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
© Timothy F. Johnsen ISBN: 978-91-7447-068-0 ISSN: 1653-7211
Paper I: © Swedish Society for Anthropology and Geography Paper II: © The Boreas Collegium
Layout: Timothy F. Johnsen (except for papers I and II)
Cover photo: View from Handöl, Sweden to north-northeast over the shoreline of Lake Ånnsjön and to Mt. Åreskutan at background right (Timothy F. Johnsen, May 2005)
Printed in Sweden by PrintCenter US-AB
Late Quaternary ice sheet history and dynamics in central and southern Scandinavia
Doctoral dissertation 2010 Timothy F. Johnsen Department of Physical Geography and Quaternary Geology Stockholm University
Abstract
Recent work suggests an emerging new paradigm for the Scandinavian ice sheet (SIS); one of a dynamically fluctuating ice sheet. This doctoral research project explicitly examines the history and dynamics of the SIS at four sites within Sweden and Norway, and provides results covering different time periods of glacial his- tory. Two relatively new dating techniques are used to constrain the ice sheet history: the optically stimulated luminescence (OSL) dating technique and the terrestrial cosmogenic nuclide (TCN) exposure dating tech- nique. OSL dating of interstadial sediments in central Sweden and central Norway indicate ice-free conditions during times when it was previously inferred the sites were occupied by the SIS. Specifically, the SIS was absent or restricted to the mountains for at least part of Marine Isotope Stage 3 around 52 to 36 kyr ago. Inland portions of Norway were ice-free during part of the Last Glacial Maximum around 25 to 20 kyr ago. Consistent TCN exposure ages of boulders from the Vimmerby moraine in southern Sweden, and their compatibility with previous estimates for the timing of deglaciation based on radiocarbon dating and varve chronology, indicate that the southern margin of the SIS was at the Vimmerby moraine ~14 kyr ago. In central Sweden, consistent TCN ages for boulders on the summit of Mt. Åreskutan and for the earlier deglaciated highest elevation moraine related to the SIS in Sweden agree with previous estimates for the timing of deglaciation around 10 ka ago. These results indicate rapid decay of the SIS during deglaciation. Unusually old radiocarbon ages of tree remains previously studied from Mt. Åreskutan are rejected on the basis of incompatibility with consistent TCN ages for deglaciation, and incompatibility with established paleoecological and paleoglaciological reconstructions. Altogether this research conducted in different areas, covering different time periods, and using compara- tive geochronological methods demonstrates that the SIS was highly dynamic and sensitive to environmental change.
Keywords: Scandinavian ice sheet, ice sheet dynamics, luminescence dating, cosmogenic exposure dating, geochronology, moraine, interstadial, deglaciation, nunatak
Late Quaternary ice sheet history and dynamics in central and southern Scandinavia
Timothy F. Johnsen
Department of Physical Geography and Quaternary Geology, Stockholm University, Sweden
This doctoral thesis consists of a summary and four appended papers. The papers are listed below and are referred to as Paper I-IV in the summary.
Paper I: Johnsen, T.F., Alexanderson, H., Fabel, D., Freeman, S.P.H.T. 2009. New 10Be cosmogenic ages from the Vimmerby moraine confirm the timing of Scandinavian Ice Sheet deglaciation in southern Sweden. Geografiska Annaler: Series A, Physical Geography, 91: 113–120. – Reprinted with permission of the Swed- ish Society for Anthropology and Geography.
Paper II: Alexanderson, H., Johnsen, T., Murray, A.S. 2010. Re-dating the Pilgrimstad Interstadial with OSL: a warmer climate and a smaller ice sheet during the Swedish Middle Weichselian (MIS 3)? Boreas, 39: 367–376. – Reprinted with permission of The Boreas Collegium.
Paper III: Johnsen, T.F., Olsen, L., Murray, A., Submitted. OSL ages in central Norway confirm a MIS 2 interstadial (25-20 ka) and a dynamic Scandinavian ice sheet. Quaternary Science Reviews.
Paper IV: Johnsen, T.F., Fabel, D., Stroeven, A. High-elevation cosmogenic nuclide dating of the last de- glaciation in the central Swedish mountains: implications for the timing of tree establishment. Manuscript.
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 dy- namic with multiple ice-free periods during the last 55 ka including around the LGM (Olsen et al. 2001a,b, 2002). Large moraine systems from the southeast portion of the ice sheet may be younger than previous estimates (Rinterknecht et al. 2006). There were possibly ice-free conditions around the LGM in southern Sweden (Alexanderson and Murray 2007) and southern Norway (Bøe et al. 2007). Based on dated stratigraphic sites and 2-D ice sheet modelling the MIS 4 ice sheet may not have persisted into the MIS 3 (Arnold et al. 2002), and more recent work suggests ice-free conditions during MIS 3 in Finland (Helmens et al. 2007a,b, Lunkka et al. 2008, Salonen et al. 2008) and Den- mark (Kjær et al. 2006) supported by numerous AMS radiocarbon dated mammoth throughout Scandinavia (Ukkonen et al. 2007, Wohlfarth 2010), although ice extents during MIS 3 are not agreed upon. Tree mega-fossils dated from high elevation areas in central Sweden suggest ice-free conditions as early as 17 cal. ka BP (Kullman 2002) although objected to (Birks et al. 2005). Modelling results of the Late Weichselian SIS
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 dat- ing 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 dy- namic 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 mar- gin 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 indi- cated, as well as several of the sites where bones of mammoth “M” have been found and dated to Mid- and early Late-Weichselian age. Curved large and small arrows are paleo-ice streams, solid line is the LGM ice margin, and the dashed line is the Younger Dryas ice margin (compiled from various sources by Olsen et al. 2001a). Ruunaa is just off the east edge of map at N 63°26' latitude.
