21846_2_Autio.indd The impact of elevation, topography and snow load damage of trees on the position of the actual timberline on the fells in central Finnish Lapland JYRKI AUTIO AND ALFRED COLPAERT Autio, Jyrki & Alfred Colpaert (2005). The impact of elevation, topography and snow load damage of trees on the position of the actual timberline on the fells in central Finnish Lapland. Fennia 183: 1, pp. 15–36. Helsinki. ISSN 0015- 0010. This study examines the impact of selected environmental variables on the posi- tion of the limit of continuous forest (actual timberline) on three fells situated in central Finnish Lapland, namely Aakenustunturi, Yllästunturi and Pyhätunturi. The factors studied are elevation, topography and snow load damage to conifer- ous trees. The potential for ascending timberlines in the region is discussed. The limit of continuous forest in the study area is hardly ever composed of a single species but rather features alternating occurrences of spruce, pine and mountain birch. The average altitudinal position of the actual timberline is highest on the southern and western slopes (366–428 m a.s.l.), while the lowest altitudes are recorded on the northern and eastern slopes (336–403 m a.s.l.). Prevalence of block fields and slope gradient are the most significant and visible factors con- trolling the position of the actual timberline. On the studied fells, from 10% to more than 50% of the total length of the actual timberline is controlled by block fields. The lowest occurrence of the limit of continuous forest, at 270 m a.s.l., results from an extremely high proportion of block fields and steep slopes. Snow patches and delayed snow melting also hamper the upward advancement of the forest. Continuous forest reaches its highest altitudes at 440 m a.s.l. on slopes with gentle inclination and land cover other than block fields, and 460 m a.s.l. in sheltered furrows. Average snow load damage to spruce and pine at the ac- tual timberline varies in the ranges 15–58% and 30–76%, respectively. Snow load damages on trees undoubtedly impede the advance of the actual timberline to higher elevation in all exposures. Seedling density (number/ha) on south-fac- ing slopes near the actual timberline is ca. 5 times greater than on the north-fac- ing slopes. Recent regeneration and seedling establishment is almost absent at the treeline near the fell tops. Results suggest that there is potential for actual timberline advance on south-facing slopes with gentle inclination and missing block fields. It is uncertain, however, whether natural regeneration will take place on felltops with few or no trees. Jyrki Autio, Department of Geography, University of Oulu, PO Box 3000, FI- 90014 Oulu, Finland. E-mail: jyrki.autio@oulu.fi. Alfred Colpaert, Department of Geography, University of Joensuu, PO Box 111, FI-80101 Joensuu, Finland. E-mail: alfred.colpaert@joensuu.fi. MS submitted 08 December 2004. Introduction Actual timberline is a generic term denoting to continuous, altitudinal mountain timberline, which is more or less distinguishable from other parts of the terrain and which can be attributable to a number of local factors (see Holtmeier 2003). This line has also been called an empirical forest line/timberline (Hustich 1966; Dahl 1998), and physiognomic forest line (Hustich 1966; Tuhkanen 1993; Autio 1995; Heikkinen et al. 2002; Autio & Hicks 2004). Areas above the actual timberline are characterised by smaller trees, which grow in groups or alone. The treeline is established by the furthest individual trees that have a tree-like form and whose minimum height is 2 metres (Hustich 16 FENNIA 183: 1 (2005)Jyrki Autio and Alfred Colpaert 1966; Tuhkanen 1993; Heikkinen et al. 2002). The terrain between the treeline and the actual timber- line is called the treeline ecotone (Holtmeier et al. 2003). The timberline ecotone, in turn, denotes to the transitional zone between the economic tim- berline and the tree-species line (Heikkinen et al. 2002). A widespread idea is that the present and fu- ture global climate warming process will raise the altitudinal forest limit (see Skre 1990; Kellomäki 1996; Kellomäki et al. 1997) especially as the trend seems to be that the climate is warming up with higher altitudes (Diaz & Bradley 1997). The advance of the coniferous treeline has already been documented on the Swedish Scandes Moun- tains (Kullman 2000; Kullman & Kjällgren 2001). New trees have emerged in timberline forests in Finnish Lapland, too (Juntunen et al. 2002). There are major local differences in the way in which the timberline reacts to climatic changes. Local differ- ences of this kind are attributable to factors such as the varied topography of fells, which in turn is reflected in local differences in snow accumula- tion and temperatures (Kullman 1979; Holtmeier 1997). No detailed regional picture is available on the relation between edaphic and topographic factors, the location of the actual timberline and the al- most complete treelessness of felltop areas (bare fell heath regions above the treeline) on the fells in Central Lapland. Although the timberline phenom- enon and its regulating factors have already been examined for more than 200 years (Holtmeier 2003), there still is no uniform theory that would exhaustively account for it. This is one reason why there is still evident need to investigate the timber- line phenomenon at the local (100 m–1 km) and regional level (1–100 km) (Holtmeier 2000, 2003; Kjällgren & Kullman 2002). The purpose of this paper is to provide, by in- terpreting the results of fieldwork and aerial pho- tographs, a maximally accurate, regionally com- prehensive picture of the positions of the actual timberline, the orographical and topographical factors affecting these positions, and snow load damage to trees on the fells Aakenustunturi, Yl- lästunturi and Pyhätunturi. An account is also given of the extent to which the warm periods of the 20th century are reflected in the actual tim- berline and how the favourable seed years of the early 1970’s are visible on the actual timberline, treeline ecotone and treeline. Finally, the question of the potential of the actual timberline to advance vertically is also dealt with. The present paper ad- dresses the following five main questions: 1. What are the altitudinal positions of the actual timberline on the studied fells in Central Lap- land? 2. To what extent is the position of the actual tim- berline attributable to the presence of block fields? 3. How does slope inclination affect the position of the actual timberline? 4. What is the impact of delayed snow deposits and snow load damage on the position of the actual timberline? 5. How does climatic change affect the position of the actual timberline and treeline? Study area The fells Aakenustunturi (67°41’N, 24°29’E; 570 m a.s.l.) in Western Lapland and Yllästunturi (67°34’N, 24°15’E; 718 m a.s.l.) and Pyhätunturi (67°01’N, 27°09’E; 540 m a.s.l.) in Central Lapland (Fig. 1) are among the oldest residual mountains in the world that contain rock types of age approxi- mately 2000 million years (Manner & Tervo 1988). Most of the rock is quartzite (Mielikäinen 1979; Rastas 1984), which effectively endures chemi- cal weathering. After the latest retreat of the ice sheet, this quartzite has been broken into jagged boulders, or block fields, by freezing water. The fells are characterised by round felltops that result from a long-term erosion effect. Visible features connected with the melting of the continental ice sheet are lateral meltwater channels and glacioflu- vial overflow channels (Kujansuu 1967; Mäkinen & Maunu 1984). There are patterned grounds (Haapasaari 1988; McCarroll et al. 1996) on their felltops, representing the effect of frost action. The fells Aakenustunturi and Yllästunturi are situated in an old-growth forest protection area, apart from Yllästunturi’s westernmost top Ylläs, where most of the area’s downhill skiing slopes are located (Ylläs-Levi… 2003). Most parts of Py- hätunturi, in turn, are included in the Pyhätunturi National Park that was established in 1938. It is only the Pyhäkero area, reserved for downhill ski- ing, that lies outside the National Park. According to Köppen’s climate classification (Köppen & Geiger 1936), the areas examined here represent the snow forest climate, and more spe- cifically the humid short-summer type, i.e. Betula climate (Dfc). The mean annual temperature at FENNIA 183: 1 (2005) 17The impact of elevation, topography and snow load damage of… Aakenustunturi and Yllästunturi, as recorded by the nearest official meteorological station in Muo- nio (67°58’N, 23°40’E; 254 m a.s.l.), was –1.7 °C in the period 1961–1990. For Pyhätunturi, the respective figure was –1.0 °C as recorded at the Sodankylä observatory station (67°22’N, 26°37’E; 179 m a.s.l.). Mean precipitation in both areas was approximately 500 mm (Climatological… 1991). It should be noted, however, that the climate con- ditions on slopes and felltop areas differ consider- ably from those of the surrounding low-lying areas, mainly due to orographic and topographic reasons (Autio & Heikkinen 2002). In the classification of northwest Europe’s veg- etation into zones and sections, the fells Aake- nustunturi, Yllästunturi and Pyhätunturi represent the northern boreal vegetation zone. In terms of its continental–marine correlations, the area is situated in the indifferent section group OC (Ahti et al. 1968; Hämet-Ahti 1988). The actual tim- berline on the above fells is mainly composed of Norway spruce (Picea abies) and Scots pine (Pi- nus sylvestris), though timberline forests can also be rich in mountain birch (Betula pubescens ssp. czerpanovii). Actual timberlines lying consider- ably below the average can also contain downy birch (Betula pubescens) (Autio 