Shifts in precipitation during the last millennium in northern Scandinavia from lacustrine isotope records

Quaternary Science Reviews 66 (2013) 22e34

Contents lists available at SciVerse ScienceDirect

Quaternary Science Reviews journal homepage: www.elsevier.com/locate/quascirev

Shifts in precipitation during the last millennium in northern Scandinavia from lacustrine isotope records Gunhild C. Rosqvist a, *, Melanie J. Leng b, c, Tomasz Goslar d, e, Hilary J. Sloane c, Christian Bigler f, Laura Cunningham f, g, Anna Dadal f, Jonas Bergman a, Annika Berntsson a, Christina Jonsson a, Stefan Wastegård a a

Department of Physical Geography and Quaternary Geology, Stockholm University, 106 91 Stockholm, Sweden Department of Geology, University of Leicester, Leicester LE1 7RH, UK NERC Isotope Geosciences Laboratory, British Geological Survey, Keyworth, Nottingham G12 5GG, UK d  , Poland Faculty of Physics, A. Mickiewicz University, ul. Umultowska 85, 61-614 Poznan e  Radiocarbon Laboratory, ul. Rubiez_ 46, 61-612 Poznan  , Poland Poznan f Department of Ecology and Environmental Science, Umeå University, SE-901 87 Umeå, Sweden g School of Geography and Geosciences, Irvine Building, University of St Andrews, North Street, St Andrews, KY16 9AL, Fife, Scotland, UK b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 25 April 2012 Received in revised form 12 October 2012 Accepted 17 October 2012 Available online 1 December 2012

Here we present d18Odiatom data from two high-latitude lakes; one has short residence time and a water isotopic composition (d18Olake) that fluctuate due to seasonal variations in precipitation and temperature, and the other has d18Olake that is influenced by longer lake water residence times and evaporation. The d18Odiatom records reveal common responses to precipitation forcing over the past millennium. Relatively wet summers are inferred from d18Odiatom between 1000 and 1080 AD, 1300 and 1440 AD, and during the early 19th century, coincided with periods of high cloud cover inferred from tree-ring carbon isotopes, and other data for high Arctic Oscillation index. While relatively dry summers with increasing influence of winter snow are indicated between 1600 and 1750 AD. The co-response between carbon isotopes in trees and oxygen isotopes in diatoms strengthens the relationship between cloud cover and precipitation and the hypothesis that these changes were the result of significant regional shifts in atmospheric circulation. Ó 2012 Elsevier Ltd. All rights reserved.

Keywords: Last millennium Lakes Isotope Diatoms Scandinavian climate

1. Introduction The atmospheric circulation over northern Scandinavia is governed by the balance between Arctic and North Atlantic air masses (Hurrell et al., 2003; Sutton and Hodson, 2005). Shifts in this balance at interannualeinterdecadal time scales have been attributed to variations in the Arctic Oscillation (AO) and the associated North Atlantic Oscillation (NAO) (Thompson and Wallace, 1998; Rigor et al., 2002; Hurrell et al., 2003). A wet and mild summer climate is a consequence of positive AO conditions (high index), a dry and cold climate the consequence of negative AO conditions (low index). Perturbations associated with NAO have the largest impact on regional winter climate which becomes mild and moist during positive phases and cold and dry during negative phases (Hurrell et al., 2003). A positive phase of summer NAO is

* Corresponding author. Tel.: þ46 (0)702293404; fax: þ46 (0)8164818. E-mail address: [email protected] (G.C. Rosqvist). 0277-3791/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.quascirev.2012.10.030

characterized by anticyclonic summer conditions in northern Scandinavia, resulting in a warm and dry climate (Folland et al., 2009). Past variation in the summer climate as a consequence of AO and NAO shifts has previously been reconstructed using treerings (D’Arrigo et al., 2003; Folland et al., 2009; Young et al., 2010, 2012). The Arctic and Atlantic derived air masses that impact Northern Europe have characteristic seasonal patterns of temperature and precipitation and distinct oxygen and hydrogen isotope compositions (Bowen and Wilkinson, 2002). The oxygen isotopic composition of lacustrine diatoms/carbonates and aquatic cellulose have previously been used to reconstruct past changes in the oxygen isotopic composition of lake water (d18Olake) and precipitation (d18OP) in northern Sweden (Hammarlund et al., 2002; Rosqvist et al., 2004, 2007; Jonsson et al., 2009, 2010a,b; Andersson et al., 2010; St. Amour et al., 2010). Many of these records show a Holocene depletion trend which has been explained in terms of long term circulation and temperature changes forced by the decrease in summer insolation at this latitude (Shemesh et al., 2001;

G.C. Rosqvist et al. / Quaternary Science Reviews 66 (2013) 22e34

Hammarlund et al., 2002, 2004; St. Amour et al., 2010). Isotopic shifts on millennial-centennial time scales have been associated with North Atlantic atmospheric and oceanic circulation changes over the last 3000 years (Rosqvist et al., 2004, 2007; St. Amour et al., 2010). Our aim is to determine past changes in seasonal variability in precipitation and moisture availability through analyses of the oxygen isotopic composition in lake water and preserved diatom silica. The results are used to infer past changes in dominant atmospheric circulation modes. While high-latitude lakes with short residence times have water isotopic compositions (d18Olake) that fluctuate due to seasonal variations in precipitation and temperature (Jonsson et al., 2009), this variation is mainly produced by the input of isotopically depleted snowmelt (winter precipitation) in May/June and to a lesser extent the subsequent influence of relatively more enriched summer precipitation. In contrast the d18Olake of semi-closed lakes, even at high latitudes, is influenced by longer lake water residence times and evaporation

23

(Leng and Marshall, 2004; Jonsson et al., 2009). Here we present two d18Odiatom records that allow us to infer changes in seasonality and amount of precipitation at much higher resolution than has been previously achieved from Northern European lakes, down to a sub decadal (in part) time scale. 2. Regional setting Lake Stuor Guossasjavri (hence forth called Guossasjavri) (67 500 4700 N, 19 400 4800 E) is located ca w30 km west of Kiruna, and 60 km east of Kebnekaise, the highest summit in Sweden at 2101 m above sea level (m. a. sl.) (Fig. 1a, b). The flat lake catchment covers 3.7 km2 between 780 and 620 m a. sl. The modern lake surface area is 0.3 km2 and maximum water depth is 6.5 m. The mean water depth is only 2.2 m and about 10% of the present day lake area is deeper than 4 m (Karlsson and Byström, 2005). Guossasjavri is a semi-closed lake as it only has one semi-permanent inlet and outlet, both of which are active during snowmelt. The water

Fig. 1. Map of Scandinavia (a), and the catchments of Lake Stuor Goussasjavri (b) and Lake Spåime (c).

24

G.C. Rosqvist et al. / Quaternary Science Reviews 66 (2013) 22e34

Table 1 Climate and lake characteristics (Alexandersson and Eggertsson Karlström, 2001; Gustavsson et al., 2011).

