Holocene timberline fluctuations in the mid-mountains of Central Europe VÁCLAV TREML, VLASTA JANKOVSKÁ AND LIBOR PETR Treml, Václav, Vlasta Jankovská & Libor Petr (2006). Holocene timberline fluc- tuations in the mid-mountains of Central Europe. Fennia 184: 2, pp. 107–119. Helsinki. ISSN 0015-0010. Central European Hercynian mountain ranges – Krkonoše Mts., Hrubý Jeseník Mts. and Vosges, Schwarzwald, Harz – represent the only islands of alpine for- est-free area between the Alps and Western Carpathians in the south and the Scandes in the north. Based on data from previously published pollen profiles and on newly taken cores, comparison of the development of the alpine timber- line position is presented. The Labský d Ö ul profile in the Krkonoše Mts. spans the whole Holocene, the Keprník profile in the Hrubý Jeseník Mts. brings informa- tion from ca 2500 BP to the present. The exceptional position of the Krkonoše Mts. in terms of permanent presence of alpine forest-free area throughout the Holocene was confirmed. In other mountain ranges, the alpine forest-free areas probably vanished or were restricted only to the exposed peaks during the peri- ods of positive temperature anomalies in the Middle Holocene. Later in the pe- riod 4000–500 BP, forest free areas reappeared, although the relative contribu- tion of climatic and anthropogenic causes to their formation remains unclear. Taking into account the supposed extent of temperature oscillations in the Mid- dle and Late Holocene and the existing pollen records the authors assume that the alpine timberline in the Hercynian mid-mountains of Central Europe varied only slightly. Václav Treml, Department of Physical Geography and Geoecology, Faculty of Science, Charles University in Prague, Albertov 6, 128 43 Prague, Czech Repub- lic. E-mail: treml@natur.cuni.cz. Vlasta Jankovská, Institute of Botany, Czech Academy of Sciences, Po vrívcí 3b, 603 00 Brno, Czech Republic. Libor Petr, Department of Botany, Faculty of Science, Charles University in Prague, Benátská 2, 128 01 Prague, Czech Republic. MS received 3 August 2006. Introduction The term alpine timberline refers to the ecotone between forest and alpine belt (Körner 1999) or subalpine shrub formations. The factor determin- ing the absence of forest is the temperature decline related to the increasing altitude (Körner 1999). The alpine timberline ecotone presents a transition zone of varying width and structure (Armand 1992), inferring differences in ecotone response to changing temperature conditions. In general, the response of the alpine timberline to changing tem- perature conditions is somewhat delayed (Slatyer & Noble 1992; Paulsen et al. 2000). Its position is determined by long-term climatic trends rather than by immediate temperature conditions (Paulsen et al. 2000). In contrast, local or regional advances of treeline or tree-species line (for more detail, see Körner 1999) as response to the short or mid-term temperature increase can be very rapid (e.g. Bug- mann & Pfister 2000; Kullman 2007). Moreover, responses to climatic changes are sometimes op- posite, such as advance of timberline due to higher temperature and decline due to higher precipita- tion or drought (Wilmking et al. 2004). Certain variations of the alpine timberline posi- tion due to climatic oscillations during the Holo- cene can be detected. In Central Europe, timber- line fluctuation is estimated at up to 200 m (Tinner & Theurillat 2003). For the High Sudetes (Krkonoše and Hrubý Jeseník Mts.) and Western Carpathians, oscillations up to 400 m are reported (Firbas 1952; 108 FENNIA 184: 2 (2006)Václav Treml, Vlasta Jankovská and Libor Petr Ložek 2001) although this phenomenon still re- mains open. This study aims at summarising the development of the timberline position in the Hercynian moun- tains of Central Europe based on detailed interpre- tation of previously published data and on newly constructed pollen diagrams from Krkonoše and Hrubý Jeseník profiles. Methods Study area Hercynian mountain ranges of Central Europe with their primary forest-free areas represent the only alpine islands between the Scandes, the Alps and Western Carpathians (Jeník 1998). They include Vosges, Harz, Krkonoše Mts. (~ Giant mountains/ Karkonosze/Riesengebirge) and Hrubý Jeseník Mts. (Fig. 1). Large forest-free area has developed also in the Schwarzwald Mts. but it is probably of secondary origin (Friedmann 2000). The same holds true for the forest-free area on the summit of Šumava (Bayerischer Wald) mountain range. All these mountains are characteristic by relatively high rainfall, from 2200 mm in the highest parts of the Vosges Mts. to 1500 mm in the Hrubý Jeseník Mts. The average July temperatures range from 8.7 °C in the highest parts of the Krkonoše Mts. (Snvežka 1602 m a.s.l.) to 11.5 °C in the Vosges Mts. (Hohneck, 1363 m a.s.l.). Apart from the Vos- ges, where European beech (Fagus sylvatica L.) is present, the timberline is formed by Norway spruce (Picea abies [L.] Karst.). The extent of forest-free area has increased in the past by deforestation in all the mentioned mountain ranges (Jeník & Lok- venc 1962; Jeník & Hampel 1991; Friedmann 2000; Schwartz et al. 2005). Pollen analysis and sedimentology This study presents two new pollen profiles that bring information about history of the alpine belt in the Krkonoše and Hrubý Jeseník Mts. Pollen cores were collected in the Labský d Ö ul valley (Krkonoše