The Holocene 15,2 (2005) pp. 161
/176 Utilizing physical sediment variability in glacier-fed lakes for continuous glacier reconstructions during the Holocene, northern Folgefonna, western Norway Jostein Bakke,1,2* Øyvind Lie,1 Atle Nesje,1,3 Svein Olaf Dahl1,2 and Øyvind Paasche1,3 1 Bjerknes Centre for Climate Research, Allégt. 55, N-5007 Bergen, Norway; Department of Geography, University of Bergen, Breiviksveien 40, N-5045 Bergen, Norway; 3Department of Earth Science, University of Bergen, Allégt. 41, N-5007 Bergen, Norway) ( 2 Received 17 February 2004; revised manuscript accepted 12 September 2004 Abstract: The maritime plateau glacier of northern Folgefonna in western Norway has a short (subdecadal) response time to climatic shifts, and is therefore well suited for reconstructing high-resolution glacier fluctuations. The reconstruction presented here is based on physical parameters of glaciolacustrine sediments retrieved from two glacier-fed lakes and a peat bog north of the ice cap. Bulk density and modelled glacier net mass balance for the last 200 years show a remarkably similar pattern, where maximum sediment yield lags the glacier net mass balance by
1/ 0 years. The record of glacier variations has been transferred into an equilibrium-line altitude (ELA) variation curve. Glaciers respond primarily to changes in summer temperature and winter precipitation. At present there is a high correlation between the North Atlantic Oscillation (NAO) index and measured (since the early 1960s) net mass balance on maritime glaciers in western Norway (r
/ 0 / .8). Reconstructed glacier variations from maritime western Norway are therefore considered indicative of the strength of the westerly airflow associated with NAO during the Holocene. The early phase of mid-Holocene glacier growth (5200 cal. yr BP) was characterized by gradual glacier expansion culminating in the first Subatlantic glacial event at 2300 cal. yr BP. The climate during the last 2200 years has favoured increased glacier activity at Folgefonna. High-amplitude shifts in ELA may be explained by unstable modes of the westerlies causing significant variability of winter precipitation. During the last 2000 years, Folgefonna expanded and decayed with significant decadal variability. During the latest period of the ‘Mediaeval Warm Epoch’, Folgefonna advanced. The Neoglacial maximum, however, was reached during the ‘Little Ice Age’ at AD 1750 and AD 1870. The northern Folgefonna glacial record is compared to other Holocene glacier records in Scandinavia. Key words: Glacier fluctuations, North Atlantic Oscillation (NAO), lake sediments, bulk density, ELA reconstructions, Folgefonna, glacier mass balance, Holocene, Norway. Introduction Small plateau glaciers like Folgefonna in southern Norway (Figure 1) are ideal for studies of Holocene climate change because they respond rapidly to mass-balance perturbations (e.g., Dahl et al., 2003), allowing the position of the equilibrium-line altitude (ELA) to reflect climatic variability (e.g., Sutherland, 1984; Dyurgerov, 2002). Because studies on *Author for correspondence (e-mail: jostein.bakke@geog.uib.no) # 2005 Edward Arnold (Publishers) Ltd modern Norwegian glaciers have shown that sediment yield in glacier-fed lakes is positively correlated with glacier size (Roland and Haakensen, 1985) the proportion of glacigenic material may provide continuous records of glacier fluctuations. The use of lake sediments in this context is widely used in Scandinavia (e.g., Karlén, 1976; 1981; Nesje et al., 1991; 1995; 2000a; 2001; Matthews and Karlén, 1992; Dahl and Nesje, 1994; 1996; Snowball and Sandgren, 1996; Matthews et al., 2000; Rosqvist et al., 2004). Various approaches use a conceptual model of glacier-meltwater induced sedimentation in which the minerogenic (nonorganic) component of the 10.1191/0959683605hl797rp 162 The Holocene 15 (2005) Figure 1 (A) Map of part of southern Norway showing the location of the glaciers Folgefonna and Hardangerjøkulen (adapted from Østrem et al ., 1988). (B) Detailed map of the moraine systems in the catchment north of the plateau glacier of northern Folgefonna. Grey areas are terminal moraines mapped in the catchment. Black dots indicate coring sites and site for the sampling of the peat bog Hestadalsmyra. The lakes Dravladalsvatn and Vassdalsvatn, cored in this study, have inflow of meltwater from northern Folgefonna. sediments automatically is related to the presence of a glacier and its size in the catchment (e.g., Karlén, 1981; Leonard, 1985; Dahl and Nesje, 1994; Nesje et al., 2000a; 2001; Dahl et al., 2003). However, only a few studies have examined the physical properties of the sediments in detail and especially the minerogenic material produced by the glacier (Leonard, 1985; Souch, 1994; Rosqvist, 1995; Snowball and Sandgren, 1996; Matthews et al., 2000; Nesje et al., 2000a; 2001; Lie et al., 2005). The most common approach is to use the organic content (loss-on-ignition (LOI) and total organic carbon (TOC)) as an inverse indicator of inorganic deposition. In lakes with high minerogenic sedimentation and/or low organic production (B/5%) this approach has its limitations due to low ‘signal-to-noise’ ratio. The climate at the west coast of Norway is influenced by advection of both warm water and air masses entering the NE Atlantic region, as well as the position of the atmospheric polar front. The heat transport of the oceans on the west coast of Norway causes large temperature anomalies (Broecker, 1991; Hopkins, 1991). Large temperature gradients across the polar front generate cyclones crossing the North Atlantic region into and across Scandinavia. A close relationship between the winter weather and the North Atlantic Oscillation (NAO) index at the western part of Norway has been demonstrated (Hurrell, 1995; Hurrell et al., 2003; Nesje et al., 2000b). Atmospheric general circulation models have shown that the NAO is probably related to long-term trends in sea-surface temperatures (SST) (Feddersen, 2003; Hurrell et al., 2003). It is also demonstrated that higher winter precipitation in western Norway is related to stronger westerlies in the North Atlantic (associated with positive NAO weather modes) (Nordli et al., 2003). Mass balance, and hence size variations, of maritime glaciers in western Norway may thus be indicative for long-term trends of the westerlies. However, a high-pressure field east of or over Scandinavia gives a ‘blocking’ situation that forces the humid air masses either to the south or to the north of southwestern Norway (Shabbar et al., 2001). Here we present a detailed, high-resolution reconstruction of the Holocene glacier variations of the maritime northern Folgefonna in western Norway. The main objectives in this paper are to: (1) refine approaches for reconstruction of ELA variations using lake sediments with low organic/high minerogenic content; (2) evaluate strengths and weaknesses of sediment parameters used to obtain high-resolution ELA reconstructions; (3) reconstruct the Holocene glacial history of the plateau glacier Folgefonna at high temporal resolution; and (4) compare the reconstructed glacial record from northern Folgefonna with measured net mass balance and modelled net mass balance at Folgefonna. Study area The ice cap of northern Folgefonna (23 km2) is the seventh largest glacier in Norway. With its circular configuration it Jostein Bakke et al.: Utilizing physical sediment variability, Folgefonna, Norway ranges from 1644 to 1200 m and has a modern mean ELA of
1/ 465 m (accumulation-area-ratio (AAR) /0.7). The ice cap has five major outlet glaciers, Jordalsbreen, Jukladalsbreen (Figure 1), Botnabreen, Dettebrea and Juklavassbreen. About 12 km2 of the northern Folgefonna glacier drain northward and nine lakes are situated within the catchment. Jordalsbreen occupyies altitudes between 1200 and 1644 m with an average gradient of 100 m/km. The climatic response time of the glacier is 10
/12 years, based on the positive mass balance years from AD 1989
/92 which resulted in a glacial advance in AD 2001. In comparison, Jukladalsbreen has an altitudinal range between 1250 and 1644 m resulting in an average gradient of 150 m/km. The response time of this glacier is unknown, but is probably shorter than Jordalsbreen considering the steeper surface gradient. Based on the gradient of the glaciers, and the iceflow velocity (
4/ 0 m/yr 1) (authors’ unpublished data), it is assumed that the sediment storage time in the glacier is short (
1/