Mean annual temperature  C Mean January temperature  C Mean July temperature  C Ice cover season Annual precipitation (mm) Annual evaporation (mm) Maximum water depth (m) Lake area (km2) Catchment area (km2) Lake volume (mm3) Residence time

Lake Spåime/Storlien

Goussasjavri/Kiruna

þ1.2 7.7 þ10.8 Mid Octeearly June 900 100 3.5 0.03 3.5 1.5  1014 2 weeks

1.7 13.8 þ12.1 Late Octelate May 500 200 6.5 (mean 2.2) 0.3 3.7 1.1  1015 >20 weeks

residence time is estimated to be 21 weeks (Table 1). The ice free period lasts about 20 weeks from mid-May to mid-October. Lake area is prone to fluctuations due to the flat catchment and the area immediately surrounding the lake is wetland. Catchment vegetation consists mainly of willow, grasses, sedges and mosses. The lake is located below the present mountain birch (Betula pubescens ssp) tree-line and approximately at the Scots Pine (Pinus sylvestris) treeline (Fig. 1b) (Table 2). Lake Spåime (63 060 4000 N; 12190 1100 E) is located at 887 m a. sl., 10 km north of the Sylarna mountains in Jämtland, west central Sweden, and w80 km from the Norwegian coast (Fig. 1a, c). The catchment encompasses ca 3.5 km2 between 887 and 1100 m a. sl. The lake surface area is 0.03 km2 and maximum water depth is 3.7 m (Hammarlund et al., 2004). The lake is located above the B. pubescens spp treeline, which occurs at ca 800 m a. sl. Lake Spåime Table 2 Radiocarbon dates of samples from Guossasjavri. Radiocarbon measurements were performed following a standard protocol for AMS dating. The 14C dates of the youngest samples are negative, reflecting a 14C bomb-peak (Hua and Barbetti, 2004). $ e Date not used in the age-depth model. Core

Material

Depth (mm)

14

II II II II II II II II II II II II II II III III III III III III III III III III III III III III III III III III III III

Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk

4 191 221 251 259 274 289 304 319 334 349 364 379 394 305 355 395 435 475 505 545 575 615 655 695 715 745 785 815 845 885 925 965 995

70$ 1015 1035 1090 1130 1115 1040 1095 1080 1030 1080 1085 1215 1150 1175 1120 1185 1185 1105 1200 1375 1625$ 1535$ 1365 1740$ 1640 1575 1630 1650 1780 1815 1860 1955 1940

organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic organic

C BP

Error 27 30 30 35 30 30 30 30 35 30 30 30 30 50 30 30 30 35 30 30 30 30 35 35 30 35 35 35 35 30 35 30 30 30

is a hydrologically open lake with one main inlet in the south and one outlet in the north. The water residence time is approximately 2 weeks (St. Amour et al., 2010; Jonsson et al., 2010b). Groundwater springs occur in the catchment. Small snow fields (<100 m2) contribute melt water during part of the summer, usually until July or August. Vegetation in the catchment is dominated by heath communities with dwarf-shrubs, willows, grasses and sedges. 3. Material and methods 3.1. Meteorological data A pronounced oceanic-continental gradient from west to east causes a steep precipitation gradient from the Atlantic coast across the Scandes mountains (Jutman, 1995; Swedish Meteorological and Hydrological Institute (SMHI) data 1961e1990). Lake Spåime is located close to the water divide and Guossasjavri is located leeward of the mountains. For both lakes summer precipitation exceeds winter precipitation as approximately half of the annual precipitation falls during the summer months (JuneeSeptember) (Fig. 2a, b). 3.2. Water sampling and isotope analysis Eighteen water samples collected during the period between March and September from Guossasjavri and nearby surface waters were analysed for oxygen (d18O) and hydrogen (dD) isotope composition at the NERC Isotope Geosciences Laboratory (NIGL). Analyses of d18O were performed by constant-temperature equilibration with CO2 (Epstein and Mayeda, 1953) and analysed using a VG SIRA mass spectrometer. For hydrogen isotopic analysis, an on-line Cr reduction method was used with an EuroPyrOH-3110 system coupled to a Micromass Isoprime mass spectrometer. Isotopic ratios (18O/16O and 2H/1H) are expressed in delta units i.e. d18O and dD and as & (parts per mille), and defined in relation to the international standard, VSMOW (Vienna Standard Mean Ocean Water). Analytical precision is typically 0.1& for d18O and 1.0& for dD (1 SD). Water samples from Lake Spåime and other surface waters nearby, were analysed for d18O and dD at the University of Copenhagen and the University of Aarhus (St. Amour et al., 2010), where analytical uncertainties were reported to be 0.2& for d18O and 2.0& for dD. 3.3. Core collection The uppermost 60 cm of Guossasjavri sediment was sampled at a water depth of 5.8 m using a freeze corer device (core I), and the uppermost 40 cm using a gravity corer (core II) (Renberg and Hansson, 2008) in March 2007. A Russian corer was used to obtain the full Holocene sequence, in total 450 cm, of which the uppermost 100 cm (core III) represents the last 1000 years. The gravity core was directly sliced in the field at contiguous 0.25 cm intervals, whereas the other cores were cleaned and sliced at 0.5 cm intervals in the laboratory. At Lake Spåime, two surface sediment cores (core I: 27 cm; core II: 15 cm long) were sampled in February 2002 and in May 2006 using the same type of gravity corer. A Russian corer was used to collect a 90 cm long core (core III) representing the last 1000 years in 2006. The gravity cores were sliced in the field and core III was sliced in the laboratory, both at 0.5 cm contiguous intervals. 3.4. Age determinations Thirteen samples from a gravity core (II) from Guossasjavri were freeze-dried and analysed for 210Pb activity and 137Cs concentration

G.C. Rosqvist et al. / Quaternary Science Reviews 66 (2013) 22e34

25

Fig. 2. Local meteorology at GuossasjavrieKiruna (a) and SpåimeeStorlien/Bredkälen (b).

at Flett Research, USA. Nineteen samples from one gravity core (2002) from Lake Spåime were freeze dried and then analysed for 210 Pb activity and 137Cs at the University of Liverpool Environmental Radiometric Laboratory to provide an age-depth model for the uppermost sediments (Davies et al., 2007). Twenty-seven samples of terrestrial plant remains (core III) from Lake Spåime and 34 bulk sediment samples (14 from core II, 20 from core III)

from Guossasjavri were radiocarbon dated at the Poznan Radiocarbon Laboratory, Poland. The lack of preserved plant remains made it necessary to use bulk sediment samples from Guossasjavri. In the radiocarbon laboratory the samples for 14C analysis were prepared with the standard AAA method, and 14C was measured using the AMS technique (Goslar et al., 2004). The cores were also screened for tephra particles using the method described in Davies

