the rela- coast, Andersen et al. 1981, 1983, Mangerud et al. tionships between climate change, sea level

1 Timothy F. Johnsen, Lars Olsen and Andrew Murray

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 former 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).

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

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 Fig. 2: Frequency distribution of uncalibrated radiocarbon dates (running mean ages ±l) of organic-bearing sediments The dynamics of the Scandinavian ice sheet have and shells from four ice-free intervals separated by glacial been explored by a number of workers (e.g., Sejrup episodes during the Middle to Late Weichselian. The sug- et al. 1994, 2000, 2009, Arnold et al. 2002, Hou- gested names, the number of dates (N) and their mean age (n in 14C ka) for each ice-free interval are indicated. The differ- mark-Nielsen and Kjær 2003, Mangerud et al. ence in mean ages between these groups of dates is significant 2004), and the work by Olsen and collaborators at a 99% confidence level (based on data presented in Olsen et (Olsen 1997, Olsen et al. 2001a,b,c, 2002) is par- al. 2001b). See text for corresponding calibrated ages. Figure from Olsen et al. 2002. ticularly relevant to this study. Based upon the regional Quaternary stratigraphy, using data from more than 50 sections from own studies and sev- change, physiography, and Earth rheology. Knowl- eral others, more than 300 dates (of which ~120 edge of very dynamic ice-margins in the past is were radiocarbon dates taken from own projects), especially pertinent given recent observations of fossil content and some paleomagnetic data, glacia- ongoing rapid but episodic glacier acceleration and tion curves have been constructed for nine tran- thinning from marine-terminating sectors of the sects from inland to shelf covering the latitudinal Greenland and West Antarctica ice sheets (Shep- range of Norway (1800 km; Fig. 1). Comparison of herd and Wingham 2007) and their possible im- these curves, including statistical analysis of the pacts on our environment. The landforms and datings, indicates that the Middle to Late Weich- sediments of the former Scandinavian ice sheet are selian glaciation in Norway was punctuated by four accessible for study, and its western marine-based major ice recessions (Fig. 2; details presented margin can provide an analogue to the rapid-paced above). These interstadials correlate well with fluctuations and deglaciation observed at the mar- peaks in the ice-rafted debris record. This recon- gins of contemporary polar ice sheets. struction of ice oscillations is based mainly (~60%) Despite the importance of the work by Olsen on AMS radiocarbon dates of glacial sediments and colleagues it has not been fully embraced by with low organic carbon content (0.2-1.5%), which the scientific community because it depends mostly gave promising results with respect to accuracy (~60%) on results from AMS radiocarbon dating of and precision (Olsen et al. 2001c). As the age esti- glacial sediments with low organic carbon content, mate of the oldest of these interstadials is close to a method which has some inherent uncertainties the age-limit of the radiocarbon technique, we will (Olsen et al. 2001c) and consequently has been met for the remaining part of this paper only deal with by some scepticism (e.g., Mangerud 2004). An- the three younger interstadials. The rapid and other reason is that results from this work represent rhythmic ice fluctuations, as reconstructed in this a significant departure from earlier views of a more model, have been fairly synchronous in most parts stable ice sheet particularly during the Late Weich- of Norway. Given the regional synchroneity and selian (Denton and Hughes 1981, Mangerud 1991, rapidity of glacial fluctuations it is suggested that 2004, Donner 1995, Ehlers 1996). Given the impli- in addition to precipitation, the mountainous fjord cations and controversy of this work it is important and valley topography, glacial isostasy and relative to assess it further, and the optically stimulated sea level changes were probably more important luminescence (OSL) dating technique has devel- for the size of the glacial fluctuations than were air oped to a point that it can be confidently applied in

3 Timothy F. Johnsen, Lars Olsen and Andrew Murray

temperature changes (Olsen et al. 2001a; cf. Ar- and in the inland areas. The ice advances along the nold et al 2002, Hubbard et al. 2009). main fjords were supported by ice flowing from Does the dynamic ice model described by Olsen most tributary valleys and fjords. Consequently, in et al. (2001a) allow for realistic ice retreat and this model, the average rate of advance for the ice advance rates? Ice retreats of 200-400 m per year sheet moving from inland to coast may have been and ice advances of the same rate would be much very rapid, possibly up to 200-400 m per year more than needed to fit with the age model for the along fjords, although each contributing glacier described Middle to Late Weichselian glacier fluc- may have advanced at much slower rates, less than tuations. If Norway topographically had a tabular 200 m per year. The west-east distance between the form with no deep fjords and valleys and no high outer coastline and the Langsmoen and Flora sites mountains, then such rates for ice retreats and ad- is ~110-113 km, and with an ice retreat rate as high vances would seem to be extremely high and quite as 200-400 m per year the ice margin could have unrealistic. However, Norway is highly dissected retreated from the coast to Langsmoen and Flora with deep fjords and valleys and with high moun- within less than 300-600 years. This rate of ice tains between. This means that the ice retreats from retreat, and a similar rate of ice advance, may be the coast and landwards occurred mainly through slowed down considerably, to half size or less, and calving along the fjords, which generally is a very the glacial system would still be very dynamic and rapid process, and the ice advances are initiated still in agreement with the model described by from numerous ice growth areas between the fjords 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.

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

Fig. 4: Location of the Flora, Langsmoen ad Drivvollen sites in the Nea River valley, with indication of the main regional ice flow direction and the position of a damming ice barrier in the valley at some point in the history of deposition of sediments at Flora. characterized as robust and not sensitive to small which drains into Selbu Lake which in turn drains and moderate changes in ice retreat and ice ad- into the Nid River which flows to the port city of vance rates. 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 Methods down-valley at 63° 8' 5"N, 11° 15' 29"E, and the Drivvollen section is located midway between Stratigraphy and sedimentology Langsmoen and Flora sections. The watershed In this study we have focused on three stratigraphic divide for the headwaters of this valley, and the sites, the Flora, Langsmoen and Drivvollen sec- Norway-Sweden border is located ~50 km up- tions, within the Selbu region of central Norway valley from the sections. (Figs. 1, 3 and 4) approximately 120 km east- The sections at the Flora site were studied for southeast of the city of Trondheim. The sections the first time in 1997 and at Langsmoen and Driv- are located in the Nea River valley (relief ~300 m) vollen three years later. The sections of the Flora

5 Timothy F. Johnsen, Lars Olsen and Andrew Murray

and Langsmoen sites were exposed by sliding and Dating, where the OSL measurements were also removal of surficial slope material due to heavy completed. rain which followed a short time after removal of The samples were analyzed using large (8 mm) trees and other slope-stabilizing vegetation. The and small (2 mm) aliquots of quartz, ranging in sections at Drivvollen were exposed by road exca- mean grain size from 83 to 170 m, on Risø vations. Sediments were studied using standard TL/OSL-readers equipped with calibrated 90Sr/90Y observational techniques – sediment texture and beta radiation sources (dose rate 0.089-0.32 Gy/s), structure, paleoflow direction indicators, bed con- blue (470±30 nm; ~50 mW/cm2) and infrared (880 tact relationships and lateral continuity and thick- nm, ~100 mW/cm2) light sources, and detection nesses of units – and lithofacies were identified and was through 7 mm of U340 glass filter (Bøtter- interpreted based on their characteristics and asso- Jensen et al. 2000). Analyses employed post-IR ciations, following sedimentologic principles (e.g., blue SAR-protocols (Murray and Wintle 2000, Middleton and Hampton 1976, Allen 1982). 2003, Banerjee et al. 2001), adapted to suit the samples based on dose recovery and preheat ex- Sampling, preparation and measurement periments (preheat 260° for 10 s, cut-heat 220°). The OSL dating technique is used to estimate the 100 s of illumination at 280° between cycles im- time elapsed since buried sediment grains were last proved recuperation (response to zero dose). Some exposed to daylight and is a reliable chronological of the small aliquots were measured using a short- tool, reflected in its growing popularity in Quater- ened IR blue SAR-protocol to increase measure- nary studies (Murray and Olley 2000, Murray and ment productivity. The shortened protocol had Wintle 2000, Duller 2004, Lian and Roberts 2006, three regenerative doses, selected to bracket the Wintle 2008). equivalent dose. The third dose was a recycled Five OSL samples were collected in August dose, and since recuperation rarely was greater 2005 and June 2007 for the Langsmoen and Flora than 5% using the full SAR-protocol, we excluded sections at the levels indicated in Figs. 5 and 6. The the zero dose measurements from the shortened sites were overgrown with herbaceous and shrub protocol. vegetation since the original detailed stratigraphic Dose rates were calculated from gamma spec- work was completed but the stratigraphic units trometry data (Murray et al. 1987) and included the could nonetheless be identified securely. In addi- cosmic ray contribution (Prescott and Hutton tion, a sample of young fluvial sediment was col- 1994). Water content (weight percent) was meas- lected from the riverbank at the mouth of the Gaula ured in the laboratory by weighing re-compacted River at Øysand (Fig. 3; 63° 20’ 27”N, 10° 13’ sediment at natural, saturated and dry water con- 44”E; ~13 km south-southwest of Trondheim; at a tent. The historical water content was estimated to depth of 25 cm) to evaluate if the OSL signals from be 0.75 the saturated water content. This seems fluvial sediments in this region are adequately reasonable given that the sections are located on bleached (i.e., zeroed). These sediments are ex- valley walls that are subject to significant ground- pected to be less than 1 ka old and to have been water flow during parts of the year. Also, sedi- transported at least 5 km from a landslide area ments in these sections have been observed at vari- upstream (Sveian et al. 2006). ous times during spring and autumn to be quite The samples were collected using opaque plas- wet, although their locations are well above the tic tubes and stored in an opaque box until opened local groundwater level and the sediments appear under darkroom conditions at the Nordic Labora- to be quite dry during the summer season. tory for Luminescence Dating (2005) or Stockholm Samples for radiocarbon dating from Flora were University (2007) where samples were apportioned taken from positions well inside the fine-grained and wet sieved. Final preparation, including heavy units and therefore well away from potentially liquid separation (2.62 g/cm3) to remove feldspars, contaminated surfaces. Water flow through the treatment with 10% HCl for 5-30 minutes, 10% sediments was not a major problem for the radio-

H2O2 for 15-30 minutes, 38% HF for 60-120 min- carbon dates, because the water flow goes mainly utes and 10% HCl again for 40 minutes was per- through the adjacent coarse grained sediments and formed at the Nordic Laboratory for Luminescence may therefore contaminate the boundary surfaces

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

Fig. 5: Generalised stratigraphic log from Flora, based on several sections which are partly separated, both in height and width, but located within a 200m by 300m area in the valley side (see Fig. 4). Shown are the location of eight radiocarbon samples and calibrated 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.

7 Timothy F. Johnsen, Lars Olsen and Andrew Murray

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.

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

of fine-grained sediment units, but probably very unity. The first step was introduced to address little inside the fine-grained units. As well, there aliquots with poor recycling ratios that would be were no visible traces of oxidation, root penetra- accepted because they had unacceptably large un- tion, or fissures within the sampled sediment. certainties (~>30%). Aliquots also had to pass the Samples were air dried and stored in a cold room following rejection criteria: recuperation <5%, with a temperature of 0-4°C to avoid bacterial equivalent dose error <50% and signal more than contamination (e.g., Wohlfarth et al. 1998). Sam- three sigma above the background. Decay and ples were measured at the Accelerator Mass Spec- growth curves also had to be regular. More conser- trometry (AMS) facility at the University of vative rejection criteria values were used for the Utrecht, which uses a 6 MV Van de Graaff tandem shortened protocol employed for some small ali- accelerator (van der Borg et al. 1997). Graphite quots; they were scaled appropriately based upon targets were used for the AMS analysis. A standard the actual structure of the sequence used. Ages for alkali-acid-alkali treatment was used to remove all samples were calculated using the mean and contamination by humic acids and carbonate. After median of the equivalent dose population of ac- combustion of the organic carbon samples, the cepted aliquots for each sample, as well as using obtained CO2 gas was routinely separated from the dry and saturated water contents. SO2 with KMnO4 (see Olsen et al. 2001c and refer- A sensitivity analysis was also performed to de- ences therein). The insoluble fraction was used to termine quantitatively which uncertainties have the derive the sample age. At Langsmoen it was more largest effect on the age. Using reasonable esti- difficult to find fine-grained units that were thick mates of uncertainty for each parameter, ages were enough to avoid a potential contamination prob- recalculated after adjustments were made to each lem, and therefore we have no radiocarbon dates of the parameters in turn. The estimated uncertain- from this site. ties (1 standard deviation) used for each parameter As we were calibrating radiocarbon dates (i.e., in these calculations were: depth below surface ±1 interstadial dates) from the literature that were well m, elevation ±50 m, grain size ±20%, water content beyond the calibration curve IntCal04 (Reimer et (gamma and beta) ±20%, dose rate gamma ±5%, al. 2004), we used Fairbanks et al. (2005) with the dose rate beta ±5%, internal dose rate ±30%, den- calibration curve Fairbanks0107 for all radiocarbon sity ±10%, cosmic ray contribution ±5%, and beta dates. Note that calibrations using IntCal04 versus source calibration ±2%. Fairbanks0107 of radiocarbon dates from Flora section produced calibrated ages differing by less than one percent. Results

OSL data analysis The Flora, Langsmoen and Drivvollen stratigraphic sites Simple component analyses of the continuous wave OSL data from selected aliquots was under- The Flora site taken using SigmaPlot 10.0, based on the parame- The Flora site consists of several sections within a ters and formulas of Choi et al. (2006). The results 200 m by 300 m area in the valley side (Fig. 4). of the component analysis are discussed further The four main sections are situated 337-340 m below, but based on such information from dose a.s.l., 301-310 m a.s.l., 270-280 m a.s.l., and 252- recovery measurements we chose the 0-0.64 s and 265 m a.s.l. (Fig. 5 and Table 1). The lowest-lying the 0.96-2.08 s portions of the signals for peak and of these was later extended to bedrock by excavat- background integration limits for all aliquots. The ing mainly through a boulder-rich lower till and a equivalent doses were then calculated in Risø Lu- coarse-grained diamicton, probably a till, with minescence Analyst 3.24 (using exponential curve heavily weathered material at the base. The vertical fitting) and in Microsoft Excel. intervals between the main sections have not been Aliquots were initially accepted when (1) the fully excavated and cleaned, but have been absolute value of the recycling ratios was within checked with the help of a spade and a 1 m-long the range 0.7 to 1.3, and (2) when accounting for drilling rod with a 1 inch-wide track for sediment their uncertainty their value was within 15% of sampling. This is done to make sure that no inter-

9 Timothy F. Johnsen, Lars Olsen and Andrew Murray

Table 1: Sedimentary diamict and other lithofacies used in Figures 5 and 7. Modified from Olsen et al. (1996). Code Facies description Interpretation Dcs Clast supported diamicton, stratified, occurrence often in combination with S-Ablation melt-out till and poorly sorted glacifluvial debris flow/ debris fall , G-, sDm-, gDm-, and Dms-facies. material. Dmm Matrix supported, massive, normal silty-sandy composition. Variable clast Till or debris flow deposits (and also some glacimarine sediments). fabric from random to strong parallel or occasionally transverse. 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 Mainly lodgement and basal melt-out till. parallel clast fabric. Dms Matrix supported, stratified diamicton, with sand lenses and linings around Mainly basal melt-out till, but often combined with intervening or alternating clasts. Medium to relatively strong parallel clast fabric, transverse fabric may zones of lodgement till. occur in some areas. Dms(r) Dms with evidence of resedimentation, for instance with rafts of deformed Mainly basal melt-out, but often combined with zones of lodgement till. sand/silt/clay laminae and abundant silt/clay stringers and rip-up clasts. Clasts fabric variable from random to fairly strong. sDm Sand-enriched matrix supported diamicton, mainly massive (sDmm), but Tillised sediment/deformation till containing sand eroded from local substrate; occasionally heterogeneous or stratified (sDms). and melt-out till. gDm Gravel-enriched matrix supported diamicton, mainly massive, but Mainly melt-out till, but in some cases possibly also tillised sediment or occasionally heterogeneous or stratified. deformation till with gravel eroded from local substrate. cDm Cobble- and boulder-enriched matrix supported diamicton, mainly massive, Mainly melt-out till, but in some cases possibly also tillised or deformed but occasionally heterogeneous or stratified. sediment reworked from local substrate. FSi Fine silt Vertical accretion by settling of sediment from suspension in quiet water lakes Si Silt (glaci-/lacustrine) and marine basins. This include FSi, Si, CSi, FS, and often CSi Coarse silt also S/SSi facies. FS Fine sand S/SSi Sand/ Sand, silt Lateral accretion, sediments supplied by traction with fallout, glaci-/fluvial and CS Coarse sand marine coarse-grained sediments. This include S/SSi, CS, and G/GSi/GS facies. G/GSi/GS Gravel/ Gravel, silt/ Gravel, sand D 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.

