The Holocene 15,1 (2005) pp. 141-148

Quantitative reconstruction of Holocene climate from the Chuna Lake pollen record, Kola Peninsula, northwest Russia Nadia Solovieva,l* Pavel E. Tarasov2 and Glen MacDonald3

('Environmental Change Research Centre, Department of Geography,

University of WCIH London 2Department OAP, College London, 26 Bedford Way, UK, Geography, Moscow State University, Vorobievy Gory 119899 Moscow, Russia; 3Departments of Geography and Organismic Biology, Ecology and Evolution, UCLA, Los Angeles, California 90095-1524, USA) Received 13 November 2003; revised manuscript accepted 25 January 2004

HOLOCENE RESEARCH

REPORT

Abstract: July mean temperatures and annual precipitation during the last 9000 years were inferred using the pollen record from Chuna Lake, Kola Peninsula, Russia. A quantitative pollen-climate model was generated using the best modem analogues method from a training set of 99 surface pollen spectra from the Kola Peninsula, northern Fennoscandia and Karelia. According to the evidence from Chuna Lake, the early and mid-Holocene (c. 9000-5000 cal. BP) was warm and dry in the central Kola Peninsula with July temperatures being 1.5-2°C higher than at present. The onset of warm and dry early to mid-Holocene occurred in the Kola Peninsula earlier than in the rest of northern Fennoscandia. July temperature started to decrease and annual precipitation increased from c. 5000 cal. yr BP and climate became cool and moist.

Key words: Climate change, best modem analogues, pollen, multiproxy record, Arctic, Kola Peninsula, Russia, Holocene.

Introduction The major patterns of Holocene climate history of Fennoscandia and adjacent Russia revealed by pollen and palaeolimnological analysis are typified by warmer than present climatic conditions during the early Holocene and climatic cooling starting between c. 6000 and 4000 cal. yr BP (e.g., Karlen, 1976; Hyvarinen, 1975; 1976, Briffa et al., 1990; Kullman, 1995a; Seppa, 1996; Dahl and Nesje, 1996; Eronen et al., 1999; Barnekow, 2000; Seppa and Hammarlund, 2000; Snyder et al., 2000; Seppa and Birks, 2001; 2002; Gervais et al., 2002; Solovieva and Jones, 2002; Heikkila and Seppa, 2003). However, the precise timing and the amplitude of climate changes are yet to be established for all regions of Fennoscandia as climate changes may have been somewhat variable and diachronous. While Holocene temperature changes have similar general patterns for Fennoscandia, humidity changes show more regional variations. The early Holocene (c. 9000 cal. BP) was oceanic with high annual precipitation in northern Fennoscandia (Seppa, 1996; Seppa and Hammarlund, 2000; Seppa and Birks, 2001; Seppa et al., 2002) whereas southern Sweden was dry (e.g., Digerfeldt, 1988; Vassiljev and Harrison, *Author for correspondence

(e-mail: nsolovieQgeog.ucl.ac.uk)

©D 2005 Edward Arnold (Publishers) Ltd

1998). From c. 8000 cal. yr BP climate became more oceanic in southern Sweden and southern Finland and relatively dry in the northeast, e.g., Finnish Lapland (e.g., Digerfeldt, 1988; Hyvarinen and Alhonen, 1994; Eronen et al., 1999). A recent multiproxy study of the Holocene sediment record from Chuna Lake implies that the warm and dry climate was established in the Kola Peninsula from c. 9000 cal. BP (see map in Figure 1, Solovieva and Jones, 2002). The aim of the current paper is to quantify the palaeoclimatic findings of the Chuna Lake pollen record using the method of best modern analogues (Guiot, 1990). This method was developed for and used to reconstruct temporal and spatial patterns of Holocene climate from European pollen records (e.g., Prentice et al., 1992; Guiot et al., 1993; Cheddadi et al., 1997; Digerfeldt et al., 1997) and have suggested that vegetation and corresponding pollen assemblages are closely related to 'bioclimatic' variables, such as mean temperature of the warmest month and annual precipitation. The method of best modern analogues is used to infer past values of the above climate parameters by assuming that if a fossil pollen assemblage has a modern analogue then the past climate represented by the fossil assemblage was similar to the present climate associated with its analogue modern pollen assemblage. A training set of 99 surface pollen spectra and 10.1 191/0959683605hl793rr

142 The Holocene 15 (2005) climate data from the Kola Peninsula, northem Fennoscandia, Karelia and the Archangel region provided modern analogues for the palaeoclimate reconstruction based on the Chuna Lake fossil pollen assemblages.