1997). Material and methods Data collection The location of the actual timberline on Aake- nustunturi was determined from 367 observation points (Fig. 2) and on Yllästunturi from 76 points (Fig. 3). Other parameters measured at the points were exposure, slope inclination, presence of block fields, degree of snow load damage to spruce and pine, and relations between tree species. The observations were made in summer 2000. On Py- hätunturi’s westernmost bare felltop, Peurakero, exposition and the location of the actual timber- line were measured from a total of 344 points (Fig. 4) in summer 2001. At the actual timberline, the crown coverage of the continuous forest was still approximately 30% (Holtmeier 1974; Påhlsson 1995). The geographical position was recorded on a handheld GPS receiver GARMIN 12XL using the average position function, which calculates a moving average of the coordinates until the aver- age stabilises, in approx. five minutes (Owner’s… 1998). The number of satellites visible is usually six to eight on these exposed open mountain tops, producing on average five to ten metre accuracy. Due to the fact that GPS measurement of the verti- cal component is less accurate we decided to ob- tain altitude from the digital elevation model (25 metre resolution raster, one decimetre resolution altitude). The elevations of the actual timberline points were determined using the overlay analysis in ArcGis programme, and the DEM of National Land Survey of Finland. The points indicating the actual timberline were transformed into a line rep- resenting the actual timberline for each fell. The method devised by Perttu (1972) was employed in order to find out the exposures of the actual tim- berline points. The method divides the felltop or felltops into four sectors each of which stands for one cardinal point (Figs. 2–4). The various sectors are delimited by boundaries drawn according to half-cardinal points. Thus each sector covers 90° of the circular perimeter. The ranges of influence between the felltops were delimited by gullies and 100 km27° Norway MB Finland Norway 67° 68° 27° 67° Finland 68° Aakenustunturi Russia Sweden Muonio OA NB MB Yllästunturi Pyhätunturi Fig. 1. Location of the areas of investigation, i.e. Aakenus- tunturi, Yllästunturi and Pyhätunturi in the northern boreal zone (NB) between the middle boreal (MB) and oroarctic (OA) vegetation zones. 18 FENNIA 183: 1 (2005)Jyrki Autio and Alfred Colpaert 24o23'44" 67o39'04" Z Z Z Z Z Vaulokuru M oloslaki 530 Vasalaki 460 430 Vareslaki 485 Pallilaki 570 21 kilometres 0 " "" "" " "" """ "" "" " " "" 88888 88888888 "" " " " " """ 888888 """"""""888888 88888 88 88888 88 888 8 8 8 88888 888 8888888 8 88 88 88888 88 88888 88 888888888888 H actual timberline sites on the S-facing slope actual timberline sites on the W-facing slope contours with 5 m inter- vals summit's highest point and sector divisions actual timberline sites on the N-facing slope actual timberline sites on the E-facing slope Vareslaki 485 m a.s.l. " " " " " 8 8 8 8 8 Fig. 2. Timberline points by points of the compass on five felltops of Aakenustunturi. Copyright National Land Survey of Finland 2004. Un- authorised copying prohib- ited. 24o11'44" 67o32'56" 88 8 8 8 8 8 8 88 8 8 8 8 8 8 8 8 Z Z ZKellostapuli 502.7 Keskinenlaki 610 Varkaankuru Ylläs 718.5 (((((((( kilometres 210 contours with 5 m inter- vals actual timberline sites on the E-facing slope actual timberline sites on the S-facing slope actual timberline sites on the W-facing slope actual timberline sites on the N-facing slope H summit's highest point and sector divisions Ylläs 718.5 m a.s.l. 88888 Fig. 3. Timberline points by points of the compass on Yllästunturi. Copyright Na- tional Land Survey of Finland 2004. Unauthorised copying prohibited. N W E S l N W E S l FENNIA 183: 1 (2005) 19The impact of elevation, topography and snow load damage of… depressions that separate the felltops. The eleva- tion curves could also be used for assessing the range of influence of each felltop (Figs. 2–4). In ad- dition, the length of the border lines that stand for the sectors indicate the boundaries of the ranges of influence between the felltops. Thus the points of compass of the timberline points were established in relation to the nearest felltop. The final expo- sition of each timberline point was determined on the basis of the cardinal point of the sector in which the target was situated. Slope inclination at the actual timberline was measured using an optical hypsometer, model PM- 5/1520. The proportion of block fields was meas- ured by assessing the part of the land area that was covered by block fields at the actual timberline. Snow load damage was assessed by calculating, in percentages, how many of the spruces and pines making up the actual timberline had broken crowns, or branches in the crown, and ramified crowns. Snow load damage to birches was not examined. The relative proportions of tree species making up the actual timberline were also exam- ined. The numbers of spruce and pine seedlings and young growth (height less than 2 m) per hectare were calculated immediately above the actual tim- berline and from the middle of treeline ecotone on the southern and northern slopes, and from Aak- enustunturi’s Vareslaki at the treeline (Fig. 5). The term seedling is used for small individuals know- ing that young growth nearly 2 metres high might be several decades old. The mortality of conifer seedlings was also calculated on the same sites. Circular experimental plots of size 300 m2 (radi- us 9.8 m) at the actual timberline, and of 500 m2 (radius 12.5 m) at the treeline ecotone and at the treeline were established totalling 15 in number. There were three parallel experimental plots at each measurement point. The distance between the central points of two adjacent plots was 40 m (see Juntunen et al. 2002). 67° 00' 25" Peurakuru Peurakero 504 27° 04' 35" Z 888 888 8 8 8 88 88 8 8 88888888 8 8 8888888 8 88888888888 10.50 kilometres H summit's highest point and sector divisions actual timberline sites on the N-facing slope actual timberline sites on the S-facing slope actual timberline sites on the W-facing slope contours with 5 m inter- vals Peurakero 504 m a.s.l. 88888 Fig. 4. Timberline points by points of the compass on Peurakero on Pyhätunturi. Copyright National Land Survey of Finland 2004. Un- authorised copying prohib- ited. N W E S l 20 FENNIA 183: 1 (2005)Jyrki Autio and Alfred Colpaert Statistical analyses Differences in average altitudinal location of the actual timberline in different exposures were tested using the non-parametric Mann-Whitney’s U test. The direction and intensity of the correla- tion between the elevations of the actual timber- line, the proportion of block fields on the slopes, slope inclination and snow load damage to spruce and pine were examined using a non-parametric Spearman’s correlation coefficient (SPSS… 1998). Before testing, the validity of the normal distribu- tion assumption for the material was checked us- ing the Kolmogorov-Smirnov test (SPSS… 1998). As the data did not follow the normal distribution, it was decided to use the non-parametric Mann- Whitney U test and Spearman’s correlation coef- ficient. Aerial photographs The proportions of the actual timberline where it converges with the lower limits of the block fields were detected using digital infrared ortho- graphic colour aerial photos for all the three fells. The same principle was employed for measuring the proportions of convergence between delayed snow deposits and the actual timberline on Aak- enustunturi. In addition, the infrared orthographic images were used to illustrate the relation between the actual timberline and various environmental factors that drop or elevate the timberline position. The digital images were scanned from aerial pho- tographs taken from an altitude of approximately 4600 metres, and had a ground resolution of one metre. Results Altitudinal position of the actual timberline The length of the actual timberline is 28 km on Aakenustunturi (Fig. 2), 17 km on Yllästunturi (Fig. 3) and 6 km on Peurakero at Pyhätunturi (Fig. 4). The altitudinal position of the actual timberline on Pyhätunturi is higher (400 m a.s.l.), on average, than on other fells (Table 1). The altitudinal position is by far lowest on Aakenustunturi, but on average higher on the south-facing slopes of Aakenustun- turi and Yllästunturi, as compared with the other slopes. The continuous forest reaches its highest vertical limit at ca. 460 metres on the west-facing slope of Yllästunturi. On Pyhätunturi, the average and absolute altitudinal positions of the actual tim- berline are highest on the west-facing slopes and lowest on the south-facing slopes. The lowest loca- tion is ca. 270 m a.s.l., situated on the north-facing slope of Yllästunturi. The differences between the highest and lowest altitudinal positions of the ac- tual timberline on Aakenustunturi and Yllästunturi are greatest at the northern exposure. This difference is greatest on Yllästunturi, where the actual timberline at the southern exposure is situated on average 62 metres higher than at the northern exposure (Table 2). The smallest elevation difference is 2 metres between the northern and eastern exposures and the southern and western exposures on Aakenustunturi. Average elevation differences are not statisti- cally significant between the actual timberlines situated at the northern and eastern exposures and southern and western exposures on Aakenustun- turi (Table 2). However, the elevation differences between other exposure combinations are statisti- cally very significant (p < 0.001) on Aakenustun- turi. On Yllästunturi, in turn, the average elevation differences between timberlines situated at the northern and eastern exposures is statistically