vening main sediment unit is missing in the strati- deformation, occur also in unit E3. This unit is graphic record. overlain by a brownish-grey till of, at least, 10 m The ~25 m high lowermost section starts with a thickness. It is separated in three parts, D1, D2 and 1-4 m thick coarse-grained gravelly sandy diamic- D3, by intervening silt and sand horizons E1 and ton, which is inferred to be a till, at the bedrock E2. The matrix of tills D1, D2 and D3 changes between ~240 and 245 m a.s.l. (Fig. 5; unit G). The with a general fining-upwards trend from gravelly bedrock is mica schist, with a glacially striated sandy in till D3 to sandy in till D2 and more silty in surface partly preserved. The striation indicates ice till D1. The compactness of these tills increases as flow towards 280°. In other parts the bedrock is the matrix changes upwards to a more fine-grained weathered and some of the weathered material has character. Overlying these tills follows again sev- been reworked in the till. Overlying the lowermost eral metres (50-60 m) of till in beds (subunits B1, till follows a 7-8 m-thick silty, sandy, compact, B2 and B3) alternating with laminated sand and silt brownish-grey to greenish-grey till (unit F) with (subunits C1, C2 and C3) of thicknesses from a many cobbles and boulders. Clast fabrics in these few centimetres to one or several metres. Tills B1, two lower tills indicate preferred maxima between B2 and B3 are all relatively compact and mainly 280° and 305°, which correspond with striation on silty and dark grey in colour, with a bluish-grey bedrock as well as on a big boulder in the contact tone in most parts of tills B1 and B2. zone between the tills. Overlying these tills follows The Flora site is completed on top by a 3 m- unit E3 (Fig. 5), which is a 4.5 m-thick series of thick sequence of two coarse-grained, brownish- alternating layers of laminated sand and silt in a grey and medium to loose compacted tills (Fig. 5; number of fining-upwards sediment successions. unit A). Full fabric analyses have not been per- Thin irregular layers of diamict material or grav- formed in the tills younger than D3, but the orien- elly sand, as well as some indications of glacial tation of some 15-20 clasts and a few fold-axes

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

Table 2: Flora section AMS radiocarbon ages of sub- tively. Unit D could alternatively be described as a till waterlain sediments (from Olsen et al. 2001c) . four fining-upwards sequences, with other bounda- Lab no. Elevation 14C age Calibrated age ries for each sub-unit than illustrated in Fig. 6. The (m) (yr BP) (yr BP)b gravel and sand are deposited in layers sloping in

1 UtC 5977 309 17800 ±400 21100 ±1000 the downstream direction towards the north. De- 2 UtC 5978 309 15920 ±260 19100 ±400 formation structures like sediment injections and 3 UtC 5979 305 17800 ±400 21100 ±1000 folding are increasing in size and frequency up- 4 UtC 5981 273 16700 ±220 19800 ±600 5 UtC 5982 272 15620 ±200 18800 ±400 wards in the upper three subunits D3-D2-D1, with 6 UtC 5984 260 19600 ±280 23400 ±800 only small deformation represented in the bounda- 7 UtC 6042 255 19050 ±120 22600 ±200 8 UtC 5985 253 18000 ±400 21400 ±1200 ries of subunit D3, whereas overturned layers of mean = 20900 sand dipping steeply, almost vertically towards the stdev = 1600 northwest occur in subunit D1. The silt in subunit

Notea: All were measured using the insoluble fraction. D3 contains foraminifers of e.g. Cibicides lobatu- Noteb: Calibrated using Fairbanks et al. 2005. Note that calibrations lus and Elphidium excavatum (Olsen et al. 2002). 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 All the C. lobatulus foraminifers indicate some at 1. 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- from glacial deformation at three levels in the up- sion and may belong to the last phase of deposition per 65 m of the Flora stratigraphy have helped to and, therefore, indicate marine influence in these indicate approximate ice flow directions (stippled sediments. arrows in Fig. 5). All these are trending towards Overlying the waterlain sediments, which are the NW. thought to be fluvial deltaic in origin, follows two Radiocarbon dating of bulk organics from the till beds with less than 1 metre of sandy slope ma- silty sediments was performed at eight levels from terial on top (Fig. 6). The lower till, unit C is rela- five of the silt and sand horizons between 309 and tively compact, with a sandy-silty matrix and some 252 m a.s.l. (Fig. 5). The mean and standard devia- boulders, cobbles and pebbles, and a brownish- tion of the ages from Flora are 20.9 ±1.6 cal. ka BP grey colour. The upper till (unit B) is more silty, (Table 2). Also, based on microscopic and stereo- darker grey to bluish-grey and very compact. Till scopic examination of parallels to most of the dated fabrics indicate preferred clast orientation towards bulk-organic sediment samples, terrestrial plant the NW and NNW for these tills, quite similar to remains are the main organic component for these the regional ice flow direction based on striation samples. OSL results for the Flora and Langsmoen and drumlin orientation (Figs. 4 and 6). sites are presented in detail in section OSL characteristics and ages. The Drivvollen site Fresh road cuts along a local road at Drivvollen The Langsmoen site (Fig. 4; Olsen et al. 2002), midway between A 15-17 m high section which was comprised by Langsmoen and Flora, revealed in the year 2000 a several 1-2 m wide and 2-3 m high sections that stratigraphy comprising both the till units recorded were overlapping in height and width is represent- at Langsmoen and an underlying glacifluvial/ flu- ing the Langsmoen site (Fig. 6). The lowermost vial/ fluvial marine gravelly and sandy sediment sedimentary unit here, unit D, is divided in four unit that may correlate with the Langsmoen sub-till subunits D4, D3, D2 and D1, where D4 is mainly deltaic sediments. At Drivvollen, these sediments gravel and sand, and D2 is gravel. Subunits D3 and reach up to 215-220 m a.s.l. (Figs. 4). D1 are silt and sand with some silt-layers, respec-

11 Timothy F. Johnsen, Lars Olsen and Andrew Murray

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.

Water content (weight %) Sample Elev (m) Sampling Lithofacies sampled Saturated Selected a Mean grain depth (m) size used

071302 210 8.0 Deformed, laminated sand 34.7 26.0 85 081357 200 18.0 Diffusely graded sand with silty clay rip-ups 33.2 24.9 94 081358 208 10.0 Deformed, laminated sand 31.0 23.3 122 081373 210 8.0 Deformed, laminated sand with massive, matrix-supported diamicton 24.2 18.2 170 071301 252 35.0 Deformed, diffusely graded sandy silt 17.0 12.8 83

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

OSL characteristics and ages indicate that the selected time intervals for peak and background integration resulted in a signal that The major sample properties, settings and results was dominated by the fast component. Doses cal- are listed in Tables 3 and 4, and Figs. 7 to 11. The quartz from Langsmoen and Flora sections is in- culated with this peak and background selection gave good dose recovery (1.06 ±0.03, n = 18, un- sensitive with photon counts less than 1000 for the certainty as the standard error; Fig. 9) and demon- first 0.08 s of stimulation (Fig. 7, insets), which leads to larger uncertainties (cf. Alexanderson and strate that the SAR protocols used were able to accurately recover a known dose given before any Murray in press). Careful selection of peak and heating. The signals from the Flora section were background channels was made (0-0.64 s and 0.96- 2.08 s portions, respectively) to best isolate the fast close to saturation, from the flattish part of the extended growth curve, while signals from the component (Fig. 8). For the eighteen accepted dose Langsmoen section were from the steep, near- recovery experiments the average proportion of the net component signal to the total net signal used linear, early part of the extended growth curve (Fig. 7). This has implications for the reliability of for dose calculation was: 95 ±6% (fast), 5 ±5% the final ages, see Discussion. (medium), and 0.3 ±0.6% (slow). These results

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

Age

Sample Aliquot nDose rate Estimated Recycling Mean Median Drya Saturated Skew 2 x SE of size (mm) (Gy/ka) dose (Gy) ratio 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 1.04 071302 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 0.60 081357 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 0.96 081357 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 0.70 081358 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 1.15 081358 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 1.10 081373 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

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.

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

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 selection effectively isolated the fast component and the selection of peak and background channels provided the optimum recycling and dose recovery.

sample is 165 Gy and 1.6 Gy/ka respectively (Ta- Fig. 7: Single-aliquot regenerative-dose (SAR) OSL growth ble 4). Recycling is exceptionally good with a curves for Flora section sample 071301, and Langsmoen section samples 071302 and 081358; first two samples are mean and standard deviation for all samples of average of two aliquots with uncertainties as standard devia- 1.02 ±0.04, n = 8. The age of the Flora section tion. The range of estimated doses used to calculate the age of sample is 102 ka. The seven ages from large and each sample is indicated by the width of the grey shading; the small aliquots from the Langsmoen section range mean equivalent dose is also shown (black triangle). Note that equivalent doses for 071301 (from the Flora section) have a between 25.4 to 20.6 ka with a mean and standard large range and are from the flatter part of the growth curve. deviation of 22.3 ±1.7 ka. The ages from the large Repeated dose points (open squares) and a zero dose point and small aliquots for the same sample are similar (open circle) are also shown. The inset shows the natural decay and overlap at 1. With the exception of the Flora curve; the decay curve for the young sediment sample from Øysand (081360) is also shown in inset for 081358 (sediment section sample, the distributions of the equivalent dose rate of 1.7 Gy/ka). doses used to calculate the age are not significantly skewed (Table 4, see footnote). The sensitivity analysis demonstrates that the calculated age is For the Langsmoen section samples, equivalent most sensitive to changes in the equivalent dose, doses range between 34 and 51 Gy and dose rates beta source calibration and dose rate, followed by between 1.5 and 2.0 Gy/ka, while for the Flora water content (Fig. 11).

13 Timothy F. Johnsen, Lars Olsen and Andrew Murray

no marine fossils have been found in these sediments. The radiocarbon datings, with a mean and standard deviation of the ages of 20.9 ±1.6 cal. ka BP (Table 2), indicate that all the glacial and glaciolacustrine 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 bracketed 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 Fig. 9: Histogram of dose recovery tests, excluding aliquots phases of glacial advances during the last glacia- that did not pass rejection criteria. As the mean is close to unity, this indicates that the samples can be used for OSL- tion. The coarse-grained lower till and the fine- dating with our analytical protocols. 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, bluish-grey tills, respectively, at Flora. However, the Discussion waterlain, sub-till sediments at these two localities are different, with ice-free conditions at Stratigraphy and paleoenvironment Langsmoen and ice proximal and ice-damming The Flora site conditions represented at Flora. These different environments may have occurred during the same The interpretation of the Flora stratigraphy is that it interval, and even possibly with ice temporarily represents a succession of repeated glacial ad- acted as a damming barrier between the localities vances alternating with repeated phases of ice- (Figs. 4 and 12). The content of plant remains and damming with pulses of sand and silt flowing into other organics is, however, not sufficiently high in lateral or subglacial glaciolacustrine basins. All the the Langsmoen sub-till sediments to allow testing waterlain sediments (E and C subunits) are thought of this hypothesis by radiocarbon dating. The rela- to represent the same sedimentary environment, a tive sea level during deposition of the Langsmoen glaciolacustrine/ ice-dammed lake environment. waterlain sediments was higher than 210 m a.s.l., For example, the 4.5 m-thick series of alternating which corresponds with the elevation of the eroded layers of laminated sand and silt in a number of top of the sub-till deltaic foreset beds at fining-upwards sediment successions that comprise Langsmoen. The relative sea level was probably at unit E3 are thought to represent repeated melt- least 220 m a.s.l., in agreement with the elevation water-pulses into an ice-dammed lake environ- of the top of similar and possibly correlative fluvial ment. Small amounts of plant remains in the sedi- beds, also with marine influence at the neighbour- ments indicate that at least some vegetation may ing Drivvollen site (Fig. 4, Drivvollen; Olsen et al. have been established during the phases of ice 2002). This is ~10% higher than the late- retreat. The relative sea level during deposition of /postglacial marine limit in this area. the Flora waterlain sediments is not known, except that it was probably lower than 252 m a.s.l. since

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

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, and saturated water content. Error bars represent one standard error. The mean of the ages is 22.3 ±1.7 ka (uncertainty as standard deviation 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.

Ice flow directions which is in agreement with the main regional ice During initial ice growth and final decay of the last flow direction (Figs. 4 and 6). ice sheet, the ice flow has followed the main trends The general glacial stratigraphy model of most of the valleys and fjords. In the region east of valleys in Norway is characterized, not only with Trondheim (Fig. 3) this resulted in a dominating more of local materials, including more weathered ice flow direction during these phases towards the material in the lower tills, but also often with more west and northwest. During maximum phases, the coarse-grained material. The general increase of ice sheet was thicker and the ice flow was less fine-grained material in younger tills is thought to influenced by the underlying topography, and was derive from more long-transported material which trending towards the northwest and north- has been crushed and abraded during a longer northwest in this region. transport history. During deglaciation the youngest Westerly to northwesterly trending ice flow di- tills may again include more coarse-grained mate- rections are inferred from striation on bedrock and rials, which may derive from ice-margin oscilla- boulders at Flora and till fabrics in the lower tills tions and a changed pattern of erosion, transporta- and deformation in the sub-till sediments at both tion and deposition. The total Flora glacial strati- Flora and Langsmoen. These ice flow directions graphy shows this trend, and therefore seems to be are in best agreement with phases during ice in concert with the general model. Therefore, it growth or final decay of the ice sheet. Till fabrics seems likely that the lower tills have been depos- indicate that the fine-grained, dark bluish-grey ited during ice growth, whereas the younger tills upper tills at both sites may be a result of a more derive from later phases. Of course there may be northwest to north-northwesterly trending ice flow, exceptions from the general model in some loca-

15 Timothy F. Johnsen, Lars Olsen and Andrew Murray

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. tions, but to really test this hypothesis carefully at plete bleaching, as it is not reasonable to expect each site requires much more detailed work. sediments from different portions of the section to suffer the same degree of incomplete bleaching to OSL and ages produce similar ages. Also, if there were bleaching Of primary importance in employing OSL dating problems we would expect measurements on small in glacial and glacially-related sediments is to de- aliquots to produce positively skewed dose distri- termine whether the sediments were adequately butions or at the very least a significantly younger exposed to light prior to deposition; otherwise the age (Fuchs and Owen 2008, and references ages will be maximum ages (Murray and Olley therein); neither is the case (Table 4), and the ages 2000, Fuchs and Owen 2008, Alexanderson and from large and small aliquots from the same given Murray in press). The good correspondence in age sample overlap within 1. In addition the very low between samples, between large and small aliquots, OSL signals from young fluvial sediments at Øy- between mean and median values, and the insig- sand indicate that fluvial sediments in this region nificant skewness strongly indicate that the can be properly bleached (see sample 081360 in Langsmoen sampled sediment was adequately bottom inset of Fig. 7). exposed to light (i.e., “well-bleached”, or “zeroed”) Based on that sample 081373 contained some prior to deposition (Table 4, Fig. 10). This also diamict sediment, which may have been derived justifies the use of the mean age and not for exam- from subaqueous debris flow or till tectonized into ple a minimum age model for the small-aliquot the fluvial sediments, there is a risk for incomplete data (Galbraith et al. 1999). Besides, it is very bleached grains to be incorporated. This sample is unlikely that all the samples suffer from incom- also slightly older than the others (Table 4, Fig.