26

G.C. Rosqvist et al. / Quaternary Science Reviews 66 (2013) 22e34

et al. (2007). The objective was to identify the depth level in different cores where tephra from the 1875 eruption of the Icelandic volcano Askja could be detected. 3.5. Carbon and nitrogen content Total Elemental Carbon (TC) and Total Nitrogen (TN) content in the sediments from core III from Guossasjavri were analysed using a PerkineElmer 2400 Series II CHNS/O-Analyzer. TOC and TN content were determined on samples from Lake Spåime sediment cores using a Carlo Erba Instruments NC2500 elemental analyser. TC data are expressed as percentages of elemental carbon in relation to total dry weight. The reproducibility is within 0.5% for both measurements. 3.6. Diatom analysis Standard techniques involving hot H2O2 (30%) and HCl (10%) treatment were applied for diatom preparation from both lakes (Battarbee, 1986; Renberg, 1990). Samples were permanently mounted on slides using Naphrax. Identification and classification of diatoms were done under a light microscope with phasecontrast optics at 1000 magnification. A minimum of 400 valves were identified and classified in each sample. Diatom taxonomy largely followed the European freshwater diatom flora (Krammer and Lange-Bertalot, 1986e1991). In total, 275 and 219 samples were analysed from Guossasjavri and from Spåime, respectively. Zonation of the diatom data was performed using optimal partitioning and the sum-of-squares criteria (Birks and Gordon, 1985) within the program ZONE (Lotter and Juggins, 1991). The number of significant zones was assessed by comparison with the brokenstick model (Bennett, 1996) and only statistically significant zones are presented.

fluorination technique to extract oxygen isotopes described in Leng and Barker (2006) and Leng and Sloane (2008). The oxygen was then converted to CO2 (Clayton and Mayeda,1963), and 18O/16O ratios were measured using a dual inlet mass spectrometer. Duplicate analysis were performed to ensure good reproducibility. The precision of the d18Odiatom technique is w0.3&. The diatom stable isotope results are expressed as delta (d) values, representing deviations in & from VSMOW. 4. Results 4.1. Isotope hydrology Modern lake water d18O of Guossasjavri plot along a local evaporation line (LEL) indicating that the lake evaporates in the summer months and is recharged in the winter (Figs. 2a and 3). The nearby river Aitejohka has an isotopic composition (14& d18O, e 105& dD) similar to most through flow waters measured in northern Lapland (Jonsson et al., 2009). The modern water residence time has been calculated (using catchment area, lake volume and annual runoff) to be 21 weeks which is almost a full summer season (JuneeSeptember). Thus the lake d18O depends on the water balance of the previous years, i.e. evaporation and input ratio (E/I ratio), and the amount and isotope composition of summer and winter precipitation. The d18O and dD of water sampled from Lake Spåime plot along the global meteoric water line (GMWL) (Fig. 3) indicating that the lake waters represent seasonal variation in precipitation (St. Amour et al., 2010), with lower values in the winter compared to summer precipitation, due to the short lake water residence time. 4.2. Sediment characteristics and chronologies

3.7. Oxygen isotope analyses of diatoms

The sediments in Guossasjavri consist of brown homogenous clay-gyttja. Mean TC is 20% (s 0.9) and C/N 13.4 (s 0.3) (Fig. 4). An

The oxygen isotopic composition of diatom silica (d18Odiatom) was performed on 0.5 cm samples from Guossasjavri cores II and III (74 samples) and from Spåime cores I and III (79 samples). The preparation of isotope samples was modified after Morley et al. (2004). Approximately 30 mg of freeze-dried sediment was used for preparation. Hydrochloric acid (32% HCl) and hydrogen peroxide (35% H2O2) were added and sediment samples were heated to 90  C on a water bath and left for >8 h until carbonate and organic material were oxidized. Subsequently samples were washed with distilled water to remove remaining chemicals and organic residue. Remaining material was sieved through different mesh sizes to test which fractions could be removed without losing diatom frustules. In Guossasjavri the <63 mm fraction contained most diatom frustules, in Spåime almost all diatom frustules were in the <20 mm fraction. To remove remaining minerogenic particles the individual samples were split into different specific density fractions using heavy liquid separation with sodium polytungstate (SPT). Pure biogenic silica samples were obtained by discarding densities above 2.3 and below 1.8 g/cm3 for Spåime and below 1.9 g/cm3 for Guossasjavri. Samples were then repeatedly centrifuged in distilled water and then wet sieved over a 5 mm mesh to remove traces of SPT, chrysophyte cysts and fragments of diatom frustules. Every sample was visually inspected for impurities under a light microscope at 400 magnification. Some samples were additionally inspected in a scanning electron microscope (SEM). None of the samples contained any visible phytoliths, clay or silt particles. Samples from Guossasjavri contained a small amount of chrysophyte cysts. Excess water was removed with pipettes and the pure biogenic silica was dried at 40  C for 24 h. Biogenic silica d18O was analysed at NIGL using the stepped

Fig. 3. The isotopic composition of present day waters from Guossasjaure, Lake Spåime and neighbouring surface waters. The Global Meteoric Water Line (GMWL) and the Local Evaporation Line (LEL) are given.

G.C. Rosqvist et al. / Quaternary Science Reviews 66 (2013) 22e34

27

Fig. 4. Total carbon (TC), Total Organic Carbon (TOC) and carbon/nitrogen ratio (C/N) for Guossasjavri (a and b), and Lake Spåime (c and d).

age-depth model based on 210Pb and 137Cs dating was developed for the uppermost 14 cm of sediments from Guossasjavri core I (Fig. 5). We used the constant rate of supply (CRS) model (Appleby, 2008) as it takes into account changes in the sediment accumulation rate. Two peaks in 137Cs at 4.00e4.25 cm and 6.75e7.00 cm depth are associated with the 1986 Chernobyl accident and the 1963 atomic nuclear bomb testing, respectively (Fig. 5). The presence of 137Cs below the lower peak is thought to be due to downward diffusion. Results show that sediment accumulated at a rate of 1 mm/yr over the past 100 years, except during the last ca 20 years when the rate seems to have been 3 mm/yr, probably due to the looser density structure of the uppermost sediments.

Fig. 5. Age determination using

210

Pb and

137

Most 14C dates from Guossasjavri core II cluster around 1100 14C BP, despite the fact the material was sampled from a wide section (19e36 cm depth, Table 1), and the uppermost sediment in this section should according to 210Pb data represent the 19th century. The distribution of the 14C dates (Fig. 6) agrees well with the radiocarbon calibration curve prior to 1950 AD (Reimer et al., 2009) if we agree that the 14C dates were affected by a reservoir effect of almost 1000 years (955  50 14C years, cf. shift between two vertical scales in Fig. 6). Using reservoir-corrected 14C dates, we constructed the age-depth model with the free-shape algorithm (Goslar et al., 2009) designed to generate (using the Monte-Carlo approach) large sets of age-depth curves being monotonic, as smooth as

C of the uppermost sediments from Goussasjavri core I.