Study site and study area Chuna is an upland oligotrophic lake (32°29'E and 67°57'N) situated in a small depression on the top of the Chuna Tundra mountain plateau at 475.3 m a.s.l. (Figure 1). The lake area is 0.125 kM2, the catchment area is c. 2 km2 and maximum depth is 18 m. One small stream flows into the lake and there is one outflow. The most common types of rock in the area are quartz diorite and grano-diorite. During the last glaciation, the Kola Peninsula was covered by the Kola ice sheet except for the Rybachi Peninsula and Kildin island (Figure 1). Small ice-free areas might have been present in the northeast sector of the Kola, at Svyatoi Nos Cape (Lavrova, 1960). Ice retreat began from the northwest (Lavrova, 1960; Comner et al., 1999). The ice remained in the upland areas of the central Kola Peninsula (e.g., Khibiny mountains, Monche Tundra, Chuna Tundra, Lovozerskie Tundra) until c. 10500-9500 cal. BP (Lebedeva et al., 1989). According to the radiocarbon and sediment evidence, Chuna Lake was deglaciated at c. 9000 cal. BP (Solovieva, 2000). The lake is located above the modem treeline which is formed by birch forest-tundra and reaches c. 380-400 m a.s.l. in the Chuna Tundra mountain plateau. The alpine tundra in the catchment is dominated by lichens (Cladonia alpestris, C. rangeferina, Cetraria nivalis, Alectoria nigricans and A. ochroleuca) and shrubs (mainly Betula nana and Empetrum nigrum subsp. hermaphroditum). Arctostaphylos uva-ursi, Lycopodium spp., Diphasiastrum alpinum, Vaccinium myrtillus, V. vitis-idaea and Gymnocarpium dryopteris are also common on the relatively drier parts of the catchment. Salix spp., Cyperaceae (Carex bigelowii) and Juncaceae (Juncus trifidis) hummocks occur at the edge of the lake along with

patches of Sphagnum spp. (further details are given in Solovieva, 2000, and Solovieva and Jones, 2002). Climate in the Kola Peninsula is maritime cold-temperate. It is highly impacted by the North Atlantic and the Polar Front. Climate is particularly influenced by the northern branch of the Gulf Stream, which keeps winter temperatures anomalously warm and prevents ice development in the coastal areas of the Barents Sea. The North Atlantic Oscillation determines the intensity of cyclone activity on the Kola Peninsula (Lebedeva et al., 1989). The cyclones are generated at the boundary between the Polar Front and the warm air masses of the North Atlantic and are more intensive in winter (Yakovlev, 1961). In the centre and on the west of the Kola Peninsula, the climate is most continental with the lowest winter temperatures reaching, on average, -14C and summer temperatures averaging 13.1°C. Precipitation varies from 1400 mm/year in the Khibiny mountain range to 376 mm/year in Murmansk (Annual Reports on Meteorology 1930-90).

Material and methods Chuna Lake The lake was sampled in April 1996 from ice using a Russian peat sampler (Jowsey, 1966) and a freeze corer (Renberg and Hanson, 1993). The details of the sampling, laboratory preparations and chronology development are given in Solovieva (2000) and Solovieva and Jones (2002). Pollen slides were prepared only between 163 and 10 cm as insufficient material was available within the top 10 cm of the core. Preparation of pollen slides followed Andersen (1960) and Bates et al. (1978). Pollen was mounted using silicone oil. Eucalyptus grains were added to estimate pollen accumulation rate. Pollen grains were identified using keys in Faegri and Iversen (1989) and Moore et al. (1991). Pollen nomenclature follows Moore et al. (1991). When certain identification was impossible, counts for Corylus, Myrica and Betula were

Figure 1 Location and bathymetry of Chuna Lake together with locations of 99 surface pollen samples.

Nadia Solovieva et al.: Climate reconstruction from poDen: Kola Peninsula 143

amalgamated and assigned as 'Betula-Corylus-Myrica' (B-C-M) on the pollen diagram. 'Herbs' comprises a sum of percentages of terrestrial pollen, which were found in the sediments but were not included in the diagram individually because of their low abundance. The group comprises Artemisia, Rumex, Asteraceae (mainly Cirsium group), Rosaceae (Potentilla and Filipendula type), Chenopodiaceae, Caryophyllaceae and Plantaginaceae. Terrestrial spore taxa were included in the total pollen sum. Sphagnum was excluded from the pollen sum as aquatic. Core chronology A core chronology was established using 210Pb analyses and AMS 14C. 210Pb analysis was carried out at the University of Liverpool, UK. The CRS dating model, in which a constant net rate of supply is assumed, was used to calculate the 210Pb dates (Appleby and Oldfield, 1978). AMS 14C analysis was conducted at the AMS 14C facility at University of Arizona, Tuscon, USA. AMS "4C dates were obtained for both bulk sediment and aquatic bryophyte remains (see Tables 2 and 3 in Solovieva and Jones, 2002). Radiocarbon dates were calibrated using CALIB 3.0 (Stuiver and Reimer, 1993) by the simple intercept method with a linear interpolation of the data points. Curve smoothing by 10 samples was employed to avoid several linear intercepts (T6mqvist and Bierkens, 1994). All calibrated 14C dates are denoted as 'cal. BP' and uncalibrated 14C dates are denoted as 'years BP'. The age-depth model was constructed with a polynomial producing the best fit to all points (Figure 2).