/3 yr 1). The plateau glacier of northern Folgefonna has no supraglacial and limited englacial transport of glacial sediments. The bedrock in the upper Jondal catchment consists mainly of acid meta-andesite, metadacite, quartzite, migmatite and migmatitic schist of Precambrian age (Sigmond, 1985; Askvik, 1995). The combination of acid rocks and the treeline situated at 600 m a.s.l. makes it a desolate landscape poor in both vegetation and superficial deposits. Except for some marginal moraines in front of the major outlet glaciers, there is only a sparse cover of colluvium and till in the area. The absence of superficial deposits reduces the influence of a paraglacial contribution to the glacial-fed lakes (Ballantyne and Benn, 1994; Ballantyne, 2002). Using two meteorological stations along Hardangerfjorden (Station no. 4949, Ullensvang Forsøksgård, 12 m a.s.l., 1962
/ 88; Station no. 5013, Omastrand, 1 m a.s.l., 1962
/90) (DNMI, 1993b), the present mean summer temperature (Ts) from 1 May to 30 September is suggested to be 12.78C at sea level in Jondal. With an environmental lapse rate of 0.68C/100 m (e.g., Sutherland, 1984) the mean Ts is close to 4.08C at the modern ELA (1465 m) of northern Folgefonna. A local meteorological station 11 km from the present glacier terminus (Station no. 5696, Kvåle, 342 m a.s.l., 1961
/90) (DNMI, 1993a) records the 1961
/90 mean winter (1 October to 30 April) precipitation (Pw) to have been 1434 mm. Based on an empirical, exponential increase in winter precipitation of 8%/100 m (Haakensen, 1989), the corresponding Pw at the ELA of northern Folgefonna is c. 3350 mm. A peat bog, Hestadalsmyra, with bedrock thresholds covers an area of 0.05 km2 and is situated between the river from northern Folgefonna in the east and a small river from Hestadalsbotnen in the south (Figure 1). Whenever there is glacial activity in the cirque Hestadalsbotnen, the small river draining through the mire transports and deposits glacially derived material. The mire was cored with a 110 mm PVC tube that was hammered into the mire and then excavated. The core was brought to the laboratory for radiocarbon dating and for magnetic susceptibility (MS) measurements. Dravladalsvatn (938 m) covers an area of 1.35 km2 (Figures 1 and 2) (bathymetry from Statkraft, unpublished). The lake is situated in a glacially eroded bedrock basin with the longest axis (2.5 km) orientated north/south. This particular glacialfed lake is bound to receive meltwater whenever the glacier on the northern Folgefonna plateau is present. The distal, eastern basin, which was cored for this study, only receives the finest fractions (all sediments passed through a 125 mm sieve) of these sediments, as the basin is sheltered from the main river current. Dravladalsvatn is the first basin to receive sediments from 163 the glacier Jordalsbreen and the third to receive sediments from the glacier Jukladalsbreen. Since AD 1974 the lake level has been artificially raised due to production of hydroelectric power. Vassdalsvatn (490 m) covers an area of 0.17 km2 (Figure 1) (bathymetry presented in Bakke et al., 2005), and is the seventh glacial-fed lake downstream from northern Folgefonna. The lake is located in a glacially eroded bedrock basin and receives input of glacier meltwater-induced sediments at the present. The site is suitable to record major flooding events in the catchment and to register whenever Folgefonna has been present. Research approach and methods The reconstruction of Holocene glacier fluctuations at northern Folgefonna is based upon the following methods. (1) Glacial-geomorphological mapping of the upper Jondal catchment, with special emphasis on former marginal moraines, glacier-meltwater channels and various ice-flow indicators. (2) Dating of the ‘Little Ice Age’ (AD 1650
/1930) glacial maximum by the use of lichenometry for absolute age estimate (data presented in Bakke et al., 2005). Lichenometry was also used to sort out the moraines older than ‘Little Ice Age’ (lichen diameter 2/ 00 mm). (3) Calculations of former equilibrium-line altitude (ELA) are obtained using an AAR of 0.7 (Porter, 1975) on the advanced glaciers reconstructed by marginal moraines. The calculation of the area distribution was carried out electronically using the vector-based GIS program MapInfo 6.0 on an N-50 map datum. (4) Ensuring a continuous record of former glacier size, two lacustrine sediment records retrieved from downstream distal glacial-fed lakes, were complemented with a peat bog stratigraphy. The two glacial-fed lakes were cored using a modified piston corer taking up to 6 m long cores with diameter 110 mm (Nesje, 1992). In Dravladalsvatn, an HTH gravity corer was used to retrieve sediments from the uppermost part of the lake sediments. The laboratory analyses of the cores from Dravladalsvatn included (sampling resolution on all parameters /0.5 cm) magnetic susceptibility (MS), weight loss-on-ignition (LOI) (Dean, 1974; Heiri et al., 2001), dry bulk density (g/ cm3) (DBD), water content and grain-size analysis using a Micromeretics Sedigraph 5100 (X-ray determination) (MasterTech, 1993). Grain-size statistics were performed by Gradistat 4.0 (Blott and Pye, 2001). Analyses of the two cores from Vassdalsvatn include LOI and MS. Twenty-five bulk samples and four plant macrofossil samples were AMS dated from Dravladalsvatn and Vassdalsvatn. Terrestrial plant macrofossils for AMS radiocarbon dating were very sparse or absent in the lakes. A hard-water reservoir effect is not considered to produce erroneous dates since both lakes are located within acid Precambrian granite gneiss (see Barnekow et al., 1998; Lowe and Walker, 2000). The radiocarbon dates are shown in Table 1, and calibrated (cal. yr BP) according to INTCAL 98 (Stuiver et al., 1998). The intercepts used are based on the mean intercept if there is more than one calculated in CALIB 4.4. 164 The Holocene 15 (2005) N Small streams without glacial meltwater 41 Small streams without glacial meltwater Core I Core HTH 56 Core II 28 Outlet Small stream without glacial meltwater the formation of a large lake to the east of the glacier that later was catastrophically drained when the glacier retreated and/or the water pressure became higher than the ice pressure. These floods can be classified as Jökulhlaups that are reported also from the southern part of the Folgefonna glacier (Tvede, 1989). During periods with extensive glaciers in Jukladalen, the drainage may have occurred randomly and to some extent also independent of climate, as ice thickness and water pressure controlled the water level in the lake. In the cirque Hestadalsbotnen (no glacier at present) there are four moraine ridges (Figure 1). H-P1 and H-P2 are assumed to be of Preboreal age (11 500
/9950 cal. yr BP) (lichen / /200 mm) and H-1 and H-2 were, according to the lichen measurements, deposited AD
/1870 and
/1750, respectively. The glacial activity in Hestadalsbotnen is, based on lichenometry, inferred to have been synchronous with the northern Folgefonna. Lithostratigraphy and radiocarbon dates 76 Hestadalsmyra Bathymetric map Lake Dravladalsvatn (938 m asl) 500 m Meltwater form northern Folgefonna Contours 10 m Figure 2 Bathymetric map of Dravladalsvatn showing the location of the retrieved piston cores and the short gravity core (HTH). Note the bedrock threshold separating the two main basins with a 28 m deep sill. Inflow and outflow of glacial meltwater from northern Folgefonna are both in the western main part of the lake. Results Moraine chronology Marginal moraines in front of the outlet glaciers from northern Folgefonna indicate up to eight successively smaller glacier halts or advances/readvances (Figure 1). The moraine chronology is not synchronous around northern Folgefonna. This may be due to differences in aspect and slope at Jordalsbreen and Jukladalsbreen. The terminal moraines are marked with site names and numbers in Figure 1, and all moraines with numbers 1
/3 were formed during the LIA (AD
/1750,
/1870 and
/1930 respectively). Calibrated against the ‘Little Ice Age’ moraines, Schmidt-hammer rebound values indicate that two marginal moraines may have formed c. 3000
