Late-Holocene temperature and precipitation changes in Vindelfjällen,mid-western Swedish Lapland,inferred from chironomid and geochemical data

Abstract

In this article, we present results from a palaeolimnological study from Lake Vuoksjávrátje in the mountain tundra region in the Vindelfjällen Mountains, northwest Sweden. We suggest that the influence of precipitation may be one of the factors causing discrepancies between chironomid-based late-Holocene July temperature (JulyT) reconstructions from Fennoscandia. We combine quantitative temperature reconstruction using chironomids for the last 5100 years with qualitative analysis of chironomid composition and geochemical analyses, such as x-ray fluorescence (XRF), total organic carbon (TOC) and C/N analysis. The studied sequence is dated by ²¹⁰Pb, ¹³⁷Cs and 11 ¹⁴C datings from terrestrial macrofossils. The aim of the study was to use chironomids to reconstruct late-Holocene summer temperature variation on a multi-centennial to centennial timescale and to use geochemical data to identify periods during which the changes in chironomid composition might have been forced by environmental variables other than temperature, such as within lake processes or precipitation. Based on ordination techniques, and a comparison between chironomid-inferred JulyTs and changes in minerogenic sedimentation with regional temperature and wetness records, it is concluded that the JulyT signal was modulated by precipitation. The proxies indicate that both JulyT and annual precipitation have influenced the chironomid communities in Lake Vuoksjávrátje, and that catchment-related processes caused by enhanced precipitation have overridden the summer temperature signal between 3000 and 2200 cal. yr BP, and between 1050 and 100 cal. yr BP.

Keywords

chironomids late Holocene northern Sweden precipitation temperature reconstruction x-ray fluorescence

Introduction

Several intervals of abrupt climate change have been identified during the late Holocene in northern Europe (e.g. Velle et al., 2005a; Wanner et al., 2008). Although the amplitudes of these climate excursions were small compared to glacial–interglacial cycles, they significantly affected ecosystems and human societies (Bakker et al., 2012; Büntgen et al., 2011; Larsson et al., 2012; Tallavaara and Seppä, 2012). Most climate reconstructions from Fennoscandia are consistent on the millennial timescale, with a warm period centred around c. 6000 cal. yr BP followed by a cooling; however, on the centennial timescale, results are less in agreement, especially for the late Holocene (Seppä et al., 2009; Velle et al., 2005a; Wanner et al., 2008). The discrepancies among records can at least partly be explained by dating uncertainties. It has also been suggested that low amplitude temperature variability influences the biological proxies to a lesser extent than other environmental variables (Velle et al., 2010), such as soil composition and nutrients. To improve the accuracy of the reconstructions, well-dated, high-resolution late-Holocene palaeoclimate records are needed, and there is also a need to improve the methods.

Palaeoclimate reconstructions from northern Sweden have mostly been conducted in the northern part of the province of Lapland (e.g. Barnekow, 1999; Bigler et al., 2002; Grudd et al., 2002; Heinrichs et al., 2005, 2006; Larocque and Bigler, 2004; Larocque and Hall, 2003; Rosén, 2005; Rosén et al., 2001, 2003). Also the province of Jämtland (Jamtli) has been subject to several palaeoclimate studies (e.g. Andersson et al., 2010; Bergman et al., 2005; Gunnarson and Linderholm, 2002; Hammarlund et al., 2004; Sundqvist et al., 2010). In contrast, few palaeoclimate studies have been conducted in the mountain tundra region between Lapland and Jamtli for example, in Vindelfjällen (e.g. Karlsson et al., 2007; Sundqvist et al., 2007). The nature reserve in Vindelfjällen Mountains is located just below the Arctic Circle, in mid-western Swedish Lapland (Figure 1a and b). An oceanic climate prevails in the western part of the mountain area, whereas the eastern part is relatively more continental (Staafjord, 2012), thus variations in the strength of the westerlies could be expected to cause changes in local climate gradients within the area. Despite its key geographical position and the presence of numerous lakes, the Vindelfjällen area has so far not been a target for palaeoclimatic lake sediment studies. Soil chemistry studies and finds of subfossil trees in the vicinity of Hemavan have shown that the treeline was once higher in the area (Earl-Goulet et al., 1998; Kullman, 2006; Mahaney et al., 1995). This is also suggested by a speleothem climate study covering the early–mid-Holocene (Sundqvist et al., 2007). Several Stone Age settlements have been discovered in the Vindelfjällen Mountains during the last decade, of which one was excavated and radiocarbon dated to 5290–5050 cal. yr BP (Andersson and Sandén, 2008; Andersson et al., 2003; Viberg et al., 2013). However, it is not known under which climate conditions prehistoric people lived in this area, since knowledge about local climate change during the last c. 5000 years is lacking.

Figure 1.

(a) Study site Lake Vuoksjávrátje (star) in Vindelfjällen Mountains, mid-western Swedish Lapland and (b) Lake Vuoksjávrátje with surrounding topography.

Lake sediments provide rich archives of past environmental conditions and often offer opportunities to reconstruct environmental change using several different independent proxies. Past variations in lake biota, such as diatoms or chironomids, can be compared with variations in allochtonous proxies, such as pollen and minerogenic sediments, to provide a more complete picture of climate impact on lake and catchment ecosystems (Velle et al., 2005b).

Chironomids have been shown to be particularly sensitive to summer temperatures and are often used as indicators of climate change (Brooks et al., 2012; Eggermont and Heiri, 2011; Walker et al., 1991). As with all biological proxies, there are several variables other than temperature that can influence the composition of chironomids over time. These include changes in pH, conductivity, oxygen concentration, nutrient status of the lake water, macrophyte communities and substratum character (Berg, 1997; Langdon et al., 2010; Lindegaard, 1992; Stewart et al., 2013; Velle et al., 2010). This has been the basis for reconstructions of hypolimnic oxygen (Quinlan et al., 1998), nutrient loading (Lotter et al., 1998), and lake depth (Engels and Cwynar, 2011; Korhola et al., 2000). Most of these environmental factors are also influenced, either directly or indirectly, by precipitation: directly through changes in the seasonality and/or amount of precipitation, and indirectly through changes in weathering and erosion of soils and vegetation in the catchment. Changes in environmental factors often associated with temperature, for example, nutrient status, may force changes in species composition that could mistakenly be interpreted as temperature change. At other times, confounding variables may weaken or enhance the reconstructed temperature signal (Brodersen and Anderson, 2002; Juggins, 2013; Velle et al., 2010). Comparisons among reconstructions from different sites in Fennoscandia have not yet provided a regional coherent summer temperature evolution during the Holocene (Velle et al., 2012). The inconsistencies are especially evident for late Holocene, suggesting strong confounding gradients compared to the gradient of interest.

Here, we present results from a palaeolimnological study of a lake located in the Vindelfjällen Mountains. We have applied quantitative temperature analysis using chironomids, in combination with qualitative analysis of chironomid composition and high-resolution geochemical analyses of well-dated lake sediment cores. Our aim was to use chironomids to reconstruct summer temperature variation on a multi-centennial to centennial timescale. Additionally, we used geochemical proxies to identify periods during which the chironomids were potentially forced by environmental variables other than temperature, such as precipitation.

Study site

Lake Vuoksjávrátje (66°14′56.50″N, 15°43′7.73″E) is situated at 850 m a.s.l., in a remote area 38 km NW of Ammarnäs, 32 km from the nearest road and 5 km from the nearest hiking trail (Figure 1). Human impact at the site is low, apart from sparse grazing by domesticated reindeer.

Lake Vuoksjávrátje (hereafter referred to as Vuoksjávrátje) is an 11.5 m deep headwater lake with a clearly defined threshold, a surface area of c. 0.059 km² and a catchment area of c. 1.08 km². The basin is fed by groundwater and seasonal meltwater streams from semi-permanent snow patches. The outlet water flows into the lower part of the brook Seärŋjiejuhka, a tributary to the Vindelälven River. Lake water pH was ~7.5 in July 2009.

The present mountain birch treeline in Vindelfjällen is located at c. 720 m a.s.l. The catchment vegetation is composed of wetland and low alpine heath at lower altitudes and high alpine heath at higher elevations. The catchment bedrock consists of Cambro-Silurian sedimentary rocks. Phyllite dominates (Gavelin and Kulling, 1955), and the occurrence of Dryas octopetala within the catchment indicates that limestone is present. The highest mountain peak in the catchment is Mount Ájvuotjåhkka, which reaches 1364 m a.s.l. The geomorphology of the area is characterized by glacial and periglacial landforms and well-developed solifluction lobes (Blomdin, 2009). Instrumental climate data registered at Ammarnäs meteorological station (440 m a.s.l.) during the reference period 1961–1990 demonstrate that climate in Vindelfjällen is characterized by cool summers (mean July temperature [JulyT] = 12.2°C) and cold winters (mean January temperature = −14.4°C) with an annual mean temperature of −0.9°C (Alexandersson and Eggertsson Karlström, 2001). Adjusted for altitudinal cooling using an environmental lapse rate (ELR) of −5.7°C/1000 m (Laaksonen, 1976), the 30-year reference mean JulyT at Lake Vuoksjávrátje was 9.9°C. The mean annual precipitation registered at the meteorological station was 566 mm/yr (Alexandersson and Eggertsson Karlström, 2001), and the mean number of days per year with snow cover >225 days with an average maximum yearly snow depth of c. 120 cm (SMHI, n.d.). Mean winter temperatures were 2.6°C higher between 1991 and 2000 compared to the reference period, whereas summers were 0.2°C warmer during the same period. Mean yearly precipitation was 10–15% higher than that of the reference period. Precipitation has increased during all seasons except during autumn, which was c. 10% less compared to the reference period (SMHI, n.d.).