Modem pollen-climate data set and method of modem analogues Most of the lakes in the training set are small to medium-sized, with the lake area being less than 0.4 km2. Sediment samples were collected from the deepest point of the lakes using a Glew corer (Glew, 1991). The combined pollen data set comprises: 23 samples collected and analysed by N. Solovieva (Solovieva, 2000); 31 samples from the Kola Peninsula collected and analysed by B. Gervais and G. MacDonald (Gervais and MacDonald, 2001); and 45 samples from Sweden, the Kola Peninsula, the Archangel region and northem Karelia which were included in the larger surface pollen data set used for the quantitative biome and climate reconstruction in the former Soviet Union and Mongolia (Tarasov et al., 1998; 1999). Pollen nomenclature follows Moore et al. (1991). Locations of surface pollen sampling sites are shown in Figure 1. 0* 20

20\XI

40

E 0 CL

0

`0

lxi

40

60

te

0

-calibrated 14C dates - 21OPb dates

N

The modem climatic variables - annual precipitation (Pann) and mean temperature of the warmest month (July is the warmest month in the Kola Peninsula, so mean temperature of the warmest month is Tj,i) - at pollen sampling sites were calculated using the updated version of the Leemans and Cramer (1991) climate data base with precise topography (W. Cramer, personal communication). It contains mean monthly temperatures, precipitation and sunshine hours for land pixels with 30' latitude and longitude resolution. Table 1 shows the summary statistics of the modem climate trainingset data and estimated climatic variables for Chuna Lake which were calculated using the Leemans and Cramer (1991) data base. The training-set climate data suggest that the pollen-climate data set is representative for the reconstruction of possible climatic fluctuations at the Chuna site since 9000 cal. yr BP. Fossil pollen spectra from Chuna Lake are well represented by modem counterparts in the assemblages of the training set. All fossil taxa occur in the modem pollen samples. The squared chord distance (d2) was used as a measure of dissimilarity between fossil and modem assemblages (Overpeck et al., 1985). A threshold value for a 'good' analogue was identified as 10% of the 4851 dissimilarities calculated between all modem samples (Birks, 1995) using ANALOG 1.6 program (Line and Birks, 1992, unpublished program). All fossil samples having a a2 below 0.0505 are therefore considered to be 'good' analogues. Eight fossil spectra lack 'good' analogues in modem pollen assemblages (see Figure 4). Details regarding the method of the best modern analogues have been described in several publications (e.g., Guiot, 1990; Guiot et al., 1993; Digerfeldt et al., 1997). PPPBase software for bioclimatic analyses (Guiot and Goeury, 1996) was used to perform all calculations. Tarasov et al. (1998; 1999) suggested using all available terrestrial pollen taxa (excluding aquatic pollen and spores) for the calculations of regional climatic variables and biome scores. A set of surface pollen records compiled in the present study consists of 34 terrestrial taxa. The calculation of relative abundance for each individual pollen taxon is based upon the total sum of 34 taxa taken as 100%. Initially, Tj,0, and Pann were reconstructed using surface pollen spectra from the Kola Peninsula and northem Karelia, excluding 39 spectra from the coastal sites and outside the Peninsula. The climate reconstruction was based on values from the eight best modem analogues (the number of analogues is chosen empirically). Logarithmic transformation log (y + 1) of taxa percentages yielded the best results compared to 'no transformation' and 'square root transformation' approaches (Guiot and Goeury, 1996). Reconstructed modem values of Tj0, and Pann were compared to Tjul and Par. estimated at each sampling site using the data base of Leemans and Cramer (1991) (see Table 2 for performance statistics). Chuna Lake data were excluded from the error estimation as the top Chuna core sample is from 10 cm depth (c. 300 cal. BP). Root mean square error of prediction (RMSEP)

80

0 0 100

Table 1 Summary characteristics of the climate data set and estimated modem values for Chuna Lake, calculated using the climate data base 19\~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ of Leemans and Cramer (1991)

120 140

160X 180 ±

0

1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Calendar years BP

Figure 2 The fitted age-depth model for Chuna Lake.

Max. Min. Mean Median Chuna Lake

Tj (°C)

Pann (mm)

15.3 9.4 12.7 12.4 11.6

679.8 355.9 444.8 424.9

640

144 The Holocene 15 (2005) Table 2 Performance statistics for the method of best modem analogues for the modem pollen-climate data set from the Kola Peninsula and northern Karelia

r2 RMSEP Mean absolute error (bias) Max. Min. Mean Median

Tiul (C)

Pa.n (mm yr

0.745 0.577 0.415 14.569 11.656 12.642 12.650

0.672 33.85 26.86 542.52 365.84 428.73 414.59

and mean absolute error (bias) of the reconstruction were estimated by the programme RMSEP (Line and Birks, unpublished program).