28

G.C. Rosqvist et al. / Quaternary Science Reviews 66 (2013) 22e34

Fig. 6. Radiocarbon dates of Core II and III from Guossasjavri (vertical axis to the right) on the background of the 14C calibration curve (INTCAL’09, vertical axis to the left). 14 C dates of two cores (II and III) are plotted with respect to the depth scales of these profiles (3 dates, rejected from the age-depth model, are displayed in grey). Relying on 210 Pb dating, 144 mm depth in core II is matched to 1910 AD. The scale of 14C dates of the Guossasjavri samples is shifted with respect to that of the calibration curve (by 955 years  50 years) to obtain the best fit to the calibration curve. Note: a shift of the depth scales of both cores is allowed, because the depth of the water-sediment interface at the site where core III was collected was not determined in detail.

possible, and at the same keeping average deviations of (reservoir corrected) 14C dates from to the radiocarbon calibration curve close to 1-sigma uncertainties of the dates. For the full set of dates, the conditions above were impossible to fulfil, unless the minimum number of most outlying dates (deviating from calibration curve by >3-sigma) were rejected. Actually, there were 3 such dates (Table 1, Fig. 6), all of them appearing much too old, and probably representing sediment reworkings. According to the model, the sediments in the interval 19e39 cm accumulated between the early 17th century and ca AD1850 (Fig. 7a, b), with a sedimentation rate that is the same (1 mm/yr) as established by the 210Pb dating of the uppermost Guossasjavri core I. The results of the 14C dating of samples from Guossasjavri core III can also be compared to the shape of 14C calibration curve (Fig. 6). Six dates from the interval between 30.5 and 50.5 cm depth in core III also cluster around an age of ca 1150 14C BP, two dates at 54.5 and 65.5 cm are close to 1350 14C BP, and four dates between 71.5 and 81.5 cm cluster around 1600 14C BP. Radiocarbon ages of these three groups are again w1000 14C years older than 14C ages of the three major wiggles of the calibration curve, between w1650e1950 AD, w1500e1600 AD and w1400e1500 AD. We therefore assume that the 14C reservoir age in this environment did not change appreciably over the last millennium. If so, sedimentation rate in Guossasjavri core III was indeed constant and similar to that of core I and II, i.e. w1 mm/yr (Fig. 7a, b). The reservoir effect probably resulted from deposition of 14C-depleted organic matter derived from the watershed (Abbott and Stafford, 1996) as there is no carbonate bedrock in the catchment (Offerberg, 1967). Detection of the Askja 1875 AD tephra in the sediments would have provided us with an independently determined age horizon, but unfortunately it was not identified at this more northern location (Carey et al., 2010).

Fig. 7. a) Age-depth model of the core II from Guossasjavri. The white silhouettes show probability distributions of calendar dates, obtained by calibration of individual 14 C dates, and the black silhouettes represent 210Pb dates. The dotted silhouettes show probability distributions of calendar dates of the radiocarbon-dated samples, derived from the sets of age-depth curves generated by the free-shape algorithm. The best-fit age-depth curve is shown by a solid line. b) Age-depth model of the core III from Guossasjavri. The white silhouettes show probability distributions of calendar dates, obtained by calibration of individual 14C dates. The best-fit age-depth model is shown by a solid line. The dotted silhouettes show probability distributions of calendar dates of the free-shape model. The thick dotted line shows the best-fit age-depth model of core II.

The Lake Spåime sediments are dark brown gyttja silt with some visible laminations. Mean TOC is 7.5% (s 1.1) and C/N is 7.5 (s 0.8). The lower carbon content in Lake Spåime is to be expected due to the different catchment characteristics compared to the Guossasjavri setting (higher altitude, through-flow, less vegetation) which decrease productivity and increase potential inwash of detrital

G.C. Rosqvist et al. / Quaternary Science Reviews 66 (2013) 22e34

29

minerogenic material and re-suspension and re-distribution of sediments in the lake. The result of the tephra analysis of sediment samples from Lake Spåime shows that rhyolitic tephra particles were present near the top in all cores, which aided core correlation. The geochemical composition of the tephra matched the Askja 1875 AD eruption (Davies et al., 2007). The age-depth model of Spåime core III was constructed using the results of radiocarbon measurements on terrestrial plant remains (Table 3). The obtained series of 14C dates revealed severe stratigraphic inversions, much exceeding uncertainties of dates and amplitudes of wiggles of 14C calibration curve (Fig. 8), thus clearly showing that many dates were not representative for the true age of the dated sediment levels. We argue that its easier to explain why a date is too old (by occurrence of rebedded macrofossils) than to explain why a date is too young. Therefore, before running the age-depth model algorithm (the same as used from Guossasjavri), we cancelled all severe inversions by rejection of 10 clearly too old 14 C dates, and supposed that the remaining (younger) 18 14C dates were truly representative for the age of the sediments. The agedepth model (Fig. 9) was additionally constrained using the Askja 1875 AD tephra, detected at 234 mm depth. 4.3. Diatoms Diatoms from the Guossasjavri core predominantly comprise periphytic taxa, with Achnanthes minutissima [agg.] dominant throughout the sequence. Nitzschia fonticola and Brachysira vitrea are also abundant. Planktonic taxa, namely Cyclotella, are also present in moderate abundances throughout the core. Overall, the diatom assemblages from Guossasjavri have remained relatively stable over the last millennium (Fig. 10a). The numerical zonation identified eight significant zones, however differences between the zones are subtle. The first (AD 750e942) and second zones (AD 942e1107) are similar, whereas the subsequent third zone (AD 1107e1266) shows relatively low abundances of N. fonticola.

Fig. 8. Age depth model of Lake Spåime. Radiocarbon dates from core III from Lake Spaime, on the background of the 14C calibration curve. The outlying dates (not used in age-depth modeling) are shown with open symbols.

No major taxonomical shifts occur within the fourth (AD 1266e 1425) and fifth zone (AD 1425e1588). The sixth zone (AD 1588e 1741) is characterized by increasing abundance of Tabellaria flocculosa [agg.] and Cyclotella pseudostelligera. Whereas the high abundances of T. flocculosa [agg.] also persist within the seventh zone (AD 1741e1889), C. pseudostelligera and also A. minutissima

Table 3 Radiocarbon dates of samples from Spåime. The 14C date of the youngest sample is negative, reflecting the 14C bomb-peak (Hua and Barbetti, 2004). $ e date not used in the age-depth model. Core

Material

III III III III III III III III III III III III III III III III III III III III III III III III III III III

Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial Terrestrial

plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plant remains plan remains

Depth (mm)

14

Error

162 165 202 217 222 239 241 260 270 273 289 291 300 320 350 370 397 421 449 450 492 525 550 587 616 645 699

629 1180$ 110 235 150 30 1005$ 390$ 410$ 460$ 120 270$ 1340$ 260 400$ 120 380 575$ 400 410 610 815 890 900 1405$ 945 950

24 60 30 35 50 50 30 30 50 40 40 30 80 30 30 50 30 30 50 40 30 35 190 40 30 30 30

C

Fig. 9. Best fit age-depth model for Lake Spåime sediment core III. The grey and white silhouettes show probability distributions of calendar dates obtained by calibration of individual 14C dates (outliers not used in age-depth modelling are distinguished by white silhouettes only). The best-fit age-depth model is shown by a solid line. The dotted silhouettes show probability distributions of calendar dates of the free-shape model. The calendar date of the Askja tephra at 234 mm, is represented by a black peak.