Mts.) and Keprník Mt. (Hrubý Jeseník Mts.). The locality Labský d Ö ul is situated in a lake of glacial origin filled with deposits (Engel et al. 2005). This profile provided the longest pollen record ever analyzed in Krkonoše: from late gla- cial period to the present. The locality lies close to the timberline which is lowered here by avalanch- es. The 1283 cm long core (cirque bottom, Fig. 1. Locations of the inves- tigated Central European Hercynian mid-mountains. FENNIA 184: 2 (2006) 109Holocene timberline fluctuations in the mid-mountains of … Table 1. Radiocarbon data from Labský d Ö ul and Keprník sites. Site Depth (cm) Lab. No. Dated material 14C Age (uncal. BP) Laboratory Labský d Ö ul 205 ERL 6295 peat 4080 ± 49 Phys. Inst. der Uni. Erlangen Labský d Ö ul 230–250 CU 1916 peat 4380 ± 148 Radiocarbon Laboratory KHIG PvrfUK Labský d Ö ul 354 ERL 6318 peat 5024 ± 53 Phys. Inst. der Uni. Erlangen Labský d Ö ul 438 ERL 6319 wood fragment 5272 ± 57 Phys. Inst. der Uni. Erlangen Labský d Ö ul 547 Poz-13708 wood fragment 5780 ± 60 Poznań Radiocarbon Laboratory Labský d Ö ul 797 ERL 7380 plant macroremnant 8216 ± 94 Phys. Inst. der Uni. Erlangen Labský d Ö ul 963 ERL 6184 plant macroremnant 9572 ± 54 Phys. Inst. der Uni. Erlangen Keprník 50–51 Poz-13744 peat 2090 ± 35 Poznań Radiocarbon Laboratory 50°45'N, 15°33'E, 990 m a.s.l.) was collected with Russian peat sampler. Except for the peat layer up in the profile, the core contained mainly of inor- ganic matter. The section between 810 and 830 cm contamined during sampling and was discard- ed from analyses. The profile from Mount Keprník (50°10'N, 17°07'E, 1429 m a.s.l.) was taken from a pit dug in an earth hummock. The organic sediment had a peaty character with sedge remains. The locality is situated on a plateau in the northern part of the mountain range at a distance of 200 m (60 m of altitude) above the timberline. The samples for pollen analysis were processed with standard acetolysis method (Moore et al. 1991), and pollen grains were identified according to the handbook by Beug (2004). Pollen diagrams including stratigraphic zones were created using TILIA (Grimm 1992) software. The sediment from Labský d Ö ul was analysed for organic matter con- tent (determined by loss-on-ignition, LOI) and par- ticle size distribution (determined by wet sieving). Based on these two indicators, the litostratigraphic units (segments with similar characteristics e.g., particle size distribution, organic matter propor- tion, colour) of the profile were determined. Radio- carbon dating (14C) was carried out by laboratories in Erlangen, Poznan (for accelerator mass spec- trometry, AMS) and at the Faculty of Science of the Charles University in Prague (for conventional 14C) (Table 1). Absolute data are expressed, if not indi- cated otherwise, as uncalibrated radiocarbon years BP. Linear age-depth model (R2 = 0.98) from un- calibrated radiocarbon data was, in the case of Labský d Ö ul, created for the depth ranging from 205 to 963 cm. Chronostratigraphic zones are used ac- cording to Lang (1994). Interpretation of the alpine timberline altitudinal shifts When interpreting the pollen diagrams in relation to the timberline position, main emphasis was put on the proportion between woody species and herbs (AP/NAP). This proportion allows approxi- mate determination between the dominance of forests and forest-free areas in the vicinity of the profile. Macroscopic remains of plant species (macrosubfossils), however, would provide better information about the timberline position (Tinner & Theurillat 2003). Unfortunately, macroremains of plant species were not mentioned in the major- ity of previously published studies, so we had to concentrate on the AP/NAP percentage when in- terpreting the timberline fluctuation. Concerning the AP/NAP rate, critical arboreal proportion was considered to be 70–80%, repre- senting the approximate value of woody species proportion in the central European profiles situ- ated at or just below timberline (e.g., Gouillé Rion, Gouillé Loéré, Grande Tsa [Tinner et al. 1996; Tinner & Theurillat 2003; Wright et al. 2003], and the northern side of Vysoké Tatry Mts. [Obidowicz 1993]). For variations of AP/NAP, it was always taken into account whether the change of the value was due not only to the change of dominant species with different (higher or lower) pollen production. In a case where timberline was constituted by spruce at the time, a propor- 110 FENNIA 184: 2 (2006)Václav Treml, Vlasta Jankovská and Libor Petr tion of ≥ 50% spruce pollen was also considered indicative of closed stands at the locality (Obido- wicz 1993). Results Krkonoše Mts. The profile from Labský d Ö ul allows reconstruction of the early Holocene timberline. The simplified pollen diagram (Fig. 2) shows data from Late Gla- cial period to ca 8000 BP corresponding to the profile depth from 1283 cm to 790 cm. According to the pollen analyses from lower parts of the pro- file, the markedly different pollen composition be- tween 810 and 830 cm (Jankovská 2004) is due to contamination from the upper parts of the core (ca. 500–650 cm). Pollen curves of Salix, Junipe- rus, Betula nana type, Ephedra distachya and E. fragilis type and also the presence of Pinus hap- loxylon type (i.e. Pinus cembra) clearly delimit the forest-free stage in the profile. Higher pollen curve of Pinus sylvestris type is produced by pollen from local Pinus mugo