Results Vegetation and climate history The pollen stratigraphy of Chuna Lake is shown in Figure 3. A detailed analysis of vegetation history of the central Kola Peninsula is given by Solovieva (2000) and Solovieva and Jones (2002). In summary, five pollen local assemblage zones represent the transition between Betula-Pinus forests (zone I, c. 9000-8000 cal. BP) to the dominance of Pinus (c. 8000-4200 cal. BP, zones II and III, regional 'Pinus-Alnus' zone) and, finally, the retreat of Pinus and establishment of mixed Picea-Pinus forests in the central Kola Peninsula (4200-300 cal. BP, zones IV and V, regional Pinus-Picea zone). A steep rise in pine pollen between 9000 and 8000 cal. BP suggests a spread of pine forests in the Kola Peninsula and therefore the establishment of a warm and dry climate favoured by pine. The spread of pine coincides with maximum

values of tree pollen which implies the advance of the treeline. From c. 5000 cal. BP the proportion of pine pollen starts to decrease, together with the gradual rise in Picea, pteridophytes and Sphagnum spores indicating the onset of the late-Holocene cooling and moister conditions in general.

Reconstructed climatic variables Pollen-based quantitative climate reconstructions (Figure 4) generally fit with the above qualitative interpretation of pollen changes. Reconstructions of Tjul show that summers were significantly warmer than today between c. 9000 and 4500 cal. yr BP with the mean Tj, for this period of 13.2°C. July temperatures reached 13.6°C between 6000 and 5000 cal. yr BP, when reconstructed mean values of Tjul were 1.5-2°C higher than at present. Although the relative abundance and pollen accumulation rates of Pinus remained high between 5000 and 4200 cal. BP, the decrease in tree pollen and retreat of Betula forest from the Chuna catchment probably indicate the onset of climatic cooling from c. 5000 cal. BP (Solovieva and Jones, 2002). An abrupt increase in Tj,l at c. 4700 cal. yr BP coincides with the sharp increase in erosion indicators and fluctuation of 6180 values (Jones et al., 2004). This relatively short-lived event is more likely to have been driven by catchment erosion, which brought in older pollen grains rather than climate change. The phase of a relatively cold climate lasted until c. 1700 cal. yr BP. Tjul rises from c. 12.2 to 13.2°C between c. 1700 and 1000 cal. yr BP, which might indicate the late-Holocene warming. However, the increase in temperature is short-lived, and with the current resolution of the analysis it is unclear whether this is a result of catchment erosion or the late-Holocene climate change. Figure 4 shows that reconstructed Tjul for Chuna Lake became close to the modern values from c. 300 cal. BP. Annual precipitation ranged between 379 and 405 mm/year during the early and mid-Holocene and steadily increased from c. 5000 cal. BP. Annual precipitation reached its maximum of

5 L.P.A.Z. 15 20. 2530

30-

*0

POL V

3540-

456

505

3000

60. 65. 70. 75-

POL IV

80-

300090 4(

5S

K. 300

POL III

POL III

0

a(A3000

POL 11

200

_F6L

_-

S( 40

-z ____

- gytta -

diatomite

gytAa-clay

Figure 3 Lithostratigraphy and percentage pollen diagram. Includes all the spore and pollen taxa identified with abundance > 1%.

Nadia Solovieva et at.: Climate reconstruction from pollen: Kola Peninsula 145 annual precipitation

mean July temperature

*

0 1000

0

-

1000 -

_

2000 -

2000 3000

3000

4000

4000 -

a. co - 5000

co -

6000-

C

7000-

7000

8000 9000 -

2

8000

9000 10000 11.0

50006000

m-n

1 ooIo

12.0

13.0

14.0

300

400

500

600

year -1

700

Figure 4 Inferred July mean temperatures and annual precipitation for Chuna Lake. Samples with poor analogues are circled. Star indicates the modem climate values at Chuna Lake estimated using the climate data base of Leemans and Cramer (1991).

490 mm/year at 300 cal. BP. However, modem Pann estimated at Chuna Lake is considerably higher reaching 640 mm/year (Figure 4). This might indicate that climate became substantially more humid in the Kola Peninsula over the last 300 years and this agrees with Lebedeva et al. (1989). This could also imply that the current reconstruction might have generally underestimated annual precipitation values at Chuna Lake during the Holocene. The development of a Sphagnum mire in the catchment of Chuna Lake also reflects increased humidity from c. 4800 cal. BP (Figure 3), and corresponds with the increase in Sphagnum spores in other areas of the Kola Peninsula (see Lebedeva et al., 1989) and in Finland during the mid-Holocene (e.g., Korhola, 1995; Seppa, 1996). The increase in Picea and decrease in Pinus from c. 4200 cal. BP is another indication of the increasing moisture (e.g., Kullman, 1995b).