/1000 yrs BP (Ju-N3 and Ju-N2), whereas the rebound values for the remaining sets of marginal moraines (with lichen sizes over 200 mm) suggest a depositional age during the Lateglacial or early Holocene (Bakke et al., 2005). Terminal moraines (Ju-P1, Ju-P2, Ju-N1, Ju-N2 and Ju-N3) north of Lake Jukladalsvatn demonstrate that Jukladalsbreen has crossed the valley several times during the Holocene. Historical sources indicate river-flooding events in Krossdalen (upper Jondal catchment) during the ‘Little Ice Age’ that caused damage to the surrounding farmland (Kolltveit, 1953). Glacier advances of Jukladalsbreen led to The lithostratigraphy in the peat bog has been divided into seven individual units (Figure 3). The lower unit (H) consists of gravel and sand with some plant macrofossils overlain by a short section of humus (unit G). The lowermost layer of fine sand and silt (unit F) is 4 cm thick and dated at 22659/45 14C yr BP (T-3602) (for details regarding the radiocarbon dates, see Table 1). The upper boundary is gradual, whereas the lower boundary is sharp with a possible erosive contact to unit G. The next unit, E, consists of homogeneous dark brown humus, whereas unit D is a 4 cm thick layer of fine sand and silt similar to unit F. A radiocarbon dating beneath the unit yielded an age of 16709/25 14C yr BP (T-13601). Unit C consists of humus similar to unit, E, whereas unit B is a third 5 cm thick layer of fine sand and silt, radiocarbon dated in the upper part to 12009/45 14C yr BP (T-13600). The upper unit (unit A) consists of humus with grass on the top. Dravladalsvatn The interpreted lithostratigraphies from the individual cores are shown in Figure 4. All three cores were taken in the deepest part of the inner basin of Dravladalsvatn (Figure 2). Core I was 97 cm long (Figure 4A) and the basal section consisted of a short sequence of grey silt and clay (unit G) below a gradually transition (unit F) into dark brown gyttja (unit E). The transition between unit G and F is dated at 86459/70 14C yr BP (TUa-3629A). Above unit E, there was another transitional layer (unit D), going from dark brown gyttja to grey clay and silt. The basal part of unit E yielded an age of 80909/40 14C yr BP (Poz-3177) and of 55309/40 14C yr BP (Poz-3176) in the upper part. Unit C consisted of grey silt and clay with some lighter grey bands. A radiocarbon date in the lower part yielded an age of 23159/45 14C yr BP (TUa-3628A). The upper part of the unit was radiocarbon dated at 20009/ 40 14C yr BP (TUa-3627A). Unit B contained browner sediments dominated by silt and clay. The youngest unit A was similar to unit C, with a radiocarbon date at the top of 20609/30 14C yr BP (Poz-3175). The LOI pattern in the core showed higher values in the section dominated by gyttja with a decrease into unit C (Figure 4A). Through unit B the LOI values were higher than below, indicating higher organic content during deposition of this unit. DBD and MS are more or less in antiphase compared with the LOI values, with some higher variability in the MS also showing anomalous values throughout unit B. Core II was 152 cm long (Figure 4B) and shows the same pattern as core I, except for the lowest part, which was missing Jostein Bakke et al.: Utilizing physical sediment variability, Folgefonna, Norway 165 Table 1 Radiocarbon dates obtained from the cores studied. When more than one calibrated intercept age is given, the mean intercept value is used. Site Lab. no. Depth (cm) Type of material Radiocarbon age9/1 sigma Intercept (cal. yr BP) 9/1 sigma (cal. yr BP) 9/2 sigma (cal. yr BP) Vassdalsvatn Core I Vassdalsvatn I Vassdalsvatn I Vassdalsvatn I Vassdalsvatn I Vassdalsvatn I Vassdalsvatn I Vassdalsvatn I Beta-102930 Beta-102931 Beta-102932 Beta-102933 Beta-102934 Beta-102935 Beta-102936 28
/31 117
/120 182
/185 250
/253 295
/298 368
/372 525
/535 Gyttja Gyttja Gyttja Gyttja Gyttja Gyttja Gyttja 11509/70 22809/60 33709/70 42709/80 52009/70 82609/80 43309/50 1060 2240 3645 4810 5960 9245 4910 1170
/970 2350
/2160 3690
/3480 4965
/4650 6170
/5890 9415
/9130 4965
/4840 1185
/930 2360
/2120 3830
/3460 5045
/4570 6170
/5750 9430
/9060 4990
/4830 Vassdalsvatn Core II Vassdalsvatn II Vassdalsvatn II Vassdalsvatn II Vassdalsvatn II Vassdalsvatn II Vassdalsvatn II Vassdalsvatn II Vassdalsvatn II T- 13607 T-13608 T- 13788A UtC-6691 UtC-6692 UtC-6693 UtC-6694 UtC-6695 19 83
/84 77
/79 123 138 142 147 171 Gyttja Gyttja Gyttja Gyttja Gyttja Gyttja Gyttja Gyttja 21009/85 19009/70 23109/60 27659/45 33199/40 34609/60 38209/50 62809/60 2050 1840 2080 2860 3550 3720 4250 7205 2295
/1970 1920
/1735 2360
/2160 2920
/2785 3630
/3475 3830
/3640 4345
/4100 7270
/7030 2210
/1890 1990
/1690 2470
/2150 2950
/2775 3640
/3465 3870
/3570 4410
/4090 7405
/7005 Hestadalsmyra Hestadalsmyra Hestadalsmyra Hestadalsmyra T-13600 T-13601 T-13602 26
/28 45.5
/47.5 75
/77 Humus Humus Humus 12009/45 16709/65 22659/60 1120 1560 2250 1255
/995 1690
/1515 2345
/2160 1280
/965 1715
/1410 2355
/2145 Dravladalsvatnet Core I Dravladalsvatn I Dravladalsvatn I Dravladalsvatn I Dravladalsvatn I Dravladalsvatn I Dravladalsvatn I Poz-3175 TUa-3627A TUa-3628A Poz-3176 Poz-3177 TUa-3629A 1 24 57 72 82 88 Macro fossil Gyttja Macro fossil Macro fossil Gyttja Gyttja 20609/30 20009/40 23159/45 55309/40 80909/40 86459/70 2030 1955 2310 6340 9055 9660 2060
/1990 1990
/1920 2355
/2305 6395
/6290 9220
/9000 9690
/9540 2115
/1950 2045
/1865 2465
/2155 6405
/6280 9130
/8980 9800
/9520 Dravladalsvatnet Core II Dravladalsvatn II Dravladalsvatn II Dravladalsvatn II Dravladalsvatn II Dravladalsvatn II Dravladalsvatn II Dravladalsvatn II Dravladalsvatn II Poz-3178 TUa-3640A Poz-3198 TUa-3630 TUa-3631A Poz-3179 Poz-3256 TUa-3632A 1 24 45 78 100 124 132 151 Gyttja Gyttja Gyttja Macro fossil Gyttja Gyttja Gyttja Gyttja 25659/30 23209/45 23159/25 19109/45 32159/60 46759/35 50509/30 63759/70 2620 2310 2330 1840 3450 5390 5810 7305 2750
/2550 2360
/2180 2350
/2330 1910
/1745 3475
/3360 5465
/5320 5890
/5805 7415
/7250 2755
/2490 2465
/2155 2355
/2310 1950
/1725 3575
/3330 5470
/5310 5900
/5725 7430
/7180 in core II (units F and G in core I). The lower unit C consisted of dark brown gyttja with a radiocarbon date at the bottom yielding 63759/70 14C yr BP (TUa-3632A) and in the upper part 50509/70 14C yr BP (Poz-3256). Unit B is a transitional unit with a change from dark brown gyttja to grey silty gyttja. A radiocarbon date in the lower part yielded an age of 46759/ 35 14C yr BP (Poz-3179). Unit A consisted of grey clay and silt with an age in the lower part of 32159/60 14C yr BP (TUa3631A). At 78 cm a radiocarbon date yielded an age of 19109/ 45 14C yr BP (TUa-3630). In the upper part of unit A, three inverted radiocarbon dates were obtained, yielding 23159/ 25 14C yr BP (Poz-3798), 23209/45 14C yr BP (TUa-3640A) and 25659/30 14C yr BP (Poz-3178), respectively, toward the top of the core. A short 39 cm gravity core (HTH) from Lake Dravladalsvatn consisted of silt and clay with LOI values below 6% throughout the entire core. Grain-size analyses, DBD and MS showed high-frequent fluctuations. No radiocarbon dates have been obtained from this core (Figure 7). Vassdalsvatn In Vassdalsvatn, the two cores showed remarkably different lithostratigraphy. Core I (550 cm) in Vassdalsvatn was retrieved in the central part of the lake, close to the main watercourse through the lake (Figures 1 and 5A). This core contained seven main units or lithological facies. Unit G consisted of a nearly 2 m long section of grey clay and silt, with a sharp transition to the overlying unit. A radiocarbon date at the upper part of the unit yielded an age of 82609/80 14C yr BP (Beta-102936). Unit F consisted of dark brown gyttja that terminated in a layer of fine sand. The next (unit E) contained brown gyttja with a minor sand layer in the upper part. Two radiocarbon dates yielded ages of 52009/70 14C yr BP (Beta-102935) and 42709/ 80 14C yr BP (Beta-102933) in the lower and upper part of the unit, respectively. A layer of fine sand and macrofossils dominated unit D. A radiocarbon date of the layer yielded an age of 33709/70 14C yr BP (Beta-102932). The unit above (unit C) consisted of brown gyttja intercalted with two layers of fine sand. A radiocarbon date in the middle part of the unit yielded an age of 22809/60 14C yr BP (Beta-102931). A layer of fine sand and plant macrofossils (similar to unit D) dominated unit B. The transition to unit A, which contained brown gyttja, was radiocarbon dated at 11509/70 14C yr BP (Beta-102930). The LOI record shows an abrupt change from the lower minerogenic section (unit G) into unit F. The MS showed low values in unit F with a gradual rise in unit E and a marked peak in units D and B. 166 The Holocene 15 (2005) consisted of grey clay and silt. Unit B was a transitional unit from silty clay to gyttja that differed from the same transition in core I (unit G in core I). This unit contained three layers of grey silt and a radiocarbon dating in the middle of unit B yielded an age of 62809/60 14C yr BP (UtC-6696). Unit A consisted of brown gyttja with two layers of macrofossils and fine sand. A layer of grey clay and silt was prominent in this bed and was radiocarbon dated at 38209/50 14C yr BP (UtC6694) below and 34609/60 14C yr BP (UtC-6693) above. From this unit several radiocarbon dates have been obtained, 33209/ 40 14C yr BP (UtC-6692), 27659/45 14C yr BP (UtC-6691), 19009/70 14C yr BP (T-13608) and 23109/60 14C yr BP (T-13788A), respectively (Figure 5B). The two layers with plant macrofossils and fine sand (unit D and B) were radiocarbon dated with one sample below each unit. Unit D yielded an age of 19009/70 14C yr BP (T-13608), whereas unit B yielded an age of 21009/85 14C yr BP (T-13607). Age-depth relationship Figure 3 Lithostratigraphy and magnetic susceptibility (MS) in a core retrieved from the peat bog from Hestadalsmyra. Units B, D and F represent periods with glacial input to the peat bog (P / peat; FSAS/fine