Material and methods

Fieldwork and sample preparation

A plateau on the lake bottom at c. 9 m depth was chosen as a suitable coring location. A 22 cm surface sediment core (Vuok A) was retrieved using a HTH corer (Renberg and Hansson, 2008) and subsampled in field at 0.5 cm resolution in March 2009. Two parallel sediment cores (Vuok B, 2 m, and Vuok C, 3 m long) were retrieved using a 3 m modified Livingstone corer (Wright, 1967). All cores were transported to Stockholm University where they were stored at c. 4°C. The long core Vuok C was split into two halves, one of which was subsampled contiguously at 0.5 cm intervals. At each 10 cm interval, a 1 cm sample was saved for radiocarbon dating. The remaining half core was described based on visual and textural properties, run through an ITRAX x-ray fluorescence (XRF) core scanner (Cox Analytical Systems, Mölndal, Sweden) and kept as a reference. Water samples for stable isotope analyses were collected directly in air-tight polypropylene test tubes at the inlets and outlet of the lake in July 2009.

Dating and age modelling

The surface sediment core Vuok A (1-cm samples) was ²¹⁰Pb and ¹³⁷Cs dated at the Gamma Dating Center, Institute of Geography at the University of Copenhagen, Denmark. The age model was produced using a modified constant rate of supply (CRS) model (Appleby, 2002). A total of 14 samples with identified terrestrial macrofossils from core Vuok C were dated by accelerator mass spectrometry (AMS) radiocarbon dating at the Ångström Laboratory in Uppsala, Sweden. The two cores were combined (estimated error range ±2 cm) based on a common trough in the total organic carbon (TOC) profiles, as no macrofossils were found in the overlap. The radiocarbon ages were calibrated using the OxCal v. 4.1 software and the IntCal09 atmospheric calibration curve, and modelled against sediment depth using BACON (Blaauw and Christen, 2011; Bronk Ramsey, 2009; Reimer et al., 2009). All calibrated radiocarbon ages are given in years before present (ad 1950), abbreviated ‘cal. yr BP’. Samples from the lower part of the core were prepared for tephra analysis with the aim of consolidating the chronology with known tephras (Hekla-3, Hekla-S/Kebister and Hekla-4; Bergman et al., 2004).

Chironomid stratigraphy

Chironomids were analysed from 44 levels from the long core (Vuok C) and five samples from the surface core (Vuok A). The sampling interval was adjusted to the sedimentation rate to ensure a good coverage of the cored time interval. The sample resolution was approximately one sample every 105 years with each 0.5–2 cm sediment slice representing between 1.5 and 32.4 years (mean = 14.8 years). The preparation of chironomid head capsules (hcs) followed the procedure of Brooks et al. (2007). For each analysed sample, 1–9 cm³ (generally 2 cm³) of wet sediment was deflocculated in 10% KOH under gentle stirring on a head plate at 75°C for 15 min. The sediment volume (as determined by water displacement) was determined to find hc concentrations and enable calculations of hc accumulation rates (hc/yr). The sediments were sieved using 212-µm and 90-µm meshes, and hcs were picked using fine forceps under stereo binoculars and mounted individually, ventral side up, in Euparal or Hydromatrix. Mounted hcs were studied under light microscope using 250×, 500× and for some details 1000× magnification (using oil immersion). Identification and taxonomy followed Brooks et al. (2007), Wiederholm (1983) and for Tanypodinae, Rieradevall and Brooks (2001). Whole hcs and incomplete hcs with the whole mentum preserved were counted as one hc, and split hc, including at least half the mentum, were regarded as halves. Other chironomid fragments were disregarded. The goal was to identify >75 hc per sample to obtain sample taxon abundances and richness representative of the abundances and richness in the full sample (Velle and Larocque, 2008). However, the hc concentration in the sediment was highly variable and in some cases too low (0.2 hc/cm³ sediment) to reach the goal within feasible time, even after merging adjacent samples. Samples with <40 hc (three samples) were omitted from the statistical analyses, including the zonation and the temperature reconstruction (Heiri and Lotter, 2001; Larocque, 2001). Most authors recommend basic sums over 50; however, some studies have found that even lower sums can be sufficient for quantitative environmental analysis (Larocque-Tobler et al., 2009; Quinlan and Smol, 2001).

Stratigraphic diagrams with chironomid taxa that constitute ≥5% in any of the subsamples were constructed using the C2 v. 1.6.8 software (Juggins, 2007). To facilitate ecological interpretation and find significant shifts in the assemblages, the stratigraphy of Vuoksjávrátje was divided into zones. The zones were based on sum-of-squares optimal partitioning criteria (Birks and Gordon, 1985) using the program ZONE (Juggins, 1992). The number of significant zones was assessed by comparison with the broken-stick model (Bennett, 1996) using the program BSTICK (JM Line and HJB Birks, unpublished software). A division of subzones to aid description was based on qualitative interpretation of species composition. Hill’s N₂ diversity index (Hill, 1973) was calculated and plotted stratigraphically to visualize changes in diversity among the samples. Diagrams and statistical analyses were based on relative abundances of chironomid taxa.

Quantitative temperature reconstruction and ordinations

Detrended correspondence analysis (DCA), using Hill’s scaling and downweighting of rare species, was performed in CANOCO v 4.5 to estimate species turnover in the Vuoksjávrátje stratigraphy (Ter Braak and Šmilauer, 2002). Past July temperatures (JulyTs) were inferred by weighted averaging–partial least squares (WA-PLS) regression using square root transformed species data, in the software C2 v. 1.6.8 software (Juggins, 2007; ter Braak and Juggins, 1993). The calibration dataset consists of chironomid taxa and environmental data from 157 Norwegian and Svalbard lakes (Brooks and Birks, 2001, unpublished data). This training set was chosen based on the investigated site’s proximity and similarity to the training set lakes. The calibration dataset covers a JulyT gradient of 3.5–16°C. The best prediction model had two WA-PLS components, a root mean square error of prediction (RMSEP) of 1.27°C, a bootstrapped r² of 0.88, a bootstrapped average bias of 0.02 and a bootstrapped maximum bias of 1.20°C. The modern and the fossil datasets were harmonized prior to analyses. A locally weighted scatterplot smoothing (LOESS; span 0.3) was added to the temperature reconstruction to highlight the major climate trends. The inferred temperatures have not been corrected for glacio-isostatic rebound. A linear isostatic rise during the last 5000 years (Lindén et al., 2006), at a rate of 4 mm/year registered in historic times (Ekman, 1996), would imply a change in altitude of c. 20 m at Vuoksjávrátje. This is equivalent to 0.11°C.

The significance of the temperature reconstruction was tested by comparison with 999 random reconstructions (http://cran.r-project.org/web/packages/palaeoSig; R Developement Core Team, 2010; Telford and Birks, 2011). The default leave-one-out cross-validation method for the WA-PLS reconstructions was replaced by bootstrapping. The relationship between the Vuoksjávrátje chironomid stratigraphy, and the environmental parameters included in the calibration dataset was analysed in CANOCO v. 4.5 (ter Braak and Šmilauer, 2002). First, a canonical correspondence analysis (CCA) was performed on the calibration dataset species and environmental data to check for a possible arch-effect in order to determine whether a CCA or a detrended canonical correspondence analysis (DCCA) would be most appropriate. As no arch-effect was detected, CCA was applied on the species and environmental data in the calibration dataset. This CCA was done in an iterative procedure, in which environmental parameters with insignificant (p > 0.05) or high inflation factors (p > 20) were successively removed until only significant parameters remained, and the significance level was adjusted to 0.01 using Bonferroni correction for multiple testing (Miller, 1991). Finally, ordination plots were produced with the Vuoksjávrátje samples added passively, to illustrate a possible relationship between the fossil samples and the environmental parameters. A time trajectory was added to visualize the relationships over time. Samples with <10% fit with the CCA axes were considered to have poor fit. For further details on the method, see Velle et al. (2005a) and references therein.

Physical and chemical analyses

To determine the water balance of the lake, water samples were analysed for stable oxygen and hydrogen isotopes, using a laser absorption spectrometer (Liquid–Water Isotope Analyzer, Los Gatos Research Inc., Mountain View, CA, USA) with a precision of 0.15‰ for O and 0.06‰ for H, at the Department of Geological Sciences, Stockholm University. δ-values are reported in ‰ relative to Vienna standard mean ocean water (VSMOW). Whether the δ-values change or not between the inlet and the outlet provides information about the evaporative status of the basin, and/or the potential influence of melting snow (Jonsson et al., 2009).

Sediment geochemistry was analysed on the reference half core using high-resolution XRF elemental analysis in an ITRAX core scanner (Cox Analytical Systems) at the Department of Geological Sciences, Stockholm University, Sweden. We used an Mb tube for the XRF analysis, dwell time was set at 25 s and the measuring step size at 1 mm. In order to account for changes in sediment density, the counts per second (cps) for each element was normalized by dividing them with the cps for coherent and incoherent scatter (Kylander et al., 2011). Means of cps were calculated for every 5 mm in order to reduce data and to facilitate comparisons with the other analyses. Here, we use the Ti cps as a proxy for minerogenic silt in-wash (Cuven et al., 2010; Kylander et al., 2011).