Discussion The quantitative reconstruction of temperature and precipitation from the Chuna pollen record is in good agreement with qualitative interpretations of different proxies from that site (Solovieva and Jones, 2002). There are a number of qualitative and quantitative Holocene climate records from Fennoscandia and the Kola with which the Chuna record can be compared to assess past climate on a larger geographic scale. Radiocarbon dated collections of fossil pine wood from beyond the present limits of Pinus sylvestris on the Kola Peninsula (MacDonald et al., 2000) and fossil pollen and stomata evidence of pine treeline advance between c. 9500 and 4000 cal. BP (Gervais et al., 2002; Snyder et al., 2000) indicate warmer, and possibly drier, conditions. Stable isotope studies of Kola lakes north of Chuna and closer to the coast suggest dry conditions from at least 10000 cal. BP, but also suggest an earlier onset of moistening than the Chuna record, commencing from around 8000 cal. BP (Wolfe et al., 2003). Palaeoclimate reconstructions suggest that the warm and dry conditions were established in the central Kola Peninsula before c. 9000 cal. BP whereas quantitative reconstructions and isotopic studies of lake sediments suggest it remained more oceanic until c. 8000 cal. BP in northern Finland (Seppa and Hammarlund, 2000; Seppa and Birks, 2001; Figure 5). In the western parts of northern Fennoscandia an oceanic climate persisted until c. 6500-5500 cal. BP (e.g., Hammarlund et al., 2002; Seppa and Hammarlund, 2000). This clearly suggests a major east-west gradient in humidity during the Holocene which was mainly determined by the cyclonic influence of the

North Atlantic. Humidity changes apparent in the isotope records from the more coastal areas of the Kola Peninsula are different again from either Chuna or other regions of Fennoscandia (Wolfe et al., 2003). The reasons for the discrepancies between climate inferences based on stable isotope data and micro- and macrofossil evidence perhaps require further evaluation. However, lake-level reconstructions from northern Finland often show dry conditions during the early to mid-Holocene (Hyvarinen and Alhonen, 1994; Eronen et al., 1999) and this suggests the need to investigate further the roles of precipitation and evaporation in the humidity records thus far available. The elevated July temperatures between c. 9000 and 5000 cal. BP at Chuna Lake might indicate that it was a time of the mid-Holocene hypsithermal in the central Kola Peninsula. During the hypsithermal, the climate appears to have been more continental with higher summer temperatures and lower humidity in the central Kola Peninsula compared to the neighbouring areas of northern Finland and perhaps the Kola coast (Seppii and Birks, 2001; 2002; Wolfe et al., 2003; Figure 5). Generally, the Kola Peninsula experienced lower precipitation than the western parts of Fennoscandia. This may indicate the impact of greater summer insolation due to a decrease in incursions of cyclonic precipitation events from the North Atlantic. It is possible that some or all of the decrease in precipitation occurred during the general winter period when high northern latitude insolation was lower than today. The period of maximum July temperatures lasted slightly longer in the central Kola Peninsula than in some westem parts of northern Fennoscandia. For instance, the highest July temperatures in northern Finland were reconstructed between c. 8200 and 6500 cal. BP (Seppa and Birks, 2001; 2002; Figure 5). However, in southern Finland the Holocene temperature maximum lasted from c. 8000 to 4500 cal. BP (Heikkila and Seppa, 2003), and in northern Sweden it occurred between c. 9500 and 5000 cal. BP (Hammarlund et al., 2002). A relatively late-lasting Holocene temperature maximum (between -8000 and 4000 cal. BP) in the northern Kola Peninsula was also shown by Gervais et al. (2002) and MacDonald et al. (2000). Although the reconstructed Tjul peaked at c. 4700 cal. BP, multiproxy data from the Chuna record (e.g., decrease in tree pollen, increase in Sphagnum and pteridophyte spores, decrease in Betula pollen accumulation rate) suggest that the cooling started earlier, i.e., from c. 5000 cal. BP. The lateHolocene cooling occurred almost a millennium later than in northern Finland (Seppa and Birks, 2001). An earlier retreat of the pine and birch treeline (starting from c. 6000 cal. BP) was found in the Nordkinn Peninsula in Norway (Seppa,

146 The Holocene 15 (2005)

E

w

A

D

lm a.)