in- significant. In other respects the elevation differ- ences are statistically very significant (p < 0.01) or almost significant (p < 0.05). On Pyhätunturi, the Treeline Actual timberline Increasing altitude Treeline ecotone 40 m 40 m The distance between plots Fell heath Fig. 5. Experimental design of spruce and pine seedlings density (number/ha) measuring. The distance between the circular plots was measured from centre to centre. FENNIA 183: 1 (2005) 21The impact of elevation, topography and snow load damage of… Table 1. Average elevation of the actual timberline, it’s standard deviation (STDV), minimum and maximum elevations, and maximum elevation difference on Aakenustunturi, Pyhätunturi and Yllästunturi by points of the compass. The highest values are indicated in bold and the smallest in italics and underlined. The number of timberline points (n) is also given. Exposure Average (m a.s.l.) STDV Min (m a.s.l.) Max (m a.s.l.) Range (vertical m) Aakenustunturi N (n = 62) 338 24.1 284 391 107 E (n = 72) 336 22.8 299 384 85 S (n = 91) 390 23.7 347 440 92 W (n = 143) 388 19.3 353 432 79 Average 370 22.5 321 410 89 Yllästunturi N (n = 19) 376 43.7 272 429 157 E (n = 23) 366 38.5 276 411 135 S (n = 16) 428 12.2 413 453 40 W (n = 18) 410 17.6 380 456 76 Average 395 28.0 335 437 102 Pyhätunturi N (n = 116) 403 8.0 387 416 29 S (n = 87) 375 14.8 335 403 68 W (n = 137) 418 11.0 383 435 52 Average 402 11.3 368 415 47 Table 2. Differences in average altitudinal position of the actual timberline in different exposures. Statistical tests on the elevation differences between the exposures were per- formed using the Mann-Whitney U-test. Abbreviations: p- value is probability. Statistical significance: *** highly signif- icant, ** significant, * almost significant and 0 insignificant. Exposure Difference in height (m) P-value Significance Aakenustunturi N - E 2 0.565 0 N - S –52 0.000 *** N - W –50 0.000 *** E - S –54 0.000 *** E - W –52 0.000 *** S - W 2 0.562 0 Yllästunturi N - E 10 0.167 0 N - S –52 0.000 *** N - W –34 0.023 * E - S –62 0.000 *** E - W –44 0.000 *** S - W 18 0.002 ** Pyhätunturi N - S 28 0.000 *** N - W –15 0.000 *** S - W –43 0.000 *** elevation differences are statistically very signifi- cant in all exposition combinations. Tree species and snow load damage at the actual timberline The relations between the various tree species were examined on all points of the actual timber- line on Aakenustunturi and Yllästunturi. The results obtained for this feature on Pyhätunturi, are based on Autio (1997). On Aakenustunturi, the actual timberline is composed mainly of spruce (Table 3), with ca. 30% proportion of mountain birch of, rising to as much as 60% on east-facing slopes. The proportion of pine on Aakenustunturi timber- line is negligible. On Yllästunturi, however, pine is the most common species (41%) to form the actual timberline (Table 4). The proportion of mountain birch is 33% and that of spruce 26% of the tree species making up the actual timberline. The most common species on the west-facing measurement points of the actual timberline on Table 3. Proportions of mountain birch, pine and spruce, and their standard deviations (STDV) of the trees forming the actual timberline on Aakenustunturi by points of the compass. The figures are average values. The highest values by tree species are indicated in bold and the smallest in italics and underlined. The number of timberline points (n) is also given. Exposure Birch (%) STDV Pine (%) STDV Spruce (%) STDV N (n = 62) 34.2 29.5 0.2 0.5 65.0 29.6 E (n = 72) 60.0 30.7 4.8 15.9 35.1 29.4 S (n = 91) 22.5 19.2 3.3 7.9 74.0 21.1 W (n = 143) 13.9 15.5 5.2 3.4 82.1 17.6 Average 32.6 23.7 3.4 6.9 64.0 24.4 22 FENNIA 183: 1 (2005)Jyrki Autio and Alfred Colpaert Peurakero at Pyhätunturi is mountain birch (37%), which is also most common at the actual timber- line of the northern slopes of Peurakero (47%). On the southern slopes of Peurakero, the actual tim- berline is composed mostly of pine. Common for all three fells is that the actual timberline is com- posed of not just a single species, but alternating occurrences of spruce, pine and mountain birch (cf. Autio 1997). Snow load damage is more common phenom- enon with pines than spruces on Aakenustunturi (Table 5) and Yllästunturi (Table 6). In this study, damages were not examined on Pyhätunturi. Snow load damage to pine and spruce is most common on the south-facing and west-facing slopes of Aakenustunturi. On Yllästunturi, pines suffer from snow load damage particularly on west-facing and east-facing slopes, whereas spruces are damaged mostly on east-facing and south-facing slopes. A statistically very significant correlation (p < 0.01) was observed between snow load damage to spruce and the altitudinal positions of the ac- tual timberline on all the slopes of Aakenustunturi (Table 7). As for pine, the same positive correlation is statistically significant (p < 0.01) on the east-fac- ing and west-facing slopes only. On Yllästunturi, snow load damage to pine correlate positively and statistically significantly (p < 0.01) with the actual timberline altitude on the east-facing slopes and south-facing slopes (p < 0.05) (Table 8). Block fields and slope inclination at the actual timberline Block fields and slope inclination at the actual tim- berline correlate positively with each other, i.e. as the inclination increases, so do usually the block fields (Table 9), and vice versa. The slopes at the actual timberline on Yllästunturi are steeper and contain more block fields than do those on Aake- nustunturi. In spite of this, the average altitudinal position of the actual timberline on Yllästunturi is higher than on Aakenustunturi. This indicates that on Yllästunturi, block fields start to occur higher, on average, than on Aakenustunturi. The higher position of the actual timberline on Yllästunturi may also be attributable to the mass-elevation ef- fect (cf. Hustich 1937). On the east-facing slopes of Aakenustunturi, the slope inclination at the actual timberline is greater than on other slopes, there are more block fields, and the average altitudinal position of the actual timberline is lower than in other points of the compass. On the western slope, however, the inclination of the actual timberline is Table 4. Proportions of mountain birch, pine and spruce, and their standard deviations (STDV) of the trees forming the actual timberline on Yllästunturi by points of the compass. The figures are average values. The highest values by tree species are indicated in bold and the smallest in italics and underlined. The number of timberline points (n) is also given. Exposure Birch (%) STDV Pine (%) STDV Spruce (%) STDV N (n = 11) 35.6 24.1 55.7 31.4 5.0 5.6 E (n = 29) 26.6 14.7 19.0 25.4 54.6 30.4 S (n = 18) 41.8 15.0 28.0 25.9 31.6 27.3 W (n = 18) 28.7 24.1 60.4 33.0 12.8 19.1 Average 33.2 19.5 40.8 28.9 26.0 20.6 Table 5. Average snow damage to pine and spruce, and their standard deviations (STDV), at the actual timberline on Aake- nustunturi by points of the compass. The highest values by tree species are indicated in bold and the smallest in italics and underlined. The number of timberline points (n) is also given. Exposure Pine (%) STDV Spruce (%) STDV N (n = 62) 62.5 47.9 15.4 9.3 E (n = 72) 59.4 31.7 17.6 19.3 S (n = 91) 75.9 30.0 25.0 9.0 W (n = 143) 67.4 23.2 27.1 8.6 Average 66.3 33.2 21.3 11.5 Table 6. Average snow damage to pine and spruce, and their standard deviations (STDV), at the actual timberline on Yl- lästunturi by points of the compass. The highest values by tree species are indicated in bold and the smallest in italics and underlined. The number of timberline points (n) is also given. Exposure Pine (%) STDV Spruce (%) STDV N (n = 19) 31.1 12.9 37.5 8.3 E (n = 23) 73.1 29.0 57.5 23.2 S (n = 16) 42.5 19.2 39.7 23.0 W (n = 18) 45.6 24.9 27.8 25.9 Average 48.1 21.5 40.6 20.1 FENNIA 183: 1 (2005) 23The impact of elevation, topography and snow load damage of… Table 7. Correlations between the actual timberline elevation, slope inclination, block fields, and snow load damage (S.l.d.) to spruce and pine on Aakenustunturi. Statistically significant correlations at risk level p < 0.01 are indicated in bold. m a.s.l. Inclination Block fields S.l.d. of spruce S.l.d. of pine North-facing slope m a.s.l. 1.000 0.078 –0.718 0.776 –0.316 Inclination 0.078 1.000 –0.114 0.144 0.056 Block fields –0.718 –0.114 1.000 –0.517 0.316 S.l.d. of spruce 0.776 0.144 –0.517 1.000 –0.316 S.l.d. of pine –0.316 0.056 0.316 –0.316 1.000 East-facing slope m a.s.l. 1.000 –0.560 –0.398 0.604 0.591 Inclination –0.560 1.000 0.455 –0.510 0.208 Block fields –0.398 0.455 1.000 –0.012 0.093 S.l.d. of spruce 0.604 –0.510 –0.012 1.000 0.627 S.l.d. of pine 0.591 0.208 0.093 0.627 1.000 South-facing slope m a.s.l. 1.000 0.478 0.612 0.746 –0.014 Inclination 0.478 1.000 0.626 0.496 0.172 Block fields 0.612 0.626 1.000 0.606 0.228 S.l.d. of spruce 0.746 0.496 0.606 1.000 0.115 S.l.d. of pine –0.014 0.172 0.228 0.115 1.000 West-facing slope m a.s.l. 1.000 0.310 0.221 0.431 0.501 Inclination 0.310 1.000 0.637 0.387 –0.174 Block fields 0.221 0.637 1.000 0.435 –0.119 S.l.d. of spruce 0.431 0.387 0.435 1.000 0.244 S.l.d. of pine 0.501 –0.174 –0.119 0.244 1.000 Table 8. Correlations between the actual timberline elevation, slope inclination, block fields, and snow load damage (S.l.d.) to spruce and pine on Yllästunturi. Statistically significant correlations at risk level p < 0.01 are indicated in bold. m a.s.l. Inclination Block fields S.l.d. of spruce S.l.d. of pine North-facing slope m a.s.l. 1.000 0.092 –0.577 0.562 0.465 Inclination 0.092 1.000 0.373 –0.116 –0.556 Block fields –0.577 0.373 1.000 0.103 –0.455 S.l.d. of spruce 0.562 –0.116 0.103 1.000 0.856 S.l.d. of pine 0.465 –0.556 –0.455 0.856 1.000 East-facing slope m a.s.l. 1.000 –0.704 –0.645 –0.041 0.536 Inclination –0.704 1.000 0.705 0.005 –0.397 Block fields –0.645 0.706 1.000 0.335 –0.286 S.l.d. of spruce –0.041 0.005 0.335 1.000 0.087 S.l.d. of pine 0.536 –0.397 –0.286 0.087 1.000 South-facing slope m a.s.l. 1.000 0.098 –0.307 0.337 0.511 Inclination 0.098 1.000 –0.043 0.289 –0.039 Block fields –0.307 –0.043 1.000 –0.302 –0.405 S.l.d. of spruce 0.337 0.289 –0.302 1.000 0.911 S.l.d. of pine 0.511 –0.039 –0.405 0.911 1.000 West-facing slope m a.s.l. 1.000 –0.411 –0.468 0.092 0.265 Inclination –0.411 1.000 0.847 0.233 –0.421 Block fields –0.468 0.847 1.000 0.466 –0.342 S.l.d. of spruce 0.092 0.233 0.466 1.000 0.390 S.l.d. of pine 0.265 –0.421 –0.342 0.39 1.000 lowest and the block fields are smallest. The aver- age altitudinal position of the actual timberline on the west-facing slopes is almost the same as on the south-facing slopes. On Yllästunturi, however, slope steepness and the presence of block fields do not reduce the altitudinal position of the av- 24 FENNIA 183: 1 (2005)Jyrki Autio and Alfred Colpaert erage actual timberline to such an extent as they do on Aakenustunturi. The actual timberline on Yl- lästunturi extend higher, on average, on the steep western slope with extensive block fields, than on the gentler and less rocky northern and eastern slopes. In cases where the actual timberline borders a block field (Fig. 6A Point 1 and Fig. 6B Point 4), its elevation remains below the average (Fig. 7). The lowest timberline points on the fells are dictated by slope gradient and block fields in particular. On the lowest actual timberline point on the northern slope of Yllästunturi (272 m a.s.l.), for instance, the block field percentage is 80% and inclination 32°. Of the total length of the actual timberline on Aakenustunturi (27.9 km), at only 9.7% (2.7 km) does the actual timberline run alongside the lower edge of a block field, as compared with 52.5% on Peurakero at Pyhätunturi and 28.4% on Ylläs- tunturi. In the latter case, 36.5% of all timberline points sit in areas where block fields make up 40% or more of the ground area, as compared with only 10.9% on Aakenustunturi (Table 10). On gentle slopes with small block fields, the for- est turns into barren fell heath gradually making the actual timberline difficult to perceive accu- rately in the field or on aerial photographs (Fig. 6A Point 2 and Fig. 6B Point 5, Fig. 8). On such slopes as well as in steep gullies (such as Varkaankuru at Yllästunturi, see Fig. 3), the timberline lies higher than elsewhere (Fig. 6A Point 3). On the other hand, in some locations, e.g. on the north-facing and east-facing slopes of Aakenustunturi, the ac- tual timberline remains at low altitude despite a gentle slope inclination. The reasons for this are sudden kinks in the slope gradient, which form fa- vourable zones for delayed snow beds. The depos- its lead to a sharp actual timberline and lower than average altitudinal position (Fig. 6C Point 6 and Fig. 9). Forests do not occur above delayed snow beds, and individual trees only survive in places moister than their surroundings (Fig. 10). In exces- sively wet places, such as paludified sites, the tim- berline occupies a lower elevation (Fig. 6D Point 8). In addition, delayed snow deposits typically accumulate in depressions, such as lateral drain- age channels, which prevent the timberline from reaching higher elevations (see Autio & Heikkinen 2002). Of the total length of the actual timberline on Aakenustunturi, 4.8% involved cases where the lower edge of delayed snow deposits adjoined the Table 9. Average block field coverage of the ground area, inclinations and standard devia- tions (STDV) at actual timberline points (n) on Aakenustunturi and Yllästunturi by points of the compass. The highest block field and inclination values are shown in bold and the lowest in italics and underlined. Exposure Block fields (%) STDV Inclination (o) STDV Aakenustunturi N (n = 62) 8.0 13.1 15.2 7.7 E (n = 72) 21.7 18.0 21.8 10.5 S (n = 91) 8.2 13.5 10.5 6.4 W (n = 143) 7.9 13.3 9.1 5.0 Average 11.5 14.5 14.2 7.4 Yllästunturi N (n = 11) 24.5 25.1 15.1 11.7 E (n = 29) 25.8 29.0 19.9 10.0 S (n = 18) 30.4 20.0 16.8 3.9 W (n = 18) 45.9 38.9 24.7 11.2 Average 31.7 28.3 19.1 9.2 Table 10. Classification of block field coverage (%) into five class intervals at actual timberline points on Aakenustunturi and Yllästunturi. The frequency of observations in each in- terval and their relative (%) proportion of the entire material are shown. Block fields coverage 40% or more is indicated as a dash line in the table. Class intervals (%) Frequency (%) Aakenustunturi 0–19 261 71.1 20–39 66 18.0 40–59 36 9.8 60–70 3 0.8 80–100 1 0.3 Yllästunturi 0–19 35 47.3 20–39 12 16.2 40–59 10 13.5 60–70 4 6.8 80–100 12 16.2 FENNIA 183: 1 (2005) 25The impact of elevation, topography and snow load damage of… Fig. 6. Four aerial photo- graphs of Yllästunturi, Py- hätunturi and Aakenustun- turi. A thick red line with a thin black line in the centre and edges indicates the ac- tual timberline. Copyright FM-kartta Oy. Publication number for all photographs is FMK002_1/2004. (A) Aerial photograph of the western and southern slope of Yllästunturi (718 m a.s.l.). The brownish forest at Point 1 borders sharply on the blu- ish lower edge of a block field. The actual timberline lies at Point 1 at 403 m a.s.l. In the left-hand upper corner the actual timberline runs in a forest (Point 2), indicating the gradual thinning of the forest on the gentle slope into a treeless felltop. In gul- lies (Point 3) the continuous forest rises above the sur- roundings. The highest actu- al timberline on Yllästunturi is situated at 456 m a.s.l. at Point 3. Photographed on 30th July 1995. (B) Aerial photograph of the western slope of Peura- kero on Pyhätunturi (504 m a.s.l.). The greenish forest clearly borders on the bluish lower edge of a block field as in Point 4 at ca. 400 m a.s.l. Where block fields do not hamper the advance of for- est, and the slope is gentle, the forest gradually becomes thinner and the actual tim- berline rises above the sur- roundings. At the tip of the southward arrow at Point 5, the actual timberline reach- es its highest elevation on Peurakero, Pyhätunturi, i.e. 435 m a.s.l. Photographed on 8th July 1989. (C) Aerial photograph of the northern slopes of Pallilaki (570 m a.s.l.) and Vareslaki (485 m a.s.l.) on Aakenustunturi. Delayed snow deposits accumulate in suitable points of the terrain on the northern and eastern slopes of Aakenustunturi (indicated by white colour in the photographs) that prevent the advance of the forest, as in places indicated by the arrow at Point 6. At those points the actual timberline is situated at 354 m a.s.l. The timberline occupies lower elevations in places indicated by the arrow at Point 7, due to frequent avalanches. At lowest the actual timberline in the avalanche slope was situated at 313 m a.s.l. Photographed on 24th June 1998. (D) Aerial photograph of the western slope of Vasalaki on Aakenustunturi (460 m a.s.l.). The lower elevation of the actual timberline in the place indicated by the arrow at Point 8 is due to paludification. At lowest the actual timberline is situated at the lower edge of a mire at 355 m a.s.l. Before the mire the actual timberline runs at 372 metres. Photographed on 24th June 1998. 26 FENNIA 183: 1 (2005)Jyrki Autio and Alfred Colpaert Fig. 7. Actual timberline on Kellostapuli, Yllästunturi. The view is from the west to the east. The actual timberline, which borders on a steep slope with block fields, is clearly and sharply visible even in the remote land- scape, as shown in the centre and left-hand corner of the photograph. Steep slope and block fields together reduce the elevation of the actual timberline. (Photo Jyrki Au- tio, 09/1996). Fig. 8. Thin, park-like treeline ecotone forest on the gentle western slope of Peurakero, Pyhätunturi. The forest grad- ually becomes thinner and the actual timberline is not sharp. (Photo Jyrki Autio, 08/2001). actual timberline. Another factor leading to lower timberline elevations is the occurrence of local av- alanches on the north-facing slopes of Kellostapuli at Yllästunturi and of Pallilaki at Aakenustunturi (Fig. 6C Point 7 and Fig. 11). At Aakenustunturi, the altitudinal position of the actual timberline correlate statistically significantly (p < 0.01) with block fields regardless of the slope aspect. The correlation is negative on the north- facing and east-facing slopes, but positive on the south-facing and west-facing slopes (Table 7). In other words, the low altitudinal position of the ac- tual timberline on the northern and eastern slopes correlates with the proportion of large block fields. On the southern and western slopes, however, the altitudinal position of the actual timberline is high, but there are also large block fields. The cor- relation between slope steepness and timberline elevation is negative and statistically significant on the east-facing slopes (p < 0.01), but positive and statistically significant on the south-facing and west-facing slopes (Table 7). Statistical correlation between the altitudinal positions of the actual timberline, the proportion FENNIA 183: 1 (2005) 27The impact of elevation, topography and snow load damage of… Fig. 9. Sharp actual timber- line on the northern slope of Pallilaki on Aakenustunturi, caused by a delayed snow deposit. The view is from the same place as that shown by the left-hand arrow at Point 6 in Figure 6C. (Photo Jyrki Autio, 08/2004). Fig. 10. Spreading of trees above the actual timberline on both sides of the seeping surface of a small spring on the eastern slope of Pallilaki on Aakenustunturi. (Photo Jyrki Autio, 08/2000). of block fields on the slope, and slope inclination is smaller on Yllästunturi (Table 8) than on Aake- nustunturi (Table 