sand and silt; GAS /gravel and sand). The MS signal shows peculiar patterns, gradually decreasing up to unit B, where it rises abruptly and remains high in the upper part. Core II (405 cm long) was retrieved in the eastern, more distal part of the lake (Figures 1 and 5B). This core was divided into three main units. The oldest (unit C) was 220 cm long and The age-depth models for Dravladalsvatn and Vassdalsvatn are shown in Figure 6. Both models are constructed by linear interpolation between radiocarbon dates or between ‘fixed’ points in the cores. A major problem establishing the age-depth relationship is that several of the radiocarbon dates during the last 2000 cal. yr BP are inverted (Figure 6). The inverted radiocarbon dates may be explained by erosion and resedimentation of terrestrial plant material during river floods. The draining of the glacier-dammed lake in Jukladalen may have led to raised lake level in Dravladalsvatn and thereby erosion along the shores. Another possible explanation for the inverted AMS bulk dates in the upper part could be the low organic content in the samples. Despite the problematic radiocarbon dates, the age-depth model in Dravladalsvatn is constrained by radiocarbon dates in Vassdalsvatn and Hestadalsmyra by correlation to flood events inferred from analyses of sorting and mean used as time markers (Arnaud et al., 2002). Periods with poorer sorting were interpreted as events with abrupt change in the input of minerogenic sediments in Dravladalsvatn (Figure 8), interpreted as river-flooding events caused by glacier damming of the valley Jukladalen. Using this approach, four major flooding events were detected and correlated Figure 4 (A) Lithostratigraphy, radiocarbon dates (14C yr BP), loss-on-ignition (% loss at 5508C), dry bulk density (DBD) and magnetic susceptibility (MS) measured at 0.5 cm intervals for Dravladalsvatn core I (Scg/grey silt/clay; Dbg /dark brown gyttja; Tst/transition from silt towards brown gyttja; Scs/silty clay with stones; Tgs /transition from gyttja to silt/clay; Gsg /grey silty gyttja). (B) Lithostratigraphy, radiocarbon dates (14C yr BP), loss-on-ignition (% loss at 5508C), DBD and MS measured at 0.5 cm intervals for Dravladalsvatn core II (Scg/grey silt and clay; Tgs /transition from brown gyttja to silt/clay; Dgb /dark brown gyttja). Jostein Bakke et al.: Utilizing physical sediment variability, Folgefonna, Norway 167 Figure 5 (A) Lithostratigraphy, radiocarbon dates (14C yr BP), loss-on-ignition (% loss at 5508C) and MS measured at 2 cm intervals for Vassdalsvatn core I (Bg /brown gyttja; Mah /macrofossil and humus; Fs /fine sand and silt; Dbg /dark brown gyttja). (B) Lithostratigraphy, radiocarbon dates (14C yr BP), loss-on-ignition (% loss at 5508C) and MS measured at 2 cm intervals for Vassdalsvatn core II (Bg /brown gyttja; Mh/macrofossil and humus; Gsc /grey clay and silt; Sbg /silty band with transition to gyttja). between Dravladalsvatn and Vassdalsvatn and used as time markers (‘fixed points’) in the age-depth models. In Vassdalsvatn, the flooding events were represented by layers of sand and inwash of plant macrofossils. The sites belong to the same drainage system, and these large floods should therefore be detectable in both lakes. In the peat bog Hestadalsmyra silt layers were interpreted to reflect periods with enhanced glacial activity in the cirque Hestadalsbotnen. It is assumed that glacial activity in Hestadalsbotnen was synchronous with periods of glacier growth at northern Folgefonna. The major flooding events occurred at
/3500 cal. yr BP (age from core I in Vassdalsvatn and core II in Dravladalsvatn),
/2250 cal. yr BP (age from core II in Dravladalsvatn, core II from Vassdalsvatn and core II in Dravladalsvatn), 1750 cal. yr BP (ages from mire Hestadalsmyra and core II in Dravladalsvatn) and
/1050 cal. yr BP (ages from peat bog Hestadalsmyra and core I from Vassdalsvatn). Beside these major floods there were several minor events observed in the sorting
/mean record from Dravladalsvatn (Figure 8). The short gravity core (HTH core) in Dravladalsvatn is suggested to overlap core II by three cm based on mean sorting, MS, DBD and LOI. The advantage of the gravity core is that the core contains an undisturbed section of the top sediments in Dravladalsvatn. The uppermost mean
/sorting anomaly is correlated to a historical documented flood in the catchment (7 cm in core/AD 1890) (Kolltveit, 1953). This gives the gravity core the same sedimentation rate as the upper part of core II (
1/ 0 yr/cm 1), the remaining 32 cm of the core is assumed to have the same sedimentation rate (Figure 7). The coring sites for these two cores are at 56 m water depths, only a few metres apart. The age-depth model for the Vassdalsvatn record is complicated by two inverted radiocarbon dates in the upper part of core II probably due to inwash of terrestrial plant macrofossil during river-flooding events. The basal date in core I is apparently too young, probably due to contamination inflow when the core was jacked out of the sediments. Discussion The minerogenic input to the eastern basin of Dravladalsvatn is suggested to represent a ‘pure’ glacial signal as there is only a sparse cover of colluvium and basal till around the lake. The small bouldery colluvial fans surrounding Dravladalsvatn are inferred to deliver different grain sizes than those produced by the glacier due to the short transport length and their possible influence is considered to have been minimal. Subaquatic erosion of previously deposited glacigenic sediments is not likely due to the low energy in this eastern part of the lake. The bedrock threshold and the position of inlet and outlet in the western main basin make the coring sites favourable for recording periods when there were minerogenic sediments in suspension. Figure 6 (A) Age-depth curves for the three cores retrieved from Dravladalsvatn. The solid line shows the linear interpolation model based on the radiocarbon dates in each core. The dotted line shows the age-depth relationships correlated against the lithostratigraphy in Vassdalsvatn and the peat in Hestadalsmyra. Error bars show9/2 sigma. (B) Age-depth models for the two cores retrieved from Vassdalsvatn. The solid line shows the linear interpolation model based on the dates in each core. Error bars show9/2 sigma. 168 The Holocene 15 (2005) Figure 7 Compiled lithostratigraphy from Dravladalsvatn based on the three cores retrieved from the eastern distal basin. Correlation between the cores is done by using the age-depth models and the DBD/MS records. Figure 8 Mean grain size plotted against sorting (standard deviation in a sample). The upper axis shows depth (cm) in the compiled stratigraphy and the lower axis shows calendar years before present. As seen from the depth scale, there was a notable change in sedimentation rate around 130 cm. Higher ‘sorting’ values mean poorer sorting of the sediments. Grey shaded areas show sorting anomalies, and dark grey shaded areas show anomalies used for constructing the age-depth models. Jostein Bakke et al.: Utilizing physical sediment variability, Folgefonna, Norway The composite chronostratigraphy from Dravladalsvatn is based on the age-depth model in Figure 7. Due to the large water depth (
7/ 0 m) (artificial rise of lake level due to hydroelectric power plant construction), it was difficult to retrieve the uppermost soft sediments. However, reliable correlations between the core stratigraphy were obtained by using MS and DBD records (Figure 7). Hence, by combining the three cores from Dravladalsvatn, a 202 cm long composite stratigraphy, starting 9660 cal. yr BP (excluding the undated deglaciation section), was established. The difference in sedimentation rates between the two piston cores is caused by difference in water depth and distance to the bedrock threshold. Core I was taken at 35 m water depth in the distal part of the basin whereas core II was taken at 55 m water depth (natural water level) closer to the bedrock threshold. Despite the inverted radiocarbon dates in Dravladalsvatn, the agedepth model of the composite stratigraphy is constrained by radiocarbon dates from Vassdalsvatn and Hestadalsmyra together with a historically documented flood event. Loss-on-ignition, grain-size distribution and magnetic susceptibility The grain-size distribution in Dravladalsvatn is shown in Figure 9. Generally, there was a high input of coarse particles during timespans when the sediments were dominated by gyttja. The very coarse silt fraction is interpreted to be sediments from the surrounding catchment and in this context regarded as ‘noise’. As the LOI values decrease, suggested glacigenic sediments consisting of clay and fine silt dominates the grain-size distribution. Negative correlations are evident between bulk density versus coarser fractions (very coarse silt and coarse