I

0

01)

I

111.12 10 1011

13.14 14 13

Pollen-inferred mean July temperature (0C) D

Ma. I

0)1

(14 1

Pollen-inferred

mean

'460''8I 0'

1200

annual precipitation (mm)

Figure 5 July temperature (I) and precipitation (II) reconstructions from (A) Lake Tibetanus (Hammarlund et al., 2002), (B) Toskaljavri (Seppa and Birks, 2002) and (C) Tsuolbmajavri (Seppa and Birks, 2001), compared with (D) the Chuna Lake climate reconstructions. Smooth lines represent five-point running averages in (A); and LOESS smoother (span = 0.25; order = 1) in (B) and (C). Sites are arranged in the west-east direction.

1996) although the older dates might be due to the use of bulk sediment dates in the area of calcareous sedimentary rocks. However, the onset of cooling is dated from c. 4000 cal. BP in the coastal areas of the Kola Peninsula (Gervais et al., 2002; MacDonald et al., 2000; Snyder et al., 2000) and eastern northern Fennoscandia (Eronen et al., 1999).

Conclusions According to the evidence from Chuna Lake, the early and mid-Holocene (c. 9000-5000 cal. BP) was relatively warm and dry in the central Kola Peninsula. During this time the central Kola Peninsula exhibited a drier and more continental climate than the western parts of northern Fennoscandia. The onset of a warm and dry early to mid-Holocene occurred in the Kola Peninsula earlier than in the rest of northern Fennoscandia. During the late Holocene which started from c. 5000 cal. BP the climate in the central Kola Peninsula became relatively cool and moist. While there appears to be rather uniform evidence for a warm mid-Holocene period across the Kola and adjacent Fennoscandia, the pattern of changes in precipitation appear to have been much more regionally variable and also reconstructions differ depending upon the proxy used. Clearly more work is required to understand the pattern of hydroclimatic changes in the region.

Acknowledgements This study was supported by grant 00-05-22000 from the Russian Foundation for Fundamental Research, ERSS 93-14 grant from Central European University, Hungary, INTAS grant 1010-CT93-0021, Dean's scholarship and ORS award from UCL, and NSF grant ATM9632926. The 14C dating was funded by NERC (ref. 676/1296). Additional help with expenses by ENSIS Ltd and Graduate School, UCL, is also gratefully acknowledged. We would like to thank Andrei Sharov, Vladimir Dauvalter, Eric, Peter Rosen, Kate Duff, Bruce Gervais, Tamsin Laing and Jeff Snyder for help with the fieldwork. We are especially grateful to Ingemar Renberg who provided fieldwork equipment, helped with fieldwork and provided 206Pb/207Pb data for core correlation. We thank Tatiana Moiseenko, Gennadi Kalabin, Luydmila Kagan and colleagues from the Kola Science Centre, Apatity, for arranging access to Chuna Lake and providing useful information and papers. We are grateful to Peter Appleby for providing 2l0Pb dating. Special thanks are due to John Birks for his advice on numerical and pollen analyses and for his comments on the manuscript, to Sylvia Peglar for her help with pollen taxonomy and hospitality in Bergen, and to Vivienne Jones. We thank the reviewers J.R.M. Allen and M. Eronen for useful comments. This is a PARCS contribution. Figure 5 is reprint with permission from Elsevier from Quaternary Research 2002 57, 194.