stands or by influx of P. sylvestris from lower altitudes, or by combination of both factors. In this period (i.e. before ca 9500 BP), al- pine timberline was probably situated clearly be- low the investigated site. The first significant oscil- lation of AP/NAP curve is recorded at around 1200 cm (LD1). Before this period (early stage after de- glaciation), closed pine forests have probably not occurred in the vicinity of the lake and thus, re- gional (climatic) rather than local driving factors are suggested. At around 9600 BP (LD2), a markedly lower proportion of arboreal pollen was recorded. It is not clear whether the decrease of AP curve was caused by local disturbance or by climatic influ- ence. During the following period, the AP curve ap- proaches 80% and passes it constantly at 8200 BP. At the same time, the proportion of organic matter in the lake sediment rises from 7–9% to 20–25%. Apparently, the timberline composing of pioneer woody species Pinus and Betula passed the level of Labský d Ö ul profile as late as around 9200–8800 BP. Another decrease of AP pollen takes place at 8200 BP (LD3), set off by resedimentation below this part of the profile. Other parameters also, such as increase of Betula nana type and Juniperus, and lower organic content in the sediment, probably indicate the increase of forest-free area either due to local disturbance or climate. The AP/NAP oscillation mentioned above repre- sents only minor wiggles in AP/NAP curve in the range from 85 to 95% arboreal pollen. These slight shifts in the AP/NAP curve probably reflect local vegetation changes and/or avalanche events. Hence, more elevated localities can give more evidence about the alpine timberline position dur- ing the most part of Middle and Late Holocene. Norway spruce has been sporadically present at the Labský d Ö ul site (based on stomata) since 7500 BP, more abundantly from 6800 BP, and as a major timberline forming species since the Atlantic peri- od (Jankovská 2004). Hrubý Jeseník Mts. The investigated pollen profile taken from an earth hummock at the summit of the Keprník Mt. covers a time span since ca 2500 BP and hence, it records a period similar to other pollen profiles analysed in the Hrubý Jeseník Mts. (Rybnívcek & Rybnívcková 2004). At the summit of Mount Keprník, the AP propor- tion in the mentioned time period was between 70 and 80% (zone K1, K2, Fig. 3). At the same time, no woody species stomata were found in the in- vestigated profile. The profile shows a regression of woody species typical for mixed oak forests and progression of beech and fir in the K2 zone. Direct indicators of human activities are present in the uppermost layers (K3). In the whole profile, there are only minor shifts in AP/NAP curve. With re- spect to the absence of spruce stomata remnants and AP < 80% (in exposed windy position with potentially high pollen influx from lower areas), it can be concluded that closed forest has not been present in the summit area of the Keprník Mt. dur- ing the last 2500 years. Other evidence includes occurrence of earth hummocks since at least 2090 BP. These landforms are usually quickly destroyed after colonisation by trees (Treml & K vrížek 2006). Discussion Evolution of the alpine timberline in the Krkonoše Mts. and the Hrubý Jeseník Mts. In the Krkonoše Mts., the alpine timberline posi- tion during the Younger Dryas can be estimated at 500–600 m according to the equilibrium line alti- FENNIA 184: 2 (2006) 111Holocene timberline fluctuations in the mid-mountains of … Fig. 2. Simplified pollen diagram, radiocarbon data, loss-on-ignition and litostratigraphic units (U1–U5) in the “Labský d Ö ul” core. Light rectangle across pollen diagram indices contaminated part of the profile (810–830 cm depth). tude at approximately 1200 m a.s.l. It follows that the timberline reached the upper locations with a time lag which corresponds to the relatively late deglaciation (Bourlés et al. 2004). Three distinct oscillations of AP/NAP curve were detected before the timberline finally passed the Labský d Ö ul cirque bottom. It is suggested that the first one (LD1) rep- resents a climatically driven timberline descent 112 FENNIA 184: 2 (2006)Václav Treml, Vlasta Jankovská and Libor Petr Fig. 3. Simplified pollen diagram, summit of Mt. Keprník (1429 m a.s.l.). rather than a result of local disturbances. The sec- ond significant oscillation of AP/NAP curve (LD2) also reflects regional rather than local cause. Be- fore this period, timberline had still remained be- low the Labský d Ö ul site and therefore, local distur- bance in closed pine-birch forest would not be significantly manifested in the AP/NAP curve. The third oscillation (LD3) could be correlated with the central European oscillation CE 2 (Haas et al. 1998), which was recorded also in the High Tatra Mts. (Kotarba & Baumgart-Kotarba 1999). The Labský d Ö ul profile gives insight to timberline position from deglaciation until ca 8000 BP, whereas other pollen profiles from the summit ar- eas of the Krkonoše Mts. (Fig. 4) cover a time pe- riod since 7600 ± 130 BP, which is the time span of the Panvcava peat bog profile (Huettemann & Bortenschlager 1987). This peat bog is located on an exposed highly elevated planated surface and thus, it is supposed a prominent part of its pollen is brought by wind from windward valleys (Jeník 1998). High proportion of herb and dwarf shrub pollen was detected (25–30% Gramineae, 