G.C. Rosqvist et al. / Quaternary Science Reviews 66 (2013) 22e34

m in ut is Ac si m hn a Ac an [a gg t hn he .] Ac an s hn the ma Am an s rg s i p th u nu Br ho es bat lat ac ra pu om a h y l i b si l o sir yc la ide a a s vi Br tre ac a C hy yc sir lo a te br lla e C ps biss yc lo eu on te do ii lla C st ro ym el s lig si C be er i ym lla a C be ce ym lla sa t b Fr e de ii ag lla sc Fr ilar mi ript us ia cro a N tul co ce av ia ns ph N icul rho true ala av a m n N icul cryp boid s [a av a to e gg .] N icul pup ce s itz a u ph l N sch rad a ala itz ia io sc a sa hi mp a fo hib nt ia Pi ic nn ol a Ta ula r be ia lla bi ria ce flo ps D cc ia to ul m os a zo [a ne gg s .]

30

Ye ar s Ac AD hn an th es

a

-2000 SGJ-8 -1900 SGJ-7

-1800 -1700

SGJ-6

-1600 SGJ-5

-1500 -1400

SGJ-4 -1300 -1200

SGJ-3

-1100 SGJ-2

-1000 -900

SGJ-1

-800

b

20

40

20

20

20

20

20

Ye ar A Ac D hn a Ac nt hn he s Ac ant cu hn he rtis an s k sim th rie Ac es ge a h le ri Ac nan v t hn he and e an s th ma ri Ac es rg hn i m n Ac a n in ula ut ta h th is Ac nan es si m hn th no a Ac an es do [a s p gg hn t h e a e .] Ac an s ter hn th pu se an es sil nii l a s t he ac Ac s cu h sc la Ac nan ot ic hn t h a e Au a n s la the su Au co s ba la se ve to Br co ira ntr moi ac se am ali de Fr hy ira big s s ag sir dis u Fr ila a v tan a ag ria itre s ila br a va r. ria ev al co istr pi ge ns iat na tru a Fr en ag s i Fr lar v a ag i a r. ila el ve ria lip nt tic Fr er p ag in a na ila ta ria Fr ps ag eu ila do ria co N vir av ns e i tru sc N cu av la en en s s N icu pup va itz la u r. l s v a N ch itio ex itz ia s ig ua Pi sch fon a nn ia tic Pi ula pe ola nn ria rm Ta ula bi inu be ria ce ta p D llar me s ia ia s to m flo olep zo ccu ta ne lo s sa [a gg .]

-700

-2000 Sp-6 -1900 Sp-5

-1800 -1700

Sp-4 -1600 -1500 Sp-3 -1400 -1300 Sp-2 -1200 -1100 -1000

Sp-1 20

20

20

20

40

20

20

20

Fig. 10. Stratigraphical record (y-axis Year AD) of abundant diatom taxa in the sediment cores, (a) Guossasjavri and (b) Spåime. Diatom taxa were selected based on their abundance (either average abundance >1%, or >5% in at least one sample) and are expressed as percentages. Y-axis show age AD.

[agg.] show again lower values. The last zone (AD 1889e2007) is characterized by pronounced shifts in the most recent samples, namely the relative abundances of N. fonticola and B. vitrea increase sharply. Diatoms from the Spåime core are dominated by small Fragilaria taxa (e.g., Fragilaria construens var. venter, Fragilaria pseudoconstruens, Fragilaria virescens var. exigua) and small Achnanthes taxa (e.g., A. minutissima [agg.], Achnanthes levanderi; Fig. 10b). Overall, the diatom taxa recorded in Spåime are typical for periphytic habitats in alpine and subarctic environments (e.g., Bigler and Hall,

2002). Even though there are no pronounced shifts in diatom composition over the studied period, the zonation approach identified six significant diatom zones. The first (until AD 1171) and the second zone (AD 1171e1345) show relatively high relative abundance of N. fonticola and Fragilaria elliptica, respectively. During the third zone (AD 1345e1546) a transient, slight decrease of A. levanderii, Fragilaria pinnata and Navicula pupula is recorded. At the transition from the third to the forth zone (AD 1546e1736), F. elliptica shows decreasing abundances, whereas Achnanthes nodosa increases. At the end of the fifth zone (AD 1736e1889),

G.C. Rosqvist et al. / Quaternary Science Reviews 66 (2013) 22e34

F. virescens var. exigua shows considerably lower values, at the same time as F. pinnata and F. elliptica increase. The main feature of the sixth zone is the high relative abundances of F. construens var. venter. 4.4. Oxygen isotope composition of diatom silica During the last 200 years samples from both sequences represent w5 yrs; while samples from sediment that accumulated prior to 1800 AD represent between 10 and 15 years. The d18Odiatom from Guossasjavri vary between þ25.7& and þ30.4& throughout the sequence, the mean d18Odiatom is þ28.4& and the variability is þ4.8&. The record minimum occurs at 1120 AD and the maximum occurs 1360 AD. The d18Odiatom from Lake Spåime varies between þ26.2& and þ29.3& throughout the sequence, the record minimum occurs at w1987 AD, the maximum at w1440 AD. The mean d18Odiatom is þ28& and the variability is 3.1&. Interestingly the d18Odiatom records are similar on the centennial time scale (Fig. 11). The magnitude of change is different between the records, which is to be expected. The potential effect on Guossasjavri d18O by changes in lake level (and surface/volume ratio) increases the potential magnitude of change in d18O lake water and diatom. Both records show decreasing d18Odiatom values between 1020 and 1120 AD. From 1120 onwards values increase successively until the Guossasjavri record maximum is reached around 1360 AD and remains high until 1440 AD. In Lake Spåime the highest value is reached 1440 AD. After 1440 AD both records display decreasing values. Values increase again in Guossasjaure between 1740 and 1890 AD. In contrast the d18Odiatom stays low in Lake Spåime until 1830 AD when a major shift towards more positive values occur. After 1890 AD d18Odiatom becomes more negative again at both sites. 5. Discussion Despite the different hydrological regimes the d18Odiatom at the two lakes appears to have responded similarly over the last millennium. This requires that past changes in seasonality and amount of precipitation significantly influenced the isotopic water balance of these lakes. Despite such changes neither the TOC, C/N

31

nor diatom assemblage records reveal any significant variation that can be ascribed directly to climate. Most of the variability of the TOC in Lake Spåime is due to variation in inwash of detrital material from the catchment. Previous work has shown that changes in d18Odiatom from open through-flow lakes in northern Sweden primarily reflect changes in the seasonality of precipitation and in the isotopic composition of dominant air masses (Shemesh et al., 2001; Rosqvist et al., 2004, 2007; St. Amour et al., 2010; Jonsson et al., 2010b). Rosqvist et al. (2004, 2007) showed that a dominance of cool dry Arctic air masses from the north causes lake water d18O depletion on the centennial time scale. In contrast more positive isotopic compositions would result from a dominance of precipitation from southwesterly derived North Atlantic air masses (GNIP Database, 2012). Both changes in the seasonality and amount of precipitation on inter-annual to inter-decadal time scales would follow upon shifts in AO and NAO phases. As our d18Odiatom records reflect changes during both summer and winter we have the opportunity to infer changes in past AO and NAO patterns. The potential effect of changing condensation temperatures also needs to be accommodated for. The relationship between increased condensation temperature and the oxygen isotopic composition of precipitation (d18OP is þ0.7&/ C (Dansgaard, 1964)). The amplitude in the d18Odiatom is 4.7& and 3.1& for Guossasjavri and Spåime, respectively, which is more than can be accounted for as a response to changing condensation temperature during the last 1000 years. We also have to consider the negative effect of lake water summer temperature on d18Odiatom (0.2&/ C; Leng and Barker, 2006)). The net effect of an increased condensation temperature is therefore only þ0.5&/ C, assuming the same temperature change in air and lake water. The amplitude of summer temperature change during the last 1000 years has only been w1  C (Esper et al., 2012). We suggest therefore that temperature changes could have only amplified or dampened the isotopic signal but were not the primary driver. Catchment hypsometry allows Guossasjavri to change its surface area and thus its surface/volume ratio (S/V) significantly. A large S/V ratio will strongly affect evaporation potential. The d18Olake and d18Odiatom therefore primarily responded to changes in

Fig. 11. d18Odiatom records from Guossasjaure (c), Lake Spåime (a), and a tree-ring d13C record from northwestern Norway (b) (Young et al., 2012). Grey bars indicate period for which a wet summer climate and a positive AO index have been inferred.