Nadia Solovieva et al: Climate reconstruction from pollen: Kola Peninsula 147

References Andersen, S.T. 1960: Silicone oil as a mounting medium for pollen grains. Danmarks Geologiske Undersogelske IV 4(1), 1-24. Appleby, P.G. and Oldfield, F. 1978: The calculation of 210Pb dates assuming a constant rate of supply of unsupported 210Pb to the sediment. Catena 5, 1-8. Annual Reports on Meteorology 1930-90: Gosudarstvennyi komitet po metereologii, Moskva. Moscow: State Committee on Hydrometeorology (in Russian). Barnekow, L. 2000: Holocene regional and local vegetation history and lake-level changes in the Tornetrask area, northern Sweden. Journal of Paleolimnology 23, 399-420. Bates, C.D., Coxon, P. and Gibbard, P.L. 1978: A new method for the preparation of clay-rich sediment samples for palynological investigations. New Phytologist 81, 459-63. Birks, H.J.B. 1995: Quantitative palaeoenvironmental reconstructions. Iu Maddy, D. and Brew, J.S., editors, Statistical modelling of Quaternary science data, Technical Guide S, Cambridge: Quaternary Research Association, 220-21. Briffa, K.R., Bartholin, T.S., Eckstein, D., Jones, P.D., Karlin, W., Schweingruber, F.H. and Zetterberg, P. 1990: A 1,400-year tree ring record of summer temperatures in Fennoscandia. Nature 346, 434-37. Cheddadi, R., Yu, G., Guiot, J., Harrison, S.P. and Prentice, I.C. 1997: The climate of Europe 6000 years ago. Climate Dynamics 13, 1-9. Corner, G.D., Yevzerov, V.Y., Kolka, V.V. and Moller, J.J. 1999: Isolation basin stratigraphy and Holocene relative sea-level change at the Norwegian-Russian border north of Nikel, northwest Russia, Boreas 28, 146-66. Dahl, S.O. and Nesje, A. 1996: A new approach to calculating Holocene winter precipitation by combining glacier equilibrium-line altitudes and pine-tree limits: a case study from Hardangerjokulen, central southern Norway. Holocene 6, 381-98. Digerfeldt, G. 1988: Reconstruction and regional correlation of Holocene lake-level fluctuations in Lake Bysjon, southern Sweden, Boreas 17, 165-82. Digerfeldt, G., de Beaulieu, J.-L., Guiot, J. and Mouthon, J. 1997: Reconstruction and paleoclimatic interpretation of Holocene lake-level changes in Lac de Saint-Leger, Haute Provence, southeast France. Palaeogeography, Palaeoclimatology, Palaeoecology 136, 231-58. Eronen, M., Hyvirinen, H. and Zetterberg, P. 1999: Holocene humidity changes in northern Finnish Lapland inferred from lake sediments and submerged Scots pines dated by tree-rings. The Holocene 9, 569-80. Faegri, K. and Iversen, J. 1989: Textbook of pollen analysis (fourth edition). Chichester: Wiley. Gervais, B.R. and MacDonald, G.M. 2001: Modern pollen and stomate deposition in lake surface sediments from across the treeline on the Kola Peninsula, Russia. Review of Palaeobotany and Palynology 114(3-4), 223-37. Gervais, B.R., MacDonald, G.M., Snyder, J.A. and Kremenetski, C.V. 2002: Pinus sylvestris treeline development and movement on the Kola Peninsula of Russia: pollen and stomata evidence. Journal of Ecology 90, 627-38. Glew, J.R. 1991: Miniature gravity corer for recovering short sediment cores. Journal of Palaeolimnology 5, 285-88. Guiot, J. 1990: Methodology of palaeoclimatic reconstruction from pollen in France. Palaeogeography, Palaeoclimatology, Palaeoecology 80, 49-69. Guiot, J. and Goeury, C. 1996: PPPBASE, a software for statistical analysis of paleoecological and paleoclimatological data. Dendrochronologia 14, 295-300. Guiot, J., Harrison, S.P. and Prentice I.C. 1993: Reconstruction of Holocene precipitation patterns in Europe using pollen and lake-level date. Quaternary Research 40, 139-49. Hammarlund, D., Barnekow, L., Birks, H.J.B., Buchardt, B. and 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, 339-51. Heikkila, M. and Seppa, H. 2003: A 11,000 yr palaeotemperature reconstruction from the southern boreal zone in Finland. Quaternary

Science Reviews 22, 541-54.