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

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

10). As we measured only large aliquots for this lowest equivalent dose for this sample gives an sample, the dose distribution unfortunately cannot apparent age of about 45 ka, which is well beyond be used to evaluate skewness-effects of incomplete the 21 ka (n = 8) radiocarbon age of the Flora bleaching (as there is a large averaging-out effect sediments. Thus it is very likely that the glacio- for large aliquots; Duller 2008). Nevertheless, on lacustrine sediment from Flora is poorly-bleached the whole we conclude that the Langsmoen sam- and therefore we reject its OSL age (Wallinga ples were well-bleached before deposition. 2002). In contrast, the Flora section sample gave an er- The sediment dose rates for the various samples roneously old age based on comparison to the eight (1.8 ±0.2 Gy/ka) were quite consistent and suggest radiocarbon dates (Fig. 5). This is anticipated as it that there is not a significant problem with the is not expected that glaciolacustrine sediments sediment dose rates measured (Table 4). Finally, would be adequately bleached prior to deposition the water content history of the sampled sediment (Alexanderson and Murray in press), especially if can only be estimated. As stated earlier, our best the ice-front was nearby or if sediments were de- estimate is that over their lifetime the sediments rived from a submerged ice-front. Unlike the well- experienced on average ~75% their saturation. bleached Langsmoen sediments, the equivalent Despite this limitation the sensitivity analysis indi- dose population for the Flora sediment is very cates that even with a 20% change of the percent broad and from the flattish, high dose, portion of water content values, ages change by less than ~3% the growth curve (Figs. 8 and 10, Table 4). Note (Fig. 11). that many aliquots gave equivalent doses well The radiocarbon dates from the Flora section beyond the highest regenerative dose as well, and are considered to be reliable as (1) sampled sedi- were not included in the determination of the ments were derived from the interior of laminae or equivalent dose; thus the dose population would be beds that were adjacent to more hydraulically con- even wider than portrayed in Figure 8. Even the ductive sediments, thus minimizing possible

17 Timothy F. Johnsen, Lars Olsen and Andrew Murray

groundwater flow and contamination through the during the Andøya – Trofors interstadial was sampled sediment, (2) the ages are consistent over probably cold to boreal (Vorren et al. 1988, Alm the 54 meter range of the section, which is unlikely 1993, Olsen et al. 2001b) with arctic to subarctic to occur if groundwater contamination was an conditions where the mollusc Mya truncata, which important process, (3) the ages fit well with similar is a temperate Atlantic water indicator, reached dates from other sites throughout Norway that northwards at least to the southern coast of Norway belong to the Trofors interstadial, and finally (4) (Olsen and Bergstrøm 2007). the ages are in agreement with the OSL ages from The geographical extensive distribution of Tro- sub-till sediments from the nearby Langsmoen fors interstadial sediments, as well as that for the section. Terrestrial plant remains, which appear to Hattfjelldal I and Hattfjelldal II interstadials (Fig. be the main organic constituent of the dated mate- 3), indicates that causes behind these interstadials rial, are considered to be reliable for radiocarbon were regional in nature. In addition, the very dy- dating, and are supported by the agreement be- namic ice sheet fluctuations exceed rates of change tween the radiocarbon and OSL ages. that could be brought on by air temperatures changes alone. Thus, in addition to precipitation, Implications for ice sheet history and the mountainous fjord and valley topography, gla- dynamics cial isostasy and relative sea level changes were The seven OSL ages of the Langsmoen section probably more important for the size of the glacial sediments of 22.3 ±1.7 ka correspond very well to fluctuations than were air temperature changes. the Trofors interstadial of ~25.2 to 20.1 ka which is This statement is simply based on that the fjord and based on 40 radiocarbon ages including eight from valley topography, considerable glacial isostatic the Flora section of 20.9 ±1.6 cal. ka BP (Fig. 10, depression and high relative Middle to Late Table 2; Olsen et al. 2001b). Thus, the OSL ages Weichselian sea levels are all documented and from the Langsmoen section support the results of obvious important factors for ice sheet margin the AMS radiocarbon dating method on glacial fluctuations (e.g., Olsen and Grøsfjeld 1999), sediments with low organic content (Olsen et al. whereas air temperature changes may be balanced 2001c), and confirm the occurrence of the Trofors with precipitation so that the glacier fluctuations interstadial in central Norway which also has been are reduced, as seen from many modern glaciers. identified elsewhere in Norway (Fig. 1). The simi- Would the ice sheet fluctuations recorded along larity in age within and between these stratigraphi- the western margin of the Scandinavian ice sheet cally-related sites and using different geochro- also occur in other areas of the ice sheet? Argua- nological techniques strongly suggests that this bly, ice streams, especially widespread in Norway, area was ice-free around ~21 or 22 ka. played a very important role in the dynamics of the The ages of the Langsmoen and Flora section Scandinavian ice sheet (Fig. 1), especially as they sediments fall within the early to middle portion of would be sensitive to relatively small changes in the Trofors interstadial (Fig. 10). We imagine a relative sea level and would be very effective at paleoenvironment whereby the Langsmoen gla- removing mass from more interior areas of the ice ciofluvial sediments (at 200 to 210 m a.s.l.) were sheet and encourage retreat from both lowlands deposited in ice-distal conditions with local sea and high elevation (Shepherd and Wingham 2007, level at least 210 m a.s.l., possibly ~215-220 m Hubbard et al. 2009); this is reflected in the re- a.s.l. if the Langsmoen sediments correlate with the gional response of the ice sheet (Olsen et al. nearby Drivvollen sediments as mentioned before. 2001a). We suspect that the south-western areas of Further advance of the ice sheet leading into LGM the ice sheet in Denmark and Sweden may also 2 resulted in ice damming down-valley (Figs. 4 and have responded in a similar fashion to the ice mar- 12) and deposition of till alternating with glacio- gin in Norway (Hillefors 1969, 1974, Lundqvist lacustrine sediments in ice-proximal conditions at 1992, 2004) because like Norway the south- the Flora site with the highest lake level up to 310 western area of the ice sheet was adjacent to the m a.s.l. Further advance of the ice sheet resulted in influencing-effect of the sea and was host to ice the moderate glaciotectonizing of these sediments streams in the North Sea and Baltic Sea (Fig. 1; and their burial beneath lodgement till. The climate

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

Houmark-Nielsen and Kjær 2003 and references topography. For example, the Middle Weichselian therein). interstadial was longer lived than the Trofors inter- In Denmark, the Kattegat Till and Mid Danish stadial and consequently is more likely to have Till may correspond to the similarly dated LGM 1 been felt in the more central areas of the ice sheet, and LGM 2 tills in Norway; and the Vennebjerg- and sites at Pilgrimstad in central Sweden, Sokli, Sejerø-Klintholm interstadial, the Lønstrup- Ruunaa, and Hitura in Finland, several sites in Uranienborg-Gärdslöv interstadial and the Ribjerg- Skåne, southernmost Sweden, as well as numerous Fegge-Spøttrup-Glumslöv-Kobbelgård interstadial AMS-radiocarbon dated mammoth finds suggest in Denmark and SW Sweden correspond to the that the ice sheet was either very small or absent Ålesund-Sandnes-Hattfjelldal I interstadial, the during MIS 3 (Figs. 1 and 2, Arnold et al. 2002, Hamnsund-Hattfjelldal II interstadial and the Svendsen et al. 2004, Kjær et al. 2006, Ukkonen et Andøya-Trofors interstadial in Norway, respec- al. 2007, Helmens et al. 2007a, 2007b, Alexander- tively (Olsen et al. 2001b; Houmark-Nielsen and son et al. 2008, in press, Lunkka et al. 2008, Sa- Kjær 2003). While one could assume that higher lonen et al 2008, Wohlfarth 2010). In contrast to elevation areas within southern Sweden may have the Early and Middle Weichselian interstadials, the been influenced less by the draw of ice streaming, Trofors interstadial was of smaller geographical thoroughly dated sediments in southern Sweden extent, shorter-lived, and occurred in the most (Småland, part of the south Swedish highland) give responsive portion of the ice sheet (Figs. 1 and 2). OSL ages with a standard deviation range of the Nevertheless, the Trofors interstadial represents mean from 23.7 to 19.1 ka (Alexanderson and acknowledgement that the Scandinavian ice sheet Murray 2007, in press), and correspond very well was indeed dynamic and that we should be open to to the Trofors interstadial (Fig. 1). the possibility that the existing ice sheets on Earth The marine records include megascale linea- could act dramatically as well. tions that indicate ice-stream activity (fast-moving Recent robust modelling of the nearby British- wet-based ice constrained by slow-moving ice) Irish ice sheet has produced a very dynamic ice localized to many channels, troughs and elongated sheet with numerous binge/purge, advance/retreat basins crossing the continental shelf off Norway (i.e., yo-yo) cycles dominated by ice streaming (e.g., Ottesen et al. 2005). Also included are large (Hubbard et al. 2009); behaviour similarly docu- trough mouth fans comprised of numerous, stacked mented for Norway. The modelled behaviour dem- sediment wedges from debris flows produced at the onstrates alternating periods of relatively cold- ice margin during maximum ice extent (e.g., Vor- based ice, with a high aspect ratio associated with ren and Laberg 1997). These features, and the net growth, and wet-based ice with a lower aspect chronology based on dates of marine fossils in ratio characterized by streaming. Phases of pre- sediments that subsequently have been overlain by dominant streaming activity coincide with periods debris flow sediments, in several stacked marine- of maximum ice extent and are triggered by abrupt debris flow sediment successions, indicate a very transitions from a cold to relatively warm climate, unstable ice margin during the MIS 3 - MIS 2 in- resulting in major iceberg/melt discharge events terval (e.g., King et al. 1996, Nygård et al. 2005, (Hubbard et al. 2009). The fjord-dominated land- Sejrup et al. 2009). scape of Norway and its shelf are in no doubt a We appreciate that it may be difficult for some record of the dominance of ice streams in draining to accept that there was an interstadial during what the Scandinavian ice sheet (Longva and Thorsnes has been traditionally thought of as the Last Glacial 1997, Ottesen et al. 2005). Perhaps the Scandina- Maximum in Scandinavia (now a two part LGM vian ice sheet (at least its western margin) behaved separated by the Trofors interstadial; Olsen et al. in an even more dynamic manner than documented 2001a). Perhaps it is easier to accept by examining so far, similar to the British-Irish ice sheet. Argua- the dynamics of the ice sheet over a longer time bly the modelling has exceeded the practicalities of scale. Large ice retreats occurred after moderate to the field science as dynamic ice sheet conditions large ice advances during the Early and Middle mean that it is difficult to find sediments from Weichselian (Mangerud 2004). These retreats and earlier ice-free periods that have survived subse- advances stretched over vast areas of low relief quent, and multiple, glacial erosion events. Fortu-

19 Timothy F. Johnsen, Lars Olsen and Andrew Murray

nately in Scandinavia there are many sub-till sites divide to the ice margin, and likely including areas that have already been discovered and that can be in Denmark and southern Sweden, was very dy- studied further to verify the timing and location of namic. These large changes in the ice margin and ice-free periods and to improve our understanding accompanying drawdown of the ice surface would of the dynamic nature of ice sheets (e.g., Roberts- have affected the eastern part of the ice sheet as son and Ambrosiani 1992, Olsen et al. 2001b, well. These conditions have important implications Lundqvist and Robertsson 2002, Lokrantz and for our understanding of ice sheet dynamics and Sohlenius 2006). paleoglaciology, vegetation dynamics, and the relationships between climate change, sea level change, physiography, and Earth rheology. The Conclusion dynamic nature of the SIS allows us to better accept and understand the rapid-paced fluctuations OSL datings of Langsmoen section sub-till sedi- and deglaciation observed at the margins of con- ments from central Norway, using large (8 mm) temporary polar ice sheets, and imagine what is and small (2 mm) aliquots, give consistent ages of possible in the future. 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 Acknowledgements similarity in age within and between these stratigraphically-related sites and using different geo- We wish to thank Helena Alexanderson of Stock- chronological techniques strongly suggests that this holm University/Norwegian University of Life area was ice-free around ~21 or 22 ka. These posi- Sciences and Jan Lundqvist of Stockholm Univer- tive results show further promise in employing the sity for valuable input into the writing; Jan-Pieter OSL dating technique for dating sub-till ice-distal Buylaert and the technical staff at the Nordic Labo- glaciofluvial/fluvial, as well as other sediments in ratory for Luminescence Dating for helping with Norway. OSL-measurements; Damian Steffen at University We imagine a local paleoenvironment whereby of Bern for introducing Johnsen to the deconvolu- glaciofluvial/fluvial sediments from the tion technique; Irene Lundquist at NGU for techni- Langsmoen site were deposited in ice-distal condi- cal assistance with maps and several drawings; tions, in a narrow fjord with relative sea level at and, field assistants Marie Koitsalu and Jakob 210 m a.s.l. or more, perhaps up to ~220 m a.s.l. Heyman. This project was supported with funding Further advance of the ice sheet leading into LGM from the Swedish Society for Anthropology and 2 resulted in ice damming down-valley from Flora, Geography, Carl Mannerfelts fund, and Ahlmanns ice-margin oscillations and deposition of glacio- fund. lacustrine sediments in ice-proximal conditions, Johnsen was in charge of and performed all as- alternating with till deposition at the Flora sections pects of the project: conception, acquiring funding, with the highest lake level up to at least 310 m OSL-sampling and measurements, data analysis, a.s.l. These sites were in-turn buried by LGM 2 till and writing. Olsen suggested sites for OSL- and glaciotectonized. sampling, guided in the field, provided important Sediments from Langsmoen and Flora sections input to the writing, and completed the original indicate ice-free conditions that are consistent with stratigraphic and sedimentologic fieldwork years other similarly-aged sediments in similar strati- earlier. Murray gave valuable contributions to graphic settings at many other sites in Norway that discussions regarding OSL-preparation and meas- collectively define the Trofors interstadial (~25-20 urement setup, and interpretation of results as well ka; 42 sites and 67 dates); a regional ice-free pe- as provided laboratory access. riod during MIS 2 that is bracketed by LGM 1 and LGM 2 stadials. The existence of the Trofors interstadial along with other interstadials during the References Middle and Late Weichselian (MIS 3 and MIS 2) Alexanderson, H., Johnsen, T.F., Murray, A.S., in press. Re- indicates that not only the western margin, but the dating the Pilgrimstad interstadial with OSL: a warmer cli- whole western part of the ice sheet, from the ice

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

mate and a smaller ice sheet during the Swedish Middle peratures in northern Scandinavia during the last glaciation. Weichselian (MIS 3)? Boreas. Geology 35, 987-990. Alexanderson, H., Murray, A.S., in press. Problems and poten- Helmens, K.F., Johansson, P.W., Räsänen, M.E., Alexander- tial of OSL dating Weichselian and Holocene sediments in son, H., Eskola, K.O., 2007b. Ice-free intervals at Sokli con- Sweden. Quaternary Science Reviews. tinuing into Marine Isotope Stage 3 in the central area of the Scandinavian glaciations. Geological Society of Finland Alexanderson, H., Eskola, K.O, Helmens, K.F., 2008. Optical Bulletin 79, 17-39. dating of a Late Quaternary sediment sequence from Sokli, northern Finland. Geochronometria 32, 51-59. Hillefors, Å., 1969. Västsveriges glaciala historia och mor- fologi. Naturgeografiska studier. Meddelanden från Lunds Allen, J. R., 1982. Sedimentary structures: their character and Universitets geografiska institution, Avhandlingar 60. 319 p. physical basis, volumes I and II. Developments in Sedimen- tology 30A and 30B, 593 and 663 p. Hillefors, Å., 1974. The stratigraphy and genesis of the Döse- backa and Ellesbo drumlins. A contribution to the knowl- Andersen, B.G., Nydal, R.,Wangen, O.P., Østmo, S.R., 1981. edge of the Weichsel-glacial history in Western Sweden. Weichselian before 15,000 years B.P. at Jaeren-Karmøy in Geologiska Föreningens i Stockholm Förhandlingar 96, southwestern Norway. Boreas 10, 297–314. 355-374. Andersen, B.G., Sejrup, H.P., Kirkhus, Ø., 1983. Eemian and Houmark-Nielsen, M. and Kjær, K.H, 2003. Southwest Scan- Weichselian deposits at Bø on Karmøy, S.W. Norway, a dinavia, 40–15 ka BP: palaeogeography and environmental preliminary report. Norges Geologisk Undersøkelse 380, change. Journal of Quaternary Science 18, 769-786. 189–201. Hubbard, A., Bradwell, T., Golledge, N., Hall, A., Patton, H., Arnold, N.S., van Andel, T.H., Valen, V., 2002. Extent and Sugden, D., Cooper, R., Stoker, M., 2009. Dynamic cycles, Dynamics of the Scandinavian Ice Sheet during Oxygen Iso- ice streams and their impact on the extent, chronology and tope Stage 3 (65,000-25,000 yr B.P.) Quaternary Research deglaciation of the British–Irish ice sheet. Quaternary Sci- 57, 38-48. ence Reviews, 28, 758-776. Banerjee, D., Murray, A. S., Bøtter-Jensen, L., Lang, A., 2001. King, E. L., Sejrup, H. P., Haflidason, H., Elverhøi, A., Equivalent dose estimation using a single aliquot of po- Aarseth, I., 1996. Quaternary seismic stratigraphy of the lymineral fine grains. Radiation Measurements 33, 73-94. North Sea Fan: glacially-fed gravity flow aprons, Bøtter-Jensen, L., Bulur, E., Duller, G.A.T., Murray, A.S., hemipelagic sediments, and large submarine slides. Marine 2000. Advances in luminescence instrument systems. Radia- Geology 130: 293-315. tion Measurements 32, 523-528. Kjær, K.H., Lagerlund, E., Adrielsson, L., Thomas, P., J., Choi, J.H., Duller, G.A.T., Wintle, A.G., 2006. Analysis of Murray, A., Sandgren, P., 2006. The first independent chro- quartz LM-OSL curves. Ancient TL 24, 9-20. nology for Middle and Late Weichselian sediments from southern Sweden and the island of Bornholm. GFF 128, Denton, G.H., Hughes, T.H., 1981. The Last Great Ice Sheets. 209-220. Wiley-Interscience, New York, 484 p. Larsen, E., Gulliksen, S., Lauritzen, S.-E., Lie, R., Løvlie, R., Donner, J., 1995. The Quaternary History of Scandinavia. Mangerud, J., 1987. Cave stratigraphy in western Norway; Cambridge University Press, 206 p. multiple Weichselian glaciations and interstadial vertebrate Duller, G.A.T., 2004. Luminescence dating of Quaternary fauna. Boreas 16, 267-292. sediments: recent advances. Journal of Quaternary Science Lian, O.B., Roberts, R.G., 2006. Dating the Quaternary: pro- 19, 183-192. gress in luminescence dating of sediments. Quaternary Sci- Duller, G.A.T., 2008. Single-grain optical dating of Quater- ence Reviews 25, 2449–2468. nary sediments: why aliquot size matters in luminescence Lokrantz, H., Sohlenius, G., 2006. Ice marginal fluctuations dating. Boreas 37, 589-612. during the Weichselian glaciation in Fennoscandia, a litera- Ehlers J., 1996. Quaternary and Glacial Geology. John Wiley ture review. SKB technical report no. TR-06-36. Swedish & Sons, Chichester. Nuclear Fuel and Waste Management Co., Stockholm. Fairbanks, R.G., Mortlock, R.A., Chiu, T.-C., Cao, L., Kaplan, Longva, O., Thorsnes, T., 1997. Skagerak in the past and the A., Guilderson, T.P., Fairbanks, T.W., Bloom, A.L., 2005. present - an integrated study of geology, chemistry, hydro- Marine radiocarbon calibration curve spanning 0 to 50,000 graphy and microfossil ecology. Norges geologiske un- years B.P. based on paired 230Th/234U/238U and 14C dates on dersøkelse, Special Publication, 8, 1-100. pristine corals. Quaternary Science Reviews, 24, 1781-1796. Lundqvist, J., 1992. Glacial stratigraphy of Sweden. Geologi- Fuchs, M., Owen, L.A., 2008. Luminescence dating of glacial cal Survey of Finland, Special Paper 15, 43-59. and associated sediments: review, recommendations and Lundqvist, J., 2004. Glacial history of Sweden. In: Ehlers, J., future directions. Boreas 37, 636-659. Gibbard, P. L.: Quaternary Glaciations – Extent and Chro- Galbraith, R.F., Roberts, R.G., Laslett, G.M., Yoshida, H., nology. Elsevier B. V., p. 401-412. Olley, J.M., 1999. Optical dating of single and multiple Lundqvist, J., and Robertsson, A.-M., 2002. Istider och mel- grains of quartz from Jinmium rock shelter, northern Austra- lanistider. In: Fredén, C. (Ed.) Sveriges Nationalatlas: Berg lia. Part I: experimental design and statistical models. Ar- och jord. 3rd ed. pp. 120-124. chaeometry 41, 339-364. Lunkka, J.P., Murray, A., Korpela, K, 2008. Weichselian Helmens, K.F., Bos, J.A.A., Engels, S., van Meerbeeck, C.J., sediment succession at Ruunaa, Finland, indicating a Mid- Bohncke, S.J.P., Renssen, H., Heiri, O., Brooks, S.J., Seppä, Weichselian ice-free interval in eastern Fennoscandia. Bo- H., Birks, H.J.B., Wohlfarth, B., 2007a. Present-day tem- reas 37, 234-244.