TOC was measured on dried and homogenized (ground) subsamples in an ELTRA Carbon Sulphur Determinator at a contiguous sample resolution of 0.5–1 cm. Samples for total C and total N were freeze-dried and homogenized prior to analysis. C and N contents were analysed in a Finnegan DeltaV advantage mass spectrometer connected to a Carlo Erba NC2500 elemental analyser through a ConFloIV open split interface at the Department of Geological Sciences, Stockholm University, Sweden. When possible, similar subsampling interval was used for C/N ratio and for chironomid analysis to enable comparisons between the chironomid composition and the C/N ratio. However, for some samples, no material was available. In such cases, the adjacent samples were analysed. The sediment was suspected to contain some carbonates; thus, double measurements were carried out, one with untreated samples and one with HCl-treated samples.

Results

Dating and age–depth modelling

The unsupported ²¹⁰Pb declined exponentially downcore in the upper 6 cm of the surface sediment core Vuok A, and ¹³⁷Cs-values were high in the upper 3 cm and declined downcore (Figure 2a). The fluctuating Pb-values below 6 cm depth were interpreted as background noise. The Cs-values were clearly elevated in the uppermost two samples. These values were interpreted as representing mainly the ad 1986 Chernobyl accident. The expected Chernobyl accident peak and the ad 1963 atomic nuclear bomb testing peak could not be separated, probably due to the low sedimentation rate in relation to the subsampling resolution of 1 cm. The values were almost identical in the two uppermost samples, suggesting mixing of the two uppermost cm of unconsolidated sediment, possibly during coring. The declining Cs-values below 3 cm depth are maybe be the result of downward diffusion. According to the CRS model, the upper 6 cm of the core comprised the last 125 years (Figure 2b). This is equivalent to a mean sediment accumulation rate of 0.5 mm/yr. The Vuoksjávrátje age–depth model is based on Vuok A and 11 AMS ¹⁴C datings on terrestrial macrofossils (Figure 2c and Table 1) from Vuok C. Because of the sediment characteristics described below, the default settings in BACON were adjusted to allow for quickly changing sedimentation rates. BACON rejected three of the 14 ¹⁴C datings (Table 1; Ua-38711, Ua-40166, Ua-38719), as age reversals were not accepted. These outliers included terrestrial macrofossils found at one organic-rich layer at 21–22 cm, and an Empetrum fruit at 240.5–241 cm. The sediment accumulation rate varied between 0.2 and 7.7 mm/yr (mean = 1.2 mm/yr, median = 0.7 mm/yr), based on the best fit model. All proxies were plotted against the best fit model produced by BACON. JulyTs were also plotted against the full greyscale assemblage age model to visualize the probability intervals of the ages for this proxy. No tephras were found.

Figure 2.

(a) Dry sediment weight, ²¹⁰Pb and ¹³⁷Cs concentrations in surface core Vuok A (the Gamma Dating Center, Institute of Geography at the University of Copenhagen), (b) Vuok A age–depth model based on ²¹⁰Pb and ¹³⁷Cs dating (the Gamma Dating Center, Institute of Geography at the University of Copenhagen). Note that Figure 2a and b have different y-scales. The age model in Figure 2b is based on the values from the uppermost 6 cm in Figure 2a. (c) The age–depth model for Vuoksjávrátje was produced in BACON (Blaauw and Christen, 2011) based on the Vuok A age model, here in green, and 14 calibrated AMS radiocarbon dates on terrestrial macrofossils. The individual calibrated ¹⁴C dates, including 95.4% probability intervals are shown in blue. Calibrations were based on the IntCal09 atmospheric radiocarbon calibration curve (Reimer et al., 2009). Three of the ¹⁴C dates were automatically rejected as they would cause age reversals if included. To the right is a simplified lithostratigraphy for cores Vuok A and Vuok C, based on visual and textural properties.

Table 1.

AMS ¹⁴C datings on terrestrial macrofossils.

Sample depth (cm)	Laboratory name	Material analysed	DW (mg)	Reported age (¹⁴C yr BP)	Calibrated age, 95.4% (cal. yr BP)	Calibrated age, 99.7% (cal. yr BP)
21–22	Ua-38711^a	Salix leaf and unidentified leaf fragments	11	1988 ± 30	1997–877	2057–1825
21–22	Ua-40166^a	Fragments of leaves	1.4	2971 ± 96	3371–2882	3454–2783
42–43	Ua-40167	Leaves, unidentified	6.4	455 ± 32	540–469	553–331
57–58	Ua-38712	Salix leaf and unidentified leaf fragments	9.9	687 ± 30	685–561	695–552
97–98	Ua-38713	Fragments of leaves	4.8	1052 ± 30	1053–925	1063–910
110–111	Ua-38714	Fragments of leaves	2.8	1423 ± 36	1381–1288	1410–1270
135–135.5	Ua-38715	Twig, unidentified	30.4	2225 ± 44	2340–2145	2351–2060
169.5	Ua-38716	Twig, unidentified	60.3	2536 ± 30	2746–2493	2755–2460
188–188.5	Ua-38717	Piece of wood, unidentified	3.7	2711 ± 44	2920–2749	2956–2741
214–215	Ua-38718	Fragments of leaves and Betula catkin scale	3.1	3031 ± 50	3362–3078	3398–2995
240.5–241	Ua-38719^a	Fruit of Empetrum nigrum ssp. hermaphroditum	5.5	4447 ± 42	5287–4879	5299–4868
244–245	Ua-40168	Fragments of leaves	1.2	3850 ± 178	4821–3833	4972–3569
260–260.5	Ua-40169	Fragments of leaves	2	3648 ± 67	4217–3731	4249–3693
285	Ua-38720	Seeds and small fragments of leaves	1.5	4499 ± 36	5304–4985	5315–4966

Calibrations of individual dates were done in OxCal v 4.1 employing the IntCal09 atmospheric calibration curve (Bronk Ramsey, 2009; Reimer et al., 2009). Calibrated ages are given in cal yr BP.

Outliers, rejected by BACON in the age-depth modelling (Blaauw and Christen, 2011).

AMS: accelerator mass spectrometry; DW: dry weight.

Chironomid stratigraphy

The mean hc count per sample in core Vuok C was 128.5. Concentration of hcs varied between 0 and 154 hc/cm³ (mean = 54.9 hc/cm³) along the sediment sequence. Accumulation rates of hcs varied between 0.0 and 22.4 hc/yr (mean = 3.5 hc/yr), while the sediment accumulation rates varied between 0.2 and 6.7 mm/yr (mean = 0.9 mm/yr) for the same samples. Some taxonomic resolution in the fossil samples was lost in the process of harmonizing with the calibration dataset (see supplementary material for details on merging). After harmonization, 16 taxa in the fossil material did not have a match in the calibration dataset and were therefore not included in the reconstruction (see supplementary material). If they represented ≥5% of the total chironomid count, they were, however, included in the stratigraphic diagram (Figure 3). This is true for Constempellina–Thienemanniola (≤10%), Tanytarsini undifferentiated (≤8%), Micropsectra sp. ‘short blunt spur’ (SBS; ≤7%) and indeterminable chironomids (≤7.5%).

Figure 3.

Stratigraphy of Lake Vuoksjávrátje with relative abundances of chironomid types ≥5% in any of the analysed subsamples, JulyT inferences (°C), sample sum of chironomids present in the training set, Hill’s N₂ effective number of species (numbers), chironomid accumulation rate (hc/yr), sediment accumulation rate (mm/yr) and zonation, plotted against age. Ages are in cal. yr BP (ad 1950). Note that subsamples with basic sums to low for statistical statement have been included in the diagram, while only subsamples with counts ≥40 hc were used for the zonation and the JulyT reconstruction.

Three chironomid composition zones were statistically significant. Based on qualitative interpretation, zone 3 was divided into subzones a and b.

Zone 1 (5080–2930 cal. yr BP)

The chironomid assemblage (Figure 3) was characterized by Micropsectra insignilobus-type, Paratanytarsus, Tanytarsus lugens-type, Abiskomyia, Heterotrissocladius grimshawi-type, H. maeaeri-type and Protanypus. H. maeaeri-type was more abundant here than in any of the following zones. Thienemannimyia, Eukiefferiella/Tvetenia (mainly Eukiefferiella claripennis-type and Tvetenia bavarica-type) and Hydrobaenus conformis-type appeared in the upper half of the zone. Accumulation rate of hcs varied between 1.1 and 8.9 hc/yr, and sediment accumulation rate varied between 0.2 and 2.6 mm/yr. The concentration of hc varied between 15 and 108.5 hc/cm³ (mean = 75.7 hc/cm³). The N₂ diversity varied between 11.8 and 23.7 (mean N₂ = 18.5) effective number of taxa.

Zone 2 (2930–2620 cal. yr BP)

This zone include four samples and was characterized by Eukiefferiella (mainly E. claripennis-type) and T. bavarica-type, Gymnometriocnemus–Bryophaenocladius, Limnophyes–Paralimnophyes, Metriocnemus, unidentified Orthocladiinae, Procladius and a high proportion of Smittia folicea-type. The chironomid accumulation rate was 0.0–0.4 hc/yr (mean = 0.2 hc/yr), whereas the sedimentation rate was 0.3–1.2 mm/yr (mean = 0.9 mm/yr), resulting in low hc concentrations (0–3 hc/cm³, mean = 1.5 hc/cm³). Three of the samples from this zone were excluded from the JulyT reconstruction because the hc sums were <40. The types present in the excluded samples were similar to the taxa found in the included sample. The N₂ diversity was between 0.0 and 11.5.