10% Calluna) at the bottom of this profile (Huettemann & Bortenschlager 1987). This suggests that at the given time (~ 7600 BP), either the timberline had not yet reached the Pan vcava peat bog (1300 m a.s.l.) or that it had already been lowered below this level. In the following period (after 7400 BP according to Huettemann & Bortenschlager 1987), the alpine timberline reached at least the level of the Panvcava peat bog and since varied apparently only a little. Nevertheless, according to Speranza et al. (2000) there was a colder period between 2640 ± 60 and 2480 ± 35 BP, but there is no evi- dence of a timberline shift due to this temperature oscillation. In the following period, no marked trend in for- estation or deforestation was detected in the high- est parts of the Krkonoše Mts. At less elevated Pan- vcava peat bog, AP reaches 90% of the pollen spec- trum (Jankovská 2001), and at Úpa peat bog (1430 m a.s.l.) it reaches approximately 80% (Svobodová 2004). Considerable proportion (20–30%) of the pollen spectrum belongs to pine, mainly from lo- cal Pinus mugo stands. At the Pan vcava peat bog, Picea pollen represents about 10–20% during 4000–800 BP (Jankovká FENNIA 184: 2 (2006) 113Holocene timberline fluctuations in the mid-mountains of … Fig. 4. Positions of described pollen profiles within the Krkonoše and Hrubý Jeseník Mts. (1) Labský d Ö ul; (2) Pan vcava peat bog (Huetteman & Bortenschlager 1987; Speranza et al. 2000; Jankovská 2001); (3) Úpa peat bog (Svobodová 2004); (4) Keprník; (5) Velký D ved; (6) Barborka; (7) Velká kotlina; (8) Velký Máj (5–8; Rybní vcek & Rybnívcková 2004). 2001, 2004), while at Úpa peat bog it reaches only 5–15% (Svobodová 2004). This is probably a con- sequence of longer distance between the Úpa peat bog and the timberline during this period. Moreo- ver, both Pan vcava and Úpa pollen profiles show significantly smaller number of Picea pollen, as compared with sites recently surrounded by spruce forest (Labský D Ö ul, Barborka). This indicates that at least Úpa peat bog was situated above the alpine timberline during 4000–800 BP. Human-induced changes in vegetation are present in the above mentioned pollen profiles since the Medieval Pe- riod (Jankovská 2004). Based on existing data, it is not possible to de- termine the exact level that the closed forest reached during the climatic optimum of the Holocene (i.e. Atlantic chronozone, Lang 1994). Most likely, a closed tall-trunk stand cannot be ex- pected at locations where well developed sorted patterned ground is present (above ca. 1430–1450 m a.s.l.). Should these landforms become over- grown by trees, they usually lose their raised cen- tre morphology, which is not the case of the above mentioned patterned ground (Sekyra et al. 2002). The maximum elevation of closed forest could, therefore, have been only about 100 m higher than today. Profiles from the Hrubý Jeseník Mts. (Fig. 4) con- tain pollen record from 4620 BP onwards. High proportion of hazel pollen indicates hazel stands at the summit locations (about 1300 m a.s.l.) dur- ing the period 4620–3500 BP (Rybní vcek & Ryb- nívcková 2004), although direct evidence in the form of macroscopic remains is missing. Some presently forest-free localities show rela- tively low proportions of arboreal pollen also dur- ing the period 1945–800 BP: approximately 60% at Velká Máj (peat bog on the summit plateau) and at Velká kotlina (peat bog on the cirque bottom), where the oscillations of AP/NAP curve were more pronounced (Rybní vcek & Rybní vcková 2004). This indicates permanently forest-free areas during this period, although significant local changes in tim- berline position probably took place in the area of 114 FENNIA 184: 2 (2006)Václav Treml, Vlasta Jankovská and Libor Petr the Velká kotlina cirque. Compared to both the above mentioned sites, the proportion of arboreal pollen was higher at Mt. Keprník (70–80%). How- ever, the presence of earth hummocks at the sum- mit of Keprník since at least ca. 2100 BP is consid- ered a proof of absence of forest, as such landforms could not persist the physical action by tree roots in a closed forest (Treml & Kvrížek 2006). Similar earth hummocks of the same age occur also on the summit of Prad ved Mt. at 1492 m a.s.l. (Treml et al. 2006). At presently forested localities such as Barborka and Velký Dved (Rybnívcek & Rybnívcková 2004), the AP proportion between 3700 and 800 BP fluctuates at around 85%. The noticeable decrease of woody species pollen, namely beech and fir, observed in most profiles at around 500 BP (zone K3 in the case of Keprník), can be ascribed to human influence. However, a synergic action of the last Little Ice Age could be involved aside human influence in the up- permost locations of the Hrubý Jeseník Mts. during the last 800–500 years (Hošek 1972). Based on the evidence from all mentioned pol- len profiles, it can be concluded that in the Hrubý Jeseník Mts. the alpine timberline did not advance to the most exposed summits (e.g. Keprník Mt., Velký Máj Mt.) during the period ca 2000–800 (500) BP and that forest free areas also persisted on steep slopes of the Velká kotlina cirque. The rest of the presently forest-free areas, however, were for- ested and remained so until as late as 800–500 BP. Alpine timberlines in the neighbouring mid- mountains: Vosges, Schwarzwald and Harz The extent of alpine forest-free area in the neigh- bouring Hercynian mid-mountains is quite