21 Timothy F. Johnsen, Lars Olsen and Andrew Murray

Mangerud, J., 1991. The Scandinavian Ice Sheet through the Olsen, L., Sveian, H. and Bergstrøm, B., 2001a. Rapid adjust- last interglacial/glacial cycle. Paläoklimaforschung, 1, 307- ments of the western part of the Scandinavian ice sheet dur- 330. ing the Mid- and Late Weichselian – a new model. Norges Geologiske Tidsskrift 81, 93-118. Mangerud, J., 2004. Ice-sheet limits on Norway and the Nor- wegian continental shelf. In: Ehlers, J. and Gibbard, P. L. Olsen, L., Sveian, H., Bergstrøm, B., Selvik, S.E, Lauritzen, (Eds.), Quaternary glaciations - extent and chronology. El- S.-E., Stokland, O., Gmsfjeld, K., 2001b. Methods and sevier, Amsterdam, pp. 271-294. stratigraphies used to reconstruct Mid- and Late Weich- selian palaeoenvironmental and palaeoclimatic changes in Mangerud, J., Gulliksen, S., Larsen, E., Longva, O., Miller, Norway. Norges geologiske undersøkelse Bulletin 438, 21- G., Sejrup, H.P., Sønstergaard, E., 1981. A middle Weich- 46. selian ice-free period in western Norway: The Ålesund In- terstadial. Boreas 10, 447–462. Olsen, L., van der Borg, K., Bergstrøm, B., Sveian, H., Lauritzen, S.-E., Hansen, G., 2001c. AMS radiocarbon dat- Mangerud, J., Løvlie, R., Gulliksen, S., Hufthammer, A.-K., ing of glacigenic sediments with low organic carbon con- Larsen, E., Valen, V., 2003. Paleomagnetic correlations be- tent-an important tool for reconstructing the history of gla- tween Scandinavian Ice-Sheet fluctuations and Greenland cial variations in Norway. Norges Geologiske Tidsskrift 81, Dansgaard-Oeschger Events, 45,000-25,000 yr B.P. Quater- 59-92. nary Research 59, 213-222. Olsen, L., Sveian, H., van der Borg, K., Bergstrøm, B., Mangerud, J., Gulliksen, S., Larsen, E., 2010. 14C-dated fluc- Broekmans, M., 2002. Rapid and rhythmic ice sheet fluctua- tuations of the western flank of the Scandinavian Ice Sheet tions in western Scandinavia 15-40 Kya – a review. Polar 45–25 kyr BP compared with Bølling–Younger Dryas fluc- Research 21, 235-242. tuations and Dansgaard–Oeschger events in Greenland. Bo- reas, 39: 328-342. Ottesen, D., Rise, L., Knies, J. Olsen, L., Henriksen, S. 2005. The Vestfjorden-Trænadjupet palaeo-ice stream drainage Middleton, G. V., Hampton, M. A., 1976. Subaqueous sedi- system, mid-Norwegian continental shelf. Marine Geology ment transport and deposition by sediment gravity flows. In: 218, 175- 189. Stanley, D. J., Swift, D. J. P. (Eds.), Marine Sediment Transport and Environmental Management. John Wiley & Prescott, J.R., Hutton, J.T., 1994. Cosmic ray contributions to Sons, New York, pp. 197-218. dose rates for luminescence and ESR dating: large depths and long-term time variations. Radiation Measurements 23, Murray, A.S., Olley, J.M., 2000. Precision and accuracy in the 497-500. optically stimulated luminescence dating of sedimentary quartz: a status review. Geochronometria 21, 1-16. Ramsey, C., 1995. Radiocarbon calibration and analysis of stratigraphy: The OxCal program, Radiocarbon 37, 425-430 Murray, A.S. and Wintle, A.G., 2000. Luminescence dating of quartz using an improved single-aliquot regenerative-dose Ramsey, C., 2001. Development of the radiocarbon calibration protocol. Radiation Measurements 32, 57-73. program OxCal, Radiocarbon 43, 355-363 Murray, A.S., Wintle, A.G., 2003. The single aliquot regenera- Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, tive dose protocol: potential for improvements in reliability. J.W., Bertrand, C.J.H., Blackwell, P.G., Buck, C.E., Burr, Radiation Measurements 37, 377-381. G.S., Cutler, K.B., Damon, P.E., Edwards, R.L., Fairbanks, R.G., Friedrich, M., Guilderson, T.P., Hogg, A.G., Hughen, Murray, A.S., Marten, R., Johnson, A., Martin, P., 1987. K.A., Kromer, B., McCormac, G., Manning, S., Bronk Analysis for naturally occurring radionuclides at environ- Ramsey, C., Reimer, R.W., Remmele, S., Southon, J.R., mental concentrations by gamma spectrometry. Journal of Stuiver, M., Talamo, S., Taylor, F.W., van der Plicht, J., Radioanalytical and Nuclear Chemistry Articles 115, 263- Weyhenmeyer, C.E., 2004. IntCal04 terrestrial radiocarbon 288. age calibration, 0-26 cal kyr BP. Radiocarbon, 46, 1029- Nygård, A., Sejrup, H.P., Haflidason, H., Bryn, P., 2005. The 1058. glacial North Sea Fan, southern Norwegian Margin: Archi- Robertsson, A.-M., García Ambrosiani, K., 1992. The Pleisto- tecture and evolution from the upper continental slope to the cene in Sweden – a review of research, 1960-1990. Geo- deep-sea basin. Marine and Petroleum Geology, 22, 71-84. logical Survey of Sweden, SGU Ca 81, 299-306. Olsen, L., 1997. Rapid shifts in glacial extension characterise a Salonen, V.-P., Kaakinen, A., Kultti, S., Miettinen, A., Eskola, new conceptual model for glacial variations during the Mid K. O., Lunkka, J. P, 2008. Middle Weichselian glacial event and Late Weichselian in Norway. Norges Geologiske Un- in the central part of the Scandinavian Ice Sheet recorded in dersøgelse Bulletin 433, 54-55. the Hitura pit, Ostrobothnia, Finland. Boreas 37, 38-54. Olsen, L., Bergstrøm, B. 2007. Glacier variations during the Sejrup, H.P., Haflidason, H., Aarseth, I., King, E., Forsberg, LGM interval in the Karmøy-Jæren district, SW Norway. C.F., Long, D., Rokoengen, K., 1994. Late Weichselian gla- Norwegian Geologial Society, Abstracts and Proceedings, ciation history of the northern North Sea. Boreas 23, 1-13. No.1, 73-74. Sejrup, H.P., Larsen, E., Landvik, J., King, E.L., Haflidason, Olsen, L., Grøsfjeld, K., 1999. High Middle and Late Weich- H., Nesje, A., 2000. Quaternary glaciations in southern Fen- selian relative sea levels in Norway – with implications for noscandia: evidence from southwestern Norway and the the glacial isostasy and the ice-retreat rates. Norges geolo- northern North Sea region. Quaternary Science Reviews 19, giske undersøkelse Bulletin 435, 43-51. 667-685. Olsen, L., Mejdahl, V., Selvik, S.F., 1996. Middle and Late Sejrup, H. P., Nygård, A., Hall, A.M., Haflidason, H., 2009. Pleistocene stratigraphy, chronology and glacial history in Middle and Late Weichselian (Devensian) glaciation history Finnmark, North Norway. Norges geologiske undersøkelse of south-western Norway, North Sea and eastern UK. Qua- Bulletin 429, 1-111. ternary Science Reviews 28, 370-380.

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

Shepherd, A., Wingham, D., 2007. Recent sea-level contribu- Van der Borg, K., Alderliesten. C., de Jong, A.F.M., van den tions of the Antarctic and Greenland Ice Sheets. Science Brink, A., de Haas. A.P., Kersemaekers, H.J.H., 315, 1529–1532. Raaymakers, J.E.M.J., 1997. Precision and Mass Fractiona- tion in 14C analysis with AMS. Nuclear Instruments and Sveian, H., Hansen, L., Solberg, I.-L., Rokoengen, K., 2006. Methods B123, 97-101. Veibygging avdekker skredhistorien (“Road-construction reveals the clay slide history”). GeoPublishing AS, Trond- Vorren, T.O., Laberg, J.S., 1997. Trough Mouth Fans – Pa- heim, GEO 3-2006, 20-25. laeoclimate and ice-sheet Monitors. Quaternary Science Re- views 16, 865-882. Svendsen, J. I., Alexanderson, H., Astakhov, V. I., Demidov, I., Dowdeswell, J. A., Funder, S., Gataullin, V., Henriksen, Vorren, T. O., Vorren, K.-D., Alm, T., Gulliksen, S., Løvlie, M., Hjort, C., Houmark-Nielsen, M., Hubberten, H. W., R., 1988. The last deglaciation (20,000 to 11,000 B.P.) on Ingólfsson, Ó., Jakobsson, M., Kjær, K. H., Larsen, E., Lok- Andøya, northern Norway. Boreas 17, 41-77. rantz, H., Lunkka, J. P., Lyså, A., Mangerud, J., Matiouch- Wallinga, J., 2002. On the detection of OSL age overestima- kov, A., Murray, A., Möller, P., Niessen, F., Nikolskaya, O., tion using single-aliquot techniques. Geochronometria 21, Polyak, L., Saarnisto, M., Siegert, C., Siegert, M. J., Spiel- 17-26. hagen, R. F., Stein, R, 2004. Late Quaternary ice sheet history of northern Eurasia. Quaternary Science Reviews 23, Wintle, A.G., 2008. Luminescence dating: where it has been 1229-1271. and where it is going. Boreas, 37, 471–482. Ukkonen, P., Arppe, L., Houmark-Nielsen, M., Kjær, K. H., Wohlfarth, B. 2010. Ice-free conditions in Sweden during Karhu, J. A, 2007. MIS 3 mammoth remains from Sweden - Marine Oxygen Isotope Stage 3? Boreas 39, 377-398. implications for faunal history, palaeoclimate and glaciation chronology. Quaternary Science Reviews 26, 3081-3098. Wohlfarth, B., Skog, G., Possnert, G., Holmquist, B., 1998. Pitfalls in the AMS radiocarbon-dating of terrestrial macro- Valen, V., Mangerud, J., Larsen, E., Hufthammer, A.K., 1996. fossils. Journal of Quaternary Science 13, 137-145. Sedimentology and stratigraphy in the cave Hamnsundhel- leren, western Norway. Journal of Quaternary Science, 11, 185-201.

Paper IV

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 summit 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 deglaciation.

Keywords: cosmogenic exposure dating, radiocarbon dating, moraine, deglaciation, nunatak

Introduction (Birks et al. 2005, 2006, Kullman 2005, 2006) there has been no attempt to directly assess these Tree remains of three species, Betula pubescens, conflicting radiocarbon data against an independ- Picea abies, and Pinus sylvestris, have been found ent dating technique. Thus, the objective of our at high-elevation alpine sites in central Sweden that study was to determine independently the timing of are 400-500 m above the modern tree-line. More- deglaciation and to discuss their implications for over, these remains are as old as 16.7 ±0.2 cal. ka the potential of tree establishment at high elevation BP (Kullman 2000, 2001, 2002a), at times when it in central Sweden and for the vertical rate of de- is commonly perceived that the sites were covered glaciation. We evaluate the radiocarbon ages of the by the Scandinavian ice sheet (SIS; Kleman et al. wood remains against new results from terrestrial 1997, Lundqvist 2002). Widespread low-elevation cosmogenic nuclide (TCN) exposure dating of radiocarbon evidence indicates, instead, that degla- glacial erratics at high elevation and discuss their ciation in the area occurred around 10.3 to 10.0 cal. divergence in terms of paleoecological and pa- ka BP (De Geer 1940, Borell and Offerberg 1955, leoglaciological implications. Lambeck et al. 1998, Lundqvist 1969, 2002). The incongruent dates of the tree remains, therefore, potentially have tremendous implications for our Study area understanding of the dynamics of the SIS, the pat- The study area in the county of Jämtland, central tern and rate of migration of tree species, the loca- Sweden, includes mountainous areas of moderate tion and role of refugia in re-establishing plants, relief (~800 m) and adjacent low-relief (~100 m) paleoclimatic conditions, and nunatak microcli- rolling hill landscape with numerous lakes and mate variability. Although the reliability of age peatlands (Fig. 1). Mountains are generally concen- constraints is central in obtaining accurate pa- trated in the southern half of the study area except leoglaciological or paleoecological reconstructions

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

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 Fig. 1: Regional map of the study area in the county of that high elevation areas around Mt Åreskutan Jämtland showing location of radiocarbon dates relevant to deglaciated thousands of years before adjacent deglaciation (cf. numbering system in Table 1); radiocarbon dates from the county of Dalarna (Table 1), approximately 80 valleys. Altogether Kullman (2002a, 2005) dated km south of the study area, are not shown. TCN sample sites sixteen well-preserved specimens found on the for Snasahögarna SIS moraine area at “S” (see Fig. 3) and for ground surface in the forefield of a receding 'per- Mt. Åreskutan summit erratics at “Å” (arrow). Dashed curves are early ice-free areas in Sweden (Borgström 1989); note Mt. ennial' snow-patch at 1360 m a.s.l. which is only Åreskutan deglaciated later. Ice movement during the late 60 m below the summit and 400-500 m above the glacial maximum was towards the west (Lundqvist 1969). modern tree-limits. The surrounding landscape on Inset map shows location of study area just east of the Swed- ish-Norwegian border in central Sweden, the counties of the summit of Mt Åreskutan is dominated by ex- Jämtland and Dalarna (dark grey), and the LGM and Younger posed and glacially sculpted bedrock. Numerous Dryas ice margins (compiled from various sources by Olsen et ponds exist that are sometimes bordered by shal- al. 2001). 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. for Mt. Åreskutan which occurs largely isolated in Within the counties of Jämtland and Dalarna only the northeast (Å in Fig. 1). The modern tree-line is four other unusually old tree remains have been at about 950-1050 m above sea level (a.s.l.) in this found (Table 1, Fig. 1, see inset map). Thus, the area. geographical distribution of unusually old tree The study area deglaciated 10.3 to 10.0 cal. ka remains is especially concentrated to the summit BP (De Geer 1940, Borell and Offerberg 1955, area of Mt. Åreskutan. Sveian 1997, Lambeck et al. 1998, Lundqvist 2002). Above the Late Glacial highest coastline and within the study area, studies on moderate Methods elevation lake sediments estimate deglaciation 10.4 ±0.1 cal. ka BP (Bergman et al. 2005). Thus, there The TCN exposure dating technique can allow is general agreement for the timing of deglaciation determination of the amount of time that a rock between these studies. However, some of the high surface has been exposed to cosmic radiation (e.g., elevation tree remains are considerably older how long ago an ice sheet deposited an erratic (Kullman 2002a, 2004b, Kullman and Kjällgren boulder). Cosmic radiation causes the accumula- 2006). While traditionally evidence has been inter- tion of 10Be in situ within quartz rock and the preted in terms of a thick ice sheet in the study area measurement of the concentration of 10Be against at that time (e.g., Lundqvist 1969, 2002, Denton the production rate of 10Be for a given site of and Hughes 1981, Kleman et al., 1997), there re- known latitude, elevation, topographic shielding,