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 gla- cially-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 sec- ond computer that controls the reader. (c) Analysis of repeated measurements of the luminescence (inset) for an aliquot using differ- ent radiation doses results in a growth curve. The dose that would be equivalent to the natural luminescence is determined by inter- polation on this curve to derive the equivalent dose (ED). Results from repeat measurements of an identical dose (recycling points) and a zero dose (recuperation), along with other criteria, are used to assess the quality of measurements from each aliquot. 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 discus- sion 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 pa- leoenvironment 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 radiocar- bon 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 explana- tion.
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 sedi- ments 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 san- durs without modification of the sandurs or deposi- tion 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, depend- ing 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 simi- larly-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 po- lymineral 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 utveck- lingen 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 chro- nology 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 gla- ciation 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 geo- logical 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
© The authors 2009 Journal compilation © 2009 Swedish Society for Anthropology and Geography 113 TIMOTHY F. JOHNSEN, HELENA ALEXANDERSON, DEREK FABEL AND STEWART P.H.T. FREEMAN
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
© The authors 2009 Journal compilation © 2009 Swedish Society for Anthropology and Geography 115 TIMOTHY F. JOHNSEN, HELENA ALEXANDERSON, DEREK FABEL AND STEWART P.H.T. FREEMAN
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
© The authors 2009 116 Journal compilation © 2009 Swedish Society for Anthropology and Geography NEW COSMOGENIC AGES FROM THE VIMMERBY MORAINE
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
© The authors 2009 Journal compilation © 2009 Swedish Society for Anthropology and Geography 117 TIMOTHY F. JOHNSEN, HELENA ALEXANDERSON, DEREK FABEL AND STEWART P.H.T. FREEMAN
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 af- fects 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 er- ratic 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
© The authors 2009 Journal compilation © 2009 Swedish Society for Anthropology and Geography 119 TIMOTHY F. JOHNSEN, HELENA ALEXANDERSON, DEREK FABEL AND STEWART P.H.T. FREEMAN
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.
© The authors 2009 120 Journal compilation © 2009 Swedish Society for Anthropology and Geography Paper II
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 lumines- cence (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 44 6ka(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
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 (deglacia- tion 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 palaeo- environmental 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, cor- related 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 mid- dle 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 thermo- luminescence 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) 38 4 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 (470 30 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 39 6 42.1 0.7 35.4 0.7 642 12 2.18 0.04 1.25 0.04 061345 34 5 42.8 0.6 34.7 0.5 668 9 2.14 0.04 1.15 0.04 061346 178 10 100.8 1.2 45 0.9 675 13 3.19 0.07 1.84 0.09 061347 33 5 49.1 0.7 37 0.5 715 10 2.38 0.04 1.37 0.04 061348 28 3 48.5 0.5 37 0.4 716 7 2.36 0.03 1.36 0.04 081318 40 8 40.2 0.8 34.3 0.8 663 15 2.22 0.05 1.24 0.04 081319 27 6 38.5 0.7 33.2 0.7 670 12 2.17 0.04 1.21 0.03 081320 29 5 31.9 0.5 30.7 0.5 680 12 2.15 0.04 1.14 0.03 081321 41 10 100.3 1.3 35.1 0.9 649 15 2.50 0.07 1.65 0.09 081322 42 7 74.4 0.9 30.7 0.7 614 11 2.26 0.05 1.39 0.06
Table 3. Water-content modelling to calculate average water 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 92 8% (fast), (b) Duration is 10% of the time 8 7% (medium) and 0.2 0.3% (slow). When these since deposition (t3 = 0.1). channels were selected for peak and background integra- tion, the resulting signal was dominated by the fast component. Doses calculated with this channel selection gave good dose recovery (1.05 0.04, 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 44 7ka (n = 5), from unit G 41 3ka mental succession from glacifluvial (unit A) to glacila- (n = 2) and single ages from unit F and C are 48 8ka custrine (unit B) to lacustrine (units C, D, E, F) and and 46 3 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 39.2 2 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 con- tent, 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 44 6 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 rea- sonable 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 calcu- lated 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 44 6ka(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 43 53736 443 5 116 12 18 2.72 0.10 1.14 0.24 061345 45 44437 446 4 118 10 19 2.62 0.10 1.06 0.26 061346 48 84932 539 7 134 23 18 2.82 0.09 1.06 0.31 061347 52 45042 453 4 145 10 16 2.77 0.10 1.08 0.25 061348 49 44940 350 4 139 10 22 2.83 0.10 1.03 0.20 081318 38 33630 338 3 101 7 28 2.69 0.10 1.17 0.09 081319 39 33429 340 397 7 31 2.51 0.09 1.07 0.12 081320 36 33527 236 389 7 32 2.52 0.09 1.05 0.17 081321 46 34437 347 3 143 6 30 3.10 0.13 1.10 0.12 081322 74 57258 475 5 202 9 30 2.74 0.11 1.07 0.12
1Age calculated with water content at time of sampling. 2Age calculated with saturated water content. 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 sa- turated 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.
55
50
45
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 (39.2 2 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