ISSN: (Print) (Online) Journal homepage:

Grana ISSN: 0017-3134 (Print) 1651-2049 (Online) Journal homepage: http://www.tandfonline.com/loi/sgra20 Annual variations in pollen deposition and ...
Author: Esther Anderson
0 downloads 2 Views 1MB Size
Grana

ISSN: 0017-3134 (Print) 1651-2049 (Online) Journal homepage: http://www.tandfonline.com/loi/sgra20

Annual variations in pollen deposition and meteorological conditions on the fell Aakenustunturi in northern Finland: Potential for using fossil pollen as a climate proxy Jyrki Autio & Sheila Hicks To cite this article: Jyrki Autio & Sheila Hicks (2004) Annual variations in pollen deposition and meteorological conditions on the fell Aakenustunturi in northern Finland: Potential for using fossil pollen as a climate proxy , Grana, 43:1, 31-47, DOI: 10.1080/00173130310017409 To link to this article: http://dx.doi.org/10.1080/00173130310017409

Published online: 21 May 2010.

Submit your article to this journal

Article views: 374

View related articles

Citing articles: 45 View citing articles

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=sgra20 Download by: [37.44.207.126]

Date: 18 January 2017, At: 02:53

Grana 43: 31–47, 2004

Annual variations in pollen deposition and meteorological conditions on the fell Aakenustunturi in northern Finland: Potential for using fossil pollen as a climate proxy JYRKI AUTIO and SHEILA HICKS

Autio, J. & Hicks, S. 2004. Annual variations in pollen deposition and meteorological conditions on the fell Aakenustunturi in northern Finland: Potential for using fossil pollen as a climate proxy. – Grana 43: 31–47. ISSN 0017-3134. Annual variation in meteorological parameters (7 years) and pollen deposition (6 years) for 4 sites on a transect across an altitudinal timberline on Aakenus, a fell in northern Finland were monitored, in order to see how pollen production, as reflected by pollen deposition (grains cm22 year21), is related to climate conditions. Wind direction and wind speed before and after estimated flowering time were determined. These indicated that, within the forest and at the physiognomic forest line, pollen deposition is primarily from plants growing within the forested area on the fell and that the contribution of windblown pollen from further south is minimal. Pollen deposition can, therefore, be taken as equivalent to pollen production. Simple linear correlation coefficients between pollen deposition and one-month and two-week mean temperature, effective temperature sums and cumulative effective temperature sums for the current and previous summers were calculated. For Pinus sylvestris the quantity of pollen deposited is affected by July mean temperature, July effective temperature sum and total effective temperature sum, for the year previous to pollen emission. Pollen deposition of Betula pubescens and Picea abies is also affected by thermal factors but of different time periods (Betula of early June and Picea of early July), but always of the year previous to flowering. However, the correlation is not as strong as for Pinus. The results suggest that annual fossil Pinus pollen quantities, if calculated from peat and/or lake sediments, are a potential climate proxy. They also demonstrate that it is possible to use temperature parameters of the current year to make forecasts of the intensity of flowering and pollen production in the following year. Jyrki Autio (Corresponding Author), Department of Geography; Sheila Hicks, Institute of Geosciences; University of Oulu, Box 3000, 90014 Oulu, Finland. E-mail: [email protected]; [email protected] (Manuscript received 4 February 2003; accepted 6 June 2003)

Investigations into the relationship between meteorological parameters and the growth factors of arboreal species frequently concentrate on areas where these species are close to the limit of their distribution, because it is in such situations that responses to changes in climate are most sensitively recorded. For this reason many tree growth factors have been investigated in timberline situations together with changes in the location of the forest and the tree line itself (Hustich 1948, 1983, Holtmeier 1974, 2000, Juntunen et al. 2002, Kullman 1990, 2000). The timberline is defined here as the transitional zone between forested and treeless areas (Heikkinen et al. 1995). The health and vigorous growth of a tree can be followed on an annual time scale from its height increment, ring width, and needle/leaf production. Two additional aspects of tree-growth which are more visible when walking through the forest are the abundance of flowers (including pollen production) and of catkins/cones (including seed production). Although foresters have traditionally recorded female flowering and seed production, less attention has been given to male flowering and pollen production. Pollen production is nevertheless a rewarding parameter in several respects. Pollen is produced # 2004 Taylor & Francis. ISSN 0017-3134 DOI: 10.1080/00173130310017409

in larger quantities than seed, it is dispersed more widely, and due to its composition (highly resistant sporopollenin), it is well preserved in wet, anaerobic environments (mires and lake sediments). There is a greater potential for obtaining a continuous record of pollen production going back thousands of years than there is for seeds. This affords the possibility of following changes in pollen production at the timberline in a way which shows long-term trends and short-term variations. At the same time, it is becoming evident (Hicks 1996, Hicks 1999, McCarroll et al. 2003) that the main factor which determines the amount of pollen that a tree produces in northern situations is temperature. If, therefore, the relationship between pollen production and temperature could be established empirically, this would be a contribution to reconstructing past climate more objectively. A pilot situation for evaluating this pollen-temperature correlation at a high level of precision was made possible through on-going research on the fell, Aakenustunturi, in Finnish Lapland, involving a well established series of four meteorological stations (measuring primarily temperature) arranged on an altitudinal transect across the timberline. Pollen production is difficult to measure directly, but annual Grana 43 (2004)

32

J. Autio and S. Hicks

pollen deposition can be measured fairly easily. The difference between the two is that deposited pollen contains both pollen produced locally by plants in the immediate surroundings and pollen blown in from elsewhere, from up to tens of kilometres away, while at the same time some proportion of the locally produced pollen may be blown away from the area and not deposited there at all. Under certain circumstances, however (within closed or partly closed forest), annual variations in pollen deposition give a good record of annual variations in pollen production. As part of the EU-funded project FOREST (ENV ENV4CT95-0063), traps for monitoring pollen deposition were located in the immediate vicinity of the Aakenustunturi meteorological stations. This monitoring arrangement provides an annual record of temperature and pollen in a sensitive timberline situation. Earlier studies on Aakenustunturi have concentrated on thermal conditions on the fell (Autio 1995, Autio et al. 1998, Autio & Heikkinen 2002). Annual temperature data were collected continuously from May 1994 to September 1998 and in the summer time between 1999 and 2001. The different plant communities, altitudinal vegetation zones and timberlines have also been studied along the same transect

(Autio 1997) and the physiognomic forest line has been determined for the whole fell area. Whereas earlier publications have focused on relating timberlines, as a visible feature on the fell, to climate, this paper is concerned with the correlation between meteorological parameters and the pollen production of the trees which grow in the timberline region.