2 High-elevation cosmogenic nuclide dating of the last deglaciation in the central Swedish mountains Kullman 2005 Kullman 2002a Kullman b b Sample numberSample Reference Beta-121827 Beta-120799 a county of Dalarna. Listed in Listed of Dalarna. county C yr BP C yr Calibrated age 14 63.43363.433 13.100 13.100 1360 1360 ±100 11440 ±90 11720 ±100 13300 ±100 13600 wood wood wood wood stem 63.027 root 63.027 12.201 63.027 12.201 63.027 12.201 12.201 1495 63.433 1495 1495 13.100 ±60 8030 63.433 stem 1495 ±60 8150 stem ±70 8430 13.100 ±100 8900 1360 ±60 8710 ±100 9100 Beta-122313 ±100 9500 Beta-172310 ±70 8850 1360 ±100 9600 Beta-172312 2004a Kullman wood Beta-172311 2004a Kullman ±90 10200 ±200 10000 63.433 2004a Kullman 63.433 Beta-152673 13.100 2004a Kullman ±200 11900 13.100 Beta-121828 2002a Kullman 1360 2002a Kullman 1360 63.178 ±70 12870 ±80 14020 12.349 ±100 15200 ±200 16700 Beta-158301 1000 Beta-133672 2002a Kullman ±100 8680 2002a Kullman ±200 9700 St-12980 1995 Kullman , stem, 63.433 13.100 1360 ±60 10380 ±100 12300 Beta-152671 2002a Kullman c c nut nut 63.167 12.367 63.167 12.367 740 740 ±50 8270 ±50 8300 ±100 9300 Beta-108756 ±100 9300 Beta-108756 1998 Kullman 1998 Kullman wood 63.167 12.367 740 ±50 8270 ±100 9300 Beta-106347 1998 Kullman wood barkwood, woodwood 63.096 63.167 12.395 12.367 63.167 855 12.367 740 ±100 9050 ±50 8100 63.178 740 ±200 10200 12.349 ±100 9100 St-12729 ±70 8220 Beta-106348 ±100 9200 720 1995 Kullman 1998 Kullman Beta-91497±120 8460 1998 Kullman ±100 9400 Beta-57645 1995 Kullman stem stem cone stem stem stem stem 63.433 13.100 wood 63.433 stem 13.100 wood 63.433 stem 1360 13.100 63.433 wood 1360 ±60 8640 63.433 wood 13.100 wood 13.100 1360 ±90 9700 ±100 9600 wood wood ±90 10250 1360 Beta-160729 ±200 11000 1360 63.181 wood Beta-121829 ±60 10700 ±200 12000 2002a Kullman 63.217 wood ±90 11020 12.326 62.917 Beta-121830 2002a Kullman ±100 12800 12.317 63.150 ±100 13000 12.483 Beta-133673 2002a Kullman 1250 12.283 62.819 Beta-121826 1180 62.961 2002a Kullman 12.142 1150 62.995 ±70 9310 cone 2002a Kullman 12.320 63.039 1030 ±90 9840 12.348 ±80 11160 62.937 ±100 10500 wood 12.769 ±80 8420 ±100 11300 990 wood 12.364 Beta-184490 63.142 ±100 13000 940 63.203 Beta-127895 925 Beta-127894 12.354 ±190 8160 2006 ±100 9400 Kjällgren and Kullman 880 12.540 ±120 8550 2000 Kjällgren and Kullman 870 Beta-108782 2000 Kjällgren ±105and 8385 Kullman ±300 9100 ±90 8030 835 ±100 9600 ±50 8150 2000 Kjällgren and Kullman St-1033 ±100 9400 810 St-1028 ±200 8900 63.167 ±120 8190 St-1099 ±100 9100 ±110 8295 St-1064 12.367 63.265 St-6280 1969 Lundqvist ±200 9200 63.308 1969 Lundqvist 11.984 ±100 9300 1969 Beta-57608 Lundqvist 12.483 1969 Lundqvist 740 St-1579 1969 Lundqvist 1995 Kullman 700 ±50 8330 526 1969 Lundqvist ±30 8600 ±100 9400 ±100 8390 Beta-108554 ±30 9500 ±100 9400 St-819 1998 Kullman LuA-5134 1969 Lundqvist Bergman2004 al. et , , Betula pubescens, Betula pubescens, Betula pubescens, Betula pubescens, Betula Picea, pubescens, Betula Pinus, pubescens, Betula Picea, pubescens Betula Pinus, Picea, Pinus Pinus pubescens, Betula pubescens, Betula Pinus, Pinus, Pinus, Pinus, pubescens, Betula Pinus, Pinus, Pinus, Pinus, Pinus, Pinus, Pinus, Pinus, Alnus, Pinus, Corylus, Betula, Corylus, Alnus, Pinus, Pinus, Pinus, Fig. 1Fig. north east (m asl) (1 error) (1 mid-point) No. inNo. material Analyzed Latitude Longitude Elevevation 13 13 13 13 1 1 1 1 1 1 1 1 1 1 1 1 6 4 17 9 7 24 15 14 12 16 11 10 5 8 8 8 8 8 8 7 3 2 Mt. Åreskutan Mt. Åreskutan Mt. Lillsylen Mt. Lillsylen Mt. Lillsylen Mt. Lillsylen Mt. Åreskutan Mt. Åreskutan Mt. Åreskutan Mt. Åreskutan Mt. Getryggen Kloppanåstugan Mt. Getryggen Mt. Getryggen Mt. Åreskutan Mt. Lillulvåfjället Mt. Getryggen Mt. Getryggen Storkluken Mt. Åreskutan Mt. Åreskutan Mt. Åreskutan Mt. Åreskutan Mt. Åreskutan Mt. Lillsnasen Mt. Helagfjället Mt. Stor-Ulvåfjället Mt. Getryggen Hammaren Hammaren Vålåstugan Hammaren Mt. Enkälen O Bunnerån Mt. Getryggen Mt. Getryggen Mt. Getryggen Klocka Bog Table 1: Summary of radiocarbon dates related to deglaciation in the county of Jämtland plus dates of mega-fossil wood from the of mega-fossil plus dates of Jämtland county the in deglaciation to related dates of radiocarbon 1: Summary Table on next page. Continued and age. elevation, type, sample county, order of Jämtland province - mega-fossil wood - mega-fossil province Jämtland Location type) (by and county

3 Timothy F. Johnsen, Derek Fabel, and Arjen P. Stroeven . r Sample numberSample Reference a art of the trunk while the second was from was second the cente ofits trunkart the while p 4.0 and the IntCal 04 calibration curve (Reimer et al. 2004). 2004). al. et (Reimer curve calibration 04 IntCal the and 4.0 er g C yr BP C yr Calibrated age 14 oun y iece of wood with the first the from the with of iece wood p leaves 63.383 13.167 700 ±65 8750 ±100 9800 Ua-21758 Ek 2004 seeds and leaf parts leaf and seeds 63.452 13.010 734 ±75 9225 ±100 10400 LuS-6447 study this , seeds and leaf parts leaf seeds , and 63.100 12.242 987 ±50 9250 ±100 10400 Poz-2750 Bergman2005 al. et and cone and 62.067 12.417 915 ±70 8160 ±100 9100 Beta-178796 2004b Kullman Ericaceae nutshell 62.117 12.450 910 ±40 8560 ±40 9500 Beta-158310 2004b Kullman leaf 63.167 12.367 740 ±50 8030 ±100 8900 Beta-108752 1998 Kullman g acorn 62.117 12.450 910 ±40 8670 ±100 9600 Beta-158309 2004b Kullman leaf 63.167 12.367 740 ±60 8510 ±60 9500 Beta-108758 1998 Kullman wood 62.117 12.450 915 ±60 8360 ±100 9400 Beta-178799 2004b Kullman leaf wood wood wood 63.167 wood wood 12.367 wood wood wood wood 740 wood wood ±50 8060 62.117 wood 62.117 12.450 62.117 ±100 8900 12.450 61.983 12.450 61.917 Beta-108753 12.850 61.917 1180 12.883 62.117 1180 1998 Kullman 12.883 61.917 1180 ±60 8500 12.450 62.117 1160 ±70 9070 12.883 61.917 1100 ±50 9230 12.450 ±60 9500 62.117 1100 ±70 8050 ±100 10300 12.883 ±60 10500 1070 62.117 ±100 10400 Beta-172317 12.450 ±60 10500 1055 Beta-172305 12.450 ±100 8900 ±100 12500 Beta-178795 1035 ±50 8380 2004b Kullman 1035 ±100 12500 ±60 7890 2004b Kullman Beta-169410 Beta-158305 2004b Kullman ±70 8050 940 Beta-158305 ±100 9400 ±60 8190 940 2004b Kullman 2004b Kullman ±200 8800 Beta-158314 ±60 8040 2004b Kullman ±100 8900 Beta-178797 ±60 8040 ±100 9100 Beta-169411 2004b Kullman ±100 8900 Beta-178794 2004b Kullman ±100 8900 2004b Kullman Beta-172316 2004b Kullman Beta-172316 2004b Kullman 2004b Kullman large cone large 62.067 12.417 850 ±70 8490 ±70 9500 Beta-108767 2001 Kullman fruits, scales, leaves and twigs and leaves scales, fruits, 63.117 12.317 887 ±160 9315 ±200 10500 Ua-16388 2004 Hammarlund al. et twi bud scales bud 63.100 12.242 987 ±120 9165 ±100 10400 LuA-5477 2005 Bergman al. et lus, y Betula, Quercus, Cor Salix, Empetrum Salix, Dryas octopetala, peatbulk peatbulk Quercus, Alnus, Ulmus, moss and bulk peatpeatbulk Moss stems leaves and Pinus, 63.167Pinus, 63.167Pinus, 12.350Pinus, 12.350 63.308Pinus, Pinus, 12.483 750Pinus, 63.233 750Pinus, 63.233Pinus, ±115 8835 12.417Pinus, ±145 8930 526 12.417Pinus, ±200 9900 Pinus, ±200 10000 ±95 8055 Larix, 680 Ua-14621 Ua-14920 680 ±200 8900 ±95 8530 2003 Stedingk von Segerström and ±105 8855 LuA-5135 2003 Stedingk von Segerström and ±100 9500 ±200 10000 Ua-13310 2004 Bergman al. et Ua-13315 2003 Stedingk Segerström von and 2003 Stedingk von Segerström and Picea, Fig. 1Fig. north east (m asl) (1 error) (1 mid-point) No. inNo. material Analyzed Latitude Longitude Elevevation 20 20 19 18 23 23 8 8 8 21 22 22 2 les Beta-120799 and Beta-121827 from are from Åreskutan Beta-121827 Mt. and same the Beta-120799 les p Lake Stentjärn Mt. Storvätteshågna Mt. Storvätteshågna Mt. Storvätteshågna Lake Stentjärn Lake Spåime Lake Ullsjön Årebjörnen Mt. Barfredhågna Mt. Storulvån Mt. Getryggen Mt. Getryggen Storsnasen II Klocka Bog Mt. Storvätteshågna Mt. Storvätteshågna Mt. Nipfjället Mt. Städjan Mt. Städjan Mt. Storvätteshågna Mt. Städjan Mt. Storvätteshågna Mt. Städjan Mt. Storvätteshågna Mt. Storvätteshågna Barfredhågna Mt. Mt. Storvätteshågna Storulvån Mt. Getryggen Storsnasen I Storsnasen Calibrated age is the midpoint of the 1 calibrated age range with the uncertainty as half this range. Calibrated using OxCal using Calibrated range. as this half uncertainty the with range age calibrated of 1 the midpoint is the age Calibrated Sam Jämtland province - sediment province lake Jämtland Dalarna province - mega-fossil wood province - mega-fossil Dalarna Location type) (by and county Jämtland province - peat province Jämtland a b Table 1: (continued). Table 1: (continued).

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

and sample thickness, provides determination of the exposure age (Lal 1991, Gosse and Phillips 2001). This technique has proven useful in numerous 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 deglaciation, 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 Fig. 2: Land uplift curve of study area adjusted for relative sea from the mountains Åreskutan and Snasahögarna level changes. Ten percent confidence interval shown. See (Å and S in Fig. 1). These data are most relevant to text for explanation. addressing the high elevation radiocarbon ages and for assessing the vertical rate of deglaciation in 9 central Sweden. We collected quartz rich samples c. 0.25 mg Be carrier, dissolved, separated by ion from glacially transported boulders from the sum- chromatography, selectively precipitated as hy- mit of Mt. Åreskutan (at 1415-1420 m a.s.l.) and droxides, and oxidized. Accelerated Mass Spec- from a newly discovered moraine on Mt. Snasa- trometry (AMS) measurements were completed at 10 9 högarna (1125-1149 m a.s.l.), which, incidentally, the SUERC AMS Facility. Measured Be/ Be is the highest elevation moraine related to the SIS ratios were corrected by full chemistry procedural 10 9 -15 discovered so far in Sweden (Heyman 2004, Hey- blanks with Be/ Be of <3 x 10 . Independent man and Hättestrand 2006). The Snasahögarna SIS measurements of AMS samples were combined as moraine stretches to an elevation of 1190 m a.s.l. weighted means with the larger of the total statisti- To minimize the risk of having ages biased by cal error or mean standard error. We calculated the processes such as cosmogenic nuclide inheritance analytical uncertainty by assuming that the uncer- (e.g., Briner et al. 2001), boulder exhumation, tainties in AMS measurement and Be carrier are surface erosion or moraine degradation (Hallet and normal and independent, adding them in quadra- Putkonen 1994, Zreda et al. 1994, Putkonen and ture in the usual fashion (e.g., Bevington and Rob- Swanson 2003), we sampled the tops of large (>1.0 inson 1992). The resulting analytical uncertainties 10 m b-axis), weathering-resistant (granitic and range from 3 to 6% (Table 2). All Be concentra- quartzitic) boulders, and in sets of three. Boulder tions were converted to exposure ages by using a surfaces and adjacent ground were carefully in- production rate linked to a calibration data set us- 10 spected for indications of the amount of differential ing a Be half-life of 1.5 Ma. erosion of erosion. The height of quartz nodules Using the CRONUS-Earth exposure age calcu- and veins above adjacent softer rock within each lator version 2.2 (http://hess.ess.washington.edu), 10 boulder were used to estimate the amount of sur- measured Be concentrations were converted to face erosion. In total six boulder samples were surface exposure ages, assuming no prior exposure processed in the Glasgow University-SUERC cos- and no erosion since deposition. The results for the 10 mogenic nuclide laboratory. different Be production rate scaling schemes used by the online calculator yielded ages that vary by Measurements and calculations about 5% for each sample. The surface exposure All samples were processed for 10Be from quartz ages were calculated using the ‘Lm’ scaling following procedures based on methods modified scheme which includes paleomagnetic corrections from Kohl and Nishiizumi (1992) and Child et al. (Balco et al. 2008; Table 2). (2000). Approximately 20 g of pure quartz was separated from each sample, purified, spiked with

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

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.

other hand, if inheritance is dominating 10Be ages 10 Be age adjustments will give maximum ages. The accuracy of the exposure age calculations can During the exposure history of a site, the air be affected by various physical factors: (1) moraine pressure will change due to the combined affect of degradation (Putkonen and Swanson 2003), (2) glacio-isostatic rebound and eustatic change; these cosmogenic nuclide inheritance (e.g., Briner et al. factors cause changes in the local air pressure over 2001), (3) the weathering and erosion of the rock time that in turn alters the cosmogenic nuclide surface during exposure, (4) the partial shielding of production rate and apparent exposure age. Thus, the rock surface from cosmic rays by topography, we estimated the land uplift corrected for sea level seasonal snow cover, and vegetation, or (5) the changes and used this information to adjust the changing elevation of the rock surface due to gla- apparent TCN exposure age for each site (cf. Stone cio-isostatic movement (Gosse and Phillips 2001). 2000, Johnsen et al. 2009). The employed land Apart from inheritance, all these factors cause 10Be uplift curve was generated for the study area by ages to appear too young and without adjusting for interpolating between shoreline displacement them, 10Be ages will be minimum ages. On the curves located to the west and east (using Dahl and Nesje 1996, and shoreline displacement curves of

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

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.