Zone 3: subzone 3a (2620–1060 cal. yr BP)

The zone was characterized by Sergentia coracina-type, Corynocera oliveri-type, M. insignilobus-type, Abiskomyia and H. maeaeri-type. In the uppermost part, the abundance of H. maeaeri-type and H. conformis-type decreased, whereas Constempellina–Thienemanniola increased. The chironomid accumulation rate was 0.2–7.4 hc/yr (mean = 3.0 hc/yr), and the sedimentation rate was 0.2–1.1 mm/yr (mean = 0.5 mm/yr). The hc concentration was 3.5–154.0 hc/cm³ (mean = 68.3 hc/cm³). The N₂ diversity varied between 16.5 and 28.0 (mean N₂ = 22.3) effective number of taxa.

Zone 3: subzone 3b (1060 cal. yr BP–ad 2009)

This subzone was characterized by higher proportions of Chaetocladius, Eukiefferiella/T. bavarica-types, Limnophyes–Paralimnophyes, S. folicea-type, S. coracina-type and T. lugens-type, as well as a marked decrease in M. insignilobus-type. The abundance of Abiskomyia and Constempellina–Thienemanniola increased in the uppermost samples. The proportion of unidentified Orthocladiinae and indeterminable chironomids was high due to poor preservation. The chironomid accumulation rate in this zone was between 0.7 and 12.8 hc/yr (mean = 4.1 hc/yr), and the sedimentation rate was 0.5–6.7 mm/yr (mean = 4.1 mm/yr). The hc concentration varied between 9.9 and 135.3 hc/cm³ (mean = 34.9 hc/cm³). The N₂ diversity varied between 16.7 and 30.0 (mean N₂ = 22.6).

Quantitative temperature reconstruction and ordinations

The DCA indicated that the Vuoksjávrátje chironomid record had a total inertia of 1.077 standard deviation units of turnover. After taxonomic harmonization (see supplementary material for details) with the calibration dataset, the 45 samples with >40 hc in the Vuoksjávrátje record included 96 identified taxa of which in total 16 types did not have a match in the calibration dataset (see supplementary material). Inferred JulyTs varied between 8.5°C and 13.2°C in the last c. 5100 years with a mean of 9.8°C (Figure 4). The chironomid-based JulyTs for the meteorological reference period ad 1961–1990 was 9.4°C. According to the significance test, the temperature inference model based on the Vuoksjávrátje chironomid stratigraphy explains c. ¾ of the variance explained by the first CCA axis and is better (p ≤ 0.01) at reconstructing temperature than 999 random temperature reconstructions based on the same dataset (Figure 5).

Figure 4.

Lake Vuoksjávrátje chironomid JulyT reconstruction. Circles represent the WA-PLS temperature estimates. Empty circles represent samples with poor fit (<10%) to the axes of a CCA based on the calibration dataset and the significant environmental parameters JulyTs, depth and alkalinity. Black vertical bars show the sample-specific error estimates. The greyscale in the background show inferred JulyTs taking dating uncertainties into account (Blaauw and Christen, 2011). A LOESS (black undulating) with span 0.3 highlights major multi-centennial trends. The line is dotted during more unreliable intervals based on poor fit samples. The grey horizontal line marks the local reference normal (ad 1960–1991) for JulyTs.

Figure 5.

Proportion of variance in the Vuoksjávrátje chironomid stratigraphy explained by the Vuoksjávrátje JulyT reconstruction (black vertical line) compared to the proportion of variance explained by 999 transfer functions calibrated with random data (grey histogram). The dotted black line marks the variance in the fossil dataset explained by the first CCA axis of the calibration dataset.

In the calibration dataset, significant (p ≤ 0.01) environmental variables were JulyT, lake depth and alkalinity, according to a CCA corrected for multiple testing. CCA axis 1 had an interset correlation to JulyT of −0.87, and CCA axis 2 had an interset correlation of 0.66 with lake depth. The projection of passively added fossil samples into the modern calibration dataset ordination space (Figure 6a and b) shows that, although there is some spread of data along CCA axis 1 and the JulyT gradient, the main variability was along CCA axis 2 and lake depth gradient. If we disregard the samples with bad fit, the spread along the two axes are approximately equal (Figure 6b). Results from the trajectory (Figure 6a) suggest that the lake was relatively deep and cold c. 5100 years ago and throughout zone 1. Shallower and warmer conditions followed in zone 2. The subzone 3a samples indicate deeper and colder conditions again, followed by relatively warm and shallow conditions in subzone 3b.

Figure 6.

Fossil Lake Vuoksjávrátje samples passively projected onto the calibration dataset CCA ordination space. The samples are marked with different symbols depending on their biostratigraphic zone or subzone. Note that the difference between subzones 3a and 3b is not statistically significant. A time trajectory (three points running mean) show the trends over time. Straight dark grey arrows represent significant (p ≤ 0.01) environmental variables. (b) Fossil Lake Vuoksjávrátje samples passively projected onto the calibration dataset CCA ordination space. Samples with a poor fit (<10%) to the ordination axes are represented by red squares. Samples that can be better explained by the axes are represented by green diamonds.

Physical and chemical analyses

The lithostratigraphy of Vuok C is summarized in Figure 2. Ti fluctuated along the stratigraphy with an increasing trend towards the present (Figure 7). Ti cps were higher than average from c. 3200 to 2500 cal. yr BP, 1900 to 1700 cal. yr BP and from c. 1000 cal. yr BP with an increase towards the end of the record (c. ad 1980). The TOC varied between 0.05% and 5.0%. The lowest values, <0.1%, correspond to a layer of sand (c. 3050–2850 cal. yr BP). The TOC values were highly fluctuating with a decreasing trend upcore until c. 100 years ago, after which the highest values were registered. The C/N ratio (Figure 7) varied between 6.3 and 20.4 (mean = 10.0, median = 9.7).

Figure 7.

Summary of the proxies in the Vuoksjávrátje record: sedimentation rate (mm/yr), Ti (XRF cps, mean = 5 mm) with mean cps as reference value, TOC (%) with 1.5% as reference value, C/N ratio, reconstructed JulyT (chironomids, °C) and CCA scores for fossil chironomid samples on axis 1 and 2 (based on the modern calibration data). Light-grey background indicates periods with increased minerogenic in-wash.

Lake water δ¹⁸O was −13.68‰ ± 0.15‰ at the inlet and −13.06‰ ± 0.15‰ at the outlet. The negligible difference of 0.62‰ ± 0.15‰ in δ¹⁸O between water from the inlet (inV) and water from the outlet (outV) suggests that there was little lake water evaporation during the summer of 2009 and that Lake Vuoksjávrátje is a flow-through lake (Jonsson et al., 2009).

Discussion

Sedimentation

Lake Vuoksjávrátje has a relatively short water residence time, as shown by the small changes in lake water δ¹⁸O between the inlet and the outlet. This could make the lake sensitive to changes in amount and frequency of precipitation. We argue that variation in Ti in the sediment provides a high-resolution record of fluctuations in in-wash of minerogenic matter from the catchment (Figure 7). Sequences with low TOC values primarily represent periods with high minerogenic dilution rather than a lowering of autochthonous lake productivity (Figure 7). From the median C/N ratio of 9.7, we infer a mixed source for the organic material in the sediments with a relatively high proportion of autochthonous carbon. In general, low C/N ratios suggest a high proportion of nitrogen from proteins left in the sediment. This is usually the case with autochthonous organic matter, such as algae, whereas higher C/N ratios suggest the analysed material has been transported into the lake (Meyers and Lallier-Vergés, 1999). The much higher C/N ratios of 20.4 and 15 found at 57 cm (c. 660 cal. yr BP) and at 95–96 cm (c. 1050 cal. yr BP), respectively (Figure 7), could have been caused by terrestrial plant material transported to the centre of the lake basin as the vegetated shores were inundated during heavy rainfalls. An alternative explanation is that there was lower algal production during times with high C/N ratios. However, it could be expected that the higher temperatures at these times caused an increase in algal growth, which would have lowered the C/N ratio.

The terrestrial macrofossils that provided the three apparently too old radiocarbon dates (Ua-38711, Ua-40166 at 21, Ua-38719) were likely washed into the lake after storage in the catchment (Table 1). The age–depth model is based on the remaining 11 radiocarbon dates and the ²¹⁰Pb and ¹³⁷Cs datings. In terms of temporal resolution, the age–depth model allows for an interpretation of changes in sedimentation rates and good control of the timing of past environmental changes.

The coarse sediment in the interval between c. 3000 and 2600 cal. yr BP (Figure 2c) suggests that it was accumulated in turbulent water or could be the result of mass wasting and sediment refocusing within the basin. Both alternatives would result in a fast sedimentation rate. However, the sedimentation rate during this time is similar to the mean for the whole core, c. 1 mm/yr. Together, this suggests that we may have a hiatus in the record, followed by a period with high sedimentation rate. The possible hiatus would be confined between the radiocarbon dates Ua-38717 (2956–2741 cal. yr BP; 99.7%) and Ua-38718 (3398–2995 cal. yr BP; 99.7%) to 0–650 years.

Quantitative analyses

The Vuoksjávrátje JulyT reconstruction is statistically significant from random (p = 0.01; Figure 5) and reconstructed temperature for the ad 1960–1991 climate normal is similar to instrumental data (−0.5°C difference). This suggests that the variation in the chironomid record from Vuoksjávrátje at least partly is forced by summer temperature. The amplitude of temperature variability in the Vuoksjávrátje reconstruction is, however, as much as 3.1°C, and even higher (4.7°C) when the 15 samples with poor fit to temperature are included. This amplitude is notably higher compared to the amplitude shown by earlier chironomid-based reconstructions (c. 1.8 °C) from northern Sweden for the same period (Heinrichs et al., 2006; Larocque and Hall, 2004; Rosén et al., 2001).