MATERIAL AND METHODS Site characteristics Aakenustunturi (570 m a.s.l.) is located in the municipality of Kittila¨, western Finnish Lapland (Fig. 1). Phytogeographically the area falls into the northern boreal zone (Ahti et al. 1968). Four meteorological stations and four pollen traps, one adjacent to each station, have been placed at different altitudes along the transect across Vareslaki (485 m a.s.l.), the western summit of the fell (Table I). The position of the continuous or almost continuous forest limit (i.e. physiognomic forest line) varies considerably on Vareslaki. The timberline terminology used is according to Hustich (1966). This terminology is well-known and appreciated (Tuhkanen 1999). The physiognomic forest line was drawn along the line where the canopy is c. 30 percent (see Heikkinen et al. 2002), and its position was

Fig. 1. A. Position of Aakenustunturi within Fennoscandia together with the major vegetation zones: OA – oroarctic, NB – northern boreal and MB – middle boreal. B. Topographic map of Aakenustunturi. The limit of commercial forest is also indicated. The category ‘saplings’ includes three tree species: Betula pubescens, Picea abies and Pinus sylvestris. Grana 43 (2004)

Pollen deposition and meteorological conditions on Aakenustunturi

33

Table I. Basic details of the four temperature monitoring sites. Monitoring site

A60

A61

A62

A63

Lat./Long.

67‡42’04@/24‡25’04"

67‡41’52@/24‡25’08"

67‡41’31@/24‡25’14@

67‡41’19@/24‡25’22@

Altitude (m a.s.l.)

330

390

482

430

Aspect

N

N

Summit

S

Timberline region

Closed forest/ Economic forest line

Physiognomic forest line

Open fell summit/ tree line

Physiognomic forest line

Vegetation type

HMT

sEMT

haMLT

EMT

Percentage cover of trees Betula pubescens Picea abies Pinus sylvestris

5 35 0

7 15 1

0 0 0

5 15 0

Total percentage cover of trees Stoniness (%)

40 0

23 15

0 45

20 7

obtained in the field using a GPS receiver. It crosses the transect at an altitude of 380 m on the north-facing slope, but is about 40 m higher on the south-facing slope. In both situations the line is formed mainly by Norway spruce (Picea abies (L.) Karst). The temperature and pollen monitoring sites include a range of situations within the timberline. Starting in closed forest, the transect crosses the economic forest line and continues upwards through the physiognomic forest line to the treeless or almost treeless area – the tree line. Monitoring site A60 is situated within mature closed Norway spruce-dominated forest on the north-facing slope at an altitude of 330 m a.s.l. The vegetation type is HMT (Hylocomium-Myrtillus type; Table I). The classification system for forest types follows basically the definitions of Cajander (1925), while that for the heath vegetation follows Haapasaari (1988). The second monitoring site, A61, lies just above the physiognomic forest line dominated by Norway spruce and has an sEMT vegetation (subalpine Empetrum-Myrtillus type). It is at a distance of 300 m from A60. Monitoring site A62 is situated in the almost treeless area on the top of Vareslaki, where the heath vegetation type is haMLT (hemiarctic Myrtillus-Lichenes type). The fourth monitoring site, A63, is situated near the physiognomic forest line on the southfacing slope, at an altitude of 430 m a.s.l., and the forest type around the site is EMT (Empetrum-Myrtillus type). The distance from the treeless site, A62, to the physiognomic forest line is 580 m (A61, northern slope) and 420 m (A63, southern slope).

Field measurements Meteorology The results used here apply to air temperature during the summer season (June – August) in the years 1995 – 2001, monitored at onehour intervals by means of data loggers at a height of 2 m above the ground surface. The data loggers were covered with a white plastic cylinder to protect them from direct solar radiation, the cylinder having holes in it to maintain ventilation. Monthly (Table II) and two-week mean temperatures were determined from the daily mean temperatures, which were in turn defined as the means of the hourly values. Temperature data are missing for the open fell top, monitoring site A62, in 1995, and were compensated for in the present work by a linear regression based on the corresponding temperature data for the two nearest monitoring sites, A61 and A63. The length of the growing season and the effective temperature sum (threshold value for daily mean temperature

z5‡C) were also determined (Table III). Monthly mean temperatures and thermal sums were compared with those for the normal period 1961 – 1990, which were calculated using the program of Ojansuu & Henttonen (1983), which estimates climate parameters for given locations on the basis of Finnish Meteorological Institute data. This program takes into account the altitude of the site, its distance from the sea and the presence of lakes in the surrounding area. It, therefore, gives a good picture of temperature conditions for the sites in question, although it might underestimate those for the south-facing monitoring site, A63. Wind direction and wind speed records were obtained from the Finnish Meteorological Institute and apply to the nearest official meteorological station, Muonio (245 m a.s.l.), situated about 45 km north-west of Aakenustunturi. Wind direction and velocity were measured at a height of 10 m above the ground every third hour as a 10-minute mean value. The wind measurement results presented in this study (Fig. 2) cover 4-week periods in each year, each period extending from 2 weeks before until 2 weeks after the time of anthesis of each of the three tree species commonly growing on the fell (Table IV). Wind records from another nearby fell, Ylla¨s (718 m a.s.l.), about 10 kilometres north-west of Aakenustunturi, were compared with those from Muonio, and the wind direction results proved to be concurrent. The wind data for Muonio were used here, however, because they provided a complete record up until 2001 while the records for Ylla¨s extend only until 1999. In order to delimit the 4-week periods mentioned above, anthesis (pollen emission) was defined separately for birch (Betula pubescens Ehrh.), Norway spruce (Picea abies (L.) Karst) and Scots pine (Pinus sylvestris L.) on the basis of the effective temperature sum threshold for flowering (Table IV). The timing of anthesis, which varies for each tree species, was adapted from Luomajoki (1993 a, 1993 b, 1999). The reason for looking at the dominant wind directions in the period 2 weeks before and 2 weeks after flowering was in order to assess the degree to which the pollen deposition samples contained pollen from beyond the area of the fell forests (i.e. pollen not reflecting the climate conditions being monitored). The strength of the wind also gives some indication of the extent of pollen dispersal to and from the monitoring stations, as it is only pollen produced in the vicinity of the temperature measuring station that can be expected to correlate with local meteorological conditions. The transect contains one situation, the fell summit (A62), where no trees are growing. Tree pollen being deposited at this site must have come from a minimum of 420 m away (the physiognomic forest line) but Grana 43 (2004)

34

J. Autio and S. Hicks

Fig. 2. Percentage distribution of mean wind direction in the period 2 weeks before and 2 weeks after the flowering of Betula, Picea and Pinus respectively, based on data from 1996 – 2001 at Muonio. The length of calm as a percentage is also shown. more likely from anything between 500 m and tens of kilometres away. A correlation with the meterological conditions monitored at A62 is, therefore, unlikely. Grana 43 (2004)