Lidén 1938, Mörner 1980, Kjemperud 1981, and vations, and GIS, detailed mapping of these mo- Sveian and Olsen 1984) for ages less than 9 cal. ka raines was completed (Fig. 3, 4, and 5). The geo- BP, and referring to regional glacio-isostatic mod- graphical pattern of the moraines, including the elling results of Lambeck et al. (1998) for ages location of the highest moraines at only 1 km from from 9 to 12 ka (Fig. 2). 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 Results (1969, 1973) and for similar moraines east of the Mt. Snasahögarna (Borgström 1979). Three boul- Winter fieldwork helped reveal a system of mo- ders from the crest of the highest moraine gave raines that were produced during decay of the SIS consistent ages (10.1 ±0.6 ka, n = 3; Table 2, Fig. within a broad u-shaped valley, named In- 3). The three ages overlap within 1 when consid- golvskalet, that cuts into the west shoulder of Mt. ering the measurement errors alone. Snasahögarna (Fig. 1, location 22, and Fig. 3). Erratic boulders from the summit of Mt. Åre- Using aerial photographs coupled with field obser- skutan also produced exposure ages that are consistent 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 errors alone. Close inspection of the boulder surfaces and adjacent ground surface indicated little signs of erosion. Boulders were either sitting on bare bedrock 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 Fig. 5: Photograph of moraine extending up the eastern, snow- spallation was not an important process. Measure- covered slope of upper-Ingolvskalet valley. This is the highest ment of the height of quartz veins and nodules elevation moraine related to the Scandinavian Ice Sheet yet identified in Sweden (1190 m a.s.l.). Note person for scale. against softer parent rock divided by the TCN ex- Inset photograph shows sample TJ-15 and arrow indicates posure age indicates a reasonable weathering rate location in photograph. of about 1 mm ka-1. Such an erosion rate would

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

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. produce less than a 1% age increase and so is ig- Further adjustments to the apparent TCN expo- nored in our calculations. For interest, a large 3 sure ages must be made when considering the ef- mm ka-1 erosion rate would increase ages less than fect of land uplift and snow shielding during the only 2.5% (~275-305 years). Similar observations exposure history of the boulders. By considering to Mt. Åreskutan were made at and adjacent to the land uplift history of the sites (Fig. 2), the TCN each boulder on the Snasahögarna moraine, indi- ages increase by 4.7% and 6.4% for Mt. Åreskutan cating that an erosion rate of 1 mm ka-1 is also and the Snasahögarna moraine, respectively. The reasonable here. difference in these adjustments is caused by their differences in apparent age and elevation. Recalcu-

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

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 erratics TJ-1 1415 63.431 13.095 granite 0.999 0.960 16.30 ±0.74 9.2 ±0.9 (0.4) 10.8 ±1.0 TJ-2 1414 63.431 13.094 quartz band 1.000 0.976 16.15 ±1.05 8.9 ±1.0 (0.6) 10.5 ±1.1 TJ-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 moraine TJ-14 1125 63.225 12.309 quartz band 0.999 0.976 14.41 ±0.50 10.1 ±0.9 (0.3) 12.1 ±1.1 TJ-15 1149 63.225 12.314 quartz band 1.000 0.976 14.80 ±0.50 10.2 ±0.9 (0.3) 12.2 ±1.1 TJ-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 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. 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.

lating the TCN ages using conservative 10% cate that the mountain top deglaciated about 10.6 bounds on the land uplift curve (Fig. 2) results in ±0.6 ka when erosion rates, glacio-isostatic re- adjustments that vary by only 0.6%. Thus the accu- bound, and snow shielding effects are taken into racy of the land uplift curve insignificantly affects consideration (Table 2). Exhumation processes that the calculated age. could cause ages to appear younger than their true Based on precipitation data from the region age are disregarded because the erratic boulders (Sweden Meteorological and Hydrological Insti- were resting on bedrock or shallow soils. If nuclide tute), field observation, and local knowledge, we inheritance was an important process then the ages estimate the snow thickness over the tops of boul- would appear older than their true depositional age; ders to be 3 m and persist for four months each which would be unreasonable to expect as at least year. This snow thickness and of medium density the valley bottoms deglaciated around the same of 0.2 g cm-3 (cf. Lundberg et al. 2006) causes a time or slightly later (Sveian 1997, Lundqvist 12% increase in the TCN ages (Gosse and Phillips 2002). Thus, the TCN exposure ages from the 2001). Note that doubling this thickness to 6 m per summit of Mt. Åreskutan are considered reliable. four months a year less than doubles the increase in We collected a lake core from Lake Ullsjön (5 the TCN ages to 22%. Thus, when adjusting for the km northwest of Mt. Åreskutan, at 734 m a.s.l.; combined affects of land uplift and snow burial, Table 1, Fig 1, location 18). Within the first organ- the apparent mean TCN exposure age for Mt. Åre- ics deposited above glacial clay we found and skutan increases by 17.6% to 10.6 ±0.6 ka, and for dated Dryas octopetala seeds and leaf parts to 10.4 the Snasahögarna moraine increases by 19.4% to ±0.1 cal. ka BP. This species is commonly associ- 12.0 ±0.6 ka. ated with the tundra paleoenvironment following deglaciation in Scandinavia (e.g., Bergman et al. 2005). Thus, this date is considered a reliable esti- Discussion mate of the timing of local deglaciation and is in good agreement with other palaeoecological evi- TCN ages dence in the area (e.g., Bergman et al. 2005) and The three consistent TCN exposure ages of erratic TCN adjusted exposure ages. boulders from the summit of Mt. Åreskutan indi-

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

Radiocarbon ages and comparison to TCN ages Figure 7 and Table 1 show a summary of radiocarbon 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 Fig. 7: Count versus age (cal. ka BP) of calibrated radiocarbon dates related to deglaciation in the counties of Jämtland and landscape, and where pioneering species (from Dalarna (See Table 1). The graph shows the contribution of lake cores) would perhaps be 200 years younger each type of date to the total per given 500 year interval. Date (Bergman et al. 2005), followed by tree and peat types are: mega-fossil wood from Dalarna, mega-fossil wood from Mt. Åreskutan, all other mega-fossil wood from species. Agreeing with this succession, dates from Jämtland, lake sediment, and peat. Also shown are TCN dat- lake cores of pioneering flora representing a tundra ings from Mt. Åreskutan and Snasahögarna SIS moraine; error bars are 1. 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 The TCN adjusted exposure ages from the Sna- 2), although central values are slightly younger. sahögarna moraine are older than those from Mt. Peat dates are in turn slightly younger than the Åreskutan (adjusted mean age of 12.0 ±0.6 ka, n = dates from lake sediment. Thus, the pattern of the 3; Table 2). This age difference probably reflects a histogram of radiocarbon dates from the region geographical pattern of deglaciation whereby the reflects the stages of ecological succession with the SIS margin was at the Snasahögarna moraine at exception of some old dates from tree remains, 1150 m a.s.l., while Mt. Åreskutan, 280 m higher principally from Mt. Åreskutan. but also 35 km east (up-ice) from Mt. Snasa- A summary of ages that relate to deglaciation högarna, still remained ice-covered. This is consis- and from a range of elevations from the study area tent with detailed ice margin reconstructions for can provide a general picture of the vertical rate of deglaciation of the area (cf. Borgström 1989, maps deglaciation (i.e., a vertical deglaciation curve). A and D). In addition, differences in average snow Using radiocarbon ages alone would indicate that thickness over the Holocene may have been impor- high elevation areas deglaciated considerably ear- tant between the sites in which case we may have lier than lower elevation areas as indicated by de- underestimated it at Mt. Åreskutan or overesti- glaciation curve 1 in Figure 8. Note that the portion mated it at Snasahögarna. For example, we would of the curve older than 11 ka is the minimum-age have to more than double the snow thickness from deglaciation curve since it is based only on tree our current estimate at Mt. Åreskutan to adjust the remains. When plotting the TCN ages from high ages to be close to the ages at Snasahögarna. Be- elevation against the radiocarbon ages, there is a cause all boulders from the Snasahögarna moraine clear conflict for those ages from Mt. Åreskutan were from the crest, this minimized the chance that (Fig. 8). How is it that the three ages for glacial exhumation may have been an important process erratics on the mountain summit (adjusted mean that could cause TCN ages to appear too young. age of 10.6 ±0.6 ka) are so much younger than Nevertheless, the TCN ages from the Snasahögarna radiocarbon ages nearby on the same mountain? moraine and Mt. Åreskutan are quite similar as The following three hypotheses attempt to accom- they overlap within one sigma of each other (Table modate both the old radiocarbon ages from high 2). 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-

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

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 and time; 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.

tional age of the erratics. If this hypothesis is true, cause 17 ka boulders to be perceived as 10.6 ka. the exposure age of the erratics should be at least Therefore hypothesis 1 is rejected. as old as the oldest tree megafossils on Mt. Åresku- Hypothesis 2: There was a small glacier or tan, ~17 cal. ka BP. When considering erosion dead-ice remnant on the mountain summit shield- only, an unrealistic erosion rate of 50 mm ka-1 ing the glacier erratics while trees were growing would be needed make 17 ka erratic boulders on nearby. Ice flow indicators (striae, and moulded Mt. Åreskutan have an apparent exposure age of and plucked bedrock) at and surrounding the sum- 10.6 ka. This extremely high rate of erosion would mit of Mt. Åreskutan indicate only westward pa- mean 85 cm of rock loss from the top surface of leo-ice flow despite variations in topography each sampled boulder. Furthermore, if the boulders (ground slope and aspect). The summit area is also experienced a high amount of erosion it is highly covered in glacial erratics. If the summit area was unlikely that this amount would be similar for all host to an alpine glacier, there would be ice flow three boulders to get similar ages. As described indicators that would reflect a thin glacier with above, careful inspection of boulder surfaces and flow that was topographically-influenced. In addi- adjacent material indicated that erosion rates were tion, it is unlikely that there would be many erratics low (Fig. 6). Similarly, as snow has a low density under a former alpine glacier due to glacier flow to and poorly shields cosmic rays, unrealistic snow the margins of the glacier. Thus, abundance of thicknesses, in excess of 10 m for a third of a year erratics in the summit area and ice flow indicators (Gosse and Phillips 2001), would be required to reveal that glacier flow was largely uninfluenced by the underlying topography (i.e., produced by an

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

ice sheet) and that there was not an alpine glacier measurements and other deglaciation evidence on the summit area of Mt. Åreskutan. Moreover, to cited above, it appears objectively much more grow an alpine glacier on the summit of the moun- reasonable to favour the TCN apparent exposure tain would require a cooler climate than today ages as representing deglaciation of the summit which would mean that the tree-line would be even area of Mt. Åreskutan over the old radiocarbon lower than today; the tree megafossils were found ages because: (1) the TCN ages are consistent with 400-500 m above modern tree-line and only 60 m each other and with well-established ages for de- below the summit (Fig. 6). One could invoke a glaciation from middle and lower elevation sites in dead-ice remnant on the mountain covering the the study area, (2) the TCN ages for the Snasa- sites of the boulders, but not the site where trees högarna moraine are consistent and overlap within were living, for a duration of approximately seven 1 with TCN ages from Mt. Åreskutan (Table 2), thousands years after deglaciation at 17 ka. Be- and (3) the radiocarbon ages are incompatible with cause dead-ice is dynamically inert, it wouldn’t other paleoecological and paleoglaciological evi- have left any traces of ice flow such as striae, dence for Scandinavia (Birks et al. 2005, 2006; see moulding or bedrock plucking, or move erratic below). boulders around. However, it would require a rather cold climate to allow for a slow melt-back Paleoglaciological and paleoecological over thousands of years. Whereas this may have considerations been the case for during the Younger Dryas, a slow While, as argued above, the strongest evidence to melt-back climate before the Younger Dryas would oppose the old radiocarbon dates from Mt. Åresku- potentially have been hostile to tree growth. As tan are incompatible TCN exposure ages, a number well, the similarly-aged Snasahögarna moraine of other arguments have been presented in the supports the TCN ages from Mt. Åreskutan (Fig. literature that address paleoecological aspects of 8). Thus, glacier erratics from the summit of Mt. high-elevation Late Glacial trees (Kullman 2002a, Åreskutan were not shielded by an alpine glacier or 2005, 2006, Birks et al. 2005, 2006). In the further dead-ice remnant to give apparent young ages. analysis, it is assumed that Mt. Åreskutan repre- Therefore, hypothesis 2 is rejected. sented a nunatak environment if trees were to have Hypothesis 3: Trees were growing at this site grown there since 17 ka. This is because many and then were overrun but not excavated by the lines of evidence show that the SIS at 17 ka ex- SIS; then erratics were deposited on the mountain tended beyond the west coast of Norway (cf. Kle- summit during the final deglaciation. Glacially man et al. 1997, Sveian 1997, Lundqvist 2002, moulded, plucked and striated bedrock found from Svendsen et al. 2004) and to southern Sweden the summit of the mountain to its base indicate that (Lundqvist and Wohlfarth 2001, Sandgren and the ice sheet was warm-based in this area at some Snowball 2001; Fig. 1, see inset map). point during glacial history. Thus the ice sheet that A nunatak environment would seem to be in- moulded, plucked and striated bedrock would cer- hospitable to tree growth, as the summer tempera- tainly also have excavated soils, peat and wood ture would most likely be below the thermal re- from its base. The almost uninterrupted range of quirement of even the hardiest species, Betula radiocarbon ages of tree remains from Mt. Åresku- pubescens, which currently requires ~9 oC mean tan from 9.6 to 16.7 cal. ka BP indicates that there July temperature in the west Norwegian mountains would not be an opportunity for the ice sheet to (Odland 1996). It appears even more strenuous to readvance over the mountain summit and deposit argue in favour of germination or growth condi- erratic boulders (Table 1) at 10.6 ±0.6 ka, even if it tions during the Younger Dryas (~12.8 to 11.7 ka, did so without erosion of the substrate. Therefore, Muscheler et al. 2008) at 1360 m a.s.l. in the Swed- hypothesis 3 is rejected. ish mountains, in light of documented rapid and The rejection of these three hypotheses implies sustained fall in temperatures during this cold in- that it is unlikely that a sequence of events or proc- terval (Coope et al. 1998, Isarin and Bohncke esses could lead to both the radiocarbon ages and 1999, Brooks and Birks 2000), yet four of the ra- TCN ages from the summit area of Mt. Åreskutan diocarbon dates from Mt. Åreskutan fall within this being compatible with each other. Based on TCN interval (Table 1). In addition, soil was skeletal or

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

non-existent on the glacially scraped Mt. Åresku- aspects of Late Glacial high elevation trees in cen- tan landscape. This would certainly not satisfy the tral Scandinavia. ecological requirements of Picea abies which requires acid soils with adequate nutrients (Nikolov Causes of radiocarbon age bias and Helmisaari 1992). Although, Kullman (2002a) Based on new TCN apparent exposure age results has reported the discovery of Picea abies saplings and paleoecological arguments, the old radiocarbon at Mt. Åreskutan (1385 m a.s.l.; ~415 m above ages from high elevation sites in central Sweden, modern Picea abies tree-line) where the standard- and particularly from Mt. Åreskutan, would seem level mean temperature for June to August is ~5°C, to be unreliable. Re-investigations of the dated tree which is a lower thermal level for sustained tree samples were not possible (Kullman, pers. com. growth and reproduction than generally understood 2006) and we therefore examine four potential (Kullman 2002b). Whatever the actual thermal explanations as to the causes of the radiocarbon requirements for tree growth, at sometime after age bias. Firstly, continuous decay processes deglaciation Betula pubescens, Picea abies, and probably occurred on exposed wood remains as Pinus sylvestris grew at 1360 m a.s.l. near the protective perennial snow banks would not always summit area of Mt. Åreskutan. have existed, especially during the Holocene cli- If Betula, Picea and Pinus occupied nunataks or matic optimum. Such decay processes would pre- other near-ice-margin areas, it is surprising that dictably lead to erroneously young ages (e.g., they did not spread out from there after deglacia- Baker et al. 1987). However, the sixteen dated tion of lower-elevation terrain (cf. Segerström and specimens were reported to be well-preserved and von Stedingk 2003, Birks et al. 2005, Hammarlund included bark for some specimens despite sitting et al. 2004). As revealed through lake sediments, on the ground surface for thousands of years the first flora in the area at lower elevation at 10.5 (Kullman 2002a). The inner and outer portions of a cal. ka BP was a pioneer, mountain-type, assem- stem from Mt. Åreskutan (lab numbers Beta- blage. Picea spread in the area thousands of years 121827 and Beta-120799, Kullman 2002a, 2005; later, in fact only 3500 years ago (Lundqvist 1969, Table 1) gave a similar age (11,440 ±100 and Giesecke and Bennett 2004, Bergman et al. 2005), 11,720 ±90 14C ka BP, respectively) indicating that probably from the west rather than from high ele- if contamination did not affect the inner portion of vation areas in central Sweden (Persson 1975; see the trunk as much as the outer portion, then it may Giesecke and Bennett 2004). not be an important process for this sample. Kullman (2002a) proposed that boreal trees sur- Secondly, another possible cause of older radio- vived the glaciation along the southwest coast of carbon ages is lightning (Libby and Lukens 1973, Norway and migrated eastward early in the Late Harkness and Burleigh 1974, Bowman 1990, Paiva Glacial to early deglaciated parts of the central 2009). Lightning is more common at high elevation Swedish mountains. However, numerous macro- and generates neutrons which may alter the proper- fossil studies show that during the Late Glacial ties of the wood. In this respect, it is noteworthy tree-Betula was absent, or occurred so localized that the old radiocarbon ages (Fig. 8) occur at high that it has remained undetected, in southwest Nor- elevation where there is a higher frequency of way (Birks et al. 2005 and references therein). lightning. Similarly, there is no macrofossil or fossil stomata Thirdly, contamination processes may lead to evidence for the presence of Picea and Pinus to radiocarbon ages being too old. It is more likely indicate that they survived the glaciation in south- that contamination of ancient samples by younger west Norway or that they were present in the Late carbon would have a much greater affect than con- Glacial (Giesecke and Bennett 2004, Birks et al. tamination of younger specimens by older carbon 2005). Hence, the refuge area for the trees thought (Bowman 1990). In this case, the dated wood to have grown on Mt Åreskutan appears to have would be ancient. Ancient trees may therefore been unpopulated with them for the time period of derive from stratigraphic units in the up-ice region the Late Glacial. We refer the reader to Kullman that date beyond the limit of radiocarbon dating (2002a, 2005, 2006) and Birks et al. (2005, 2006) (>50 14C ka BP; Lundqvist 1967) and that contain for further detailed discussion on paleoecological pollen and wood fragments of the same tree species