Hyvirinen, H. 1975: Absolute and relative pollen diagrams from northemmost Fennoscandia. Fennia 142, 1-23. 1976: Flandrian pollen deposition rates and tree-line history in northernmost Fennoscandia. Boreas 5, 163-75. Hyvarinen, H. and Alhonen, P. 1994: Holocene lake level changes in the Fennoscandian tree-line region, western Finnish Lapland: diatom and cladoceran evidence. The Holocene 4, 251-58. Jowsey, P.C. 1966: An improved sampler. New Phytologist 65, 245-48. Jones, V., Leng, M., Solovieva, N., Sloane, H. and Tarasov, P. 2004: Holocene climate on the Kola Peninsula; evidence from the oxygen isotope record of diatom silica. Quaternary Science Reviews 23, 833-39. Karlen, W. 1976: Lacustrine sediments and tree-limit variations as indicators of Holocene climatic fluctuations in Lappland, northern Sweden. Geografiska annaler 58A, 1-33. Korhola, A. 1995: Holocene climatic variations in southern Finland reconstructed from peat initiation. The Holocene 5, 43-58. Kullman, L. 1995a: Holocene tree limit and climate history from the Scandes Mountains, Sweden. Ecology 7, 2490-502. 1995b: New and firm evidence for mid-Holocene appearance of Picea abies in the Scandes Mountains, Sweden. Journal of Ecology 83, 439-47. Lavrova, M.A. 1960: Quaternary geology of the Kola Peninsula. Moscow-Leningrad: Izdatel'stvo Academii Nauk SSSR (in Russian). Lebedeva, R.M, Kagan, L.Ya. and Nikonov, S.V. 1989: Climatological study of the Holocene deposits of the Kola Peninsula. Annual report on the programme 'Evolution of biosphere'. Apatity: Kola Science Centre, Geological Institute, 9-26. Leemans, R. and Cramer, W. 1991: The IIASA climate database for mean monthly values of temperature, precipitation and cloudiness on a global terrestrial grid. RR-91-18. Laxenburg: International Institute of Applied Systems Analysis. MacDonald, G.M., Gervais, B.R., Snyder, J.A., Tarasov, G.A. and Borisova, O.K. 2000: Radiocarbon dated Pinus sylvestris L. wood beyond tree-line on the Kola Peninsula, Russia. The Holocene 10, 143-47. Moore, P.D., Webb, J.A. and Collinson, M.E. 1991: Pollen analysis (second edition). London: Blackwell Scientific Publications. Overpeck, J.T., Prentice, I.C. and Webb, T. III 1985: Quantitative interpretation of fossil pollen spectra: dissimilarity coefficients and the method of modern analogues. Quaternary Research 23, 87-108. Prentice, I.C., Cramer, W., Harrison, S.P., Leemans, R., Monserud, R.A. and Solomon, A.M. 1992: A global biome model based on plant physiology and dominance, soil properties and climate. Journal of Biogeography 19, 117-34. Renberg, I. and Hansson, H. 1993: A pump freeze corer for recent sediments. Limnology and Oceanography 38, 1317-21. Seppi, H. 1996: Post-glacial dynamics of vegetation and tree-lines in the far north of Fennoscandia. Fennia 174, 1-96. Seppii, H. and Birks, H.J.B. 2001: July mean temperature and annual precipitation trends during the Holocene in the Fennoscandian tree-line area: pollen-based climate reconstructions. The Holocene 11, 527-39. 2002: Holocene climate reconstructions from the Fennoscandian tree-line area based on pollen data from Toskaljavri. Quaternary Research 57, 191-99. Seppa, H. and Hammarlund D. 2000: Pollen-stratigraphical evidence of Holocene hydrological change in northern Fennoscandia supported by independent isotopic data. Journal of Paleolimnology 24, 69-79. Seppi, H., Nyman, M., Korhola, A. and Weckstr6m, J. 2002: Changes of treelines and alpine vegetation in relation to post- glacial climate dynamics in northern Fennoscandia based on pollen and chironomid records. Journal of Quaternary Science 17, 287-301. Snyder, J.A., MacDonald, G.M., Forman, S.L., Tarasov, G.A. and Mode, W.N. 2000: Postglacial climate and vegetation history, northcentral Kola Peninsula, Russia: pollen and diatom records from Lake Yarnyshnoe-3. Boreas 29, 261-71. Solovieva, N. 2000: A palaeoecological study of Holocene environmental change in a small upland lake from the Kola Peninsula, Russia. PhD thesis, University of London. Solovieva, N. and Jones, V.J. 2002: A multiproxy record of Holocene environmental changes in the central Kola Peninsula, northwest Russia. Journal of Quaternary Science 17, 303-18. Stuiver, M. and Reimer, P.J. 1993: Extended 14C data base and revised CALIB 3.0 14C calibration program. Radiocarbon 35, 215-30.

148 The Holocene 15 (2005) Tarasov, P.E., Guiot, J., Cheddadi, R., Andreev, A.A., Bezusko, L.G., Blyakharchuk, T.A., Dorofeyuk, N.I., Filimonova, L.V., Volkova, V.S. and Zernitskaya, V.P. 1999: Climate in northern Eurasia 6000 years ago reconstructed from pollen data. Earth and Planetary Science Letters 17(1/4), 635-45. Tarasov, P.E., Webb, T. III, T., Andreev, A.A., Afanas'eva, N.B., Berezina, N.A., Bezusko, L.G., Blyakharchuk, T.A., Bolikhovskaya, N.S., Cheddadi, R., Chernavskaya, M.M., Chernova, G.M., Dorofeyuk, N.I., Dirksen, V.G., Elina, G.A., Filimonova, L.V., Glebov, F.Z., Guiot, J., Gunova, V.S., Hafrison, S.P., Jolly, D., Khomutova, V.I., Kvavadze, E.V., Osipova, I.M., Panova, N.K., Prentice, I.C., Saarse, L., Sevastyanov, D.V., Volkova, V.S. and Zernitskaya, V.P. 1998: Present-day and mid-Holocene biomes reconstructed from pollen

and plant macrofossil data from the former Soviet Union and Mongolia. Journal of Biogeography 25, 1029-53. T6rnqvist, T.E. and Bierkens, M.C. 1994: How smooth should the curves be for calibrating radiocarbon ages? Radiocarbon 36, 11-26. Vassiljev, J. and Harrison, S.P. 1998: Simulating the Holocene lakelevel record of lake Bysjon, southern Sweden. Quaternary Research 44, 62-71. Wolfe, B.B., Edwards, T.W.D., Jiang, H., MacDonald, G.M., Gervais, B.R. and Snyder, J.A. 2003: Effect of varying oceanicity on early-to mid-Holocene palaeohydrology, Kola Peninsula, Russia: isotopic evidence from treeline lakes. The Holocene 13, 153-60. Yakovlev, B.A. 1961: Climate of the Murmansk region. Murmansk: Murmanskoe knizhnoe izdatel'stvo (in Russian).