Pollen deposition Annual pollen deposition was measured by means of modified ‘Tauber traps’ following the standard procedure of the PMP (Pollen

Pollen deposition and meteorological conditions on Aakenustunturi

35

Table II. Mean air temperatures (‡C) and frost days in June at a height of 2 m at the sites A60 – A63 during the summer between 1995 and 2001. 30-years average temperatures were calculated by using Ojansuu & Henttonen’s climate model (1983). Mean Temperature

Mean Daily Minimum

Mean Daily Maximum

Aug

Jun

Jul

Aug

Frost Days in June

6.9 7.1 6.1 6.4

5.9 6.6 5.8 6.1

15.1 15.0 14.7 16.8

15.2 14.9 13.8 15.7

14.1 14.2 13.4 14.7

0 0 0 0

5.4 4.8 4.4 4.5

8.2 8.1 7.7 8.0

9.1 9.1 9.2 8.8

14.5 12.4 11.0 12.9

15.7 16.4 14.6 16.2

17.0 17.0 17.4 18.6

0 2 3 3

13.1 12.7 12.3 13.1

6.8 6.3 6.2 6.0

11.4 10.7 10.5 10.8

9.0 8.4 8.3 8.3

17.4 17.4 17.0 16.8

20.5 20.3 19.3 20.3

17.4 17.5 16.8 18.4

1 1 2 2

14.1 13.8 12.9 13.7

9.5 9.1 8.4 8.9

3.6 4.3 4.4 4.1

10.3 10.2 10.2 10.3

6.8 6.5 6.4 6.5

13.2 13.2 11.2 12.3

18.2 18.8 16.1 17.6

12.6 12.5 10.8 12.0

3 2 2 2

12.8 13.2 12.4 13.3

13.0 13.1 12.6 13.3

8.2 8.3 8.1 8.9

8.6 8.7 8.3 8.3

9.3 9.0 8.7 8.7

4.9 4.8 4.8 4.8

17.2 18.2 16.0 18.2

17.3 17.9 16.6 18.3

11.8 12.3 11.6 14.2

0 0 0 0

A60 A61 A62 A63

9.9 9.8 9.6 10.9

14.4 14.7 14.5 16.0

11.1 11.4 11.3 12.5

6.0 6.1 5.8 5.8

9.8 10.1 10.2 9.8

7.1 7.2 7.6 7.4

14.1 14.5 13.9 17.8

19.1 20.0 18.9 23.5

15.5 16.4 15.9 20.8

0 0 0 0

2001

A60 A61 A62 A63

12.9 12.7 11.6 12.8

13.5 13.4 12.5 13.4

11.2 11.3 10.7 11.5

7.9 8.1 7.7 7.5

9.4 9.4 8.8 8.5

7.4 7.7 7.4 7.2

17.9 18.8 15.7 18.7

17.4 18.0 16.0 18.3

15.2 15.8 14.3 16.4

0 1 2 1

7-years average

A60 A61 A62 A63

11.1 10.7 10.0 11.0

13.4 13.3 12.5 13.5

10.9 10.8 10.5 11.2

6.5 6.5 6.3 6.2

9.3 9.2 8.9 8.9

7.2 7.2 7.1 7.0

15.6 15.6 14.2 16.2

17.6 18.0 16.5 18.6

14.8 15.1 14.3 16.4

0.6 0.9 1.3 1.1

30-years average

A60 A61 A62 A63

10.7 10.5 10.0 10.3

13.2 13.0 12.6 12.9

10.5 10.3 9.9 10.1

Year

Sites

Jun

Jul

Aug

Jun

1995

A60 A61 A62 A63

11.0 11.0 9.9 11.8

11.0 11.0 9.9 10.9

9.7 10.2 9.6 10.1

6.9 7.1 7.2 7.2

1996

A60 A61 A62 A63

9.7 8.4 7.6 8.5

12.0 11.8 10.9 12.0

13.6 12.9 13.1 13.5

1997

A60 A61 A62 A63

12.3 11.6 11.4 11.6

15.6 15.2 14.1 15.2

1998

A60 A61 A62 A63

9.0 8.3 7.5 8.1

1999

A60 A61 A62 A63

2000

Monitoring Programme; Hicks et al. 1996, 1999, http://wdc. obs-mip.fr/paleo/pmp/pmp.html). The criteria for the location of the pollen traps were that each should be in the immediate vicinity of a meteorological recording site (Table I). Thus one (A60) records pollen deposition actually within the forest, two (A61 and A63) at the altitudinal forest limit (i.e. open in one direction) and the fourth on the exposed fell summit (A62). The traps were in the field the whole year round, the collection being made at the end of the flowering season in September/October. Results are available for the 6-year period 1996 – 2001 inclusive, i.e. for one year less than the meteorological results. Unfortunately, no data are available for monitoring site A63 in 1998 and site A62 in 2000. The yearly content of each pollen trap was treated in the laboratory. Before treatment began, a set number of Lycopodium tablets, 1 for the year 1996 and 5 for the years 1997 – 2001, were added to the trap contents in

Jul

order to enable total pollen to be calculated (Stockmarr 1971). Laboratory treatment included filtering followed by dissolving the filter paper in an acetolysis mixture, heating for 2 minutes in a water bath, rinsing with acetic acid followed by treatment with 10% KOH, and then mounting in silicone oil. Each pollen grain was identified, and counting for each sample was continued until a pollen sum of 500 arboreal grains was achieved. The results are expressed as annual pollen deposition (grains cm22 year21). The diagram in Fig. 3 has been produced using the Tilia and Tilia.graph programs (Grimm 1992).

Statistical analyses Simple linear correlation coefficients (r-values) between pollen deposition and mean monthly temperatures and effective temperature sums, Grana 43 (2004)

36

J. Autio and S. Hicks

and also two-week mean temperatures and cumulative effective temperature sums for the current and previous summers (June to August) were calculated. The pollen and temperature series are rather short for a parametric correlation, and so the non-parametric Spearman correlation coefficient was used.