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

as those found on Mt. Åreskutan. However, the cal rates of deglaciation can be higher. The ap- well-preserved physical condition of the Mt. Åre- proximate vertical rate of deglaciation for the Mt. skutan samples (Kullman 2002a) potentially pre- Åreskutan area, using central values for TCN ad- cludes this possibility. justed exposure ages and radiocarbon dates (data Fourthly, contamination processes may lead to points 6, 18, 19, and 20; Fig. 1, and Fig. 8, curve radiocarbon ages being too young. Although less 3), may have been as high as ~5 m year-1. There is effective (Bowman 1990), there also is a possibility no evidence apart from the old radiocarbon dates for old/dead carbonate in the catchments where for a local vertical rate of deglaciation as low as tree remains have been found. Lundqvist (1969, 0.007 m year-1 (Fig. 8, curve 1). This means that figure 17) shows the occurrence of limestone out- high elevation areas within central Sweden re- cropping 4-6 km east (up-ice) and southeast of Mt. mained ice covered and could not be colonised by Åreskutan (Lundegårdh et al. 1984). We speculate trees as early as 17 ka. Interestingly, TCN exposure that the ice sheet could have incorporated lime- age dating work ~195 km south of our study area at stone erosional products in till on Mt Åreskutan Elgåhogna close to the Swedish-Norwegian border and that the old/dead carbonate could be utilised by indicates that during the LGM the ice sheet surface the plants, or absorbed after their death through was above 1460 m a.s.l., and that rapid deglacia- groundwater (Bowman 1990), leading to some tion commenced about 12 ka (Goehring et al. erroneously older radiocarbon dates from Mt Åre- 2008). This corresponds well with estimates for skutan. deglaciation from our study area. We concur with Birks et al. (2005) that the radiocarbon dated high elevation wood samples should be assessed carefully for possible sources of Conclusion contamination and have their tree-rings analysed Three terrestrial cosmogenic nuclide (TCN) expo- against the Scandinavian Holocene tree-ring se- sure ages from the summit of Mt. Åreskutan (1360 quence. In particular we suggest for these samples m a.s.l.) in central Sweden are consistent with each that cellulose be chemically extracted from both other (adjusted mean age of 10.6 ±0.6 ka) and are the inner and outer portions of all stems, and dated. similar to lower-elevation dates for deglaciation Implications of new deglaciation ages from the region. However, the TCN ages are incompatible with reported old radiocarbon dates With the rejection of radiocarbon dates from high from wood of three tree species from the summit elevation locations in central Sweden and using area (as old as 16.7 ±0.2 cal. ka BP). We cannot new high-elevation TCN apparent exposure age find a plausible hypothesis that accommodates both data, the timing of deglaciation at high elevation the radiocarbon and TCN ages from this site. and the vertical rate of deglaciation must be re- Three TCN samples were collected from the examined. These values can only be estimated highest elevation SIS moraine in Sweden. Located because of the uncertainties associated with the 35 km down-ice from Mt Åreskutan, the Snasa- TCN ages along with the difference in mean age högarna SIS moraine samples yielded consistent between the Mt. Åreskutan and Snasahögarna mo- TCN ages (sampled at 1125-1149 m a.s.l.; adjusted raine sites. Nevertheless, high-elevation areas de- mean age of 12.0 ±0.6 ka). The difference in TCN glaciated sometime after ~12.0-10.6 ka, coinciding ages between Mt. Åreskutan and the Snasahögarna approximately with the termination of the Younger SIS moraine probably reflects a geographical dif- Dryas cold interval (11.7 ka, Muscheler et al. 2008; ference in the timing of deglaciation between sites Fig. 8, curve 2). During the termination of the and possibly a difference in the actual historical Younger Dryas, the margin of the SIS was ~85 km snow depths. and ~60 km west of Mt. Åreskutan and Mt. Snasa- We reject the reported old radiocarbon ages on högarna, respectively (Reite 1994, Sveian 1997). the basis of (1) incompatibility with consistent The mean vertical rate of deglaciation for the study TCN ages and (2) incompatibility with pa- -1 area was ~0.5 m year . However, as this value is leoecological and paleoglaciological evidence for derived from data over a large area, it is averaging deglaciation (Birks et al. 2005). Nevertheless, the estimates over space and time; and thus local verti- mere presence of tree remains of three different

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

species at this high elevation, and well-above (400- mate and a smaller ice sheet during the Swedish Middle Weichselian (MIS 3)? Boreas 39, 367-376. 500 m) the modern-tree line, still allows for some speculation with regards to their climatic interpre- Arnold, N. S., van Andel, T. H., Valen, V., 2002. Extent and dynamics of the Scandinavian Ice Sheet during oxygen iso- tation. The problem lies in reliably dating these tope stage 3 (65,000–25,000 yr B.P.). Quaternary Research tree remains. Assuming these are younger tree 57, 38–48. remains, consistent with TCN age results, we sug- Baker, P. E., Harkness, D. D., Roobol, M. J., 1987. Discussion gest that contamination from calcareous bedrock or on differences in radiocarbon ages reported from St. Kitts, West Indies. Journal of the Geological Society 144, 205- neutron production from lightning may have 206. caused the age bias. We also strongly recommend Balco, G., Stone, J.O.H., Porter, S.C., Caffee, M.W., 2002. that specimens for radiocarbon dating be thor- Cosmogenic nuclide ages for New England coastal mo- oughly tested to ascertain possible sources of con- raines, Martha’s Vineyard and Cape Cod, Massachusetts, USA. Quaternary Science Reviews 21, 2127-2135. tamination and that complementary dating techniques be employed before proposing radical Balco, G., Stone, J.O.H., Lifton, N.A., Dunai, T.J., 2008. A complete and easily accessible means of calculating surface changes to the established paleoglaciological and exposure ages or erosion rates from 10Be and 26Al meas- paleoecological history. urements. Quaternary Geochronology 3, 174-195. Mt. Åreskutan within central Sweden did not Bergman, J., Hammarlund, D., Hannon, G., Barnekow, L., deglaciate as early as ~17 ka. High-elevation areas Wohlfarth, B., 2005. Deglacial vegetation succession and Holocene tree-limit dynamics in the Scandes Mountains, in this region deglaciated ~12.0-10.6 ka, coinciding west-central Sweden: stratigraphic data compared to mega- approximately with the termination of the Younger fossil evidence. Review of Palaeobotany and Palynology Dryas cold interval (11.7 ka). The vertical rate of 134, 129-151. deglaciation may have been as high as ~5 m year-1 Bergman, J., Wastegård, S., Hammarlund, D., Wohlfarth, B., -1 Roberts, S.J., 2004. Holocene tephra horizons at Klocka but was more typically 0.5 m year . We propose Bog, west-central Sweden: aspects of reproducibility in su- that sometime after deglaciation Betula pubescens, barctic peat deposits. Journal of Quaternary Science 19, Picea abies, and Pinus sylvestris grew at 1360 m 241-249. a.s.l. near the summit area of Mt. Åreskutan. Bevington and Robinson, 1992. Data Reduction and Error Analysis for the Physical Sciences. 2nd edition. New York, McGraw-Hill. Acknowledgements 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 We thank Maria Miguens-Rodriguez and Henriette Kullman (2002). Journal of Biogeography 32, 1461-1471. Linge for chemistry and preparation of AMS tar- Birks, H. H., Larsen, E., Birks, H. J. B., 2006. On the presence gets. Jan Lundqvist, Hilary Birks, and Terri La- of late-glacial trees in western Norway and the Scandes: a course provided fruitful discussions, and Daniel further comment. Journal of Biogeography 33, 376-378. Veres, Jakob Heyman, Marie Koitsalu, Helena Borell, R., Offerberg, J., 1955. Geokronologiska Alexanderson and Jonas Bergman assisted in the undersökningar inom Indalsälvens dalgång mellan Bergeforsen och Ragunda. SGU, Ca 31. fieldwork. Helena Alexanderson, Jan Lundqvist, and Hilary Birks, helped improve the manuscript. Borgström, I., 1979. De Geer moraines in a Swedish mountain area?, Geografiska Annaler 61, 35-42. Funding was provided by the Swedish Society for Borgström, I., 1989. Terrängformerna och den glaciala Anthropology, the Gerard De Geer fund, and the utvecklingen i södra fjällen. Meddelande från Carl Mannerfelt fund. Johnsen was in charge of Naturgeografiska institutionen vid Stockholms universitet, and performed all aspects of the project: concep- Nr A 234, 133 p. (PhD thesis). tion, acquiring funding, TCN-sampling, data analy- Bowman, S., 1990. Radiocarbon dating. University of Califor- sis, and writing, except for sample measurements nia Press, Berkeley. 64 p. which were completed by Fabel. Fabel and Stro- Briner, J.P., Swanson, T.W., Caffee, M., 2001. Late Pleisto- cene cosmogenic Cl-36 glacial chronology of the south- even provided feedback on the science and writing. western Ahklun Mountains, Alaska. Quaternary Research 56, 148-154. Brooks, S.J., Birks, H.J.B., 2000. Chironomid-inferred late- References glacial and early-Holocene mean July air temperatures for Kråkenes Lake, western Norway. Journal of Paleolimnology Alexanderson, H., Johnsen, T. F., Murray, A.S., 2010. Re- 23, 77-89. dating the Pilgrimstad Interstadial with OSL: a warmer cli-

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

Child, D., Elliott, G., Mifsud, C., Smith, A.M., Fink, D., 2000. Helmens, K. F., Bos, J. A. A., Engels, S., Van Meerbeeck, C. Sample processing for earth science studies at ANTARES. J., Bohncke, S.J.P., Renssen, H., Heiri, O., Brooks, S. J., Nuclear Instruments and Methods in Physics Research Sec- Seppä , H., Birks, H. J. B., Wohlfarth, B., 2007a. Present- tion B, Beam Interactions with Materials and Atoms 172, day temperatures in northern Scandinavia during the last 856-860. glaciation. Geology 35, 987-990. Clark, P.U., Brook, E.J., Raisbeck, G.M., Yiou, F., Clark, J., Helmens, K.F., Johansson, P.W., Räsänen, M. E., Alexander- 2003. Cosmogenic Be-10 ages of the Saglek Moraines, son, H., Eskola, K.O., 2007b. Ice-free intervals at Sokli con- Torngat Mountains. Labrador Geology 31, 617-620. tinuing into Marine Isotope Stage 3 in the central area of the Scandinavian glaciations. Geological Society of Finland, Coope, G.R., Lemdahl, G., Lowe, J. J., Walkling, A., 1998. Bulletin 79, 17-39. Temperature gradients in northern Europe during the last glacial–Holocene transition (14–9 14C kyr BP) interpreted Heyman, J., 2004. Pre-recent End Moraines in the Swedish from coleopteran assemblages. Journal of Quaternary Sci- Mountains. Examensarbete i naturgeografi Stockholms ence 13, 419-433. universitet, Examensarbete N-54. Dahl, S. O., Nesje, A., 1996. A new approach to calculating Heyman, J., Hättestrand, C., 2006. Morphology, distribution Holocene winter precipitation by combining glacier equilib- and formation of relict marginal moraines in the Swedish rium-line altitudes and pine-tree limits: a case study from mountains. Geografiska Annaler 88, 253-265. Hardangerjøkulan, central Norway. The Holocene 6, 381- Isarin, R.F.B., Bohncke, S.J.P., 1999. Mean July Temperatures 398. during the Younger Dryas in Northwestern and Central De Geer, G., 1940. Geochronologia Suecica Principles. Kungl. Europe as Inferred from Climate Indicator Plant Species. Svenska Vetenskapsakademiens Handlingar III : 18 : 6, 367 Quaternary Research 51, 158-173. pp. Johnsen, T.F., Alexanderson, H., Fabel, D., Freeman, Denton, G.H., Hughes, T.J. (eds.)., 1981. The Last Great Ice S.P.H.T., 2009. New 10Be cosmogenic ages from the Vim- Sheets. Wiley, New York, NY. merby moraine confirm the timing of Scandinavian Ice Sheet deglaciation in southern Sweden. Geografiska An- Ek, L.-G., 2004. The establishment of Norway spruce (Picea naler 91, 113-120. abies (L.) Karst.) on two mountains in the Åre area - a fol- low-up of the macrofossil finds on Mount Åreskutan. Johnsen, T.F., Olsen, L., Murray, A.S. submitted. OSL ages in Institutionen för skoglig vegetationsekologi, Sveriges central Norway confirm a MIS 2 interstadial (25-20 ka) and lantbruksuniversitet, Umeå, Sweden. Examensarbeten nr 1 a dynamic Scandinavian ice sheet. Quaternary Science Re- (Masters thesis) views. Fabel, D., Stroeven, A. P., Harbor, J., Kleman, J., Elmore, D., Kjær, K.H., Lagerlund, E., Adrielsson, L., Thomas, P.J., Fink, D., 2002. Landscape preservation under Fennoscan- Murray, A., Sandgren, P., 2006. The first independent chro- dian ice sheets determined from in situ produced 10Be and nology for Middle and Late Weichselian sediments from 26Al. Earth and Planetary Science Letters 201, 397-406. southern Sweden and the island of Bornholm. Geologiska Föreningens; Stockholm Förhandlingar 128, 209-220. Fabel, D., A.P. Stroeven, K.I.T. Dahlgren, J. Harbor, D. Fink, M.W. Caffee, in review. Enhanced ice sheet deglaciation Kjemperud, A., 1981. A shoreline displacement investigation chronologies using cosmogenic nuclide dating of eskers, from Frosta in Trondheimsfjorden, Nord-Trønde1ag, glaciofluvial deltas, meltwater channels and lineations. Qua- Norway. Norsk Geologisk Tidsskrift 61, 1-15. ternary Research. Kleman, J., Hättestrand, C., Borgström, I., Stroeven, A., 1997. Giesecke, T., Bennett, K.D., 2004. The Holocene spread of Fennoscandian paleoglaciology reconstructed using a glacial Picea abies (L.) Karst. in Fennoscandia and adjacent areas. geological inversion model. Journal of Glaciology 43, 283- Journal of Biogeography 31, 1523-1548. 299. Goehring, B.M., Brook, E.J., Linge, H., Raisbeck, G.M., Yiou, Kohl, C.P., Nishiizumi, K., 1992. Chemical isolation of quartz F., 2008. Beryllium-10 exposure ages of erratic boulders in for measurement of in situ-produced cosmogenic nuclides, southern Norway and implications for the history of the Geochim. Cosmochim. Acta 56, 3586-3587. Fennoscandian Ice Sheet. Quaternary Science Reviews 27, Kullman, L., 1995. Holocene tree-limit and climate history 320-336. from the Scandes Mountains, Sweden. Ecology 76, 2490- Gosse, J. C., Phillips, F. M., 2001. Terrestrial in situ cos- 2502. mogenic nuclides: theory and application. Quaternary Sci- Kullman, L., 1998. Non-analogous tree flora in the Scandes ence Reviews 20, 1475-1560. Mountains, Sweden, during the early Holocene-macrofossil Hallet, B., Putkonen, J., 1994. Surface dating of dynamic evidence of rapid geographic spread and response to palaeo- landforms: young boulders on aging moraines. Science 265, climate. Boreas 27, 153-161. 937-940. Kullman, L., 2000. The Geoecological histroy of Picea abies Hammarlund, D., Velle, G., Wolfe, B. B., Edwards, T. W.D, in northern Sweden and adjacent parts of Norway. A con- Barnekow, L. , Bergman, J., Holmgren, S., Lamme, S., trarian hypothesis of postglacial tree immigration patterns. Snowball, I., Wohlfarth, B., Possnert, G., 2004. Palaeolim- Geoöko, 21, 141-172. nological and sedimentary responses to Holocene forest re- Kullman, L., 2001. Immigration of Picea abies into North- treat in the Scandes Mountains, west-central Sweden. The Central Sweden. New evidence of regional expansion and Holocene 14, 862-876. tree-limit evolution. Nordic journal of Botany 21, 39-54. Harkness, D.D., Burleigh, R., 1974. Possible carbon-14 en- richment in high altitude wood. Archaeometry 16, 121-127.