7). A negative, statistically signifi- cant correlation (p < 0.01) was observed between block fields, slope steepness and the altitudinal position of the actual timberline on the east-facing slopes. According to Autio (1997), block fields oc- cupy as much as 50% of the ground cover at the actual timberline on the southern slope of Laaka- kero at Pyhätunturi but only 3% at the timberline on the northern slope of Laakakero and 10% on the western slope of Peurakero. The altitudinal 28 FENNIA 183: 1 (2005)Jyrki Autio and Alfred Colpaert Fig. 11. Trees on the north- ern slope of Pallilaki on Aakenustunturi suffer from frequent avalanches, which prevent the forest from ad- vancing. The view shows an avalanche accumulation site. Avalanches originate at higher points on the slope where the inclination is ap- proximately 35°. They stop at the actual timberline shown in the background (see Fig. 6D Point 7). (Photo Jyrki Au- tio, 08/2000). position of the actual timberline is lowest on the south-facing slopes, which are characterised by large block fields and steep gradient, while on the gentle western and northern slopes, where also the block fields are smallest, it reaches the highest al- titudinal position on Pyhätunturi. Seedling density and mortality at the treeline ecotone The number of spruce seedlings decline dramati- cally from the actual timberline to the treeline (Ta- ble 11). The southern and northern slopes also dif- fer greatly in this respect: on southern slopes, the number of spruce seedlings at the actual timber- line and in the middle of the treeline ecotone is ca. 5 times higher than on northern slopes. Pine seed- lings are quite scarce at the timberline and in the middle of treeline ecotone dominated by spruce. In relative terms, the proportion of pine seedlings of the conifer new growth increase with elevation so that on the treeline, for instance, pine account for half of all conifer seedlings, as opposed to zero oc- currences at the actual timberline on the southern slope. The proportion of dead conifer seedlings is 4.4% of all seedlings at the actual timberline of the south-facing slope, 16.7% in the middle of treeline ecotone and 52.6% at the treeline, as compared with 28.0% at the actual timberline of the northern slope, which exceeds the figure recorded for the treeline ecotone (22.5%). An opposite trend was observed on the southern slope. Table 11. Number of spruce and pine seedlings per hectare at the actual timberline, in the middle of treeline ecotone and tree line on the southern and northern slope of Vareslaki on Aakenustunturi. Conifer seedling mortality per hectare is also given. Exposure Zone Altitude (m a.s.l.) Spruce seedling density (number/ha) Pine seedling density (number/ha) Seedling mortality (number/ha) South Actual timberline 435 4961 0 231 Treeline ecotone 460 1672 88 352 Summit Tree line 475 12 6 20 North Treeline ecotone 395 297 44 99 Actual timberline 365 891 11 351 FENNIA 183: 1 (2005) 29The impact of elevation, topography and snow load damage of… Discussion Actual timberline elevation The average and maximum actual timberline eleva- tions on Aakenustunturi and Yllästunturi discussed here and those obtained by Hustich (1937) differ quite notably in that the present ones are clearly below those presented by Hustich. The measuring instruments (barometer) and topographic maps at his disposal were not as accurate and furthermore, the method that Hustich (1937) used for delimit- ing the actual timberline is not known. The actual timberline positions recorded here by means of a GPS receiver can be considered quite accurate (see Colpaert 1998), as indicated by the fact that on the aerial photographs the actual timberline ad- joins the lower edges of the block fields (see for example Fig. 6A Point 1), which indicate the high- est timberline elevation. According to Holtmeier (1974, 2003), the ac- tual timberline elevation varies from 400 metres up to 600 metres in the “high fjeld” area situated in Western Lapland. On Aakenustunturi and Yl- lästunturi, however, the average elevations remain below 400 metres, apart from Yllästunturi’s south- facing and west-facing slopes, where the average actual timberline elevation exceeds 400 m a.s.l., with a maximum of 460 metres. Earlier investiga- tions (e.g. Norokorpi 1994) have suggested that the coniferous timberline is situated at approxi- mately 440 m a.s.l. in the Pyhätunturi area. The results obtained here indicate, however, that al- though the actual timberline on Pyhätunturi can reach a maximum elevation of 440 metres, the average remains at 405 metres. The actual timber- line elevations recorded by Autio (1997) for Aake- nustunturi, Pyhätunturi and Yllästunturi also differ from the results obtained here. This is mainly due to the fact that the results from Autio (1997) were derived from single-transect observations whereas those discussed here are based on a much more comprehensive sample. Tree species at the actual timberline On the studied fells, the relations between tree species at the actual timberline turned out to be typical for fells situated to the south of the northern timberline (see Autio 1995, 1997; Holtmeier 2003) in that the limit composed of alternating composi- tions of spruce (Picea abies), pine (Pinus sylves- tris) and mountain birch (Betula pubescens ssp. czerpanovii). On Aakenustunturi, for instance, the actual timberline is composed entirely of spruce in only 1.5% of all cases. The birch proportion is greatest on the east-facing and north-facing slopes, attributable to a thicker snow cover (see Solantie 1974; Kullman 1979; Autio 1997). According to Solantie (1974), the thick snow cover results from the dynamics of air flows crossing the fells. In ad- dition, snow melts from these slopes later than from slopes facing other directions (Autio 1997). According to Kullman (1979), considering the snowy conditions of this kind, mountain birch is better capable of seed-based reproduction than is spruce or pine. Usually, shade-tolerant species, such as spruce, tend to form abrupt timberlines, while light-de- manding species, such as pine, tend to produce open stands near the vertical limit for continuous forest with a wide transition zone (Walter 1968). In the current study, however, the tree species did not have an impact on the nature of the actual timber- line: the forest always turns gradually into a tree- less or almost treeless fell heath region regardless of tree species expect in cases where the limit is determined by soil-related and topographic factors such as block fields, delayed snow deposits and avalanches. Factors affecting snow load damage at the actual timberline Snow load damage to spruce turned out to be smaller than to pine, which is in accordance with earlier investigations (Heikinheimo 1920; Norokorpi & Kärkkäinen 1985; Norokorpi 1994; Mustonen 1997; Korpi 2001). Although the nar- row-crowned spruce is well capable of withstand- ing snow masses, pine is still more common than spruce at the actual timberline for example on Yllästunturi. It is therefore likely that the species composition at the actual timberline is determined by soil-related factors, such as moisture and tex- ture of soil, more than by the ability of the tree to withstand snow load damage. The combined per- centages of snow load damage to spruce and pine at the upper limit of continuous forest are mark- edly higher than on the limit where snow starts to accumulate on trees (30% damage) so the latter limit on Aakenustunturi and Yllästunturi is situated below the actual timberline (cf. Norokorpi 1994). It should be noted, however, that on Aakenustun- turi, the snow load damage at the actual timberline to spruce alone is under 30% on average, indicat- 30 FENNIA 183: 1 (2005)Jyrki Autio and Alfred Colpaert ing that the snow accumulation limit for spruce is above the actual timberline. The snow accumula- tion limit for spruce is known to be higher than for pine (Norokorpi & Kärkkäinen 1985). In addition, the amount of snow load damage varies consider- ably, mainly due to major variation in the actual timberline elevation. The fact that snow load damage to spruce is less severe on Aakenustunturi than on Yllästunturi is mainly attributable to the fact that on Yllästun- turi the actual timberline is situated higher than on Aakenustunturi. According to earlier investiga- tions (Heikinheimo 1920; Norokorpi & Kärkkäi- nen 1985; Korpi 2001), the amount and frequency of snow load damage increase with higher eleva- tions. The finding is also supported by the fact that on Aakenustunturi, snow load damage to spruce and pine is more severe on the south-facing and west-facing slopes with higher actual timberline elevation, as compared with northern and east- ern slopes. Also Heikinheimo (1920), Nykänen et al. (1997) and Päätalo (2000) point out that snow loads are largest on south-facing and west-facing slopes. In earlier studies in Kuusamo by Norokorpi and Kärkkäinen (1985), however, the amount of snow load damage was not found to be dependent on slope aspect. On Yllästunturi, the average snow load damage is largest on the east-facing slopes, probably reflecting the impact of artificial snow- ing practiced in Yllästunturi’s skiing resort in early winter. Rime formation is extensive in early winter, too (Tammelin & Säntti 1992). Yllästunturi’s east- facing slopes are also well positioned for receiving the humid air masses spreading from the south so the resulting humidity tends to condensate readily on the trees. On Yllästunturi, snow load damage to pines is less extensive than on Aakenustunturi, possibly due to the fact that on the latter pines make up a small proportion of the timberline’s growing stock, thus increasing the proportion of random fluctuations in the results. On the other hand, pine crowns on Aakenustunturi