RESULTS Variations in temperature The yearly variation in temperature over the period 1995 – 2001 was greater than the variations between stations in individual years (Table II), the maximum difference between monthly mean temperatures being 5.4‡C, between the August results for the forested monitoring site A60 in 1996 and 1999, whereas the maximum difference between two monitoring sites in the same summer was recorded in June 1996, when the fell top monitoring site A62 was 2.1‡C colder than monitoring site A60. Considerable fluctuations in temperatures could occur within one summer, however, as in 1996, when the early summer was cool but the late summer the hottest of all. On the other hand, conditions in the summers of 1997 and 2001 were consistently warm throughout, with all the monthly means well above those for the normal period 1961 – 1990. Summer 2000 was also warmer than normal as far as July and August were concerned. The coldest June during the period occurred in 1998 at all the monitoring sites, the coldest July in 1995 and the coldest August in 1999. Night frosts were recorded in June in four summers, 1996 – 1998 and 2001, at the physiognomic forest line monitoring sites A61, A63 and the fell top A62, but only in two summers, 1997 and 1998, at the forest monitoring site A60. The mean monthly temperatures calculated over the 7-year period 1995 – 2001 do not differ greatly between the monitoring sites, although the coldest monthly means of all were recorded on the fell top A62, which was on average 1.1‡C cooler than the forest monitoring site A60 in June, while the highest mean temperatures in July and August were achieved at the south-facing physiognomic forest line monitoring site A63. All in all, the years 1995 – 2001 were somewhat warmer than the normal period 1961 – 1990 (note however, that the model used for calculating the norm may underestimate temperature conditions for A63). The most extreme temperature conditions prevailed at the south-facing physiognomic forest line monitoring site A63 (see Autio 1995, Autio et al. 1998) and the most stable at the fell top monitoring site A62, where the almost constant wind evens out the temperatures and lowers them somewhat, eliminating the most prominent peaks, so that the mean maximum temperatures tended to be several degrees lower there than at the forest limit on the south-facing slope. Growing season and effective temperature sum As with mean temperatures, wide annual variations are also seen in the length of the growing season (Table III). A length well in excess of that defined for the normal period, reaching 156 days, was recorded in the year 2000, when the exceptionally mild autumn meant that it came to an end Grana 43 (2004)

almost four weeks later than normal (Ilmastokatsaus 2000), while the shortest growing season was that of 1998, lasting for 93 days at the fell top monitoring site A62, 95 days at the physiognomic forest line monitoring sites A61 on the north-facing slope and A63 on the south-facing slope, and 112 days at the forest monitoring site A60. Thus the growing season on the fell top was 63 days shorter in 1998 than in 2000. The greatest difference between stations in the length of the growing season in the same year was achieved in 1995, when it was 20 days shorter at monitoring site A62 than at monitoring site A63. The mean length of growing season was 114 days on the fell top, about 10 days longer than this in the forest (123.7 days), and somewhat similar to the latter, 121 days, on the south-facing and north-facing slopes. The mean date for the beginning of the growing season over this period was 30.5. at the forest limit and on the fell top, but three days earlier in the forest, while that for the end of the season was 27.9. at all the monitoring sites except A62, where it was 21.9. The best growing seasons in terms of the effective temperature sum were those of 1997, 2000 and 2001. However the mean effective temperature sum for the whole period was higher than that for the normal period at all the monitoring sites, being particularly high at monitoring site A63, which possessed the most favourable growing conditions. The only years in which temperature sums lower than those for the normal period were obtained were 1995, 1996 and 1998. As with mean temperatures and the lengths of the growing seasons, the variation in temperature sums between years was considerable, over 300 d.d. at most, whereas the maximum difference between stations in the same year was just over 130 d.d., between monitoring sites A62 and A60 in 1998. A longer than normal growing season does not necessarily mean that the effective temperature sum will be higher than normal. Thus, although the growing season of 1999 was longer than that of 1997 at all the stations, the effective temperature sum in the latter year was substantially greater.

Pollen deposition The pollen deposition values (grains cm22 year21) for the major tree, shrub and dwarf shrub taxa are illustrated in Fig. 3, along with figures for the herbs (grouped according to their ecology), charcoal and spherical carbonaceous particles (SCP~flyash). Previous work (Jacobson & Bradshaw 1981, Prentice 1985, 1988, Sugita 1993, 1994, Sugita et al. 1997, Jackson & Lyford 1999, Davis 2000, Hicks 2001) has demonstrated that the effective source area of pollen deposited on the ground is very much greater in open treeless situations than within a forest. For this reason, the pollen assemblage in each trap can be expected to reflect the vegetation situation for a different area, which will be spatially small for the within-forest trap (A60) but could cover several kilometres for the summit trap (A62). It is also known, however, that the vast majority of the pollen emitted by a plant comes to the ground fairly close to its origin and a much smaller proportion is transported further afield. This means that numerically much more pollen can be

Pollen deposition and meteorological conditions on Aakenustunturi

37

Table III. Effective temperature sums and duration of the growing seasons at a height of 2 m at the sites A60 – A63 during the summer between 1995 and 2001. 30-years average effective temperature sums were calculated by using Ojansuu & Henttonen’s (1983) climate model.

Sites

Effective temp. sum d.d.

Start of growing season

End of growing season

Length of growing season

1995

A60 A61 A62 A63

546.4 576.0 526.1 603.2

24.5 24.5 24.5 24.5

22.9 22.9 5.9 22.9

122 122 102 122

1996

A60 A61 A62 A63

650.8 598.1 529.1 608.7

31.5 2.6 2.6 2.6

21.9 23.9 17.0 21.9

114 114 108 112

1997

A60 A61 A62 A63

886.4 807.3 763.9 821.0

5.6 5.6 5.6 5.6

26.9 26.9 12.9 26.9

114 114 100 114

1998

A60 A61 A62 A63

591.4 528.6 455.4 513.0

3.6 20.6 21.6 20.6

22.9 22.9 21.9 22.9

112 95 93 95

1999

A60 A61 A62 A63

690.5 710.3 666.3 755.7

19.5 19.5 19.5 19.5

28.9 28.9 22.9 28.9

133 133 127 133

2000

A60 A61 A62 A63

761.2 785.9 746.9 917.9

17.5 17.5 17.5 17.5

19.10 19.10 19.10 19.10

156 156 156 156

2001

A60 A61 A62 A63

820.9 813.5 712.3 807.7

30.5 30.5 30.5 30.5

21.9 21.9 21.9 21.9

115 115 115 115

7-years average

A60 A61 A62 A63

706.8 688.5 628.6 718.2

27.5 30.5 30.5 30.5

27.9 27.9 21.9 27.9

123.7 121.3 114.4 121.0

30-years average

A60 A61 A62 A63

690.0 663.0 613.0 644.0

expected in the within-forest trap (A60) than in the summit trap (A62). The trap on the open summit (A62) does, in most years, receive smaller amounts of tree and shrub pollen than the other traps but the highest amounts of charcoal and flyash. With respect to the charcoal it should be mentioned that many walkers light fires on the open summit for preparing food. It is also clear that there is great variation in pollen deposition from year to year. In the case of Betula pubescens type pollen, for example, the years 1996, 1998 and 2000 have higher values at every site and the intervening years lower values. This apparent biennial rhythm of birch pollen production has been noticed elsewhere (Ja¨ger et al. 1991). 1998 was a ‘big’ pollen year for all the tree taxa at all the traps (unfortunately no records for A63 in this year),