limited (Tables 2 and 3). In the Vosges Mts., forest-free ar- eas of natural origin are found only in the highest exposed parts of the summit plateaus (Carbiener 1963). Large part of the area previously called “chaumes primaries” originates from deforestation during the Iron Age (Schwartz et al. 2005). During Younger Dryas, timberline altitude is estimated at 500 m (Schloss 1979), ascending rapidly to at least 1100 m in the early Holocene with the advance of pine and birch (Schloss 1979; Edelman 1985). At the beginning of the Boreal period (ca. 9000 BP), timberline reached an altitude of ca. 1200 m a.s.l. (Lemée 1963). In the Vosges Mts., the early Holo- cene timberline development may have been af- fected by mesoclimatic phenomena of some val- leys with glaciation relicts (Mercier et al. 1999). According to Schwartz et al. (2005), Tilia occurred at an altitude of 1060 m at around 8000 cal. BP (ca 7200 BP), whereas at present its occurrence does not exceed 900 m. This fact suggests warmer cli- mate than at present. During the climatic optimum (8000–4500 BP), even the highest parts were for- ested. After ca. 4500 BP, beech stands in the high- est parts thinned down and the timberline reap- peared (De Valk 1981). At the most exposed loca- tions, it existed until the beginning of summer farming activities (grazing, hay making, forest clearance), which resulted in timberline lowering and formation of the majority of the present-day forest-free areas at around 1400–1200 BP (Schwartz et al. 2005). In Schwarzwald, the altitude of the timberline in the Younger Dryas is estimated at 750 m (Lang 2006). During the early Holocene it advanced rap- idly to the level of the highest peaks where closed stands of pine and birch were formed. Present for- est-free area is probably secondary and originates from the period of expansion of the summer farm- ing to the highest parts at around 1000 AD (Bogen- rieder 1982; Friedmann 2000). In the Harz Mts., the summit of Brocken Mt. (1141 m a.s.l.) was forested during the climatic op- timum of the Holocene (Firbas 1952; Beug et al. 1999). Altogether four treeless periods are docu- mented in the Brocken summit area (Beug et al. 1999): 1) from Younger Dryas to 9700 BP, 2) from ca 5700 to 5300 BP, 3) from 2900 to 2800 BP and 4) after 500 BP. The latest forest-free period is doc- umented from the 16th century, i.e. before the in- tensified human influence (Tackenberg et al. 1997). However, the evidence for a natural origin of the forest-free area at the Brocken summit is mainly floristic (Hauepler 1970) and historical (Tacken- berg et al. 1997). Although direct evidence (ar- chaeological findings, soil charcoal, etc.) of early anthropogenic impact is missing, human contribu- tion to the formation of the forest-free area at Brocken summit can not be excluded (Beug et al. 1999). Extent of the alpine belt during the Holocene Based on the present vertical extent of the alpine belt in various Hercynian mountains of Central Eu- rope (Table 3) it can be assumed that the most “en- dangered” alpine forest-free areas during the Holocene were situated in the Harz and Vosges FENNIA 184: 2 (2006) 115Holocene timberline fluctuations in the mid-mountains of … Table 2. List of sources and sites which were used for reconstruction of alpine timberline position. Mountain range Site/Source Age start of record Proxy used for treeline reconstruction Author of study Vosges Gazon de Faing, 1230, 1290 m only relative biostratigraphic dating – since Boreal pollen Lemée 1963 Vosges Altenweiher 926 a.s.l., Moselotte 1290 ca 8000 BP resp. 2500 BP pollen De Valk 1981 Vosges Sewensee 500 m late glacial pollen Schloss 1979 Vosges Rossberg 1190 m 7600 BP – the oldest charcoal charcoals – soil profiles Schwartz et al. 2005 Vosges Goutte Loiselot 850 m late glacial pollen Edelman 1985 Vosges several sites on Hautes Chaumes (above 1200 m) soil profiles Carbiener 1963 Schwarzwald 9 sites (654–1280 m) late glacial pollen Lang 2006 Harz Brocken summit area historic data Tackenberg et al. 1997 Harz Brocken summit area floristic data Hauepler 1970 Harz several sites in Brocken area (highest 1100 m) late glacial pollen Beug et al. 1999 Krkonoše Labský d Ö ul 990 m late glacial, first 14C date 9200 BP pollen, stomata this study Krkonoše Pan vcavské rašeliništ ve 1325 m 7600 BP pollen Huettemann & Borten- schlager 1987 Krkonoše Pan vcavské rašeliništ ve 1320 m 3100 BP pollen, pollen concentration Speranza et al. 2000 Krkonoše Pan vcavské rašeliništ ve 1325 m 3995 BP pollen Jankovská 2001 Krkonoše Úpské rašeliništve 1420 m 3440 BP pollen Svobodová 2004 Krkonoše summit plateaus (above 1430–50 m) persisted from late glacial soils this study Hrubý Jeseník Velký Máj 1350 m 1945 BP pollen Rybní vcek & Rybnívcková 2004 Hrubý Jeseník Velká Kotlina 1400 m ca 1700 BP pollen Rybní vcek & Rybnívcková 2004 Hrubý Jeseník Barborka 1315 m ca 3700 BP pollen Rybní vcek & Rybnívcková 2004 Hrubý Jeseník Velký D ved 1395 m 4600 BP pollen Rybní vcek & Rybnívcková 2004 Hrubý Jeseník Keprník 1423 m 2090 BP pollen this study Hrubý Jeseník Keprník 1415–1423 m, Prad ved 1450–1490 m 2100 BP earth hummocks – soils this study, Treml et al. (2006) Table 3. Recent vertical extent of the alpine belt in Hercynian mid-mountains of Central Europe – difference between eleva- tion of the highest peak and the uppermost outposts of the timberline. Values in brackets correspond to the average height of natural nondepressed