silt) and positive correlations with the finer fractions. The positive correlation probably reflects an in- 169 creased energy in the lake. As the glacier grew the water discharge increased, transporting more sediment over the bedrock threshold and into the coring site. The negative correlations indicate that the coarser grain sizes in Dravladalsvatn are transported into the lake by slopewash from the lake surroundings. Based on these results, the lithostratigraphy was divided into four phases (Figures 7 and 9). MS shows low values when gyttja dominated the sediments (phases III and II), whereas the MS values increased rapidly as the proportion of minerogenic sediments increased in phase IV. During phase I the MS signal rose as the DBD values increased, reflecting varying influx of minerogenic sediments into the lake (Figure 7). The glacial signal in Vassdalsvatn is suggested to be weaker because of longer transport length of the sediments compared to Dravladalsvatn. Grain-size and DBD analyses are not performed for the cores from Vassdalsvatn. The differences in lithostratigraphy between the two cores may be explained by the coring sites lying with different distance from the inlet and outlet of the main river. The lithostratigraphy in Vassdalsvatn was used to complement Dravladalsvatn regarding major flooding events, represented by sandy organic-rich layers. Bulk density as a proxy for glacier size Loss-on-ignition has traditionally been used as an inverse indicator for inorganic lake sedimentation. The approach is widely used to reconstruct glacier variations (e.g., Karlén, 1976; 1981; Leonard, 1985; Nesje et al., 1991; Dahl and Nesje, 1994; Rosqvist, 1995; Matthews et al., 2000; Nesje et al., 2000a; 2001). However, when the organic content is low (B/5%), it is difficult to solve the amplitude of the glacial signal, as the signal-to-noise ratio becomes very low. This approach has therefore natural limitations in high alpine and Figure 9 Bulk density, clay, very fine silt, fine silt, medium silt, coarse silt, very coarse silt and very fine sand. The compiled lithostratigraphy is divided into four phases based on the presented parameters where they indicate notable changes in sedimentation environment. 170 The Holocene 15 (2005) Arctic lakes. In Dravladalsvatn, the LOI values were below 5% during periods when northern Folgefonna was at its present size (Figure 10). It was therefore difficult to obtain a continuous ELA reconstruction based on the LOI as an inverse indicator of the minerogenic sedimentation. Several physical sediment parameters describing the sediments produced by the glacier (e.g., bulk density and grain-size distribution) have therefore been taken into account. As seen in Figure 10, the overall patterns were reflected in both LOI and bulk density, but the bulk-density record has a larger amplitude than the LOI record in the minerogenic (low LOI) end of the spectrum, demonstrated by the exponential fit (Figure 10). As the nature of glacial erosion is reflected by the supply of insoluble particles to a river system, analyses of physical properties of the glacial sediments may be a diagnostic parameter for variations in glacier size. Warm-based glaciers produce abundant clay-silt size fractions that are transported downstream to produce characteristic signatures in glaciolacustrine sediments (e.g., Østrem, 1975). The use of grain-size variations have, however, not been widely used in this context. An important factor concerning the grain-size distribution in glacial-fed lakes is that glaciers normally produce more than one dominating grain-size fraction. As seen from till studies, glaciers produce a composition of more or less all grain sizes (e.g., Vorren, 1977). The glacial transport length and the size of the glacier do not seem to strongly influence the grain-size distribution of glacigenic sediments (e.g., Jørgensen, 1977; Haldorsen, 1981; 1983). The grain-size variations in glacial-fed lakes are therefore mainly reflecting changes in fluvial and lacustrine systems. High-energy streams deposit coarser sediments, and vice versa (e.g., Hjulström’s diagram; Sundborg, 1956). In ‘open-ended’ lakes, the finest grain sizes will be transported further downstream because of stronger currents and slow settling. In a small, almost closed sediment basin, the grain-size distribution commonly consists of grain sizes suitable for suspension (1
/63 mm), commonly giving more sediments per time than an ‘open-ended’ lake basin. By definition, bulk density expresses the ratio of the mass of dry solids to the bulk volume of the sediment (Blake and Hartge, 1986). Commonly, this parameter defines how granular, fibrous and powdery materials pack or consolidate under a variety of conditions and can be used to calculate the porosity of the sediment. Changes in flux and packing (reflected in grain-size composition) are probably the most important parameter in a glacial-fed lake (Webb and Orr, 1997). Organic sediments should potentially be reflected by the lowest bulk values, whereas the highest values are expected in sediments consisting of fine-grained poorly sorted minerogenic sediments (Figure 11). Water content is a parameter strongly linked to the bulk-density parameter, as water fills the pores and expresses the porosity of the sediment (Menounos, 1997). In cores I and II from Dravladalsvatn this relationship is very strong (r2 /0.97). ELA variations at northern Folgefonna A relationship between grain-size variations (sorting
/mean anomalies), DBD and glacier size based on the analysis has been established (Figure 12). The altitudinal position of the moraines Ju-N1, Ju-N2, Jo-1, Jo-2 and Jo-3 have been used to calibrate the ELA curve by a correlation between ELA and DBD (Figure 12B). The moraines Jo-1, Jo-2 and Jo-3 are independently ‘dated’ by the use of lichenometry and historical sources, whereas Ju-N1 and Ju-N2 are relatively dated by Schmidt hammer (Bakke et al., 2005). Periods with sorting anomalies (due to flooding events) have been removed from the ELA reconstruction (open squares in Figure 12). Figure 10 (A) Residue after 5508C ignition in% giving the minerogenic proportion after the organic content is removed. This parameter has traditionally been used as an indicator of inorganic sedimentation. Dotted line shows DBD. (B) Regression between residue (%) and DBD, showing a close relationship (r2 /0.8) between the two parameters. Jostein Bakke et al.: Utilizing physical sediment variability, Folgefonna, Norway 171 Bulk density and net-balance modelling DBD as a proxy for former glacier size is a new approach, and the validity is tested against net mass-balance data for the last 200 years (model; n/200, DBD; n/40) at Folgefonna (Figure 13). The uppermost gravity core is undated; however, age control is established using linear interpolation between the sediment surface (present) and a historically dated flood (AD 1890) (seen as a sorting
/mean anomaly) as a time marker. Tvede (1979) established some equations for modelling of the glacier mass balance (Bw/Bs and Bn) of Folgefonna based on temperature and precipitation from the Bergen-Florida meteorological station (station. no. 5054/56). The equation was later reformulated (Elvehøy, 1998) and established also for the northern part of Folgefonna: Bn 4442:16  P54  T3 Figure 11 Schematic figure explaining the relationship between bulk density and water content related to type of sediment. Angular minerogenic particles give higher porosity than rounded glacially derived minerogenic particles. Lowest bulk-density values are obtained from sediments dominated by gyttja and angular minerogenic particles. The moraines in front of Jukladalsbreen indicate that the glacier had two advances (
2/ 250 cal. yr BP and 1750 cal. yr BP) where the glacier front was lying beyond the ‘LIA’ maximum. In addition to the relative age estimate based on Schmidt hammer, the flood events detected in Vassdalsvatn and Dravladalsvatn are used for absolute age estimate of the moraines Ju-N1 and Ju-N2. Damming of Jukladalen is not possible without the glacier expanding across the valley. If the prevailing wind direction changes toward S
/SE, it is assumed that Jukladalsbreen expands more than Jordalsbreen due to its aspect and surface geometry. However, the DBD record from Dravladalsvatn is regarded as most representative for Jordalsbreen, as part of the glacially derived sediments produced by Jukladalsbreen probably are deposited in Jukladalstjørn and Jukladalsvatn before entering Dravladalsvatn (Figure 1). The ELA reconstruction has been divided into six phases. (1) Between 9600 and 5200 cal. yr BP the ELA at northern Folgefonna was above the highest mountain (1/ 550 m) and there was no glacier present in the catchment. (2) ELA dropped around 5200 cal. yr BP and the Folgefonna glacier was reformed after the ‘thermal optimum’. (3) From
4/ 600 to