20.5

30-years mean from Munio 20.9

124.0

and was an extremely high pollen year for Betula in the within-forest trap (A60) and the only substantial one for Picea. In contrast to the other two tree species, Pinus had a high pollen year in 2001, and this was reflected in all the traps except that on the open summit (A62). It is important to note, however, that although high and low years are obvious, the actual quantity of pollen varies between tree species, with Picea producing much less pollen than either Betula or Pinus. Pollen of forest herbs is more abundant at the within-forest trap (A60) than at either the forest limit or the open forest sites, while the pollen of Juniperus and the dwarf shrubs is most abundant in the south-facing physiognomic forest line trap (A63). Two features emerge from the results, (1) for the majority of individual years there is a fall in the amount of tree pollen with Grana 43 (2004)

38

J. Autio and S. Hicks

Fig. 3. Annual pollen deposition as number of grains cm22 of ground surface at each trap location. The six years for which results are available are shown individually. The taxa recorded are either shown individually or grouped ecologically (note differences in scale). The annual deposition of charcoal (in two size classes) and spherical carbonaceous particles (fly ash) are also illustrated.

movement away from the forest limit towards the unforested area (emphasized when the average value for the six years is considered) and (2) there is a big variation in pollen deposition from one year to the next, which is consistent at all sites. Feature (1) suggests that what is being recorded is indeed largely pollen production of the trees growing on the fell. In a publication based on pollen trap results from Finnish Lapland, Hicks (2001) presented threshold values of annual pollen deposition for three tree species, which she related to the abundance of the trees within different distances from the site of pollen deposition. The value for Pinus was 300 – 500 grains cm22 year21 when the tree was not present within 1km and 500 – 1500 grains cm22 year21 when the tree was sparsely present. The comparable values for Picea were 25 – 50 and 50 – 100 respectively. The 6-year average values for Pinus and Picea at trap A62, on the open summit 420 and 680 m from the forest limit are 1280 grains cm22 year21 for Pinus and 70 for Picea, which is well in keeping with the Lapland results. Feature (2), however, suggests that the actual amount of pollen produced in any one year is affected by a more regional form of control (climate?). It is this second feature of annual variation which will be considered in most detail here.

Wind conditions The wind directions calculated for the period extending from two weeks before to two week after anthesis in each of the tree species show a high degree of similarity (Fig. 2), in spite Grana 43 (2004)

of the fact that the timing of anthesis varied greatly between species (Table IV). The dominant wind direction at Muonio (245 m a.s.l.) during these periods for the various trees species in 1996 – 2001 was from the east, while north-east and northerly winds were also common. Southerly winds were somewhat less frequent, however, with the clear exception of 1999, and the least common of all were westerly winds. Windless conditions prevailed for a mean of 3.4% of the time, and the mean wind speed was 2.6 m/s. The data for Ylla¨s (718 m a.s.l.) differed to the extent that the mean wind speed was almost three times as great, 7.7 m/s, and windless conditions prevailed for only 1% of the time. Relation of pollen deposition to wind conditions The wind directions support the assumption that the pollen deposited in the traps is coming primarily from the fell and its surroundings rather than from a completely different geographical region. For pollen from several kilometres away to form a significantly high proportion of that deposited in the within forest and physiognomic forest line traps, one would need winds blowing from areas where the same trees were already flowering and maybe even flowering more prolifically, i.e. areas further south with more favourable climate conditions. However, southerly winds are exceptional in the present data. In the one year when southerly winds were more common, 1999, tree pollen amounts are generally below average, with the exception of the treeless

Pollen deposition and meteorological conditions on Aakenustunturi

39

Fig. 3. (Cont.)

Table IV. Variation in timing of (the estimated) anthesis for Betula pubescens, Picea abies and Pinus sylvestris at sites A60 – A63 during the period 1996 – 2001. The latest timing of the anthesis occured in the years 1998 and 1996. The effective temperature sum thresholds for anthesis are as follows: Betula 69.7 d.d., Picea 127.7 d.d. and Pinus 198.6 d.d.

Site

Year

Elevation

Site description

Time of Betula anth.

Time of Picea anth.

Time of Pinus anth.

A60

1996 1997 1998 1999 2000 2001

330

Closed forest/ economic forest line North-facing slope

17.6 10.6 25.6 12.6 4.6 2.6

27.6 15.6 1.7 17.6 21.6 19.6

9.7 29.6 10.7 26.6 29.6 28.6

A61

1996 1997 1998 1999 2000 2001

390

Physiognomic forest line North-facing slope

21.6 11.6 29.6 12.6 3.6 2.6

4.7 21.6 7.7 16.6 21.6 19.6

14.7 1.7 13.7 24.6 29.6 28.6

A62

1996 1997 1998 1999 2000 2001

482

Open fell summit/Tree line

24.6 11.6 1.7 13.6 6.6 2.6

8.7 21.6 9.7 18.6 23.6 19.6

22.7 2.7 17.7 27.6 30.6 28.6

A63

1996 1997 1998 1999 2000 2001

430

Physiognomic forest line South-facing slope

21.6 11.6 30.6 11.6 30.5 2.6

4.7 21.6 8.7 16.6 11.6 19.6

15.7 1.7 14.7 23.6 25.6 28.6

Grana 43 (2004)

40 J. Autio and S. Hicks

Grana 43 (2004)

Fig. 4. Cross section of Aakenustunturi showing the four monitoring sites marked (A60 – A63). The inset graphs for each site show annual pollen deposition for the tree taxa and the length of the growing season and effective temperature sum in the year preceding flowering.