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

Kullman, L., 2002a. Boreal tree taxa in the central Scandes Lundqvist, J., 1973. Isavsmältningens förlopp i Jämtlands län. during the Late-Glacial: implications for Late-Quaternary Sveriges Geologiska Undersökning (Ser. C.) 681, 187 p. forest history. Journal of Biogeography 29, 1117-1124. Lundqvist, J., 2002. Istider och mellanistider. In: Fredén, C. Kullman, L., 2002b. Rapid recent range-margin rise of tree (Ed.), Berg och jord, Sveriges Nationalatlas. Stockholm, p. and shrub species in the Swedish Scandes. Journal of Ecol- 124-135. ogy 90, 68-77. Lundqvist J., Wohlfarth, B., 2001. Timing and east-west Kullman, L., 2004a. Early Holocene appearance of mountain correlation of south Swedish ice marginal lines during the birch (Betula pubescens ssp. tortuosa) at unprecedented high Late Weichselian. Quaternary Science Reviews 20, 1127- elevations in the Swedish Scandes: megafossil evidence ex- 1148 posed by recent snow and ice recession. Arctic, Antarctic, Lunkka, J.P., Murray, A., Korpela, K., 2008. Weichselian and Alpine Research 36, 172-180. sediment succession at Ruunaa, Finland, indicating a Mid- Kullman, L., 2004b. Tree-limit landscape evolution at the Weichselian ice-free interval in eastern Fennoscandia. Bo- southern fringe of the Swedish Scandes (Dalarna province) - reas 37, 234-244. Holocene and 20th century perspectives. Fennia 182, 73-94. Middleton, R., Brown L., Dezfouly-Arjomandy, B., Klein, J., Kullman, L., 2005. Reply - On the presence of late-glacial 1993. On 10Be standards and the half-life of 10Be. Nuclear trees in the Scandes. Journal of Biogeography 32, 1499- Instruments and Methods in Physics Research B, 82, 399- 1500. 403. Kullman, L., 2006. Late-glacial trees from arctic coast to Mörner, N.-A., 1980. The Fennoscandian uplift: Geological alpine tundra: response to Birks et al. 2005 and 2006. Jour- data and their geodynamical implications. In: Mörner, N.-A. nal of Biogeography 33, 376-378. (ed.), Earth rheology, isostasy and eustasy. Chichester: John Wiley, 251-284. Kullman, L., Kjällgren, L., 2000. A coherent postglacial tree- limit chronology (Pinus sylvestris L.) for the Swedish Scan- Muscheler, R., Kromer, B., Bjorck, S., Svensson, A., Frie- des: aspects of paleoclimate and "recent warming," based on drich, M., Kaiser, K.F., Southon, J., 2008. Tree rings and ice megafossil evidence. Arctic, Antarctic, and Alpine Research cores reveal 14C calibration uncertainties during the 32, 419-428. Younger Dryas. Nature Geoscience 1, 263-267. Kullman, L., Kjällgren, L., 2006. Holocene pine tree-line Nikolov, N., Helmisaari, H., 1992. Silvics of the circumpolar evolution in the Swedish Scandes: Recent tree-line rise and boreal forest trees. In: A systems analysis of the global bo- climate change in a long-term perspective. Boreas 35, 159- real forest. Shugart, H., Leemans, R., Bonan, G.B., Eds. 168. Cambridge University Press, Cambridge. 13-84 p. Lal, D., 1991. Cosmic-ray labeling of erosion surfaces: in situ Odland, A., 1996. Differences in the vertical distribution nuclide production rates and erosion models. Earth and pattern of Betula pubescens in Norway and its ecological Planetary Science Letters 104, 424-439. significance. Paläoklimaforschung, 20, 43-59. Lambeck, K., Smither, C., Johnston. P., 1998. Sea-level Olsen, L., Sveian, H., Bergstrøm, B., Selvik, S.F., Lauritzen, change, glacial rebound and mantle viscosity for northern S.-E., Stokland, Ø, Grøsfjeld, K., 2001. Methods and strati- Europe. Geophysical Journal International 134, 102-144. graphies used to reconstruct Mid- and Late Weichselian palaeoenvironmental and palaeoclimatic changes in Norway. Libby, L.M., H.R. Lukens, 1973, Production of Radiocarbon Norges geologiske undersøkelse Bulletin 438, 21-46. in Tree Rings by Lightning Bolts, J. Geophys. Res. 78, 5902-5903. Olsen, L., Sveian, H., van der Borg, K., Bergstrøm, B., Broekmans, M., 2002. Rapid and rhythmic ice sheet fluctua- Licciardi, J.M., Clark, P.U., Brook, E.J., Pierce, K.L., Kurz, tions in western Scandinavia 15-40 Kya – a review. Polar M.D., Elmore, D., Sharma, P., 2001. Cosmogenic He-3 and Research 21, 235-242. Be-10 chronologies of the late Pinedale northern Yellow- stone ice cap, Montana, USA. Geology, 29, 1095-1098. Paiva, G.S., 2009. Terrestrial gamma-ray flashes caused by neutron bursts above thunderclouds. Journal of Applied Lidén, R., 1938. Den senkvartara strandförskjutningens Physics 105, 083301. förlopp och kronologi i Angermanland. Geologiska Foreningens i Stockholm Forhandlingar 60, 397-404. Persson, C., 1975. Speculations on the immigration of spruce into Sweden. Geologiska föreningens förhandlingar. Lundberg, A., Richardson-Näslund, C., Andersson, C., 2006. Snow density variations: consequences for ground- Phillips, F.M., Zreda, M.G., Evenson, E.B., Hall, R.D., penetrating radar. Hydrological Processes 20, 1483-1495. Chadwick, O.A., Sharma, P., 1997. Cosmogenic Cl-36 and Be-10 ages of Quaternary glacial and fluvial deposits of the Lundegårdh, P.H., Gorbatschev, R., Kornfält, K.-A., Wind River Range, Wyoming. Geological Society of Amer- Strömberg, A., Karis, L., Zachrisson, E., Sjöstrand, T., ica Bulletin 109, 1453-1463. Skoglund, R., 1984. Karta över berggrunden i Jämtlands län (utom förutvarande Fjällsjö k n). Bedrock geology of Putkonen, J., Swanson, T., 2003. Accuracy of cosmogenic Jämtland County. Sveriges Geologiska Undersökning Ca ages for moraines. Quaternary Research 59, 255-261. 53. Map. Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, Lundqvist, J., 1967. Submoräna sediment i Jämtlands län. J.W., Bertrand, C., Blackwell, P.G., Buck, C.E., Burr, G., Sveriges Geologiska Undersökning C 618. 267 p. Cutler, K.B., Damon, P.E., Edwards, R.L., Fairbanks, R.G., Friedrich, M., Guilderson, T.P., Hughen, K.A., Kromer, B., Lundqvist, J., 1969. Beskrivning till jordartskarta över Mc-Cormac, F.G., Manning, S., Bronk Ramsey, C., Reimer, Jämtlands län. Sveriges Geologiska Undersökning (Ser. Ca.) R.W., Remmele, S., Southon, J.R., Stuiver, M., Talamo, S., 45, 418 p. Taylor, F.W., van der Plicht, J., Weyhenmeyer, C.E., 2004.

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

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 production. 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 history 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 råáîÉêëáíó açÅíçê 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 ëçìíÜïÉëíÉêå s®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 ~åÇ mêÉÄçêÉ~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 ~ë êÉÅçêÇÉÇ 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~ëí URMM óÉ~êë ~ë áåÇáÅ~íÉÇ Äó ãáÅêçÑçëëáäë ~åÇ ÅÜÉãáÅ~ä é~ê~ãÉíÉêë áå ëÉÇáãÉåíëK aáëëÉêí~íáçå kç RK _àêÅâI gçå~ëI NVVVW bîÉåí ëíê~íáÖê~éÜó Ñçê íÜÉ i~ëí dä~Åá~äJeçäçÅÉåÉ íê~åëáíáçå áå É~ëíÉêå ãáÇÇäÉ pïÉÇÉåK aáëëÉêí~íáçå kçK SK hêáëíá~å pÅÜçåáåÖI OMMNK j~êáåÉ ÅçåÇáíáçåë áå ãáÇÇäÉ pïÉÇÉå ÇìêáåÖ íÜÉ ä~íÉ tÉáÅÜëÉäá~å ~åÇ É~êäó eçäçÅÉåÉ ~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~å é~êí çÑ íÜÉ _~äíáÅ pÉ~ Ä~ëáåK aáëëÉêí~íáçå kçK VK ^åå~ eÉÇÉåëíêãI OMMNK b~êäó eçäçÅÉåÉ ëÜçêÉ Çáëéä~ÅÉãÉåí áå É~ëíÉêå pîÉ~ä~åÇI pïÉÇÉåI Ä~ëÉÇ çå Çá~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~å ÇÉîÉäçéãÉåí 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 råáîÉêëáíó açÅíçê 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 ÇáÑÑÉêÉåí íóéÉë çÑ î~êîÉ ëÉêáÉë Ñêçã íÜÉ _~äíáÅ _~ëáåK aáëëÉêí~íáçå kçK NK c~âìäíÉíëçééçåÉåíW mêçÑK _àêå j~äãÖêÉå K o~íÜå~ëáêá mêÉã~íÜáä~âÉI OMMPW i~íÉ nì~íÉêå~êó é~ä~ÉçÉÅçäçÖáÅ~ä ÉîÉåí ëíê~íáÖê~éÜó áå íÜÉ eçêíçå mä~áåëI ÅÉåíê~ä pêá i~åâ~ J ïáíÜ ÅçåíêáÄìíáçåë íç íÜÉ êÉÅÉåí éçääÉå Ñäçê~K aáëëÉêí~íáçå kçK OK c~âìäíÉíëçééçåÉåíW mêçÑK cê~åÅçáëÉ d~ëëÉK ^åÖÉäáÅ~ cÉìêÇÉ~åI OMMQW m~ä~ÉçÉåîáêçåãÉåí áå åçêíÜJïÉëíÉêå oçã~åá~ ÇìêáåÖ íÜÉ ä~ëí NRIMMM óÉ~êëK aáëëÉêí~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~íÉ ÅÜ~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 ÇÉîÉäçéãÉåí áå íïç pïÉÇáëÜ íçïåëI â 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 dblildv OMMSÓ e™â~å dêìÇÇI OMMSK qêÉÉ êáåÖë ~ë ëÉåëáíáîÉ éêçñáÉë çÑ é~ëí Åäáã~íÉ ÅÜ~åÖÉK aáëëÉêí~íáçå kçK NK c~âìäíÉíëçééçåÉåíW mêçÑK _êá~å iìÅâã~åK räÑ gçåëÉääI OMMSK pìäÑìê áå éçä~ê áÅÉ ~åÇ ëåçïK fåíÉêéêÉí~íáçåë çÑ é~ëí ~íãçëéÜÉêÉ ~åÇ Åäáã~íÉ íÜêçìÖÜ 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 eçäçÅÉåÉ ëéÉäÉçíÜÉãë ~åÇ íÜÉáê ÖêçïíÜ ÉåîáêçåãÉåíë áå pïÉÇÉåK aáëëÉêí~íáçå kçK PK c~âìäíÉíëçééçåÉåíW mêçÑK cê~åâ jÅaÉêãçííK m~íêáâ häáåíÉåÄÉêÖI OMMTK jçêÉ ï~íÉêI äÉëë Öê~ëë\ ^å ~ëëÉëëãÉåí çÑ êÉëçìêÅÉ ÇÉÖê~Ç~íáçå ~åÇ ëí~âÉÜçäÇÉêëÛ éÉêÅÉéíáçåë çÑ ÉåîáêçåãÉåí~ä ÅÜ~åÖÉ áå lãÄìÖ~ Öê~ëëä~åÇI åçêíÜÉêå k~ãáÄá~K aáëëÉêí~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 ~åÇ ä~âÉ äÉîÉä î~êá~Äáäáíó áå bãé~â~~á `ê~íÉêK aáëëÉêí~íáçå kçK RK c~âìäíÉíëçééçåÉåíW mêçÑK eÉåêó i~ãÄK a~åáÉä pK sÉêÉëI OMMTK qÉêêÉëíêá~ä êÉëéçåëÉ íç a~åëÖ~~êÇJlÉëÅÜÖÉê ÅóÅäÉë ~åÇ eÉáåêáÅÜ ÉîÉåíëW íÜÉ ä~ÅìëíêáåÉ êÉÅçêÇ çÑ iÉë bÅÜÉíëI ëçìíÜJÉ~ëíÉêå cê~åÅÉK aáëëÉêí~íáçå kçK SK c~âìäíÉíëçééçåÉåíW mêçÑK gçÜå gK içïÉK vçëÜáÜáêç pÜáÄìçI OMMTK jçÇÉääáåÖ ï~íÉê ~åÇ ëçäìíÉ Ñäçïë ~í ä~åÇJëÉ~ ~åÇ ä~åÇJ~íãçëéÜÉêÉ áåíÉêÑ~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 éêçÅÉëëK aáëëÉêí~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~íáçå kçK 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 íç áåä~åÇ ~åÇ Åç~ëí~ä ï~íÉêëK aáëëÉêí~íáçå kçK NMK c~âìäíÉíëçééçåÉåíW mêçÑK ^åÇêÉ~ oáå~äÇçK bäáå kçêëíêãI OMMUK i~íÉ nì~íÉêå~êó Åäáã~íÉ ~åÇ ÉåîáêçåãÉåí~ä ÅÜ~åÖÉ áå íÜÉ ëìããÉê ê~áåÑ~ää êÉÖáçå çÑ pçìíÜ ^ÑêáÅ~ J ^ ëíìÇó ìëáåÖ íêÉÉë ~åÇ ïÉíä~åÇ éÉ~í ÅçêÉë ~ë å~íìê~ä ~êÅÜáîÉëK aáëëÉêí~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 çÑ fåÑçêã~íáçå póëíÉãë Ñçê fåíÉÖê~íÉÇ t~íÉê j~å~ÖÉãÉåíK aáëëÉêí~íáçå kçK NOK c~âìäíÉíëçééçåÉåíW mêçÑK aÉååáë iÉííÉåã~áÉêK j~ííá~ë ÇÉ tçìäI OMMUK oÉëéçåëÉ çÑ Öä~ÅáÉêë íç Åäáã~íÉ ÅÜ~åÖÉK aáëëÉêí~íáçå kçK NPK c~âìäíÉíëçééçåÉåíW mêçÑK oçÖÉê _ê~áíÜï~áíÉK _ê~ÇäÉó tK dççÇÑÉääçïI OMMUK oÉäáÅí åçåJÖä~Åá~ä ëìêÑ~ÅÉë ~åÇ ~ìíçÅÜíÜçåçìë ÄäçÅâÑáÉäÇë áå íÜÉ åçêíÜÉêå 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 åçêíÜÉêå pïÉÇÉå Ó áåíÉêéêÉí~íáçåë Ñêçã Ñçëëáä ~åÇ ãçÇÉêå éçääÉå êÉÅçêÇëK aáëëÉêí~íáçå kçK NRK c~âìäíÉíëçééçåÉåíW mêçÑK açå~íÉää~ j~ÖêáK iáåÇ~ ^ãéÉäI OMMUK a~åëÖ~~êÇJlÉëÅÜÖÉê ÅóÅäÉë ~åÇ eÉáåêáÅÜ ÉîÉåíë áå ïÉëíÉêå bìêçéÉ Ó ^ Çá~íçã éÉêëéÉÅíáîÉK 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~âÉ ëÉÇáãÉåíëK aáëëíÉêí~íáçå kçKNTK c~âìäíÉíëçééçåÉåíW mêçÑK fåÖã~ê oÉåÄÉêÖK `Üêáëíáå~ bK gçåëëçåI OMMVK eçäçÅÉåÉ Åäáã~íÉ ~åÇ ~íãçëéÜÉêáÅ ÅáêÅìä~íáçå ÅÜ~åÖÉë áå åçêíÜÉêå cÉååçëÅ~åÇá~ J áåíÉêéêÉí~íáçåë Ñêçã ä~ÅìëíêáåÉ çñóÖÉå áëçíçéÉ êÉÅçêÇëK aáëëÉêí~íáçå kçK NUK c~âìäíÉíëçééçåÉåíW aêK mÜáäáé _~êâÉêK h~êáå bÄÉêíI OMMVK `ÉåçòçáÅ ä~åÇëÅ~éÉ Éîçäìíáçå áå åçêíÜÉêå pïÉÇÉåK dÉçãçêéÜçäçÖáÅ~ä áåíÉêéêÉí~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 OMK c~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 ONK c~âìäíÉíëçééçåÉåíW mêçÑK cê~åâ iÉÜãâìÜäK qáãçíÜó gçÜåëÉåI OMNMK i~íÉ nì~íÉêå~êó áÅÉ ëÜÉÉí Üáëíçêó ~åÇ Çóå~ãáÅë áå ÅÉåíê~ä ~åÇ ëçìíÜÉêå pÅ~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 different time periods, and using comparative geochronological methods demonstrates that the SIS was highly dynamic and sensitive to environmental change.

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.

ISBN 978-91-7447-068-0 ISSN 1653-7211

Department of Physical Geography and Quaternary Geology