The projection of the fossil Vuoksjávrátje samples on the training set CCA (Figure 6a) suggests that the Vuoksjávrátje chironomids were mainly influenced by depth and by JulyT. The samples that cause the stronger depth gradient in the CCA also have a poor fit with the CCA (Figure 6a and b). Therefore, their position in the CCA should not be used to infer changes in depth, temperature, or other environmental variables. The depth and temperature gradients are similar if these samples are disregarded.

The chironomid composition in 11 of the 16 samples in biostratigraphic zone 2 and subzone 3b (2930–2620 cal. yr BP and 1060 cal. yr BP to ad 2009) had a poor fit with the CCA axes (Figure 6a and b), including axis 1, which was strongest correlated to JulyT (interset correlation = −0.87), and axis 2, which was strongest correlated to lake depth (interset correlation = 0.66). The results of the ordination with fossil samples added passively suggest that the taxa in the abovementioned samples responded to some other environmental variables in addition to JulyT and depth, a response that seems to change over time.

These samples include high abundances of the rheophilic chironomid types Limnophyes–Paralimnophyes, Eukiefferiella and T. bavarica-type and terrestrial/shallow littoral types like Gymnometriocnemus–Bryophaenocladius, Metriocnemus and S. folicea-type (Brooks et al., 2007; Cranston et al., 1983). Self et al. (2011) found a positive correlation between Limnophyes and mean annual precipitation. The high abundance of rheophilic and terrestrial/shallow littoral chironomids suggests the lake was influenced by stream flow and catchment surface erosion. These stream flow and erosion indicating taxa also have higher JulyT optima in the calibration dataset than the majority of the fossil taxa recorded at Vuoksjávrátje. Thus, the abundance of these taxa possibly enhanced the warm temperatures in parts of the sediment record.

Mean sedimentation rates were higher between 2900 and 2100 cal. yr BP (1.1 mm/yr), and between 1050 cal. yr BP and ad 2009 (2.0 mm/yr) compared to the periods between 2100 and 1050 (0.9 mm/yr), and between 5100 and 3000 cal. yr BP (0.6 mm/yr). The higher sedimentation rates concur with relatively high Ti peak areas and low TOC, changes in the chironomid composition characterized by higher relative abundance of taxa indicating stream flow and catchment erosion, lower mean hc concentrations and poor fit between chironomids and the CCA axes (including temperature). Together these proxies suggest that the temperature signal from the Vuoksjávrátje record partly is biased by an in-wash signal caused by increased precipitation.

Comparison with other palaeoclimate records

We found that the periods between 3200 and 2500, 1900 and 1700, 1100 and 400 cal. yr BP, that are characterized by high Ti and low TOC values indicative of high (summer) precipitation and/or snow melt, correspond to periods with low δ¹⁸O in a speleothem (SG 93) from Soylegrotta (66°N13°E), Rana, Norway (Lauritzen and Lundberg, 1999). Low δ¹⁸O at this cave site is thought to represent warm and wet climate conditions (Linge et al., 2009; Sundqvist et al., 2010). The similarity between the Vuoksjávrátje Ti and SG 93 δ¹⁸O suggests a common response to changes in precipitation patterns, most likely due to shifts in the atmospheric circulation.

According to the Vuoksjávrátje chironomid-based reconstruction, JulyTs decreased from c. 9.7 to 8.6°C between 5100 and 4100 cal. yr BP. This cool period lasted until c. 3600 cal. yr BP. Results from several other studies have indicated that a pronounced summer cooling started c. 4000 years ago and lasted about one thousand years (e.g. Antonsson et al., 2006; Jessen et al., 2005; Velle et al., 2005a). The cooling caused the treeline to descend in northern Sweden (Barnekow, 1999; Bergman et al., 2005) and, in combination with increased winter precipitation, is thought to have led to Neoglacial activity in Norway and Sweden (Dahl and Nesje, 1996; Davis et al., 2009). The climate from c. 4000 was also generally moister and more variable than previously (Hammarlund et al., 2004; Jonsson et al., 2010; Kullman, 1995; Rosqvist et al., 2004; Väliranta et al., 2007). At c. 3600 cal. yr BP, the JulyTs started to rise towards modern levels (c. 10°C).

Few chironomids (0.0–0.4 hc/yr) were preserved in the sediment between c. 3000 and 2600 cal. yr BP, but relatively many hc fragments occurred. The coarse sediment in this interval suggests high kinetic energy of the water, which may have led to hc fragmentation and out-wash. Up to c. 10% abundance of M. insignilobus and a smaller abundance of S. coracina indicate that the chironomids from this sediment sequence are at least partly of profundal origin; however, the main part of the chironomid assemblage between c. 3000 and 2600 cal. yr BP belongs to rheophilic and terrestrial/semi-terrestrial taxa. Relative wetness documented from humification changes in a peat bog in Jamtli, Sweden, and from methane efficiency in peat bogs in southern Norway suggests increased surface wetness between c. 3300 and 2700 cal. yr BP (Andersson et al., 2010; Nichols et al., 2009).

Temperatures >1°C higher than the modern reference value of 9.9°C were reconstructed for the period between c. 2750 and 2200 cal. yr BP. Most of the samples originating from this sediment sequence have a poor fit to temperature (Figures 4 and 6a and b). Interestingly, increased JulyT at this time was also reconstructed from chironomids in Lake Sjuuodjijaure (67°22′, 18°04′), northern Swedish Lapland and Lake Södra Gilltjärn (60°05′, 15°50′), central Sweden, although with amplitudes of c. 1°C (Antonsson et al., 2006; Rosén et al., 2001). If the associated age–depth models are correct, this similarity suggests that the temperatures were indeed higher during this period. However, the poor fit of the Vuoksjávrátje samples implies that the actual temperature inferences should be treated with caution. As mentioned above, Ti values and sedimentation rates were higher and TOC was lower at these stratigraphic levels, suggesting that the catchment erosion changed simultaneously with temperature. Also, the chironomid assemblage at this time suggests that the chironomid JulyT reconstruction was likely influenced by increased surface erosion.

From the high Ti values recorded between 1900 and 1700 cal. yr BP, we infer increased wetness. The Soylegrotta Speleothem record also indicates a relatively wetter climate at this time. At c. 1000 cal. yr BP, the chironomids infer a drastic shift towards higher temperatures. All but five samples representing the period from 1060 cal. yr BP to the present had a poor fit to temperature (Figure 4). Still, the reconstruction provides relatively higher temperatures at the time of the ‘Medieval Climate Anomaly’ around ad 1000 (950 cal. yr BP; Graham et al., 2011; Grudd, 2008). Other chironomid-based JulyT reconstructions from alpine lakes in northern Sweden also show raised temperatures at this time, including Lake 850 (68°15′, 19°07′), Lake Sjuuodjijaure and Lake Spåime (63°07′, 12°19′; Hammarlund et al., 2004; Larocque and Hall, 2004; Rosén et al., 2001). Higher JulyT was also registered in a pollen-based JulyT reconstruction from Lake Svanåvatnet (66°44′, 14°05′), Rana, Norway (Bjune and Birks, 2008). The high JulyT inferred at Vuoksjávrátje at c. 500 and 250 cal. yr BP were also reconstructed at Lake 850 and Lake Svanåvatnet. Keeping in mind that among-site similarities are only based on a few samples and uncertainties associated with comparisons among records with radiocarbon-based age models (Bennett, 2002), the match between Vuoksjávrátje JulyT and the other JulyT reconstructions mentioned above may suggest that there is indeed a temperature response at this time in Vuoksjávrátje. The temperatures are, however, unexpectedly high. We should also note that similarities among the temperature records could be the result of chance. Based on our record only, the combined evidence from the ordination, the taxonomic information, and from the TOC, C/N and Ti variability suggest that Vuoksjávrátje was more sensitive to changes in catchment surface erosion and less sensitive to JulyT.

Based on tree-ring carbon isotope data from northern Sweden and Norway, Young et al. (2011) inferred cloudy summers between 1000 and 830 cal. yr BP and between 550 and 400 cal. yr BP, and linked the cloudiness to shifts in regional atmospheric circulation. Rosqvist et al. (2013) concluded based on higher δ¹⁸O registered in diatom silica from Lake Spåime and Lake Stuor Guossasjavri (67°50′, 19°40′), that these shifts in atmospheric circulation caused wet summers in both Lapland and Jamtli in northern Sweden. We note a higher but less reliable inferred JulyT and raised Ti values in the Vuoksjávrátje stratigraphy during these times. This match among records emanating from different types of climate proxies, such as carbon isotopes in tree rings, oxygen isotopes in lake sediments and minerogenic in-wash in lake sediments, suggests a common response to regional changes in climate. The Vuoksjávrátje Ti record indicates that the intensity of in-wash has varied over periods of decades or centuries, strengthening the interpretation of variations in main precipitation patterns. The variations in precipitation patterns were likely caused by changes in atmospheric circulation, with shifts between a dominance of wet/warm Atlantic air masses and dry/cold Arctic air masses (Loader et al., 2013).