Pollen deposition and meteorological conditions on Aakenustunturi fell-top site (A62) where Pinus, and Picea values are higher than in the forest and northern physiognomic forest line traps (but not the southern physiognomic forest line trap). The fact that the winds blowing throughout the period before and after local tree flowering are mostly from the north and east means that they are unlikely to be bringing with them any large quantity of tree pollen from an area with different meteorological parameters. The wind is likely to be least strong within the closed forest (monitoring site A60) and strongest on the open summit (monitoring site A62, cf. the situation at Ylla¨stunturi), while intermediate wind strength can be expected at the forest edge. In view of the Aakenustunturi wind direction results, any pollen transported from outside the fell area is likely, to originate from the east or north rather than the south or west, an assumption that is supported by the fact that pollen deposition was regularly greater at the northfacing physiognomic forest line monitoring site A61 than at the south-facing site A63. In other words, pollen coming from the east and north, whether from the immediate vicinity or from further away, will have had better access to the trap at monitoring site A61 than at site A63, where the summit of the fell will have served as a topographical barrier. At the same time, the dominant wind direction at site A61 will have been from across a forested area, whereas that at site A63 will not only have been from across the treeless summit area but will also have carried away pollen produced at the forest limit. Pollen produced in forests in the lowlands surrounding the fell can be regarded as making up only a small proportion of the total pollen deposition in the forest and physiognomic forest line traps. At site A62 on the fell summit, however, all the tree pollen deposited must have come from elsewhere. The wind directions suggest that primarily the source is forests on the north slope of the fell and further to the north and east. The distance involved could be as little as 580m (the distance from the forest limit) but is probably several kilometres as the 1999 data indicate. The total average quantity of tree pollen on the summit is only slightly less than that at the southern forest limit but generally about half of that at the northern forest limit and one third of that within the forest. This is in keeping with the dominant wind direction.

Relation of pollen deposition to temperature factors Comparison of the pollen deposition figures with temperature conditions in the previous year indicates that each good pollen year, at least for pine, such as 1998 or 2001, was preceded by one with a high effective temperature sum for the growing season (Fig. 4) and high mean temperatures for July (Table II). This relationship is most marked in the case of pine, a good flowering year for which is regularly preceded by warm temperatures and a bad flowering season by cool temperatures. Birch, on the other hand, would seem to be capable of a good pollen yield only every second year, but this will be enhanced if the good year happens to follow one with a warm July mean temperature and a high effective temperature sum and reduced by less favourable conditions. Spruce pollen

41

production was low in all the years, except at monitoring sites A60 and A61 in 1998. Pollen deposition does not seem to be influenced by temperatures during the same summer. Thermal conditions in 1998, a good pollen year, are the least favourable of the whole series. The early summer was far cooler than normal, there were several night frosts in June, the growing season began later than usual, particularly at the physiognomic forest line and on the summit, and the effective temperature sum fell well short of the normal figure. The length of the growing season would similarly not appear to have any obvious effect on pollen deposition. Of the sites studied here, station A63, at the physiognomic forest line on the south-facing slope of the fell, had the most favourable meteorological conditions, but pollen deposition in this trap was lower than at site A61, at the physiognomic forest line on the north-facing slope, or at site A60, in the forest. The above observations that pollen deposition in a given flowering season was influenced by the previous summer’s July mean temperature and effective temperature sum and that this effect was most noticeable in the case of pine were confirmed by the correlation analyses (Fig. 5 A). The deposition of pine pollen at site A60 showed statistically significant correlations with the mean temperature and temperature sum for the July of the preceding year (n~6, r~0.83, pv0.05) and with the temperature sum for the whole of the preceding growing season (n~6, r~1.00, pv0.01), while that at site A61 gave a significant correlation only with the effective temperature sum for the whole summer (n~6, r~0.89, pv0.05). No statistically significant relation between the deposition of pine pollen and the indicators of temperature could be perceived at the other sites. Birch pollen deposition had statistically significant correlations only with the June mean temperature in the preceding year and the effective temperature sum for the preceding growing season, and then only in the case of site A61 (n~6, r~0.83, pv0.05) (Fig. 5 A), while spruce pollen deposition did not show any statistically significant relations to temperature variables (Fig. 5 A). The coefficients for the correlations between pollen deposition and temperature conditions in the same year were low throughout, and no positive relationships were detectable between any of the tree species and any of the temperature variables (Fig. 5 B). Significant negative relationships, on the other hand, were observed between the effective temperature sum for the growing season and the deposition of pine pollen at site A62 (n~6, r~20.89, pv0.05) and between both the mean temperature and the effective temperature sum in June and the deposition of birch pollen at site A60 (Fig. 5 B; n~6, r~20.89, pv0.05). Statistically significant correlations between the two-week temperature variables and the deposition of pine pollen were found only in the case of the values for the late summer of the preceding year at monitoring site A60 (Fig. 6) and those for the middle of the summer at station A63. Birch pollen deposition was statistically significantly correlated with twoweek mean temperatures in the early summer of the preceding year at site A61 (n~6, r~0.83, pv0.05) (Fig. 6), and Grana 43 (2004)

42

J. Autio and S. Hicks

Fig. 5. A & B. Correlation coefficients between Pinus sylvestris, Betula pubescens and Picea abies pollen deposition and the mean monthly temperature, monthly effective temperature sum and total effective temperature sum from the four (A60 – A63) monitoring sites: (A) of the previous year; (B) of the current year. Values significant at pv0.05 are marked grey and pv0.01 black.

the temperature sum during the period 1.6. – 13.7. at monitoring sites A60 and A61. Spruce pollen deposition was statistically significantly correlated with the temperature sum during the period 1.6. – 13.7. in the preceding year at site A60 (n~6, r~0.80, pv0.05) (Fig. 6). On the other hand, a statistically significant negative correlation was found between spruce pollen deposition and the mean temperature for the period 15 – 29.6. at monitoring site A61 (n~6, r~20.80, pv0.05). The correlations of pollen deposition with the temperature variables calculated on a two-week basis were qualitatively similar to those for the monthly variables, and again no Grana 43 (2004)

statistically significant positive correlations with variables for the year of flowering were detectable. These results are therefore not presented here in diagram form. Examination of the correlations between the deposition of pollen of each tree species and the two-week temperature variables in terms of the sums of deposition at all four sites taken together (n~22) pointed to a clear temporal sequence for the species, in that birch pollen appeared to be statistically significantly correlated with temperature variables in the early summer, spruce with those for the midsummer period and pine with the late summer variables of the year previous to flowering (Fig. 7).

Pollen deposition and meteorological conditions on Aakenustunturi

43

Fig. 5. (Cont.)

DISCUSSION AND CONCLUSIONS The results obtained here are consistent with those of Hicks (1999) and McCarroll et al. (in press), who found that the deposition of pine pollen was affected by temperatures during the previous summer. No such obvious relationship was detectable for spruce pollen, however, partly because the amounts deposited were small, a fact which in itself is in keeping with earlier findings that spruce produces less pollen than birch or pine (Hicks 2001) and has fewer prolific flowering years in this northern situation. The fact that many of the spruces in the region were infected with the rust fungus,

Chrysomyxa ledi, in 1996 and 1999 only serves to compound this effect. Birch appears to achieve good pollen production every second year, as observed earlier. The present results should be viewed with certain minor reservations, however, largely for two reasons. Firstly, the birch, spruce and pine pollen deposition data cover only six years and the temperature data seven, and secondly, pollen deposition data are lacking for monitoring site A62 in 2000 and site A63 in 1998. A consideration of how wind conditions influence pollen deposition leads to the conclusion that the majority of the pollen originates from the forest on the fell, so that pollen Grana 43 (2004)

44

J. Autio and S. Hicks

Fig. 6. Correlation coefficients between Pinus sylvestris, Betula pubescens and Picea abies pollen deposition of the current year and 2 weeks time periods for mean temperatures and effective temperature sums and cumulative temperature sums of the previous year from the four (A60 – A63) monitoring sites. Values significant at pv0.05 are marked grey and pv0.01 black.