alpine timberline (Treml & Banaš 2000). Maximal elevation of the alpine timberline (m a.s.l.) T** (°C) Vertical extent of the alpine belt (m) Corresponding gradient of summer temperature (°C) Vosges 1360 8.0 60 0.3–0.4 Harz 1125* 7.8 20 0.1–0.2 Krkonoše 1370 7.1 210 (300) 1.2–1.3 (1.8) Hrubý Jeseník 1430 7.2 60 (140) 0.3–0.4 (0.8) * … sensu Tackenberg et al. (1997) ** … average temperature April–September (growing season, 1981–1990) at the maximal elevation of the alpine timberline, calculated from data published by Migala (2005), with temperature lapse 0.6°C/100 m. 116 FENNIA 184: 2 (2006)Václav Treml, Vlasta Jankovská and Libor Petr Mts. During the latter half of the Holocene, when the woody species constituting the timberline to- day were already present in the Harz and Vosges Mts., there were probably no naturally treeless ar- eas during periods 0.5–1°C warmer than today. The existence of such positive temperature anoma- lies is likely, considering the recent Holocene tem- perature estimates from Central Europe (Haas et al. 1998; Hieri et al. 2003). This hypothesis is supported also by the pollen analyses. They show that during the Holocene, the alpine timberlines in Hercynian mountain ranges of the Central Europe developed in different ways. In the Krkonoše Mts. large forest-free areas were present throughout the Holocene, whereas in the Harz and Vosges Mts., even the uppermost loca- tions (except for steep slopes, rocks, block fields or exposed peaks) were forested during the climatic optimum (De Valk 1981; Beug et al. 1999). The alpine forest-free areas in the Vosges reappeared after 5000 BP (De Valk 1981). In the Hrubý Jeseník Mts., according to recent temperatures and Holo- cene temperature estimates (e.g., Hieri et al. 2003), the alpine forest-free area probably had a very lim- ited extent during the warmer periods of the Holocene. It expanded most likely before 2000– 2500 BP, which is the age of the part of the pollen profile documenting forest-free areas at summit lo- calities (Velký Máj, Velká Kotlina – Rybní vcek & Rybnívcková 2004, Keprník). Presence of treeless areas at the summits of the Hrubý Jeseník is also proved by an earth hummock rise around 2100 BP. In the Schwarzwald Mts. the actual relatively large forest-free area is of anthropogenic origin (Fried- mann 2000). The history of alpine areas can be related to present-day biodiversity. For example, the butterfly communities of the Krkonoše Mts. differ greatly from those present in the Harz, Hrubý Jeseník and Králický Sn vežník Mts. They are more similar to those of the West Carpathian mountain ranges (Mavrák & Kuras 2006). This fact could confirm the notable distinctions of the forest-free area devel- opment in the Krkonoše Mts. as compared to other mentioned Hercynian mountain ranges. The pres- ence of many plant species and diversified com- munities dependent strictly on forest-free areas in all those mountain ranges (Jeník 1998) indicates that long lasting forest-free enclaves have been in existence. Nevertheless, the presence of these for- est-free patches depend on soil conditions, water regime or slope inclination rather than on temper- ature. Rate of timberline fluctuation The maximum amplitude of timberline oscillations in Central European mid-mountains is governed by their low altitude and a very limited space be- tween their summits and the timberline. In the Krkonoše Mts., the presence of a closed forest can- not be expected at sites with well developed sorted forms of patterned ground at altitudes above 1450 m a.s.l. Sorted forms of patterned ground formed here at the end of the last glacial stage (Traczyk & Migovn 2003). The maximum difference in the timberline position compared to the present is, therefore, less than 100 m. In other Central Eu- ropean Hercynian mountain ranges, the timberline ascended to its highest locations. Minimum differ- ence in the timberline position ranged therefore from 25 m (Harz Mts.) to 40–150 m (Hrubý Jeseník Mts.). However, it is possible that during some of the cold oscillations recorded in the Central Eu- rope (for example CE 8, Haas et al. 1998) the tim- berline was situated lower than today and the total fluctuation would therefore be several tens of me- ters. As for the Vysoké Tatry Mts., Obidowicz (1993) argues that during the climatic optimum, the alpine timberline was only 50 to 100 m higher than today. The above mentioned smaller timber- line oscillations correspond to temperature recon- structions (Haas et al. 1998; Hieri et al. 2003) that estimate the extent of summer temperature fluctu- ations in mid and late Holocene to 1 °C. Neverthe- less, the changes in timberline position in this pe- riod may not have been as vigorous as earlier in the Holocene (Tinner & Kaltenrieder 2005) due the well established stable communities strongly influenced by competition. This is true for the Krkonoše Mts. even today: as temperature increas- es, the timberline ascends first at disturbed locali- ties with low herb cover (such as old debris flow tracks) and higher rates of seedling establishment (Treml 2004). Conclusions During the Holocene, Harz, Vosges, Schwarzwald and Hrubý Jeseník Mts. were prone to disappear- ance of alpine forest-free areas in the periods of favourable climatic conditions. Most likely, a tem- perature dependent alpine belt was absent from these areas during the climatic optimum. In con- trast, a large alpine area was maintained