32

G.C. Rosqvist et al. / Quaternary Science Reviews 66 (2013) 22e34

summer precipitation and evaporation rather than temperature as also shown in the modern isotope hydrology. A higher lake level, relative to the present, increases the S/V and hence evaporation, causing an increase in d18Olake. In contrast, a reduced lake level and area will cause a trend towards more negative d18Olake. Lake Spåime is hydrologically open and we therefore expect a large response of d18Olake and d18Odiatom to changes in seasonality of precipitation (winter/summer domination) as shown in the modern isotope hydrology. Melt water from snow fields in the Lake Spåime catchment potentially influences d18Odiatom in that it depletes the lake water during the summer. The significance of this influence is determined by winter snow abundance and meltseason temperatures. In contrast, the d18Olake data from Guossasjavri indicate that the direct influence of snow-melt in the summer on the d18Olake and d18Odiatom of Guossasjavri is restricted and unlikely to change the whole lake water composition. A reconstruction of cloud cover based on carbon ratios (d13C) in tree-rings from north-western Norway exhibits considerable variation over the past millennium (Young et al., 2010, 2012; Gagen et al., 2011). Because we expect both changes in cloud cover and lake isotopes to co-vary with specific shifts in atmospheric circulation, which are in turn related to AO and NAO patterns, a comparison between these two climate proxies provides a unique opportunity to detect responses to shifts in climate in both terrestrial and aquatic environments. The relatively positive d18Odiatom values in the Spåime record show that summer precipitation dominated over winter precipitation resulting in low winter/summer (W/S) precipitation ratio and that the influence of snowmelt on the d18Olake must have been small between 1000 and 1080 AD (Fig. 11). The most likely explanation for the positive d18Odiatom values in Guossasjavri over the same time period is high summer precipitation increasing the lake area, the S/ V ratio and, therefore evaporation. According to tree-ring data, summers were indeed warm and cloudy at this time (Young et al., 2012). High summer temperatures probably resulted in an early melt of winter snow in the Spåime catchment limiting the potential influence of isotopically depleted melt water on d18Odiatom over the summer. We suggest that summer precipitation had a southwesterly source at this time consistent with a high AO index. The Spåime d18Odiatom record shows that the relative importance of winter precipitation had increased (high W/S ratio) by 1080 AD (Fig. 11), a significant depletion is also seen in the Guossasjavri d18Odiatom record around this time. We suggest that a significant decrease in summer precipitation must have occurred, lowering the lake level and thereby reducing evaporation. The record minimum in Guossasjavri is reached at w1120 AD. The tree-ring data indicate that the summers were cold and sunny during this period (Young et al., 2012). Low summer temperatures would have allowed snow-melt to deplete d18Odiatom in Spåime and reduce evaporation further in Guossasjavri. We suggest that cool and dry Arctic air dominated at this time consistent with a low AO index. The Guossasjavri record indicates that the S/V ratio and evaporation increased from 1120 AD until 1360 AD and remained very high until 1440 AD. High summer precipitation must have increased the lake level, the S/V ratio, and evaporation significantly. The Spåime record also suggests more positive d18Odiatom over this time period, however, the magnitude of change is smaller. Tree-ring data show an increase in cloudiness over the same period (Young et al., 2012). Moist North Atlantic air probably dominated in this region, at least during the summer season. Wet air masses with relatively positive d18OP originating over the North Atlantic likely dominated in northern Scandinavia at this time (high AO). Both lake records show a shift towards lower d18Odiatom values after 1440 AD, which is at the transition between the Medieval Climate Anomaly (MCA) and the Little Ice Age (LIA) (e.g. Trouet

et al., 2012). The Spåime record shows that the influence of winter precipitation became increasingly more important until 1820 AD, either because of higher winter precipitation or lower summer precipitation, or a combination of both. The Guossasjavri record indicates that evaporation and the S/V ratio decreased, probably because of falling lake levels (until 1740 AD) indicating that summer precipitation and evaporation decreased. Tree-ring data suggest that summers were cold and sunny, implying dry summers (Young et al., 2012). Low melt-season temperatures probably increased the depleted d18Odiatom in Lake Spåime. We argue that cool, dry and d18O depleted air again dominated during summers at this time, consistent with many records that indicate that a transition to more negative NAO conditions and a low AO index occurred at this time (e.g. Trouet et al., 2012). Shorter term changes in the summer water balance of Guossasjavri most likely caused the two smaller peaks detected at w1590 AD (peak in Spåime) and at w1640 AD. The influence of snow-melt, perhaps due to both high amounts of winter snow and to low summer temperatures, kept the d18O in Spåime low (Gagen et al., 2011; Young et al., 2012). The isotope records show opposing shifts between 1750 and ca 1820 AD. The Guossasjavri d18Odiatom becomes more enriched (1&) by 1750 AD meanwhile the depletion continues in Lake Spåime until 1820 AD. As summers were relatively warm and sunny, implying dry conditions between 1750 and 1800 AD (Gagen et al., 2011; Young et al., 2012) we conclude that there must have been enough winter snow to deplete Spåime lake water over the summer, indicating that winters were snow rich consistent with a high NAO index. This increase in winter snow coincides with historical glacier advances in Norway which have been attributed to increased winter snow accumulation (Nesje et al., 2008) and with higher winter precipitation in the Alps (Vincent et al., 2005). An abrupt and prominent shift of 2& to more positive isotope values occurs in Lake Spåime between 1820 and 1830 AD indicating a shift towards a low W/S ratio. Tree-ring data show that summers were cloudy and that summer temperatures stayed low, so summer warming would have had less of an influence on snowmelt influencing d18Odiatom. We suggest that winters probably got dryer and summers wetter at this time implying a high AO index. An isotope depletion trend over the last 150 years has been detected by Jonsson et al. (2010b) in d18Odiatom records from Lake Spåime and Vuolep Allakasjaure (68 100 2500 , 18 100 3700 ); located 100 km northwest of Guossasjavri. An increase in the amount of winter precipitation is thought to have affected these two throughflow lakes in a similar way (Jonsson et al., 2010b). Here we identify this depletion also at Guossasjavri, confirming the regional extent of this shift towards increased contribution of winter precipitation, with a high NAO index around 1990’s. Several European famines occurred during the 14th, 15th and 19th centuries (e.g. Ruddiman, 2005). Tree-ring based temperature reconstructions show that some of these famines coincided with low, but not unusually low, growing season temperatures (Melvin et al., 2012). We infer increased summer wetness and a dominance of North Atlantic derived precipitation in northern Sweden for these periods, indicating that shifts in precipitation strongly influenced human society at these times. The recurrence of such changes would greatly affect future regional climate conditions in the North Atlantic region. 6. Conclusions Despite the hydrological differences and the w700 km distance between the lakes, the d18Odiatom records presented here reveal common responses to precipitation forcing over the past 1000 years. The relatively wet summers inferred from the d18Odiatom records