Grana 43 (2004)

Pollen deposition and meteorological conditions on Aakenustunturi

45

Fig. 7. Correlation coefficient between pollen deposition of the three tree species and 2 weeks time periods for different temperature parameters. Values significant at pv0.05 are marked grey and pv0.01 black. Data are for all four sites combined.

deposition does indeed reflect local pollen production. It is, therefore, relevant to look at the relationship between local temperature and pollen deposition. The statistically significant correlation between pollen deposition in a given flowering season and local temperature variables for the preceding summer, (especially at monitoring site A60), further confirmed the concept that within the forest and at the physiognomic forest line the deposited pollen originates from the immediate vicinity. It is, therefore, understandable that the tree pollen deposited at the fell summit site, A62, did not correlate significantly with temperatures at the same site, i.e. the tree pollen entering the traps there has come from much further away (up to tens of kilometres). Of the correlations between pollen deposition and temperature variables representing various lengths of time during the summer of flowering and the preceding summer, it was for the most part only the variables for the preceding summer that showed statistically significant correlations with pollen deposition, and scarcely ever with the variables for the summer of flowering. It was also noticeable that the variables calculated on a two-week basis were more suitable for this purpose than monthly variables, and provided more accurate information on which part of the previous year’s growing season was relevant for pollen production in each tree species. Examined in this way, combined data for all four sites served to indicate that each species has its own specific temperature period during the preceding summer which influences its pollen production. Total deposition of birch, spruce and pine pollen was naturally highest in the closed forest and lowest on the open fell top, while the figures for the physiognomic forest line lay in between these. In the case of pine pollen, however, there were some years (1996, 1997, 2001) in which deposition at the physiognomic forest line on the north and south-facing slopes of the fell was higher than in the forest. Pinus pollen, being smaller in size and lighter than Picea pollen is more

easily dispersed at the physiognomic forest line. The same feature is seen for Betula pollen (which is equally well dispersed) in 1996 and 2000. The importance of wind direction was emphasized at the physiognomic forest line, where pollen deposition was greater on the north-facing slope, since the prevailing winds at the time of anthesis were from the forest towards the trap, whereas on the southfacing slope they were blowing away from the trap. This is one of the very few instances where correlations have been sought between temperature factors and pollen deposition at exactly the same spot, and the results clearly indicate a relationship between them. This opens up possibilities for the quantitative reconstruction of past temperatures at a high temporal resolution (annual), potentially extending back before the time of instrumental records. If the same annual pollen deposition variations can be extracted from sediments (fast-growing peat profiles or annually laminated lake sediments), the pollen record can potentially provide a proxy for summer temperatures which could be set alongside the dendroecological record. Longer palaeoseries of this nature would enable a distinction to be made between long-term trends, short-term variations and abrupt changes, which are not always covered by shorter instrumental series. In addition, if close correlations can be discerned between pollen deposition and temperature factors during the preceding summer, it would be possible, in principle, to make annual forecasts of the intensity of flowering and pollen production in the coming year. This could be of use to people suffering from pollen allergies, particularly in the case of birch. ACKNOWLEDGEMENTS Among the numerous people who have helped in the gathering of the field data, we would especially like to express our thanks to Pertti Vuolteenaho and Pekka Kauppila. Thanks are also due to Raija-Liisa Huttunen for counting the pollen and to Anja Kaunisoja Grana 43 (2004)

46

J. Autio and S. Hicks

for drawing several of the figures and maps. We would additionally like to thank Marie-Jose´ Gaillard and an un-named reviewer for extremely helpful and constructive comments on an earlier version of the manuscript. The work has been financed by the European Union, under project ENV4-CT95-0063, ‘‘Forest Response to Environmental Stress at Timberlines: sensitivity of Northern, Alpine and Mediterranean forest limits to climate (FOREST)’’.

REFERENCES Ahti, T., Ha¨met-Ahti, L. & Jalas, J. 1968. Vegetation zones and their sections in northwestern Europe. – Ann. Bot. Fenn. 5: 161 – 211. Autio, J. 1995. Local temperature differences and their relation to altitudinal vegetation zones and the location of the timberline: an example from the fell of Aakenustunturi in Finnish Lapland. – Nordia Geogr. Publ. 24: 103 – 111. Autio, J. 1997. Pera¨pohjolan tuntureiden kasvillisuustyypit, kasvillisuuden korkeusvyo¨hykkeet ja metsa¨nrajat. (Vegetation types, altitudinal vegetation zones and timberline at the fells of Pera¨pohjola in Finnish Lapland). – Ph. Lic. Thes. Geogr. Dept. Oulu Univ., Oulu. Autio, J. & Heikkinen, O. 2002. The climate of northern Finland. – In: Finland: Nature, Society and Regions (ed. P. Raento & J. Westerholm), pp. 61 – 66. – Sp. Iss. (with CD-ROM). Fennia 180 (1 – 2), Vammala. Autio, J., Colpaert, A. & Hara, Y. 1998. Vertical summer temperature variation and phytogeographical zonation at the fell of Aakenustunturi in Finnish Lapland. – Proc. Gen. Meet. Assoc. Jpn. Geogr. 54: 296 – 297. Cajander, A. 1925. The theory of forest types. – Acta Forest. Fenn. 29: 1 – 108. Davis, M. B. 2000. Palynology after Y2K- understanding the source area of pollen in sediments. – Ann. Rev. Earth Planet. Sci. 28: 1 – 18. Finnish Meteorological Institute. 1995 – 2001. Meteorological data from the Muonio meteorological station. – FMI, Helsinki. Grimm, E. C. 1992. Tilia and Tilia-graph: Pollen spreadsheet and graphics programs. – In: 8th Int. Palynol. Congr. Aix-enProvence 1992. Progr. & Abstracts (ed. A. Pons et al.), p. 56. – Org. Comm., Aix-en-Provence. Haapasaari, M. 1988. The oligotrophic heath vegetation of northern Fennoscandia and its zonation. – Acta Bot. Fenn. 135: 1 – 219. Heikkinen, O., Obrebska-Starkel, B. & Tuhkanen, S. 1995. Introduction: the timberline – a changing battlefront. – In: Environmental aspects of the timberline in Finland and in the Polish Carpathians (ed. O. Heikkinen, B. Obrebska-Starkel & S. Tuhkanen), pp. 7 – 16. – Prace Geogr. 98. Pol. Acad. Sci., Krakow. Heikkinen, O., Tuovinen, M. & Autio, J. 2002. What determines the timberline? – In: Finland: Nature, Society and Regions (ed. P. Raento, & J. Westerholm), pp. 67 – 74. – Sp. Iss. (with CD-ROM). Fennia 180 (1 – 2), Vammala. Hicks, S. 1996. The feasibility of using pollen deposition data as climate indices. – Pala¨oklimaforschung 20: 173 – 187. Hicks, S. 1999. The relationship between climate and annual pollen deposition at northern tree-lines. – Chemosphere. Global Change Sci. 1: 403 – 416. Hicks, S. 2001. The use of annual arboreal pollen deposition values for delimiting tree lines in the landscape and exploring models of pollen dispersal. – Rev. Palaeobot. Palynol. 117: 1 – 29. Hicks, S., Ammann, B., Latalowa, M., Pardoe, H. & Tinsley, H. (Eds.) 1996. European Pollen Monitoring Programme: Project description and guidelines. – Oulu Univ. Press, Oulu (http:// wdc.obs-mip.fr/paleo/pmp/pmp.html ). Hicks, S., Tinsley, H., Pardoe, H. & Cundill, P. 1999. Supplement to Grana 43 (2004)