Chironomids, temperature and precipitation

There is no doubt that many chironomid taxa are influenced by summer temperature. Results from many studies show a good statistical relationship between species composition and July air temperature in calibration datasets (e.g. Brooks and Birks, 2000). Many of the environmental variables that influence chironomids, such as substratum, water pH, conductivity, oxygen saturation and nutrient availability are influenced, either directly or indirectly, by temperature. Unfortunately, it is not easy to identify past changes in confounding factors and their relative importance for the species composition. Changes in the amount, intensity and seasonality of precipitation influence stream flow and water temperature. Indirectly, changes in precipitation can change weathering and erosion of soils and vegetation in the catchment, which in turn influence the terrestrial sediment contribution in the lake. Changes in precipitation and erosion characteristics, such as water chemistry and sediment grain size may force changes in species composition that could mistakenly be interpreted as a temperature response. At times, they may weaken or enhance the signal. Experimental studies on artificial data have shown that introducing a secondary environmental gradient to a dataset may create peaks and troughs in a reconstruction even when there was no change in the environmental variable being reconstructed (Juggins, 2013).

The influence of other environmental parameters than temperature on chironomid composition, such as habitat, can provide valuable qualitative information about, for example, changes in bottom substrate or hypolimnic oxygen. We use qualitative information together with physical and geochemical sedimentary data to infer periods with increased surface wetness and stream flow.

We suggest that changes in JulyT and precipitation, both forced by regional shifts in atmospheric circulation, influenced the chironomids in Vuoksjávrátje during the last 5100 years. The data suggest that precipitation caused an enhancement of the temperature signal between 3000 and 2100 cal. yr BP and between 1050 and c. 100 cal. yr BP, by favouring rheophilic taxa that have relatively high temperature optima. Sediment accumulation has caused a 3 m decrease in lake depth over the past 5100 years at the coring location. This infilling has led to a gradual shallowing of the lake, a general trend that can be discerned in the time trajectory in the CCA (Figure 6a). The samples that indicate shallower conditions in the CCA are mainly the same as the samples with higher relative abundances of rheophilic taxa. This running water influence indicate that an increased precipitation caused the faster sedimentation rates, rather than a lowering of lake level, due to changes in the lake water balance (decrease in precipitation and/or increase in evaporation). The lake basin configuration is simple and restricted by the surrounding topography. Therefore, we find it unlikely that changes in lake depth and lake bathymetry would have had any significant influence on the chironomid composition in the sediments. The samples from zone 2 and subzone 3b contain a shallowing signal that most likely was caused by in-washed allochtonous chironomids that originated from more surficial habitats, such as puddles and wetlands surrounding the lake. This is supported by the simultaneous increased erosion and in-wash inferred from the Ti values. As the chironomids that indicate shallower conditions do not originate from the lake itself, they should not be used to infer changes in lake depth but rather changes in catchment erosion.

We propose that shifts in precipitation likely have influenced the chironomid assemblages at Vuoksjávrátje. This was also suggested for the early Holocene in Lake Spåime in mid-Sweden, where a high percentage of Chaetocladius piger-type indicated increased in-wash caused by increased precipitation (Hammarlund et al., 2004). Because precipitation is locally more variable compared to temperature (Trenberth et al., 2007), due to, for example, topographical effects, among-site inconsistencies in inferred JulyT are to be expected at sites and times when precipitation influences the chironomid records. Previously inconsistencies among chironomid-based Holocene temperature reconstructions in Fennoscandia have been demonstrated by Velle et al. (2010, 2012). Future chironomid-based climate studies may benefit from increased emphasis on lake catchment erosion as it may influence temperature inferences. A possible way forward to more reliable temperature reconstructions could be to limit these to chironomid taxa with well-known temperature dependencies and exclude those who are strongly affected by other parameters. Different statistical approaches to reduce the effects of secondary environmental gradients in reconstructions have been tested for diatoms and for chironomids (Hausmann and Kienast, 2006; Racca et al., 2003; Velle et al., 2011). Such studies improve the knowledge of the limitations and opportunities of using training sets for environmental reconstructions.

Conclusion

The similarity in timing between events identified in the Lake Vuoksjávrátje sedimentary record and those in other regional palaeoclimate records indicates a chironomid response in Lake Vuoksjávrátje to regional climate change. This response was likely not strictly a response to temperature, but a combination of responses to changes in dominant atmospheric circulation related to the Arctic Oscillation, causing shifts between wet/warm and dry/cool air masses.

Based on a general correlation between regional wetness records and samples with poor fit to temperature in our JulyT reconstruction, and also with changes in minerogenic sedimentation at Vuoksjávrátje, we conclude that catchment-related processes caused by enhanced precipitation at times have overridden the summer temperature signal. Such periods occurred from between 3000 and 2200 cal. yr BP, and between c. 1050 and c. 100 cal. yr BP. Thus, the temperature signal in the Vuoksjávrátje chironomid record was partly modulated by precipitation.

Precipitation is spatially more variable than temperature. In addition, the influence of precipitation is likely to be catchment- specific, causing discrepancies among chironomid-based Holocene JulyT reconstructions. The influence of precipitation does not exclude the influence from other confounding factors, such as dating uncertainties, changes in continentality and internal lake processes.

Careful dating and high-resolution TOC and XRF analyses allowed us to identify the variable sedimentation rate at Vuoksjávrátje. These findings suggest that high-resolution sedimentary studies and thorough dating should always accompany biological proxy–based palaeoclimate studies. This study also shows the potential of using qualitative chironomid data to infer past changes in surface wetness and stream flow.

Footnotes

Acknowledgements

We are grateful to the following persons: Andrew Mercer and Ulf Eskilsson, who assisted with the sediment coring; Robin Blomdin and Krister N Jansson, who assisted with the summer field mapping and water sampling; Steve Brooks and Lars Eriksson who helped with identification of critical subfossil chironomids; Malin Kylander and Ludvig Löwemark, who were responsible for the XRF core scanning; Carolin Georges, who did part of the TOC measurements; Helen Dahlke, who assisted with the water isotope measurements; and Ewa Lind and Stefan Wastegård who analysed samples for tephra. Age modelling of the surface core was performed by Thorbjørn J Andersen at the Gamma Dating Center, University of Copenhagen. Special thanks to Steve Brooks and H John B Birks for kindly letting us use their Norwegian and Svalbard chironomid calibration dataset. We would also like to thank an anonymous reviewer for greatly improving the manuscript.

Funding

This work was supported by the Swedish Polar Research Secretariat (IPY/Arctic Sweden), Andréefonden (Svenska Sällskapet för Antropologi och Geografi), Göran Gustafssons stiftelse för forskning i Lappland, Mannerfeldts Fond för geografisk forskning and De Geers Stiftelse för kvartärgeologisk forskning.

References

Alexandersson

Eggertsson Karlström

(2001) Temperaturen och nederbörden i Sverige 1961–1990 (Referensnormaler, utgåva 2). Norrköping: Sveriges Meteorologiska och Hydrologiska Institut, p. 71.

Andersson

Sandén

(2008) Arkeologisk undersökning av RAÄ Sorsele 1372:1, vid Luspasjaure, Sorsele kn, Västerbottens län. Rapport Västerbottens Museum/Uppdragsverksamheten. Umeå: Västerbottens Museum.

Andersson

Heinerud

Robertsson

(2003) 2001 års fornminnesinventering i Västerbottens län. Stockholm: Riksantikvarieämbetet & Västerbottens Museum.

Andersson

Rosqvist

Leng

. (2010) Late Holocene climate change in central Sweden inferred from lacustrine stable isotope data. Journal of Quaternary Science 25: 1305–1316.

Antonsson

Brooks

Seppa

. (2006) Quantitative palaeotemperature records inferred from fossil pollen and chironomid assemblages from Lake Gilltjarnen, northern central Sweden. Journal of Quaternary Science 21: 831–841.

Appleby

(2002) Chronostratigraphic techniques in recent sediments. In: Last

Smol

(eds) Tracking Environmental Change Using Lake Sediments. Dordrecht: Springer, pp. 171–203.

Bakker

Kaniewski

Verstraeten

. (2012) Numerically derived evidence for late-Holocene climate change and its impact on human presence in the southwest Taurus Mountains, Turkey. The Holocene 22: 425–438.

Barnekow

(1999) Holocene tree-line dynamics and inferred climatic changes in the Abisko area, northern Sweden, based on macrofossil and pollen records. The Holocene 9: 253–265.

Bennett

(1996) Determination of the number of zones in a biostratigraphical sequence. New Phytologist 132: 155–170.

10.

Bennett

(2002) Comment: The Greenland 8200 cal. yr BP event detected in loss-on-ignition profiles in Norwegian lacustrine sequences – Correspondence. Journal of Quaternary Science 17: 97–99.

11.

Berg

(1997) Larval food and feeding behaviour. In: Armitage

Cranston

Pinder

LCV

(eds) The Chironomidae: Biology and Ecology of Non-Biting Midges. London: Chapman & Hall, pp. 194–224.

12.

Bergman

Hammarlund

Hannon

. (2005) Deglacial vegetation succession and Holocene tree-line dynamics in the Scandes Mountains, west-central Sweden: Stratigraphic data compared to megafossil evidence. Review of Palaeobotany and Palynology 134: 129–151.

13.

Bergman

Wastegård

Hammarlund

. (2004) Holocene tephra horizons at Klocka Bog, west-central Sweden: Aspects of reproducibility in subarctic peat deposits. Journal of Quaternary Science 19: 241–249.

14.

Bigler

Larocque

Peglar

. (2002) Quantitative multiproxy assessment of long-term patterns of Holocene environmental changes from a small lake near Abisko, northern Sweden. The Holocene 12: 481–496.