G.C. Rosqvist et al. / Quaternary Science Reviews 66 (2013) 22e34

between 1000 and 1080 AD, 1300 and 1440 AD, and during the early 19th century coincided with periods for which high cloud cover has been inferred from tree-ring carbon isotopes which is consistent with a high AO index. The d18Odiatom records indicate a relatively dry summer climate with increasing influence of winter snow between 1600 and 1750 AD. According to the tree-ring records this period was sunny and the coldest part of the LIA, both reconstructions are consistent with a low AO index at this time. The co-response between carbon isotopes in trees and oxygen isotopes in diatoms establish the relationship between cloud cover and precipitation and strengthens the hypothesis that these changes were the result of significant regional shifts in atmospheric circulation.

Acknowledgements This work was funded by the European Union Millennium project (017008). We thank Danny McCarroll, Sheila Hicks and Giles Young for helpful discussions and Arjen Stroeven for improving the manuscript.

References Abbott, M.B., Stafford, T.W., 1996. Radiocarbon geochemistry of modern and ancient Arctic lake systems, Baffin Island, Canada. Quaternary Research 45 (3), 300e311. Alexandersson, H., Eggertsson Karlström, C., 2001. Temperaturen och nederbörden i Sverige 1961e1990, Referensnormaler. Meteorologi 99, utgåva 2. Sveriges Meteorologiska och Hydrologiska Institut, Norrköping. Andersson, S., Rosqvist, G., Leng, M.J., Wastegard, S., Blaauw, M., 2010. Late Holocene climate change in central Sweden inferred from lacustrine stable isotope data. Journal of Quaternary Science 25 (8), 1305e1316. Appleby, P.G., 2008. Three decades of dating recent sediments by fallout radionuclides: a review. The Holocene 18 (1), 83e93. Battarbee, R.W., 1986. Diatom analysis. In: Berglund, B.E. (Ed.), Handbook of Holocene Palaeoecology and Palaeohydrology. John Wiley and Sons, Chichester, pp. 527e570. Bennett, K.D., 1996. Determination of the number of zones in a biostratigraphical sequence. New Phytologist 132, 155e170. Bigler, C., Hall, R.I., 2002. Diatoms as indicators of climatic and limnological change in Swedish Lapland: a 100-lake calibration set and its validation for paleoecological reconstructions. Journal of Paleolimnology 27 (1), 97e115. Birks, H.J.B., Gordon, A.D., 1985. The analysis of pollen stratigraphical data. Zonation. In: Birks, H.J.B., Gordon, A.D. (Eds.), Numerical Methods in Quaternary Pollen Analysis. Academic Press, London, pp. 47e90. Bowen, G.J., Wilkinson, B., 2002. Spatial distribution of d18O in meteoric precipitation. Geology 30 (4), 315e318. Carey, R.J., Houghton, B.F., Thordarson, T., 2010. Tephra dispersal and eruption dynamics of wet and dry phases of the 1875 eruption of Askja Volcano, Iceland. Bulletin of Volcanology 72 (3), 259e278. Clayton, R.N., Mayeda, T.K., Jan 1963. The use of bromine pentafluoride in the extraction of oxygen from oxides and silicates for isotopic analysis. Geochimica et Cosmochimica Acta 27, 43e52. Dansgaard, W., 1964. Stable isotopes in precipitation. Tellus 16, 436e468. D’Arrigo, R.D., Cook, E.R., Mann, M.E., Jacoby, G.C., 2003. Tree-ring reconstructions of temperature and sea-level pressure variability associated with the warm-season Arctic Oscillation since AD 1650. Geophysical Research Letters 30 (11), 1549. Davies, S.M., Elmquist, M., Bergman, J., Wohlfarth, B., Hammarlund, D., 2007. Cryptotephra sedimentation processes within two lacustrine sequences from west central Sweden. The Holocene 17, 319e330. Epstein, S., Mayeda, T., 1953. Variation of O18 content of waters from natural sources. Geochimica et Cosmochimica Acta 4, 213e224. Esper, J., Büntgen, U., Timonen, M., Frank, D.C., 2012. Variability and Extremes of Northern Scandinavian summer temperatures over the Past Two Millennia. Global and Planetary Change 88e89, 1e9. Folland, C.K., Knight, J., Linderholm, H.W., Fereday, D., Ineson, S., Hurrell, J.W., 2009. The summer North Atlantic Oscillation: past, present, and future. Journal of Climate 22, 1082e1103. Gagen, M., Zorita, E., McCarroll, D., Young, G.H.F., Grudd, H., Jalkanen, R., Loader, N.J., Robertson, I., Kirchhefer, A., 2011. Cloud response to summer temperatures in Fennoscandia over the last thousand years. Geophysical Research Letters 38. GNIP database, 2012. http://www-naweb.iaea.org/napc/ih/IHS_resources_gnip.html. Goslar, T., Czernik, J., Goslar, E., 2004. Low-energy 14C AMS in Poznan radiocarbon laboratory, Poland. Nuclear Instruments and Methods in Physics Research B 223-224, 5e11. Goslar, T., Van der Knaap, W.O., Kamenik, C., Van Leeuwen, J.F.N., 2009. Free-shape (14)C age-depth modelling of an intensively dated modern peat profile. Journal of Quaternary Science 24 (5), 481e499.