the guidelines. European Pollen Monitoring Programme. – Oulu Univ. Press, Oulu (http://wdc.obs-ip.fr/paleo/pmp/pmp.html ). Holtmeier, F-K. 1974. Geoo¨kologische Beobachtungen und Studien an der subarktischen und alpinen Waldgrenzen in vergleichender Sicht. – Fr. Steiner, Wiesbaden. Holtmeier, F-K. 2000. Die Ho¨hengrenze der Gebirgswa¨lder. – Arbeit. Inst. Landschaftso¨kol. Bd 8. Westf. Wilhelms-Univ., Mu¨nster. Hustich, I. 1948. The Scots pine in northernmost Finland and its dependence on the climate in the last decades. – Acta Bot. Fenn. 42., Helsinki. Hustich, I. 1966. On the forest-tundra and the northern tree lines. – Rep. Kevo Subarctic Res. St. 3: 7 – 47. Hustich, I. 1983. Tree-line and tree growth studies during 50 years: some subjective observations. – Nordicana 47: 181 – 188. Ilmastokatsaus Lokakuu 2000. Ilmatieteen laitos. – (Monthly Review/October. Finn. Meteorol. Inst.), Helsinki. Jackson, S. T. & Lyford, M. E. 1999. Pollen dispersal models in Quaternary plant ecology: assumptions, parameters and prescriptions. – Bot. Rev. 65: 39 – 75. Jacobson, G. L. & Bradshaw, R. H. W. 1981. The selection of sites for palaeoenvironmental studies. – Quat. Res. 16: 80 – 96. Ja¨ger, S., Spieksma, F. Th. M. & Nolard, N. 1991. Fluctuations and trends in airborne concentrations of some abundant pollen types, monitored at Vienna, Leiden and Brussels. – Grana 30: 309 – 312. Juntunen, V., Neuvonen, S., Norokorpi, Y. & Tasanen, T. 2002. Potential for timberline advance in northern Finland, as revealed by monitoring during 1983 – 99. – Arctic 55: 348 – 361. Kullman, L. 1990. Dynamics of altitudinal tree-limits in Sweden: a review. – Norw. J. Geogr. 44: 103 – 116. Kullman, L. 2000. Tree-limit rise and recent warming: a geoecological case study from the Swedish Scandes. – Norw. J. Geogr. 54: 49 – 59. Luomajoki, A. 1993 a. Climatic adaptation of Scots pine (Pinus sylvestris L.) in Finland based on male flowering phenology. – Acta Forest. Fenn. 237: 1 – 27. Luomajoki, A. 1993 b. Climatic adaptation of Norway spruce (Picea abies (L.) Karsten) in Finland based on male flowering phenology. – Acta Forest. Fenn. 242: 1 – 28. Luomajoki, A. 1999. Differences in the climatic adaptation of silver birch (Betula pendula) and downy birch (Betula pubescens) in Finland based on male flowering phenology. – Acta Forest. Fenn. 263: 1 – 35. McCarroll, D., Jalkanen, R., Hicks, S., Tuovinen, M., Pawellek, F., Eckstain, D., Schmitt, U., Autio, J. & Heikkinen, O. 2003. Multi-proxy dendroclimatology: a pilot study in northern Finland. – Holocene 13: 829 – 838. Ojansuu, R. & Henttonen, H. 1983. Kuukauden keskila¨mpo¨tilan, la¨mpo¨summan ja sadema¨a¨ra¨n paikallisten arvojen johtaminen Ilmatieteen laitoksen mittaustiedoista /Estimation of local values of monthly mean temperature, effective temperature sum and precipitation sum from the measurements made by the Finnish Meteorological Office. – Silva Fenn. 17: 143 – 160. Prentice, I. C. 1985. Pollen representation, source area, and basin size: towards a unified theory of pollen analysis. – Quat. Res. 23: 76 – 86. Prentice, I. C. 1988. Records of vegetation in time and space: the principles of pollen analysis. – In: Vegetation history (ed. B. Huntley & T. Webb, III), pp. 17 – 42. – Kluwer, Dordrecht. Stockmarr, J. 1971. Tablets with spores used in absolute pollen analysis. – Pollen Spores 13: 615 – 621. Sugita, S. 1993. A model of pollen source area for an entire lake surface. – Quat. Res. 39: 239 – 244. Sugita, S. 1994. Pollen representation of vegetation in Quaternary sediments: theory and method in patchy vegetation. – J. Ecol. 82: 881 – 898.

Pollen deposition and meteorological conditions on Aakenustunturi Sugita, S., MacDonald, G. M. & Larsen, C. P. S. 1997. Reconstruction of fire disturbance and forest succession from fossil pollen in lake sediments: potential and limitations. – NATO ASI Series I 51: 387 – 412. Tuhkanen, S. 1999. Sustainable development in northern timberline

47

forests. The northern timberline in relation to climate. – In: Timberline Workshop, Whitehorse 1998. Proc. (ed. S. Kankaanpa¨a¨, T. Tasanen, & M-L. Sutinen), pp.29 – 61. – Res. Pap. 734. Kolari Res. St. Finn. For. Res. Inst., Saarija¨rvi.

Grana 43 (2004)