15.

Birks

HJB

Gordon

(1985) Numerical Methods in Quaternary Pollen Analysis. London: Academic Press, 317 pp.

16.

Bjune

Birks

HJB

(2008) Holocene vegetation dynamics and inferred climate changes at Svanåvatnet, Mo i Rana, northern Norway. Boreas 37: 146–156.

17.

Blaauw

Christen

(2011) Flexible paleoclimate age-depth models using an autoregressive gamma process. Bayesian Analysis 6: 457–474.

18.

Blomdin

(2009) Deglaciationsförloppet och isdämda sjöar i Vindelälvens källområde, norr om Ammarnäs. Examensarbete på Kandidatnivå inom Biogeovetenskap, Institutionen för Naturgeografi och Kvartärgeologi, Stockholms universitet, Stockholm, 51 pp.

19.

Brodersen

Anderson

(2002) Distribution of chironomids (Diptera) in low arctic West Greenland lakes: Trophic conditions, temperature and environmental reconstruction. Freshwater Biology 47: 1137–1157.

20.

Bronk Ramsey

(2009) Bayesian analysis of radiocarbon dates. Radiocarbon 51: 337–360.

21.

Brooks

Birks

HJB

(2000) Chironomid-inferred late-glacial and early-Holocene mean July air temperatures for Krakenes Lake, western Norway. Journal of Paleolimnology 23: 77–89.

22.

Brooks

Birks

HJB

(2001) Chironomid-inferred air temperatures from Lateglacial and Holocene sites in north-west Europe: Progress and problems. Quaternary Science Reviews 20: 1723–1741.

23.

Brooks

Langdon

Heiri

(2007) The Identification and Use of Palaearctic Chironomidae Larvae in Palaeoecology (QRA Technical Guide No. 10). London: Quaternary Research Association, 276 pp.

24.

Brooks

Axford

Heiri

. (2012) Chironomids can be reliable proxies for Holocene temperatures: A comment on Velle et al. (2010). The Holocene 22: 1495–1500.

25.

Büntgen

Tegel

Nicolussi

. (2011) 2500 years of European climate variability and human susceptibility. Science 331: 578–582.

26.

Cranston

Oliver

Saether

(1983) The larvae of Ortocladiinae (Diptera, Chironomidae) of the Holarctic region – Keys and diagnoses. In: Wiederholm

(ed) Chironomidae of the Holartic region – Keys and diagnoses. Part 1. Larvae. Entomologica Scandinavica, Supplement 19: 149–291.

27.

Cuven

Francus

Lamoureux

(2010) Estimation of grain size variability with micro X-ray fluorescence in laminated lacustrine sediments, Cape Bounty, Canadian High Arctic. Journal of Paleolimnology 44: 803–817.

28.

Dahl

Nesje

(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: 381–398.

29.

Davis

Menounos

Osborn

(2009) Holocene and latest Pleistocene alpine glacier fluctuations: A global perspective. Quaternary Science Reviews 28: 2021–2033.

30.

Earl-Goulet

Mahaney

Sanmugadas

. (1998) Middle-Holocene timberline fluctuation: Influence on the genesis of Podzols (Spodosols), Norra Storfjället Massif, northern Sweden. The Holocene 8: 705–718.

31.

Eggermont

Heiri

(2011) The chironomid-temperature relationship: Expression in nature and palaeoenvironmental implications. Biological Reviews 87: 430–456.

32.

Ekman

(1996) A consistent map of the postglacial uplift of Fennoscandia. Terra Nova 8: 158–165.

33.

Engels

Cwynar

(2011) Changes in fossil chironomid remains along a depth gradient: Evidence for common faunal thresholds within lakes. Hydrobiologia 665: 15–38.

34.

Fjällkartan 1:100 000 ^©Lantmäteriet [i2012/899].

35.

Gavelin

Kulling

(1955) Berggrundskarta över Västerbottens län. Pre-Quaternary rocks of the Västerbotten county, 1:400000. Stockholm: Sveriges Geologiska Undersökning.

36.

Graham

Ammann

Fleitmann

. (2011) Support for global climate reorganization during the ‘Medieval Climate Anomaly’. Climate Dynamics 37: 1217–1245.

37.

Grudd

(2008) Torneträsk tree-ring width and density AD 500–2004: A test of climatic sensitivity and a new 1500-year reconstruction of north Fennoscandian summers. Climate Dynamics 31: 843–857.

38.

Grudd

Briffa

Karlén

. (2002) A 7400-year tree-ring chronology in northern Swedish Lapland: Natural climatic variability expressed on annual to millennial timescales. The Holocene 12: 657–665.

39.

Gunnarson

Linderholm

(2002) Low-frequency summer temperature variation in central Scandinavia since the tenth century inferred from tree rings. The Holocene 12: 667–671.

40.

Hammarlund

Velle

Wolfe

. (2004) Palaeolimnological and sedimentary responses to Holocene forest retreat in the Scandes Mountains, west-central Sweden. The Holocene 14: 862–876.

41.

Hausmann

Kienast

(2006) A diatom-inference model for nutrients screened to reduce the influence of background variables: Application to varved sediments of Greifensee and evaluation with measured data. Palaeogeography, Palaeoclimatology, Palaeoecology 233: 96–112.

42.

Heinrichs

Barnekow

Rosenberg

(2006) A comparison of chironomid biostratigraphy from Lake Vuolep Njakajaure with vegetation, lake-level, and climate changes in Abisko National Park, Sweden. Journal of Paleolimnology 36: 119–131.

43.

Heinrichs

Peglar

Bigler

. (2005) A multi-proxy palaeoecological study of Alanen Laanijärvi, a boreal-forest lake in Swedish Lapland. Boreas 34: 192–206.

44.

Heiri

Lotter

(2001) Effect of low count sums on quantitative environmental reconstructions: An example using subfossil chironomids. Journal of Paleolimnology 26: 343–350.

45.

Hill

(1973) Diversity and evenness: A unifying notation and its consequences. Ecology 54: 427–432.

46.

Jessen

Rundgren

Bjorck

. (2005) Abrupt climatic changes and an unstable transition into a late Holocene Thermal Decline: A multiproxy lacustrine record from southern Sweden. Journal of Quaternary Science 20: 349–362.

47.

Jonsson

Andersson

Rosqvist

. (2010) Reconstructing past atmospheric circulation changes using oxygen isotopes in lake sediments from Sweden. Climate of the Past 6: 49–62.

48.

Jonsson

Leng

Rosqvist

. (2009) Stable oxygen and hydrogen isotopes in sub-Arctic lake waters from northern Sweden. Journal of Hydrology 376: 143–151.

49.

Juggins

(1992) TRAN (Version 1.8) and ZONE (Version 1.2) User Manual. University of Newcastle, 15pp.

50.

Juggins

(2007) C2 Version 1.5 User guide. Software for ecological and palaeoecological data analysis and visualisation. Newcastle University, 73pp.

51.

Juggins

(2013) Quantitative reconstructions in palaeolimnology: New paradigm or sick science? Quaternary Science Reviews 64: 20–32.

52.

Karlsson

Hörnberg

Hannon

. (2007) Long-term vegetation changes in the northern Scandinavian forest limit: A human impact-climate synergy? The Holocene 17: 37–49.

53.

Korhola

Olander

Blom

(2000) Cladoceran and chironomid assemblages as quantitative indicators of water depth in subarctic Fennoscandian lakes. Journal of Paleolimnology 24: 43–53.

54.

Kullman

(1995) Holocene tree-limit and climate history from the Scandes Mountains, Sweden. Ecology 76: 2490–2502.

55.

Kullman

(2006) Long-term geobotanical observations of climate change impacts in the Scandes of West-Central Sweden. Nordic Journal of Botany 24: 445–467.

56.

Kylander

Ampel

Wohlfarth

. (2011) High-resolution X-ray fluorescence core scanning analysis of Les Echets (France) sedimentary sequence: New insights from chemical proxies. Journal of Quaternary Science 26: 109–117.

57.

Laaksonen

(1976) The Dependence of Mean Air Temperatures upon Latitude and Altitude in Fennoscandia (1921–1950). Annales Academiæ Scientiarum Fennicæ A III 119: 1–18.

58.

Langdon

Ruiz

Wynne

. (2010) Ecological influences on larval chironomid communities in shallow lakes: Implications for palaeolimnological interpretations. Freshwater Biology 55: 531–545.

59.

Larocque

(2001) How many chironomid head capsules are enough? A statistical approach to determine sample size for palaeoclimatic reconstructions. Palaeogeography, Palaeoclimatology, Palaeoecology 172: 133–142.

60.

Larocque

Bigler

(2004) Similarities and discrepancies between chironomid- and diatom-inferred temperature reconstructions through the Holocene at Lake 850, northern Sweden. Quaternary International 122: 109–121.

61.

Larocque

Hall

(2003) Chironomids as quantitative indicators of mean July air temperature: Validation by comparison with century-long meteorological records from northern Sweden. Journal of Paleolimnology 29: 475–493.

62.

Larocque

Hall

(2004) Holocene temperature estimates and chironomid community composition in the Abisko Valley, northern Sweden. Quaternary Science Reviews 23: 2453–2465.

63.

Larocque-Tobler

Grosjean

Heiri

(2009) High-resolution chironomid-inferred temperature history since AD 1580 from varved Lake Silvaplana, Switzerland: comparison with local and regional reconstructions. The Holocene 19: 1201–1212.

64.