33

Gustavsson, H., Stensen, B., Wern, L., 2011. Regional klimatsammanställning d Norrbottens län. Swedish Meteorological and Hydrological Institute. SMHI Rapport Nr 2011-20. Hammarlund, D., Barnekow, L., Birks, H.J.B., Buckardt, B., Edwards, T.W.D., 2002. Holocene changes in atmospheric circulation recorded in the oxygen-isotope stratigraphy of lacustrine carbonates from northern Sweden. The Holocene 12 (3), 339e351. Hammarlund, D., Velle, G., Wolfe, B.B., Edwards, T.W.D., Barnekow, L., Bergman, J., Holmgren, S., Lamme, S., Snowball, I., Wohlfarth, B., Possnert, G., 2004. Palaeolimnological and sedimentary responses to Holocene forest retreat in the Scandes Mountains, west-central Sweden. The Holocene 14 (6), 862e876. Hua, Q., Barbetti, M., 2004. Review of tropospheric bomb C-14 data for carbon cycle modeling and age calibration purposes. Radiocarbon 46, 1273e1298. Hurrell, J.W., Kushnir, Y., Ottersen, G., Visbeck, M., 2003. An overview of the North Atlantic Oscillation. In: Hurrell, J.W., Kushnir, Y., Ottersen, G., Visbeck, M. (Eds.), The North Atlantic Oscillation. Climatic Significance and Environmental Impact. Geophysical Monograph, vol. 134. American Geophysical Union, Washington, DC, pp. 1e35. Jonsson, C.E., Leng, M.J., Rosqvist, G.C., Seibert, J., Arrowsmith, C., 2009. Stable oxygen and hydrogen isotopes in sub-Arctic lake waters from northern Sweden. Journal of Hydrology 376 (1e2), 143e151. Jonsson, C.E., Andersson, S., Rosqvist, G.C., Leng, M.J., 2010a. Reconstructing past atmospheric circulation changes using oxygen isotopes in lake sediments from Sweden. Climate of the Past 6 (1), 49e62. Jonsson, C.E., Rosqvist, G.C., Leng, M.J., Bigler, C., Bergman, J., Tillman, P.K., Sloane, H.J., 2010b. High-resolution diatom d18O records, from the last 150 years, reflecting changes in amount of winter precipitation in two sub-Arctic high-altitude lakes in the Swedish Scandes. Journal of Quaternary Science 25 (6), 918e930. Jutman, T., 1995. Runoff. In: Raab, B., Vedin, H. (Eds.), Climate Lakes and Rivers, National Atlas of Sweden. SNA Publishing, Stockholm, pp. 106e111. Karlsson, J., Byström, P., 2005. Littoral energy mobilization dominates energy supply for top consumers in subarctic lakes. Limnology and Oceanography 50, 538e 543. Krammer, K., Lange-Bertalot, H., 1986-1991. Bacillariophyceae. In: Süsswasserflora von Mitteleuropa, vol. 2. G. Fischer, Stuttgart, pp. 1e4. Leng, M.J., Barker, P.A., 2006. A review of the oxygen isotope composition of lacustrine diatom silica for palaeoclimate reconstruction. Earth-Science Reviews 75, 5e27. Leng, M.J., Marshall, J.D., 2004. Palaeoclimate interpretation of stable isotope data from lake sediment archives. Quaternary Science Reviews 23, 811e831. Leng, M.J., Sloane, H.J., 2008. Combined oxygen and silicon isotope analysis of biogenic silica. Journal of Quaternary Science 23 (4), 313e319. Lotter, A.F., Juggins, S., 1991. POLPROF, TRAN and ZONE: Programs for Plotting, Editing and Zoning Pollen and Diatom Data. In: Newsletter: INQUASubcommission for the Study of the Holocene Working Group on Datahandling, Methods 6, pp. 4e6. Melvin, T.M., Grudd, H., Briffa, K.R., 2012. Potential bias in ’updating’ tree-ring chronologies using regional curve standardisation: re-processing 1500 years of Torneträsk density and ring-width data. The Holocene. http://dx.doi.org/ 10.1177/0959683612460791. Morley, D.W., Leng, M.J., Mackay, A.W., Sloane, H.J., Rioual, P., Battarbee, R.W., 2004. Cleaning of lake sediment samples for diatom oxygen isotope analysis. Journal of Paleolimnology 31 (3), 391e401. Nesje, A., Bakke, J., Dahl, S.O., Lie, Ø., Matthews, J.A., 2008. Norwegian mountain glaciers in the past, present and future. Global and Planetary Change 60, 10e27. Offerberg, J., 1967. Beskrivning till berggrundskartbladen Kiruna NV, NO, SV, SO. Description of the Geological Maps Kiruna NV, NO, SV, SO. SGU Serie Af 1-4. Swedish Geological Survey, Stockholm. Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Ramsey, C.B., Buck, C.E., Burr, G.S., Edwards, R.L., Friedrich, M., Grootes, P.M., Guilderson, T.P., Hajdas, I., Heaton, T.J., Hogg, A.G., Hughen, K.A., Kaiser, K.F., Kromer, B., McCormac, F.G., Manning, S.W., Reimer, R.W., Richards, D.A., Southon, J.R., Talamo, S., Turney, C.S.M., van der Plicht, J., Weyhenmeye, C.E., 2009. IntCal09 and Marine09 radiocarbon age calibration curves, 0e50,000 years cal BP. Radiocarbon 51 (4), 1111e1150. Renberg, I., Hansson, H., 2008. The HTH sediment corer. Journal of Paleolimnology 40 (2), 655e659. Renberg, I., 1990. A procedure for preparing large sets of diatom slides from sediment cores. Journal of Paleolimnology 4 (1), 87e90. Rigor, I.G., Wallace, J.M., Colony, R.L., 2002. Response of sea ice to the Arctic Oscillation. Journal of Climate 15, 2648e2663. Rosqvist, G., Jonsson, C., Yam, R., Karlen, W., 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 (7e8), 851e 859. Rosqvist, G.C., Leng, M.J., Jonsson, C., 2007. North Atlantic region atmospheric circulation dynamics inferred from a late-Holocene lacustrine carbonate isotope record, northern Swedish Lapland. The Holocene 17 (7), 867e873. Ruddiman, W.F., 2005. Plows, Plagues and Petroleum: How Humans Took Control of Climate. Princeton University Press, Princeton, 224 pp. Shemesh, A., Rosqvist, G., Rietti-Shati, M., Rubensdotter, L., Bigler, C., Yam, R., Karlén, W., 2001. Holocene climatic changes in Swedish Lapland inferred from an oxygen-isotope record of lacustrine biogenic silica. The Holocene 11 (4), 447e454.

34

G.C. Rosqvist et al. / Quaternary Science Reviews 66 (2013) 22e34

St. Amour, N.A., Hammarlund, D., Edwards, T.W.D., Wolfe, B.B., 2010. New insights into Holocene atmospheric circulation dynamics in central Scandinavia inferred from oxygen-isotope records of lake-sediment cellulose. Boreas 39 (4), 770e782. Sutton, R.T., Hodson, D.L.R., 2005. Atlantic ocean forcing of North American and European summer climate. Science 309 (5731), 115e118. Thompson, D.W.J., Wallace, J.M., 1998. The Arctic Oscillation Signature in the wintertime geopotential height and temperature fields. Geophysical Research Letters 25, 1297e1300. Trouet, V., Scourse, J.D., Raible, C.C., 2012. North Atlantic storminess and Atlantic Meridional Overturning Circulation in the last Millennium: reconciling

contradictory proxy records of NAO variability. Global and Planetary Change 8485, 48e55. Vincent, C., Le Meur, E., Six, D., Funk, M., 2005. Solving the paradox of the end of the Little Ice Age in the Alps. Geophysical Research Letters 32 (9). Young, G.H.F., McCarroll, D., Loader, N.J., Kirchhefer, A.J., 2010. A 500-year record of summer near-ground solar radiation from tree-ring stable carbon isotopes. The Holocene 20 (3), 315e324. Young, G., McCarroll, D., Loader, N., Gagen, M., Kirchhefer, A., Demmler, J., 2012. Changes in atmospheric circulation and the Arctic Oscillation preserved within a millennial length reconstruction of summer cloud cover from northern Fennoscandia. Climate Dynamics 39, 495e507.