2/ 300 cal. yr BP there was a gradual buildup of northern Folgefonna towards its present size. (4) Around 2200 cal. yr BP there was a double, short-lived glacier advance followed by a rapid rise in ELA (
1/ 500 m) around 2000 cal. yr BP. (5) From 2000 until
1/ 400 cal. yr BP the ELA was lowered from 1440 m to 1360 m as the glacier area gradual became bigger. (6) From 1200 cal. yr BP until present the ELA variations led to high-frequent changes in glacier size before a rather long period (from
6/ 00 cal. yr BP until AD 1930) with large glaciers during the ‘Little Ice Age’. (1) where P is winter precipitation in Bergen (01.10
/31.05) and T3 is average summer temperature (01.06
/31.08). The equation gives high predictability compared to the net mass balance from 1963 to 1997 (r2 /0.84). In the reconstruction, temperature and precipitation records from Bergen back to AD 1841 (data from Meteorological Institute) were put into the equation, whereas a temperature record from Ullensvang was used from AD 1800 to 1840 (Birkeland, 1932). As there is a lack of precipitation records for this timespan, a linear regression model between the January, February and March temperatures (r/0.6) to reconstruct the winter precipitation was used. Both models reproduce the AD 1870 (late LIA) glacier advance and the AD 1930 glacier advance, which correspond to moraines Jo2 and Jo-3 at Jordalsbreen, respectively. The low Bn values from AD 1800 to 1840 reflect the retreat of the glacier after the AD 1750 glacier event, and may explain the high DBD values during the timespan. If the age model is correct, another interesting feature is that there apparently exist lags in the bulk-density record with
/10 years from a change in net balance to increased bulk-density values. This is suggested to reflect the response time of the glacier to mass-balance perturbations. Implications of climatic importance from the northern Folgefonna record The approach shown in this paper provides new methodological tools for reconstruction of glacier variations from lake sediments with low organic content. The approach is appropriate in high alpine and Arctic regions, where high-resolution reconstructions of former glacier variations from lake sediments are sparse. In Figure 14, the reconstructed glacier variations at northern Folgefonna are compared with other studies from southern Norway. Based on the adopted ELA curve in this study, several implications follow from the temporal pattern of Holocene glacier variations at northern Folgefonna. (1) In Bakke et al. (2005), the early-Holocene event chronology is discussed on the basis of lake sediments from Vetlavatn (Figure 1), indicating three episodes of glacier advance subsequent to the Younger Dryas. The second Erdalen Event glacier readvance did not cross the threshold to Vetlavatn, and it is therefore not recognized as a glacial readvance at northern Folgefonna. However, the data from Dravladalsvatn indicate a rapid retreat of the glacier, as the initial gyttja-dominated sediments are dated to 9660 cal. yr BP. This is in accordance with the data from Matthews and Karlén (1992) in the Jostedalsbreen
/Jotunheimen area. At Hardangerjøkulen and Jostedalsbreen, the glaciers existed continuously until the end of the Finse Event (
/8.2 ka cal. yr BP). 172 The Holocene 15 (2005) Figure 12 (A) TP-ELA curve for northern Folgefonna based on the bulk-density record. The lower part of the figure is a summary of sources for validating the ELA record; the peat bog in Hestadalsmyra, floods in Vassdalsvatn, sorting anomalies in Dravladalsvatn, glacial input in Vassdalsvatn and record of glacial input to Dravladalsvatn. (B) Regression between periods with known ELA (calibrated against the moraine chronology) and bulk-density values. The regression model is used to transfer the bulk-density record to a continuous TP-ELA curve. (2) The record from northern Folgefonna does not support glacier readvances during the (double) Finse Event, which were first recorded in a peat bog at Finse (Dahl and Nesje, 1994; 1996). The event has later been reproduced from other peat sections, glacial-fed and nonglacial lakes (Dahl and Nesje, 1996; Matthews et al., 2000; Nesje et al., 2000a; 2001; Nesje and Dahl, 2001). A possible explanation for this discrepancy may be the altitudinal range of the glacier. In Matthews and Karlén (1992), the importance of glacier altitude regarding temperature changes was examined, and they concluded that the highest-lying glaciers existed longer into the ‘thermal optimum’ than lower-lying glaciers. As the glaciation threshold due to the topography around northern Folgefonna is 1550 m a.s.l., the lowering of the ELA during the Finse Event may not have crossed the glaciation threshold or the altitude of instantaneous glacierization (AIG) (Lie et al., 2003). During the maximum of the Finse Event at Hardangerjøkulen, the TPELA was 1580 m (Dahl and Nesje, 1996), which is above the highest part of the subglacial mountain plateau beneath northern Folgefonna. The plateau glacier Hardangerjøkulen lies 80 km to the northeast of northern Folgefonna, and the TP-ELA is therefore regarded as comparable. (3) The first Neoglaciation at northern Folgefonna started around 5200 cal. yr BP. This is different from the other reconstructions from southern Norway where there were several shorter and longer glacial periods between 9660 and 5200 cal. yr BP. The reason for this is most likely the altitudinal range of northern Folgefonna. This is also indicated for the glacier Ålfotbreen that has a modern mean TP-ELA of
/1200 m. Here, the Neoglaciation started
/23309/60 cal. yr BP by some smaller glacier events, before the glacier recovered and existed continuously from around 850 cal. yr BP (Nesje et al., 1995). (4) The first Neoglaciation at northern Folgefonna has some notably consistent modes. The glacial advance from 5200 cal. yr BP until
/2300 cal. yr BP was a phase with gradual glacier growth. Such a gradual transition is known from other palaeoclimatic archives in the North Atlantic region, especially prominent in the reconstructed sea-surface temperatures (SST) at the Vøring Plateau (Calvo et al., 2002). The SST record shows marked drops in temperature at 5400 and 2500 cal. yr BP, which correspond to marked changes in glacier size at northern Folgefonna. It is therefore assumed that the boundary conditions for glacier growth in southwestern Norway indicate a change in the atmospheric and oceanic conditions, rather than abrupt climate changes as recorded during the early Holocene (Dahl et al., 2002; Nesje et al., 2004). The reduced SST in the North Atlantic and the expansion of the Folgefonna glacier may indicate that the ocean has a major forcing on the precipitation distribution in the North Atlantic Realm. (5) The Holocene glacier record from northern Folgefonna indicates high-frequently changes in glacier size during the last 2300 years, with century- to millennial-scale glacier expansions and some less extensive decadal glacier fluctuations. Of special interest are three relatively large glacial readvances dated at 2200, 1600 and 1050 cal. yr BP. Periods with glacier expansion are also recognized at Jostedalsbreen (Nesje et al., 2001), Hardangerjøkulen (Dahl and Nesje, 1994) and at Bøvertunsbreen (Matthews et al., 2000) in southern Norway during the Jostein Bakke et al.: Utilizing physical sediment variability, Folgefonna, Norway 173 Figure 13 Dry bulk density (DBD) compared to two different glacier net mass-balance models for the Folgefonna and the northern Folgefonna glaciers. The dotted line shows a model developed by Tvede (1979) for the southern part of the Folgefonna glacier, whereas the black line was produced for northern Folgefonna in this study. Both models are compared with a glacier net mass-balance model and net balance measurements from AD 1963 to 1997 (Elvehøy, 1998) at Folgefonna. The reconstruction is based on temperature and precipitation records from Bergen and Ullensvang (Birkeland, 1932; DNMI, 1993a; 1993b). Thick grey shaded areas indicate periods with moraine formation. same timespan. From northern Norway, two late-Holocene glacier readvances are recognized at Okstindan dated at 3000
/ 2500 14C yr BP and 1250
/1000 14C yr BP (Griffey and Worsley, 1978). At Okstindan, these glacial readvances were larger than during the ‘Little Ice Age’. This is similar to the record of Jukladalsbreen at northern Folgefonna. Based on the wide geographical distribution of the late-Holocene glacier advances, it is assumed that these high-frequency climatic shifts, leading to glacier expansion and decay, are representative for at least western Scandinavia. Folgefonna is a maritime glacier where