Larsson

Rosqvist

Ericsson

. (2012) Climate change, Moose and humans in northern Sweden 4000 cal. yr BP. Journal of Northern Studies 6: 9–30.

65.

Lauritzen

S-E

Lundberg

(1999) Calibration of the speleothem delta function: An absolute temperature record from the Holocene in northern Norway. The Holocene 9: 659–669.

66.

Lindegaard

(1992) Zoobenthos ecology of Thingvallavatn: Vertical distribution, abundance, population, dynamics and production. Oikos 64: 257–304.

67.

Lindén

Möller

Björck

. (2006) Holocene shore displacement and deglaciation chronology in Norrbotten, Sweden. Boreas 35: 1–22.

68.

Linge

Lauritzen

Andersson

. (2009) Stable isotope records for the last 10,000 years from Okshola cave (Fauske, northern Norway) and regional comparisons. Climate of the Past 5: 667–682.

69.

Loader

Young

GHF

Grudd

. (2013) Stable carbon isotopes from Torneträsk, northern Sweden provide a millennial length reconstruction of summer sunshine and its relationship to Arctic circulation. Quaternary Science Reviews 62: 97–113.

70.

Lotter

Birks

HJB

Hofmann

. (1998) Modern diatom, cladocera, chironomid, and chrysophyte cyst assemblages as quantitative indicators for the reconstruction of past environmental conditions in the Alps. II. Nutrients. Journal of Paleolimnology 19: 443–463.

71.

Mahaney

Kalm

Julig

(1995) Geoecology of the Norra Storfjället area, northern Sweden. Mountain Research and Development 15: 165–174.

72.

Meyers

Lallier-Vergés

(1999) Lacustrine sedimentary organic matter records of Late Quaternary paleoclimates. Journal of Paleolimnology 21: 345–372.

73.

Miller

RGJ

(1991) Simultaneous Statistical Inference. New York: Springer-Verlag.

74.

Nichols

Walcott

Bradley

. (2009) Quantitative assessment of precipitation seasonality and summer surface wetness using ombrotrophic sediments from an Arctic Norwegian peatland. Quaternary Research 72: 443–451.

75.

Quinlan

Smol

(2001) Setting minimum head capsule abundance and taxa deletion criteria in chironomid-based inference models. Journal of Paleolimnology 26: 327–342.

76.

Quinlan

Smol

Hall

(1998) Quantitative inferences of past hypolimnetic anoxia in south-central Ontario lakes using fossil midges (Diptera: Chironomidae). Canadian Journal of Fisheries and Aquatic Sciences 55: 587–596.

77.

R Developement Core Team (2010) R: A Language and Environment for Statistical Programming. Vienna: R Foundation for Statistical Computing.

78.

Racca

JMJ

Wild

Birks

HJB

. (2003) Separating wheat from chaff: Diatom taxon selection using an artificial neural network pruning algorithm. Journal of Paleolimnology 29: 123–133.

79.

Reimer

Baillie

MGL

Bard

. (2009) IntCal09 and Marine09 radiocarbon age calibration curves, 0–50,000 years cal BP. Radiocarbon 51: 1111–1150.

80.

Renberg

Hansson

(2008) The HTH sediment corer. Journal of Paleolimnology 40: 655–659.

81.

Rieradevall

Brooks

(2001) An identification guide to subfossil Tanypodinae larvae (Insecta: Diptera: Chironomidae) based on cephalic setation. Journal of Paleolimnology 25: 81–99.

82.

Rosén

(2005) Total organic carbon (TOC) of lake water during the Holocene inferred from lake sediments and near-infrared spectroscopy (NIRS) in eight lakes from northern Sweden. Biogeochemistry 76: 503–516.

83.

Rosén

Segerström

Eriksson

. (2001) Holocene climatic change reconstructed from diatoms, chironomids, pollen and near-infrared spectroscopy at an alpine lake (Sjuodjijaure) in northern Sweden. The Holocene 11: 551–562.

84.

Rosén

Segerström

Eriksson

. (2003) Do diatom, chironomid, and pollen records consistently infer Holocene July air temperature? A comparison using sediment cores from four alpine lakes in northern Sweden. Arctic Antarctic and Alpine Research 35: 279–290.

85.

Rosqvist

Jonsson

Yam

. (2004) Diatom oxygen isotopes in pro-glacial lake sediments from northern Sweden: A 5000 year record of atmospheric circulation. Quaternary Science Reviews 23: 851–859.

86.

Rosqvist

Leng

Goslar

. (2013) Shifts in precipitation during the last millennium in northern Scandinavia from lacustrine isotope records. Quaternary Science Reviews 66: 22–34.

87.

Self

Brooks

Birks

HJB

. (2011) The distribution and abundance of chironomids in high-latitude Eurasian lakes with respect to temperature and continentality: Development and application of new chironomid-based climate-inference models in northern Russia. Quaternary Science Reviews 30: 1122–1141.

88.

Seppä

Bjune

Telford

. (2009) Last nine-thousand years of temperature variability in Northern Europe. Climate of the Past 5: 523–535.

89.

Staafjord

(2012) Vindelfjällens naturreservat. Grundutredning om natur, kultur, nyttjande och förvaltning. Länsstyrelsen Västerbotten Meddelande 1–2012, January. Umeå: Länsstyrelsen Västerbotten, p. 125.

90.

Stewart

Michelutti

Blais

. (2013) Contrasting the effects of climatic, nutrient and oxygen dynamics on subfossil chironomid assemblages: A paleolimnological experiment from eutrophic High Arctic ponds. Journal of Paleolimnology 49: 205–219.

91.

Sundqvist

Holmgren

Lauritzen

(2007) Stable isotope variations in stalagmites from northwestern Sweden document climate and environmental changes during the early Holocene. The Holocene 17: 259–267.

92.

Sundqvist

Holmgren

Moberg

. (2010) Stable isotopes in a stalagmite from NW Sweden document environmental changes over the past 4000 years. Boreas 39: 77–86.

93.

Tallavaara

Seppä

(2012) Did the mid-Holocene environmental changes cause the boom and bust of hunter-gatherer population size in eastern Fennoscandia? The Holocene 22: 215–225.

94.

Telford

Birks

HJB

(2011) A novel method for assessing the statistical significance of quantitative reconstructions inferred from biotic assemblages. Quaternary Science Reviews 30: 1272–1278.

95.

ter Braak

CJF

Juggins

(1993) Weighted averaging partial least squares regression (WA-PLS): An improved method for reconstructing environmental variables from species assemblages. Hydrobiologia 269–270: 485–502.

96.

Ter Braak

CJF

Šmilauer

(2002) CANOCO Reference Manual and CanoDraw for Windows User’s Guide: Software for Canonical Community Ordination (Version 4.5). Ithaca, NY: Microcomputer Power.

97.

Trenberth

Jones

Ambenje

. (2007) Observations: Surface and atmospheric climate change. In: Solomon

Qin

Manning

. (eds) Climate Change 2007: The Physical Science Basis, the Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge and New York: Cambridge University Press, pp. 235–336.

98.

Väliranta

Korhola

Seppä

. (2007) High-resolution reconstruction of wetness dynamics in a southern boreal raised bog, Finland, during the late Holocene: A quantitative approach. The Holocene 17: 1093–1107.

99.

Velle

Larocque

(2008) Assessing chironomid head capsule concentrations in sediment using exotic markers. Journal of Paleolimnology 40: 165–177.

100.

Velle

Kongshavn

Birks

HJB

(2011) Minimizing the edge-effect in environmental reconstructions by trimming the calibration set: Chironomid-inferred temperatures from Spitsbergen. The Holocene 21: 417–430.

101.

Velle

Brodersen

Birks

HJB

. (2010) Midges as quantitative temperature indicator species: Lessons for palaeoecology. The Holocene 20: 989–1002.

102.

Velle

Brodersen

Birks

HJB

. (2012) Inconsistent results should not be overlooked: A reply to Brooks et al. (2012). The Holocene 22: 1501–1508.

103.

Velle

Brooks

Birks

HJB

. (2005a) Chironomids as a tool for inferring Holocene climate: An assessment based on six sites in southern Scandinavia. Quaternary Science Reviews 24: 1429–1462.

104.

Velle

Larsen

Eide

. (2005b) Holocene environmental history and climate of Ratasjoen, a low-alpine lake in south-central Norway. Journal of Paleolimnology 33: 129–153.

105.

Viberg

Berntsson

Lidén

(2013) Archaeological prospection of a high altitude Neolithic site in the Arctic mountain tundra region of northern Sweden. Journal of Archaeological Science 40: 2579–2588.

106.

Walker

Smol

Engstrom

. (1991) An assessment of Chironomidae as quantitative indicators of past climate change. Canadian Journal of Fisheries and Aquatic Sciences 48: 975–987.

107.

Wanner

Beer

Bütikofer

. (2008) Mid- to late Holocene climate change: An overview. Quaternary Science Reviews 27: 1791–1828.

108.

Wiederholm

(ed.) (1983) Chironomidae of the Holarctic Region – Keys and Diagnoses. Part 1: Larvae. Södra Sandby: Entomologica Scandinavica.

109.

Wright

HEJ

(1967) A square-rod piston sampler for lake sediments. Journal of Sedimentary Petrology 37: 975–976.

110.

SMHI (n.d.) Homepage. Available at: http://www.smhi.se (accessed 7 June 2013).

111.

Young

McCarroll

Loader

. (2011) Changes in atmospheric circulation and the Arctic Oscillation preserved within a millennial length reconstruction of summer cloud cover from northern Fennoscandia. Climate Dynamics 39: 495–507.