may be more asymmetric and broader, thus accumulating more snow than those on Yllästunturi (see Nykänen et al. 1997). Scattered pines on Aakenustunturi suf- fer from snow load damage more than pines that occur in denser growths. Individual pines growing in spruce-dominated timberline forests may accu- mulate larger snow loads than those standing in pine-dominated growths (cf. Kangas 1959). The relationship between slope angle and snow load damage is apparent on the south-facing and west- facing slopes of Aakenustunturi, where damage is most widespread on steep slopes (cf. Nykänen et al. 1997). On the other hand, at these points also the timberline elevation is higher and hence, it is possible that the amount of the snow load damage is partially due to higher timberline elevation. Orographic and topographic factors affecting the position of the actual timberline The mass-elevation effect resulting from a large mass of broad, high mountains (Hustich 1937; Holtmeier 1974, 2003) may indeed affect the maximum elevation of the actual timberline in the study area, as the highest occurrence of continu- ous forest limit was found on the highest fell, Yl- lästunturi. The mass-elevation effect does not seem to have as obvious impact on the average actual timberline elevation, however, as the actual tim- berline reaches its highest average elevation on Pyhätunturi, which is lower than Yllästunturi. The proportion and distribution of block fields on the slope have an impact on timberline eleva- tion, as indicated in previous research, too (Hus- tich 1937; Piirola 1969; Holtmeier 1974; Autio 1995, 1997). It is difficult for trees to become rooted in a nutrient-poor, rocky soil (Heikkinen et al. 1995), and they are also sensitive to wind dam- age (Eronen 1979). Thin humus layer and thin soil horizon (Autio & Kinnunen 1992) give rise to high temperatures and low moisture content in summer (Tranquillini 1979; Holtmeier 2000; Holtmeier et al. 2003). Areas above the actual timberline are exposed to high wind speeds (Lehtonen 1992; Tammelin 1998), which also accelerates the dry- ing of the surface soil. Excessive drying of the soil surface in areas above the actual timberline is en- hanced by coarse, highly permeable soils, particu- larly on steep slopes (Mäkitalo et al. 1994). The low soil moisture content hampers or even pre- vents the sprouting of tree seeds and the ability of trees to obtain nutrients (Kallio & Lehtonen 1973; Holtmeier 2000, 2003). The formation of the actual timberline is affect- ed also by slope inclination. On gentle slopes, the uppermost vertical limit of continuous forest usu- ally alters into a treeless area gradually through a transition zone (Fig. 8; see also Autio 1995; Holt- meier 2003). It may be difficult to indicate the ac- tual timberline under such conditions. The actual timberline is usually sharpest on steep slopes (Fig. 7) where the uppermost vertical limit of continu- ous forest is also fairly easy to point out. On steep FENNIA 183: 1 (2005) 31The impact of elevation, topography and snow load damage of… slopes in the fells of Lapland, freeze-thaw cycles have often caused quartzite to crack, giving rise to block fields. Steep gradient seems to encourage the occurrence of block fields, a situation also not- ed elsewhere (Piirola 1969; Autio 1995). Holtmei- er (2003) named the timberline dictated by rocky fields and steep topography as the orographic tim- berline. If we use a block field limit value of 40% or more of the ground area covered with blocks, these make up 36.5% of the timberlines on Ylläs- tunturi and 10.9% on Aakenustunturi. These limits, regulated by block fields and steep topography, are thus orographic timberlines. On Peurakero at Py- hätunturi, the proportion of orographic timberlines is over 50%. Steep inclination and the presence of block fields together force the actual timberline to remain at lower elevations. On sheltered north-fac- ing slopes with steep topography, the local timber- line elevation is also reduced by avalanches (Fig. 11; see also Autio & Heikkinen 1999). On gentle slopes, mires have the same impact (Fig. 6D; see also Holtmeier 1973, 1974; Wardle 1993). Local landforms have an impact on the structure of the actual timberline. Depressions, such as later- al meltwater channels, are treeless and may accu- mulate several metres of snow. In these landforms, the absence of trees is due to excessive snow cover that only melts in late summer and the resulting short growing season (Autio & Heikkinen 2002). The varied topography of fells is also closely reflected in their microclimate, snow distribution and mass movements. Topographically induced microclimatic variations have been found to have an impact on the location of the actual timberline and treeline (Autio 1995). In parallel with earlier investigations (Perttu 1972; Mayer & Ott 1991; Odland 1996; Autio 1997; Kjällgren & Kullman 2002), the timberlines in the study area reach their maximum elevations on the south-facing and west-facing slopes while the lowest elevations were recorded on the north- facing and east-facing slopes. The only expection to this rule is Pyhätunturi, where the actual tim- berline elevation is lowest on the south-facing slopes. Factors affecting seedling density and mortality at the treeline ecotone On the south-facing slope of Aakenustunturi, spruce seedlings form vital, dense growths im- mediately above the actual timberline and in the treeline ecotone (Fig. 12). Yet the number of spruce seedlings is slightly below that recorded by Juntunen et al. (2002) on Pallastunturi (68°00’N, 24°09’E) but well above that in the corresponding zones on Pyhätunturi. The number of spruce seed- lings is much smaller at the actual timberline and in the middle of treeline ecotone on the northern slope of Vareslaki at Aakenustunturi than on Pal- lastunturi (Juntunen et al. 2002). Their number is larger on the south-facing slope of Aakenustunturi than on the north-facing slope, which is attribut- able to a more favourable local climate, earlier snow melt and higher effective temperature sum (Autio & Hicks 2004). Delayed snow deposits ac- cumulate immediately above the actual timber- line on the northern slope (see Fig. 6C Point 6). In these points, the local growing season is shorter, the ground becomes waterlogged and ground tem- peratures are low at the beginning of the growing season. In addition, pockets of descending cold air are likely to form at the dense, continuous actual timberline on Aakenustunturi during the growing season, causing damage to tree seedlings and an- nual shoots. The reasons for the small number of seedlings at the treeline and for their high mortality and fading on Aakenustunturi might be multiple, including the occurrence of sub-zero temperatures at the beginning of the growing season, variation in temperature factors, unfavourable soil tempera- ture conditions, drought and poor nutrient supply, minor or non-existent snow protection in winter, reindeer grazing in a sensitive environment, and mechanical abrasion and drought caused by harsh winds (see Holtmeier et al. 2003). Many of the above factors also contribute to the scantiness of seedlings in the treeline ecotone of the northern slopes. Potential advance of the actual timberline The size of spruce seedlings and young growth on Aakenustunturi varies from 10 cm to almost 2 m. It is likely that the smallest seedlings date back to the 1980’s and 1990’s when according to Juntunen et al. (2002) spruce seedling regeneration was suc- cessful on the fells of Lapland. Summer seasons warmer than usual with conditions favourable for seedling regeneration occurred in 1994–2001 on Aakenustunturi. The best years with this respect were 1997 and 1999–2001, when temperatures stayed above the long term average and high pol- len deposition at the actual timberline was record- ed (Autio & Hicks 2004). Most of the seedlings ger- minated in 1972–1973, which is known to be one 32 FENNIA 183: 1 (2005)Jyrki Autio and Alfred Colpaert Fig. 13. Most of the trees shown in the background are mountain birches of height less than 2 metres, with some spruces and pines, above the actual timberline at 450 metres on Peurakero, Pyhä- tunturi. (Photo Jyrki Autio, 08/2001). Fig. 12. To the front in the picture are spruce seedlings and young growth having started their growth in the past three decades, which are situated at 435 metres and above the actual timber- line on the southern slope of Aakenustunturi. (Photo Jyrki Autio, 08/2004). of the most favourable conifer regeneration years of the last century (Sirén 1993), thanks to warm summer seasons (Juntunen et al. 2002). Conifer seedlings seem to be able to react vividly to cli- matic changes (cf. Kellomäki et al. 1997), which is reflected by the strongly stocked young growth at the actual timberline and treeline ecotone of south-facing slopes. The rapid reaction of conifer seedlings to climate warming indicates a potential for the advance of the treeline and subsequently also of the actual timberline. It is possible that climate warming will be re- flected in just a few decades in the position of the actual timberlines that lie at topographically and edaphically favourable sites on Aakenustun- turi, Pyhätunturi and Yllästunturi. This is suggested by the fact that there are several places on Aak- enustunturi (Aho 1999) and on Pyhätunturi and Yl- lästunturi (Sonninen 1993; Tuovinen 1997) where spruces and pines having started their growth in the warm years of the 1920’s and 1930’s form vi- tal actual timberlines. At the treeline, the situation FENNIA 183: 1 (2005) 33The impact of elevation, topography and snow load damage of… is somewhat different, as the spruces and pines of the same favourable generation almost invari- ably show signs of suffering mainly as a result of damage caused