/80% of its modern Bn is forced by changes in Bw. Possible explanations for the changes around 2200 cal. yr BP may therefore be in the record of winter precipitation (associated with positive NAO weather mode) that appears in a less stable mode than during the period from 5200 to 2200 cal. yr BP. Another explanation may be a stronger effect of the Russian High, giving variable patterns of the westerlies and hence the precipitation pattern along the west coast of Norway (Hurrell, 1995; Shabbar et al., 2001; Hurrell et al., 2003). (6) The period termed ‘the Mediaeval Warm Epoch’ (MWE) is without any significant signature in the glacial record form northern Folgefonna. The MWE is referred to as the time interval between AD 800 and 1300 (Cronin et al., 2003). It is no evidence for the temperature record exceeding the present temperature range (Crowley and Lowery, 2000; Bradley et al., 2003). If a MWE temperature rise was followed by an increase in winter precipitation, the glacier at northern Folgefonna would have expanded. This is suggested from the ELA reconstruction in this study, as the time interval for the MWE includes both periods of glacier expansion and decay. A possible explanation for the glacier decay could be higher winter temperatures, which could give rain instead of snow at the glacier. (7) The ‘Little Ice Age’ at northern Folgefonna had three periods of glacier growth peaking at AD
/1750,
/1870 and
/ 1930 with successively smaller glacier advances, but marked by distinct marginal moraines. This is in accordance with earlier studies from Folgefonna and also with historical sources at the southern parts of Folgefonna. It seems like the southern part of Folgefonna had increased net mass balance during the latest glacial expansion phase that culminated in AD 1940 (Tvede, 1972). (8) The record of late-Holocene glacial fluctuations may contribute to increased understanding of the coupling between oceanographic and atmospheric processes that led to the observed late-Holocene decadal and millennial climatic variability. Thus, it is apparent that high-resolution glacier reconstructions, especially from the last two millennia, should be adapted to a wider geographical area, involving glaciers in the range from continental to maritime climate regimes. Summary and conclusions (1) By using grain-size analysis and bulk density as proxies for former glacier-size variations, it is shown that there is a potential for high-resolution glacier reconstructions in lakes where the LOI has its limitations (B/
/5%) due to low signalto-noise ratio. 174 The Holocene 15 (2005) Figure 14 Compilation of some selected ELA reconstructions from southern Norway throughout the Holocene period (Matthews and Karlén, 1992; Dahl and Nesje, 1996; Matthews et al ., 2000; Nesje et al ., 2001; Bakke et al ., 2005). Upper panel with arrows shows glacial events (abrupt events of decadal to millennial duration) derived from the compiled glacial records. During the early Holocene, most of the events are named (J1 /Jondal Event 1; J2 /Jondal Event 2; E1/Erdalen Event 1; E2/Erdalen Event 2; F1/Finse Event 1; F2 /Finse Event 2), whereas the mid-Holocene has several unnamed events. Late-Holocene events are named from Jotunheimen (BI /Bøvertun Event 1 and BII /Bøvertun Event 2). (2) Sorting
/mean anomalies can be used to track abrupt changes in the sedimentation environment in a lake and thereby validate the use of lake sediments for reconstruction of former glacier fluctuations. (3) Basal radiocarbon dates from Dravladalsvatn indicate that glaciers were absent from the catchment shortly after 9600 cal. yr BP and that they reformed at 5200 cal. yr BP. (4) The early phase of mid-Holocene glacier growth was characterized by gradual glacier expansion leading to the first Subatlantic glacial event dated at 2300 cal. yr BP. This was a centennial-scale glacial readvance. (5) At 2200 cal. yr BP there was a significant change in glacier size, from a small glacier to glacier size larger than at present. The record from the last 2200 years shows high-frequenly glacial fluctuations at decadal and centennial timescales. It is indicated that the so-called ‘Mediaeval Warm Epoch’ was a humid phase at northern Folgefonna, as glacier growth and decay during this timespan was recorded. Altogether, the climate during the last 2200 years has been favourable for glacier growth at Folgefonna. The high-amplitude variation in ELA is therefore interpreted as a consequence of a more variable mode of the westerlies at the west coast of Norway. (6) The glacier net mass balance for northern Folgefonna is modelled by using instrumental temperature and precipitation records from Bergen and Ullensvang back to AD 1800. A comparison between DBD and modelled glacier net mass balance shows a remarkably similar pattern. (7) Dry bulk density (DBD) has the potential to resolve even small changes in silt production caused by interference in mass balance over short periods (subcentennial). Acknowledgements This is a contribution from NORPEC, a strategic university program headed by H.J.B. Birks and funded by the Norwegian Research Council (NFR). We are grateful to Åsmund Bakke, Joachim Riis Simonsen and Jorun Seierstad for help during the Jostein Bakke et al.: Utilizing physical sediment variability, Folgefonna, Norway fieldwork. We offer our sincere thanks to John Matthews, Gunhild Rosqvist and an anonymous reviewer for valuable comments. This is publication no. A72 of the Bjerknes Centre for Climate Research. References Arnaud, F., Lignier, V., Revel, M., Desmet, M., Beck, C., Pourchet, M., Charlet, F., Trentesaux, A. and Tribovillard, N. 2002: Flood and earthquake disturbance of 210Pb geochronology (Lake Antern, NW Alps). Terra Nova B072, 1
/8. Askvik, H. 1995: Oversikt over Norges Prekambriske og Paleozoiske berggrunn , Bergen: Geologisk Institutt, 1
/75. Bakke, J., Dahl, S.O. and Nesje, A. 2005: Lateglacial and earlyHolocene palaeoclimatic implications based on reconstructed glacier fluctuations and equilibrium-line altitudes at northern Folgefonna, Hardanger, western Norway. Journal of Quaternary Science 20, in press. Ballantyne, C.K. 2002: Paraglacial geomorphology. Quaternary Science Reviews 21, 1935
/2017. Ballantyne, C. and Benn, D.I. 1994: Paraglacial slope adjustment and resedimentation following recent glacier retreat, Fåbergstølsdalen, Norway. Arctic and Alpine Research 26, 255
/69. Barnekow, L., Possnert, G. and Sandgren, P. 1998: AMS C-14 chronologies of Holocene lake sediments in the Abisko area, northern Sweden
/a comparison between dated bulk sediment and macrofossil samples. Geologiska Föreningens i Stockholm Förhandlingar 120, 59
/67. Birkeland, B.J. 1932: Altere meteorologische beobachtungen in Ullensvang, luftdruck und temperatur seit 100 Jahren. Geofysiske publikasjoner. Vol. IX, 6. Det norske videnskaps-akademi i Oslo. Blake, G.R. and Hartge, K.H. 1986: Bulk density. In Klut, A. editor, Methods of soil analysis. Madison, WI: American Society of Agronomy-Soil Science Society of America, Agronomy Monograph 9, 363
/75. Blott, S.J. and Pye, K. 2001: Gradistat: a grain size distribution and statistics package for the analysis of unconsolidated sediments. Earth Surface Processes and Landforms 26, 1237
/48. Bradley, R.S., Hughes, M.K. and Diaz, H.F. 2003: Climate in Medieval time. Science 302, 404
/405. Broecker, W.S. 1991: The great ocean conveyor. Oceanography 4, 79
/89. Calvo, E., Grimalt, J. and Jansen, E. 2002: High resolution Uk/37 sea surface temperature reconstruction in the Norwegian Sea during the Holocene. Quaternary Science Reviews 21, 1385
/94. Cronin, T.M., Dwyer, G.S., Kamiya, T., Schwede, S. and Willard, D.A. 2003: Medieval Warm Period, Little Ice Age and 20th century temperature variability from Chesapeake Bay. Global and Planetary Change 36, 17
/29. Crowley, T.J. and Lowery, T.S. 2000: How warm was the medieval warm period? Ambio 29, 51
/54. Dahl, S.O. and Nesje, A. 1994: Holocene glacier fluctuations at Hardangerjøkulen, central southern Norway: a high-resolution composite chronology from lacustrine and terrestrial deposits. The Holocene 4, 269
/77. ____ 1996: A new approach to calculating Holocene winter precipitation by combining glacier equilibrium-line altitudes and pine-tree limits: a case study from Hardangerjøkulen, central southern Norway. The Holocene 6, 381
/98. Dahl, S.O., Bakke, J., Lie, O. and Nesje, A. 2003: Reconstruction of former glacier equilibrium-line altitudes based on proglacial sites: an evaluation of approaches and selection of sites. Quaternary Science Reviews 22, 275
/87. Dahl, S.O., Nesje, A., Lie, O., Fjordheim, K. and Matthews, J.A. 2002: Timing, equilibrium-line altitudes and climatic implications of two early-Holocene glacier readvances during the Erdalen Event at Jostedalsbreen, western Norway. The Holocene 12, 17
/ 25. Dean, W.E. 1974: Determination of carbonate and organic matter in calcareous sediments and sedimentary rocks by loss on ignition: 175 comparison with other methods. Journal of Sedimentary Petrology 44, 242
/58. DNMI 1993a: Nedbørsnormaler 1961