through- out the Holocene in the Krkonoše Mts. In the Hrubý FENNIA 184: 2 (2006) 117Holocene timberline fluctuations in the mid-mountains of … Jeseník Mts., a temperature dependent forest-free area existed at least since 2000 BP to the present. In the Krkonoše Mts., alpine timberline gradu- ally advanced from 500–600 m in the Younger Dryas to 1000 m (9200–8800 BP). After 7400 BP timberline ascended at least to 1320 m, which is the altitude of the Pan vcava peat bog. Maximum timberline position in the Krkonoše Mts. has not exceeded 1450 m a.s.l., as this is the lower limit of well developed sorted patterned ground formed in the late glacial. Therefore, it has passed the present maximum by 60 m only, and the average positions of the natural timberline by 150 m. During 4000– 800 BP, the alpine timberline was situated proba- bly below the Úpa peat bog (1420 m a.s.l.). In the Labský d Ö ul profile, three distinct oscillations of AP/ NAP curve were recorded during the early Holo- cene. Of these, at least two (LD1 and LD2 – 9600 BP) oscillations are believed to be timberline de- scends due to climatic shifts rather than local dis- turbances. ACKNOWLEDGEMENTS This study would not have been possible without the support from grant projects GACR 205/06/0587, VaV MZP SM/6/70/05 and MSM 0021620831. We also like to thank the two anonymous reviewers for valu- able comments on the manuscript, and Tereza Chý- lová for improving the English translation. REFERENCES Armand AD (1992). Sharp and gradual mountain tim- berlines as a result of species interaction. In Hansen AJ & F di Castri (eds). Landscape bounda- ries: consequences for biotic diversity and eco- logical flows. Ecological Studies 92, 360–378. Beug HJ (2004). Leitfaden der Pollenbestimmung: für Mitteleuropa und angrenzende Gebiete. 263 p. Verlag Dr. Friedrich Pfeil, München. Beug HJ, I Henrion & A Schmüsser (1999). Land- schaftgeschichte im Hochharz. 454 p. Papierflie- ger Verlag, Clausthal-Zellerfeld. Bogenrieder A (1982). Die Flora der Weidfelder, Moore, Felsen und Gewässer. In Der Feldberg im Schwarzwald. Die Natur und Landschaftschutzge- biete B.-W. 12, 244–316. Bourles DL, R Braucher, Z Engel, J Kalvoda & JL Mer- cier (2004). Deglaciation of the Giant Mountains indicated by 10Be dating. In Drbohlav D, J Kalvoda & V Voženílek (eds). Czech geography at the dawn of the millenium, 25–39. Nakladatelství Univerzi- ty Palackého, Olomouc. Bugmann H & C Pfister (2000). Impacts of interan- nual climate variability on past and future forest composition. Regional Environmental Change 1, 112–125. Carbiener R (1963). Les sols du massif du Hohneck, leurs rapports avec les tapis végétal. In R Carbie- ner. Le Hohneck, 103–152. Ass. philomatique d’Alsace et Lorraine, Strasbourg. De Valk EJ (1981). Late holocene and present vegeta- tion of the Kastelberg (Vosges, France). Disserta- tion. 120 p. University of Utrecht, Utrecht. Edelman HJ (1985). Late glacial and holocene vegeta- tion development of la Goutte Loiselot (Vosges, France). Dissertation. 152 p. University of Utrecht, Utrecht. Engel Z, V Treml, M K vrížek & V Jankovská (2005). Lateglacial/holocene sedimentary record from the Labe source area, the Krkonoše Mts. Acta Univer- sitatis Carolinae, Geographica 39: 1, 73–88. Firbas F (1952). Spät- und nacheiszeitliche Waldge- schichte Mitteleuropas nördlich der Alpen 2. 282 p. Verlag von Gustav Fischer, Jena. Friedmann A (2000). Die Spät- und Postglaziale Land- schafts- und Vegetationsgeschichte des südlichen Oberrheintieflands und Schwarzwalds. Freiburger Geographische Hefte 62. 222 p. Grimm EC (1992). TILIA 1.11 and TILIA GRAPH 1.17. 65 p. Illinois state museum, Research and Collec- tion center, Springfield. Haas JN, I Richoz, W Tinner & L Wick (1998). Syn- chronous holocene climatic oscillations recorded on the Swiss plateau and at timberline in the Alps. The Holocene 8: 3, 301–309. Hauepler H (1970). Vorschläge zur Abgrenzung der Höhenstufen der Vegetation im Rahmen der Mit- teleuropakartierung. Göttinger Floristische Rund- briefe 4, 54–62. Hieri O, AF Lotter, S Hausmann & F Kienast (2003). A chironomid-based Holocene summer air temper- ature reconstruction from the Swiss Alps. The Holocene 13: 4, 477–484. Hošek E (1972). Present evolution of alpine timber- line in the Jeseníky Mts. Ochrana pvrírody 27, 110–113. Huettemann H & S Bortenschlager (1987). Beitrage zur Vegetationsgeschichte Tirols VI: Riesengebir- ge, Hohe Tatra – Zillertal, Kühtai. Berichte des Na- turwissenschaftlich-Medizinischen Vereins in Inns- bruck 74, 81–112. Jankovská V (2001). Vegetation development in the western part of the Giant Mts. during the Holocene (Pan vcavské rašeliništ ve mire – palaeoecological re- search). Opera Corcontica 38, 11–19. Jankovská V (2004). Giant Mountains in Postglacial – vegetation and landscape. Opera Corcontica 41, 111–123. Jeník J (1998). Biodiversity of the Hercynian moun- tains of central Europe. Pireneos 151–152, 83–89. Jeník J & R Hampel (1991). Die waldfreien Kammla- gen des Altvatergebirges (Geschichte und Ökolo- gie). 128 p. MSSGV, Stuttgart. 118 FENNIA 184: 2 (2006)Václav Treml, Vlasta Jankovská and Libor Petr Jeník J & T Lokvenc (1962). Die alpine Waldgrenze im Krkonoše Gebirge. Rozpravy vCeskoslovemské Akademie vved 72: 1, 1–65. Körner Ch (1999). The alpine plantlife. 