by snow, wind-blown ice crystals, extreme climate and soil temperature fluctuations (Heikkinen et al. 2002). According to Holtmeier (1974, 1993) and Kullman (1983), many of the trees from the exceptionally warm period of the 1920’s and 1930’s have already died or suffer from damage caused by harsh, extreme climate. Decisive with regard to the successful advance of the actual timberline is that there are sufficiently many successive years with conditions that favour the development of tree seedlings and seeds, as was apparently the case in the 1920’s, 1930’s, early 1940’s, and again during the past three decades. Treelines climates are characterised by extreme, fluctuating temperatures and soil condi- tions, which in part impair the possibilities of seed- lings and adult trees to survive. The mere rise in air temperature alone does not guarantee the advance of the altitudinal forest limit, not to mention the stocking of treeless or almost treeless felltop areas with young stock. On the contrary, many phenom- ena introduced by climate change slow down or may even prevent the advance of the actual tim- berline (see Holtmeier et al. 2003). As summer seasons become warmer and maxi- mum temperatures rise (Carter & Saarikko 1996; Holopainen et al. 1996), the already high tempera- ture fluctuations on the ground surface and in the soil may give rise to drying of the soil surface and thus impair the ability of tree and seedling roots to obtain water and nutrients from the ground (Holt- meier 2003; Holtmeier et al. 2003) and the sprout- ing of seeds (Kallio & Lehtonen 1973; Holtmeier 2003) in the treeline ecotone and at the treeline. The projected possible increase in the occur- rence of harsh winds and storms (Holopainen et al. 1996) from the already high level would ac- celerate evaporation and enhance wind damage in felltop areas. The reduction of protective snow cover due to higher winds would increasingly ex- pose seedling roots to frost damage in winter (cf. Holtmeier et al. 2003). The predicted increase of winter precipitation due to global cimate change (Holopainen et al. 1996) may accelerate the accumulation of snow on trees, and thus slow down or even prevent the advance of the actual timberline. The ice-free waters of Bothnian Bay release warm humid air particularly in early winter, giving rise to extensive frost formation also on the fells of Lapland when the air cools down (Solantie 1974). The open-wa- ter season on the Bothnian Bay is predicted to be- come longer (Kauppi & Kämäri 1996) and hence, the snow load formation on trees may increase. These would mean an increased risk of heavy snow load damage to trees on the fells of Lapland (see Jalkanen & Konôpka 1998). On the fells examined here, it is likely that as a result of climate warming, the actual timberline will advance in the coming decades on gentle slopes where ground-related factors resemble those of a closed forest (see Fig. 8). This process requires, however, that there are no extreme weather con- ditions, and that the competition between conifer seedlings and dwarf shrubs and mosses for light and nutrients is not too intense, at least not dur- ing the first growing seasons (Kallio & Lehtonen 1973; Holtmeier 2003). Thus the altitudinal posi- tion of the actual timberline can ascend through a process in which the seedling stocks currently standing at or above the actual timberline develop into forests, and many of the birches, spruces and pines growing above the altitudinal timberline reach a height of 2 metres, which is considered the minimum for trees (Fig. 13) (see Holtmeier 2003). The vertical ascend of the actual timberline from the present maximum elevations (ca. 440–460 m a.s.l.), mainly at climatic timberlines, will be some tens of metres at best. In felltop areas, it is most uncertain whether trees can grow there success- fully at all. Orographic timberlines at steep slopes and abundant block fields can hardly be expected to advance as a result of climate warming – not even after long periods of time (see Holtmeier 2003). The potential advance of actual timberline will vary substantially from place to place due to variations in topography, edaphic factors and mi- croclimates. Conclusions The key conclusions of the research can be sum- marised as follows: • The average actual timberline elevation varies from ca. 370 metres to 400 metres a.s.l., de- pending on slope aspect. On Aakenustunturi and Yllästunturi, the timberline reaches the aver- age and maximum elevations on the west-facing and south-facing slopes, whereas the limit for continuous forest is lowest on the steep south- ern slope of Pyhätunturi. The lowest vertical for- est limit in the study area lies at approximately 34 FENNIA 183: 1 (2005)Jyrki Autio and Alfred Colpaert 270 m a.s.l., and the highest continuous forest reaches almost 460 m a.s.l. The actual timberline is hardly ever composed of a single tree species but rather featuring alternating occurrences of spruce, pine and mountain birch. • The proportion of block fields and the slope gra- dient are the most significant and visible factors controlling the altitudinal position of the actual timberline. Timberlines dictated by block fields make up at ca. 10% of the total length of the actual timberline on Aakenustunturi, ca. 50% at Peurakero on Pyhätunturi and ca. 30% on Yl- lästunturi. North-facing slopes with inclination 30°–44° are susceptible to avalanches, which are a substantial agent affecting the actual tim- berline position. On slopes with a gentle incli- nation and no block fields, the actual timberline reaches its highest elevation. Here the limit of continuous forest is controlled by climate. • Delayed snow deposits effectively block the ad- vance of the actual timberline mainly on the north-facing and east-facing slopes. On Aak- enustunturi, for instance, the actual timberlines bordered by delayed snow deposits constituted ca. 5% of the length of the entire actual timber- line. • The amount of snow load damage depends on tree species, slope aspect, to some extent on slope inclination and most importantly on ter- rain elevation. The average snow load damages at the actual timberline on Aakenustunturi and on Yllästunturi, respectively, are 66% and 48% to pine, as compared with 21% and 40% for spruce, respectively. Snow load damage evi- dently debilitates the trees at and above the actual timberline and therefore hampers the potential advance of the actual timberline and treeline. • Recent successful regeneration, seedling estab- lishment and young growth just above the ac- tual timberline and in the treeline ecotone on gentle southern and western slopes with no block fields suggest that forests are in the proc- ess of ascend due the climate warming. This will take place more slowly on slopes with northern and eastern aspect. It is, however, unlikely that timberlines determined by steep slopes and the presence of block fields would make any major advance. Results suggest that natural regenera- tion is most uncertain at the treeline near the felltops. The projected climate change may bring along negative phenomena such as increasing snow loads on trees and greater occurrence of strong winds, which may slow down or even prevent the advance of timberline. ACKNOWLEDGEMENTS We wish to thank Anja Kaunisoja for drawing the Fig. 1 and Antti Rönkkö for translating this text into English. Thanks are also due to Olavi Heikkinen and two anonymous referees for their valuable comments on the manuscript. REFERENCES Aho R (1999). Ilmaston vaikutus kuusen (Picea abies) paksuuskasvuun Aakenustunturin metsänrajavyö- hykkeellä. Unpublished Master’s thesis. Depart- ment of Geography, University of Oulu. 83 p. Ahti T, L Hämet-Ahti & J Jalas (1968). Vegetation and their sections in northwestern Europe. Annales Botanici Fennici 5, 169–211. 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. Nor- dia Geographical Publications 24: 2, 103–111. Autio J (1997). Peräpohjolan tuntureiden kasvil- lisuustyypit, kasvillisuuden korkeusvyöhykkeet ja metsänrajat. (Vegetation types, altitudinal vegeta- tion zones and timberline at the fells of Peräpoh- jola in Finnish Lapland). Unpublished Phil. Lic. thesis. Department of Geography, University of Oulu. 164 p. Autio J & O Heikkinen (1999). Lumi vyöryy laattana, pallona, pulverina. Tiede 2000 3, 48–49. Autio J & O Heikkinen (2002). The climate of north- ern Finland. Fennia 180: 1–2, 61–66. Autio J & S 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, 31–47. Autio J & M Kinnunen (1992). Pyhätunturin kansal- lispuiston kangasmaiden kasvillisuutyypit ja kor- keusvyöhykkeet. Unpublished Master’s thesis. Department of Geography, University of Oulu. 160 p. Carter T & R Saarikko (eds) (1996). Maatalous. In Kuusisto E, L Kauppi & P Heikinheimo (eds). Il- mastonmuutos ja Suomi, 127–143. Helsinki Uni- versity Press, Helsinki. Climatological statistics in Finland 1961–1990 (1991). Supplement to the meteorological year- book of Finland 90: 1, 1–125. Colpaert A (1998). Satellite data and environmental GIS, from remotely sensed data to geographical information. Acta Universitatis Ouluensis A 307. 36 p. FENNIA 183: 1 (2005) 35The impact of elevation, topography and snow load damage of… Dahl E (1998). The phytogeography of northern Eu- rope. (British Isles, Fennoscandia and adjacent areas). 