/1990. Klimaavdelingen: Det Norske Meteologiske Institutt. ____ 1993b: Temperaturnormaler 1961
/1990. Klimaavdelingen: Det Norske Meteologiske Institutt. Dyurgerov, M. 2002: Glacier mass balance and regime: data of measurements and analysis (edited by M. Meier and R. Armstrong). Occasional Paper no. 55, Institute of Arctic and Alpine Research, University of Colorado, 10
/15. Elvehøy, H. 1998: Samanlikning av massebalanse på Hardangerjøkulen og Folgefonna. In Elvehøy, H., editor, Oppdragsrapport, Norges vassdrag- og energiverk (Norwegian Survey for Hydrology and Energy Supply), 1
/27. Feddersen, H. 2003: Predictability of seasonal precipitation in the Nordic region. Tellus 55A, 385
/400. Griffey, N.J. and Worsley, P. 1978: The pattern of Neoglacial variations in the Okstindan region of northern Norway during the last three millennia. Boreas 7, 1
/17. Haldorsen, S. 1981: Grain-size distribution of subglacial till and its relation to glacial crushing and abrasion. Boreas 10, 91
/105. ____ 1983: Mineralogy and geochemistry of basal till and their relationship to till-forming processes. Norsk Geologisk Tidsskrift 63, 15
/25. Heiri, O., Lotter, A.F. and Lemcke, G. 2001: Loss on ignition as a method for estimating organic and carbonate content in sediments: reproducibility and comparability of results. Journal of Paleolimnology 25, 101
/10. Hopkins, T.S. 1991: The GIN-Sea-a synthesis of its physical oceanography and literature review 1972
/1985. Earth-Science Reviews 30, 175
/318. Hurrell, J.W. 1995: Decadal trends in the North Atlantic Oscillation: regional temperatures and precipitation. Science 269, 676
/79. Hurrell, J.W., Kushnir, Y., Ottersen, G. and Visbesk, M. 2003: The North Atlantic oscillation: climatic significance and environmental impact. In Hurrell, J.W., Kushnir, Y., Ottersen, G. and Visbesk, M., Geophysical Monograph series, 134, 1
/35. Haakensen, N. 1989: Akkumulasjon på breene i Norge vinteren 1988
/89. Været 13, 91
/94. Jørgensen, P. 1977: Some properties of Norwegian tills. Boreas 6, 149
/57. Karlén, W. 1976: Lacustrine sediments and tree-line variations as indicators of climatic fluctuations in Lappland, northern Sweden. Geografiska Annaler 58 A, 1
/34. ____ 1981: Lacustrine sediments studies. A technique to obtain a continous record of Holocene glacier variation. Geografiska Annaler 63A, 273
/81. Kolltveit, O. 1953: Jondal i gamal og ny tid. Jondal, Norway: Jondal Historielag. Leonard, E.M. 1985: Glaciological and climatic controls on lake sedimentation, Canadian Rocky Mountains. Zeitscrift für Gletcherkunde und Glazialgeologie 21, 35
/42. Lie, O., Dahl, S.O. and Nesje, A. 2003: A theoretical approach to glacier equilibrium-line altitudes using meteorological data and glacier mass-balance records from southern Norway. Holocene 13, 365
/72. Lie, O., Dahl, S.O., Nesje, A., Matthews, J.A. and Sandvold, S. 2005: Holocene fluctuations of a polythermal glacier in high-alpine eastern Jotunheimen, central-southern Norway: a multi-site, multiparameter approach on lacustrine sediments. Quaternary Science Reviews 23, 1925
/45. Lowe, J.J. and Walker, M.J.C. 2000: Radiocarbon dating the last glacial-interglacial transition (Ca. 14
/9 C-14 ka BP) in terrestrial and marine records: the need for new quality assurance protocols. Radiocarbon 42, 53
/68. MasterTech 1993: Sedigraph particle size analysis system. In Mastertech 5100 Operator’s manual. Matthews, J.A. and Karlén, W. 1992: Asynchronous neoglaciation and Holocene climatic change reconstructed from Norwegian glaciolacustrine sedimentary sequences. Geology 20, 991
/94. Matthews, J.A., Dahl, S.O., Nesje, A., Berrisford, M.S. and Andersson, C. 2000: Holocene glacier variations in central 176 The Holocene 15 (2005) Jotunheimen, southern Norway, based on distal glaciolacustrine sediment cores. Quaternary Science Reviews 19, 1625
/47. Menounos, B. 1997: The water content of lake sediments and its relationship to other physical parameters: an alpine case study. The Holocene 7, 207
/12. Nesje, A. 1992: A piston corer for lacustrine and marine sediments. Arctic and Alpine Research 24, 257
/59. Nesje, A. and Dahl, S.O. 2001: The Greenland 8200 cal. yr BP event detected in loss-on-ignition profiles in Norwegian lacustrine sediment sequences. Journal of Quaternary Science 16, 155
/66. Nesje, A., Dahl, S.O., Andersson, C. and Matthews, J.A. 2000a: The lacustrine sedimentary sequence in Sygneskardvatnet, western Norway: a continuous, high-resolution record of the Jostedalsbreen ice cap during the Holocene. Quaternary Science Reviews 19, 1047
/65. Nesje, A., Dahl, S.O. and Bakke, J. 2004: Were abrupt Lateglacial and early-Holocene climatic changes in northwest Europe linked to freshwater outbursts to the North Atlantic and Arctic Oceans? The Holocene 14, 299
/310. Nesje, A., Dahl, S.O. and Løvlie, R. 1995: Late Holocene glacier and avalanche activity in the Ålfotbreen area, western Norway: evidence from a lacustrine sedimentary record. Norsk Geologisk Tidsskrift 75, 120
/26. Nesje, A., Kvamme, M., Rye, N. and Løvlie, R. 1991: Holocene glacial and climate history of the Jostedalsbreen region, western Norway; evidence from lake sediments and terrestrial deposits. Quaternary Science Reviews 10, 87
/114. Nesje, A., Lie, O. and Dahl, S.O. 2000b: Is the North Atlantic Oscillation reflected in Scandinavian glacier mass balance records? Journal of Quaternary Science 15, 587
/601. Nesje, A., Matthews, J.A., Dahl, S.O., Berrisford, M.S. and Andersson, C. 2001: Holocene glacier fluctuations of Flatebreen and winter-precipitation changes in the Jostedalsbreen region, western Norway: evidence from pro-glacial lacustrine sediment records. The Holocene 11, 267
/80. Nordli, P.Ø., Lie, Ø., Nesje, A. and Dahl, S.O. 2003: Springsummer temperature reconstruction in western Norway 1734
/ 2003: A data-synthesis approach. International Journal of Climatology 23, 1821
/41. Østrem, G. 1975: Sediment transport in glacial meltwater streams. In Jopling, A. and Mc Donald, B., editors, Glaciofluvial and glaciolacustrine sedimentation , Society of Economic Palaeontologists and Mineralogists Special Publication 23, 101
/ 122. Østrem, G., Dale Selvig, K. and Tandberg, K. 1988: Atlas of glaciers in South Norway. Oslo: Norges vassdrags- og energiverk, vassdragsdirektoratet. Porter, S.C. 1975: Equilibrium-line altitudes of late quaternary glaciers in the Southern Alps, New Zealand. Quaternary Research 5, 27
/47. Roland, E. and Haakensen, N. 1985: Glasiologiske undersøkelser i Norge 1982 (with English summary). Norges vassdrags- og elektrisitetsvesen, Vassdragsdirektoratet, Hydrologiskavdeling, Report 1, 1
/102. Rosqvist, G.C. 1995: Proglacial lacustrine sediments from El Altar, Ecuador: evidence for late-Holocene climate change. The Holocene 5, 111
/17. Rosqvist, G.C., Jonsson, C., Yam, R., Karlén, W. and Shemesh, A. 2004: Diatom oxygen isotopes in pro-glacial lake sediments from northern Sweden: a 5000 year record of atmospheric circulation. Quaternary Science Reviews 23, 851
/59. Shabbar, A., Huang, J.P. and Higuchi, K. 2001: The relationship between the wintertime North Atlantic Oscillation and blocking episodes in the North Atlantic. International Journal of Climatology 21, 355
/69. Sigmond, E.M.O. 1985: Bedrock map of Norway. Trondheim: Norwegian Geological Survey. Snowball, I.F. and Sandgren, P. 1996: Lake sediment studies of Holocene glacial activity in the Kårsa valley, northern Sweden: contrasts in interpretation. The Holocene 6, 367
/72. Souch, C. 1994: A methodology to interpret downvalley lake sediments as records of Neoglacial activity: coast mountains, British Colombia, Canada. Geografiska Annaler 76A, 169
/85. Stuiver, M., Reimer, P.D., Bard, P.J., Beck, J.W., Burr, G.S., Hughen, K.A., Kromer, B., McCormac, G., van der Plicht, J. and Spurk, M. 1998: INTCAL98 radiocarbon age calibration, 24, 00
/ 0 cal BP. Radiocarbon 40, 1041
/83. Sundborg, A. 1956: The River Klarälven, a study of fluvial processes. Geografiska Annaler 38, 125
/316. Sutherland, D. 1984: Modern glacier characteristics as a basis for inferring former climates with particular reference to the Loch Lomond stadial. Quaternary Science Reviews 3, 291
/309. Tvede, A. 1972: En glasio-klimatisk undersøkelse av Folgefonni. Unpublished thesis, Department of Geography, University of Oslo. ____ 1979: Likninger til beregning av nettobalansen fra værdata. In Wold, B. and Repp, K., editors, Glasiologiske undersøkelser i Norge 1978 , Norges Vassdrags og elektrisitetsvesen. Report 4
/19, 1
/71. ____ 1989: Floods caused by a glacier dammed lake at the Folgefonni ice cap, Norway. Journal of Glaciology 13, 262
/64. Vorren, T.O. 1977: Grain-size distribution and grain-size parameters of different till types on Hardangervidda, south Norway. Boreas 6, 219
/27. Webb, P.W. and Orr, C. 1997: Analytical methods in fine particle technology. Norcross, GS: Micromeritics Instrument Corporation.