338 p. Spring- er, Berlin. Kotarba A & M Baumgart-Kotarba (1999). Problems of glaciation of the High Tatra Mountains – Josef Partsch synthesis in the light of current knowledge. Zeitschrift für Geomorphologie 113, 19–31. Kullman L (2007). Tree line population monitoring of Pinus sylvestris in the Swedish Scandes, 1973– 2005: implications for tree line theory and cli- mate change ecology. Journal of Ecology 95, 41–52. Lang G (1994). Quartäre Vegetationgeschichte Euro- pas. 371 p. Gustav Fischer Verlag, Jena. Lang G (2006). Late-glacial fluctuations of timberline in the Black Forest (SW Germany). Vegetation His- tory and Archeobotany 15, 373–375. Lemée G (1963). L’Évolution de la végetation et du climat des Hautes Vosges centrales depuis la der- niére glaciation. In Carbiener R. Le Hohneck, 185–193. Ass. philomatique d’Alsace et Lorraine, Strasbourg. Ložek V (2001). Malá Fatra Mts. and alpine timber- line oscillations. Ochrana pvrírody 56: 2, 35–40. Mavrák P & T Kuras (2006). Influence of alochtonous dwarf pine (Pinus mugo) to invertebrates commu- nities, case study epigeic arachnids in the Prad ved reserve (Hrubý Jeseník). In Proceedings of Zoolog- ické dny 2006 conference, 55–56. Czech Associa- tion of Zoologists, Brno. Mercier JL, J Kalvoda & D Bourles (1999). Utilisation du 10Be produit pour datter la derniére sequence glaciaire dans les monts du centre de l’Europe. Acta Universitatis Carolinae, Geographica 34: 2, 137–142. Migala K (2005). Climatic belts in the European mountains and the issue of global change. Geo- graphical Studies 78. Moore PD, JA Webb & ME Collingson (1991). Pollen analysis. 216 p. Blackwell, Oxford. Obidowicz A (1993). Fluctuation of the forest limit in the Tatra Mts. during the last 12,000 years. Doku- menty Geograficzne 4–5, 31–43. Paulsen J, UM Weber & Ch Körner (2000). Tree growth near treeline: abrupt or gradual reduction with al- titude? Arctic, Antarctic and Alpine Research 32: 1, 14–20. Rybnívcek K & E Rybní vcková (2004). Pollen analyses of sediments from the summits of the Prad ved range in the Hrubý Jeseník Mts (Eastern Sudetes). Preslia 76: 4, 331–348. Schloss S (1979). Pollenanalytische und stratigra- phische Untersuchungen im Sewensee. Disserta- tiones Botanicae 52. 138 p. Schwartz D, M Thinon, S Goepp, Ch Schmitt, J Casner, T Rosique, P Wuscher, A Alexandre, E Dambrine, Ch Martin & B Guillet (2005). Premières datations directes de défrichements protohistoriques sur les chaumes secondaires des Vosges (Rossberg, Haut- Rhin). Approche pédoanthracologique. Comptes Rendus Geoscience 337, 1250–1256. Sekyra J, M Kociánová, H Štursová, J Kalenská, I Dvo vrák & M Svoboda (2002). Recent cryogenic processes. In Soukupová L, M Kociánová, J Jeník & J Sekyra (eds). Arctic-alpine tundra in the Krkonoše, the Sudetes. Opera Corcontica 32, 31– 37. Slatyer RO & IR Noble (1992). Dynamics of montane treelines. In Hansen AJ & F di Castri (eds). Land- scape boundaries: consequences for biotic diver- sity and ecological flows. Ecological Studies 92, 327–345. Speranza A, J Van der Plicht & B Van Geel (2000). Improving the time control of the Subboreal/Sub- atlantic transition in a Czech peat sequence by 14C wiggle-matching. Quarternary Science Reviews 19, 1589–1604. Svobodová H (2004). Development of the vegetation on Úpské rašeliništve mire in the Holocene. Opera Corcontica 41: 1, 124–131. Tackenberg O, P Poschold & G Karste (1997). Verän- derungen der subalpinen Vegetation und Land- schaft des Brockens (Harz). Verhandlungen der Gesellschaft für Ökologie 27, 45–51. Tinner W, B Amman & P Germann (1996). Treeline fluctuations recorded for 12,500 years by soil pro- files, pollen, and plant macrofossils in the Central Swiss Alps. Arctic, Antarctic and Alpine Research 28: 2, 131–147. Tinner W & P Kaltenrieder (2005). Rapid responses of high-mountain vegetation to early Holocene envi- ronmental changes in the Swiss Alps. Journal of Ecology 93: 5, 936–947. Tinner W & JP Theurillat (2003). Uppermost limit, ex- tent, and fluctuations of the timberline and treeline ecotone in the Swiss central Alps during the past 11,500 years. Arctic, Antarctic and Alpine Re- search 35: 2, 158–169. Traczyk A & P Migovn (2003). Cold-climate landform patterns in the Sudetes – effects of lithology, relief and glacial history. Acta Univesitatis Carolinae, Geographica 25, 185–210. Treml V (2004). Recent tendencies of alpine timber- line shifts in the Krkonoše (Giant) Mts., High Su- detes. In Drbohlav D, J Kalvoda & V Voženílek (eds). Czech geography at the dawn of the mille- nium, 151–162. Nakladatelství Univerzity Palack- ého, Olomouc. Treml V & M Banaš (2000). Alpine timberline in the High Sudetes. Acta Universitatis Carolinae, Geo- graphica, 15: 2, 83–99. Treml V & M Kvrížek (2006). Effects of Dwarf pine (Pi- nus mugo) on patterned ground in the Czech part of the High Sudetes. Opera Corcontica 43, 45–56. Treml V, M Kvrížek, Z Engel & L Petr (2006). Patterned ground in the High Sudetes – morphology and the period of origin. Geomorfologický sborník 5, 68– 74. Wilmking M, GP Juday, VA Barber & HS Zald (2004). Recent climate warming forces contrasting growth FENNIA 184: 2 (2006) 119Holocene timberline fluctuations in the mid-mountains of … responses of white spruce at treeline in Alaska through temperature thresholds. Global Change Biology 10, 1274–1736. Wright HE, I Stefanova, JFN Van Leeuwen & B Am- mann (2003). The patterned fen of La Grand Tsa at 2330 m in the Swiss Alps. In Ravazzi C & W Tin- ner (eds). Excursion guide to XXVII moor-excursion of the institute of plant sciences, 46–50. Univer- sity of Bern, Bern.