297 p. Cambridge University Press, Cam- bridge. Diaz HF & RS Bradley (1997). Temperature variations during the last century at high elevation sites. Cli- matic Change 36, 253–279. Eronen M (1979). The retreat of pine forest in Finnish Lapland since Holocene climatic optimum: a gen- eral discussion with radiocarbon evidence from subfossil pines. Fennia 157: 2, 93–114. Haapasaari M (1988). The oligotrophic heath vegeta- tion of northern Fennoscandia and its zonation. Acta Botannica Fennica 135, 1–219. Hämet-Ahti L (1988). Kasvillisuus ja kasvisto. In Ala- lammi P (ed). Atlas of Finland, Folio 141: Vegeta- tion and flora, 1–3. National Board of Survey and Geographical Society of Finland, Helsinki. Heikinheimo O (1920). Suomen lumituhoalueet ja niiden metsät (Referat: Die Schneeschadengebiete in Finnland und ihre Wälder). Communicationes Instituti Forestalis Fennici 3, 1–134. Heikkinen O, B Obrębska-Starkel & S Tuhkanen (1995). Introduction: the timberline – a changing battlefront. In Heikkinen O, B Obrębska-Starkel & S Tuhkanen (eds). Environmental aspects of the timberline in Finland and in the Polish Carpathi- ans. Prace Geograficzne 98, 7–16. Heikkinen O, M Tuovinen & J Autio (2002). What de- termines the timberline? Fennia 180: 1–2, 67–74. Holopainen E, M Heikinheimo & M Kulmala (eds) (1996). Ilmakehä. In Kuusisto E, L Kauppi & P Hei- kinheimo (eds). Ilmastonmuutos ja Suomi, 11–69. Helsinki University Press, Helsinki. Holtmeier F-K (1973). Geoecological aspects of tim- berline in northern and central Europe. Arctic and Alpine Research 5, 45–54. Holtmeier F-K (1974). Geoökologische Beobachtun- gen und Studien an der subaktischen und alpinen Waldgrenze in vergleichender Sicht (nördliches Fennoskandien/Zentralalpen). Erdwissenschatfli- che Forschung VIII. 130 p. Holtmeier F-K (1993). Timberlines as indicators of climatic changes: problems and research needs. Paläoklimaforchung 9, 211–222. Holtmeier F-K (1997). Timberlines: research in Europe and North America. In Lovén L & S Salmela (eds). Pallas-Symposium 1996. Proceedings of the re- search symposium held in the Pallas-Ounastunturi National Park on 10.–11.10.1996. Finnish Forest Research Institute, Research Papers 623, 23–36. Holtmeier F-K (2000). Die Höhengrenze der Gebirgs- wälder. Arbeiten aus dem Insititut für Landschafts- ökologie 8. 337 p. Holtmeier F-K (2003). Mountain timberlines – ecol- ogy, patchiness, and dynamics. Advances in Glo- bal Change Research 14. 369 p. Holtmeier F-K, G Broll, A Müterthies & K Anschlag (2003). Regeneration of trees in the treeline eco- tone: northern Finnish Lapland. Fennia 181: 2, 103–128. Hustich I (1937). Pflanzengeographische Studien im Gebiet der niederen Fjelde im westlichen finn- ischen Lappland I. Acta Botanica Fennica 19, 1–156. Hustich I (1966). On the forest tundra and northern tree-lines. Reports from the Kevo Subarctic Re- search Station 3, 1–47. Jalkanen R & B Konôpka (1998). Snow-packing as a potential harmful factor on Picea abies, Pinus syl- vestris and Betula pubescens at high altitude in northern Finland. European Journal of Forest Pa- thology 28, 373–382. Juntunen V, S Neuvonen, Y Norokorpi & T Tasanen (2002). Potential for timberline advance in north- ern Finland, as revealed by monitoring during 1983–99. Arctic 55: 4, 348–361. Kallio P & J Lehtonen (1973). Birch forest damage caused by Oporinia autumnata (Bkh.) in 1965– 1966 in Utsjoki, N Finland. Reports from the Kevo Subarctic Research Station 10, 55–69. Kangas E (1959). Lumenmurrot ja metsänhoito. Met- sätaloudellinen aikakauslehti 1, 3–4. Kauppi L & J Kämäri (eds) (1996). Vedet. In Kuusisto E, L Kauppi & P Heikinheimo (eds). Ilmastonmuu- tos ja Suomi, 145–178. Helsinki University Press, Helsinki. Kellomäki S (ed) (1996). Metsät. In Kuusisto E, L Kauppi & P Heikinheimo (eds). Ilmastonmuutos ja Suomi, 71–106. Helsinki University Press, Hel- sinki. Kellomäki S, H Väisänen & T Kollström (1997). Model computations on the effects of elevating tempera- ture and atmospheric CO2 on the regeneration of Scots pine at timber line in Finland. Climatic Change 37, 683–708. Kjällgren L & L Kullman (2002). Geographical pat- terns of tree-limits of Norway spruce and Scots pine in southern Swedish Scandes. Norwegian Journal of Geography 56, 237–245. Köppen W & R Geiger (1936). Handbuch der Klima- tologie. 556 p. Verlag von Gebrüder Borntraeger, Berlin. Korpi M (2001). Tykyn muodostuminen ja vaikutuk- set puustoon Pallastunturin alueella. Unpublished Master’s thesis. Department of Geography, Uni- versity of Oulu. 74 p. Kujansuu R (1967). On the deglaciation of west- ern Finnish Lapland. Bulletin de la Comission Géologique de Finlande 232, 1–98. Kullman L (1979). Change and stability in the altitude of birch tree-limit in the southern Swedish Scan- des 1915–1975. Acta Phytogeographica Suecica 65, 1–121. Kullman L (1983). Past and present tree-lines of differ- ent species in the Handölan valley, Central Swe- den. Collection Nordicana 47, 25–45. Kullman L (2000). Tree-limit rise and recent warm- ing: a geoecological case study from the Swed- ish Scandes. Norwegian Journal of Geography 54, 49–59. 36 FENNIA 183: 1 (2005)Jyrki Autio and Alfred Colpaert Kullman L & L Kjällgren (2001). A coherent postgla- cial tree-limit chronology (Pinus sylvestris L.) for the Swedish Scandes. Aspects of paleoclimate and ‘recent warming’ based on megafossil evi- dence. Arctic, Antarctic and Alpine Research 43, 419–428. Lehtonen P (1992). Wind measurements at Ylläs broadcasting station 1982–1991. In Tammelin B, K Säntti, E Peltola & H Neuvonen (eds). BOREAS: an international experts’ meeting on wind power in icing conditions, Hetta 10.–13.2.1992, North wind – Pohjatuuli, 1–10. Finnish Meteorological Institute, Helsinki. Mäkinen K & M Maunu (1984). Pohjois-Suomen maaperä. Geologinen Pohjois-Suomi. Acta Lap- ponica Fenniae 12, 51–84. Mäkitalo K, R Sutinen, E Hyvönen, E Pulkkinen, M Pänttäjä & M-L Sutinen (1994). Määrääkö maaperä kuusen polaarisen metsänrajan Pohjois- Lapissa? Metsäntutkimuslaitoksen tiedonantoja 539, 35–48. Manner R & T Tervo (1988). Lapin geologiaa, hiekka- rannoista tuntureiksi, tulivuorista tasangoiksi, mannerjäätiköistä maaperäksi. 188 p. Lapin maakuntaliitto ry & Lapin lääninhallitus, Rova- niemi. Mayer H & E Ott (1991). Gebirgswaldbau Schutz- waldpflege. 587 p. Gustav Fisher Verlag, Stutt- gart. McCarroll D, J Autio, O Heikkinen & L Koutaniemi (1996). Degree of rock surface weathering on fjell summits in northern Finland: implications for thermal regime of last ice sheet. Boreas 25, 1–7. Mielikäinen P (1979). Geological map of Finland, Quaternary deposits 1:100,000, sheet 3642 Pyhä- tunturi. Geological survey of Finland. Mustonen T (1997). Tykkyvauriot Kittilän Aakenustun- turilla. (Snow load damage at the fell Aakenustun- turi, Kittilä municipality). Unpublished Master’s thesis. Department of Geography, University of Oulu. 55 p. Norokorpi Y (1994). Havumetsänrajan määräytymi- nen Suomessa. Metsäntutkimuslaitoksen tiedon- antoja 539, 7–15. Norokorpi Y & S Kärkkäinen (1985). Maaston korkeu- den vaikutus puusto- ja kasvupaikkatunnuksiin sekä tykkytuhoihin Kuusamossa. Folia Forestalia 632, 1–26. Nykänen M-L, H Peltola, CP Quince, S Kellomäki & M Broadgate (1997). Factors affecting snow load damage of trees with particular reference to Euro- pean conditions. Silva Fennica 2, 193–213. Odland A (1996). Differences in the vertical distribu- tion pattern of Betula pubescens in Norway and its ecological significance. Paläoklimaforschung 20, 42–59. Owner’s manual & reference (1998). 60 p. GARMIN Corporation, Taiwan. Päätalo M-L (2000). Snow load damage to Scots pine, Norway spruce and birch: modell approach. 92 p. Joensuun yliopisto, Joensuu. Påhlsson L (ed) (1995). Vegetationtyper i Norden. Tema Nord 1994, 1–665. Nordiska Ministerrådet, København. Perttu K (1972). Skogsgänsens beroende av olika kli- matologiska och topografiska faktorer. Skoghög- skulan, Institutionen för skogsföryngring, Rap- porter och Uppsatser 34, 1–91. Piirola J (1969). Rakkakenttien suhde metsänrajoihin Saariselällä (Summary: Realtionship between Block Fields and Forest Limits in Saariselkä, Finn- ish Lapland). Terra 81: 3, 206–208. Rastas P (1984). Geological map of Finland, Quater- nary deposits 1:100,000, sheet 2732 Kittilä. Geo- logical survey of Finland. Sirén G (1993). Advances and retreat of pine tree and timber lines in the far north of Finland. World Re- source Review 5: 1, 104–110. Skre O (1990). Consequences of possible climatic temperature change for plant production and growth in alpine and sublpine areas in Fenno- scandia. NINA Notat 4, 18–37. Solantie R (1974). Pohjois-Suomen lumipeitteestä (Summary: On Snow Cover in Northern Finland). Acta Lapponica Fenniae 8, 74–89. Sonninen M (1993). Puiden kasvumuodot ja kasvun vaihtelut korkeusvyöhykkeittäin Pyhätunturin Laakakerolla. Unpublished Master’s thesis. Depart- ment of Geography, University of Oulu. 141 p. SPSS Base 8.0, Applications guide (1998). 365 p. SPSS Inc, Chigago. Tammelin B (1998). WASP predictions of wind pow- er potential upon arctic hills. In Tammelin B & K Säntti (eds). Wind energy production in cold cli- mate. BOREAS IV. Proceeding of an International Meeting, 31 March–2 April 1998, Hetta Finland, 164–170. Finnish Meteorological Institute, Hel- sinki. Tammelin B & K Säntti (1992). Huurrekertymät tuntu- rien lakialueilla. Rime accretation at the top areas of fells. Meteorologisia julkaisuja 19, 1–39. Tranquillini W (1979). Physiological ecology of the alpine timberline. Ecological studies 31. 137 p. Tuhkanen S (1993). Treeline in relation to climate, with special reference to oceanic areas. In Alden J, JL Mastrantonio & S Ødum (eds). Forest devel- opment in cold climates, 115–134. Plenum Press, New York. Tuovinen M (1997). Ilmaston vaikutus männyn (Pinus sylvestris L.) paksuuskasvuun Yllästunturin alti- tudinaalisella metsänrajavyöhykkeellä. Unpub- lished Phil. Lic. thesis. Department of Geography, University of Oulu. 113 p. Walter H (1968). Die Vegetation der Erde, II. Die Die gemässigten und arktischen Zonen. 1001 p. VEB Gustav Fisher Verlag, Jena. Wardle P (1993). Causes of alpine timberline: a re- view of the hypothesis. In Alden J, JL Mastrantonio & S Ødum (eds). Forest development in cold cli- mates, 89–103. Plenum Press, New York. Ylläs-Levi, Outdoor Map 1:50,000 (2003). Genimap Oy, Vantaa.