afg?hi%gfahj%.kllmh?j?i>n%a%>j a j ?>j?. a f%a>>i>>lihj%ko...

137

Physical Geography, 2006, 27, 2, pp. 137–154.Copyright © 2006 by V. H. Winston & Son, Inc. All rights reserved.

ALPINE PLANT COMMUNITIES: A STATISTICAL ASSESSMENT OF THEIR RELATION TO MICROCLIMATOLOGICAL, PEDOLOGICAL,

GEOMORPHOLOGICAL, AND OTHER FACTORS

C. M. Vonlanthen, A. Bühler, and H. VeitInstitute of Geography

University of BernHallerstrasse 12

3012 BernSwitzerland

P. M. KammerLLB S1 Biology

Canton and University of BernGertrud-Wokerstrasse 5

3012 BernSwitzerland

W. EugsterInstitute of Plant SciencesETH Zürich, LFW C55.2

8092 ZürichSwitzerland

Abstract: Floristic composition and environmental factors vary widely among plant communities in the alpine belt. Thus far no study has attempted to measure all relevant site conditions in a larger number of alpine communities. Here we show (1) which environ-mental factors were highly correlated with the floristic composition of the 14 plant com-munities investigated in the Swiss Alps and (2) which plant communities have similar environmental affinities. In every plant community investigated, the main factors poten-tially having an impact on plant life were measured and the floristic composition was defined. We used nonmetric multidimensional scaling (NMDS) to determine linkage between plant communities and complex environmental gradients. The first axis of the NMDS corresponds to a climate gradient (temperature/wind speed), and the second axis corresponds to a soil gradient (soil suction/pH/Ca content). With the exception of the Nardus grassland and Carex curvula turf, plant communities belonging to the same phy-tosociological class are exposed to very similar combinations of environmental factors. Our study shows that the variation between phytosociological classes is much larger than within classes. Still, the variation of environmental factors within individual classes leads to a further differentiation of the floristic composition. Thus, our study reinforces the valid-ity of the phytosociological classification. [Key words: phytosociology, Alps, nonmetric multidimensional scaling NMDS, environmental factors, floristic composition.]

INTRODUCTION

Alpine vegetation near and above the treeline resembles a fine-scale mosaic com-posed of numerous, generally well delimited plant communities (e.g., Ellenberg,

138 VONLANTHEN ET AL.

1996). The phytosociological method of vegetation classification assumes that each community is characterized by a site-specific range of environmental factors leading to differences in species composition (e.g., Dierschke, 1994). Floristic compositions of alpine plant communities and their ecological conditions were described for the first time at the end of the 19th and beginning of the 20th century (e.g., Stebler and Schröter, 1892; Brockmann-Jerosch, 1907; Rübel, 1911–1912). Since then, many phytosociological studies have contributed to classifying alpine vegetation (e.g., Grabherr and Mucina, 1993), and the ecological conditions of alpine plant commu-nities have been described in numerous textbooks (e.g., Favarger and Robert, 1995; Ellenberg, 1996; Mertz, 2000).

Comparative studies that involve measured environmental factors, however, often include only a small number of environmental factors or communities. Previ-ous investigations include micro-climatic studies (e.g., Gigon, 1971; Cernusca, 1976, 1989; Mosimann, 1985; Cernusca and Seeber, 1989a, 1989b) and analyses of soil nutrients and water status (e.g., Rehder, 1970, 1976; Gigon, 1971; Rehder and Schäfer, 1978; Körner et al., 1980, 1989; Galland, 1982; Fries, 1985; Mosimann, 1985; Wieser et al., 1989) for some alpine plant communities. Other studies have investigated the influence of herbivory (e.g., Cernusca, 1989; Dullinger et al., 2003), growing-season length (e.g., Klug-Pümpel, 1982; Wallossek, 1990; Kudo, 1991), solifluction (e.g., Johnson and Billings, 1962), and avalanches (e.g., Erschbamer, 1989) on vegetation.

Although the phytosociological relationships of the dominant plant communities in the Alps are well known, the microclimatological, pedological, and geomorpho-logical factors governing these plant communities are not well documented. To the best of our knowledge, no study so far has attempted to measure all relevant site conditions in a larger number of communities. Thus, the aim of our field survey was to investigate the linkage between environmental conditions and floristic composi-tion of 14 widely distributed, representative, and well-documented plant commu-nities occurring in the alpine belt of the European Alps.

In this study, the main factors influencing plant life, as selected by Kammer and Möhl (2002), were measured or observed in each of the 14 plant communities. These factors are: air temperature, relative air humidity, wind speed, global radia-tion, UV-B radiation, growing-season length, soil moisture, main soil nutrients, pH, waterlogging, soil movement, denudation, avalanches, herbivory, wind damage, and freezing damage. The 14 plant communities investigated cover the main vege-tation types of the area above treeline in the European Alps, including extreme hab-itats such as wind-exposed communities, scree slopes, and snowbeds. Thus, our study is expected to be representative for most of the different habitats occurring in the European Alps, with the exception of rock-surface, bog, and fen communities.

In this study, we first explored how the floristic composition of the plant commu-nities is related to the different environmental gradients. Our aim was not to asses the variation of environmental factors within each community (this would have required a replicated experimental design) but rather to analyze the more general response curves along environmental gradients. Thus, we discuss the environmen-tal variables that we found to be highly correlated with the floristic composition. Second, we determined how the plant communities belonging to the same

ALPINE PLANT COMMUNITIES 139

phytosociological class can be characterized by their specific combinations of envi-ronmental factors.

RESEARCH AREA

The present study was restricted to plant communities that occur widely through-out the Alps and are clearly defined in phytosociological literature (characteristic species) and habitat characteristics (Landolt and Urbanska, 1989; Grabherr and Mucina, 1993; Ellenberg, 1996). Fourteen plant communities were chosen in the field based on the close correspondence of their floristic composition (characteris-tic species, dominant species) and site factors (Table 1) with those of plant associa-tions described in the existing phytosociological literature. Syntaxa names and classification follow Grabherr and Mucina (1993).

Field work was carried out in the western part of the Central Swiss Alps at Gemmi Pass and Grimsel Pass (Fig. 1). Grimsel Pass is situated in the crystalline Aare massif. Rankers and Podzols have developed on its siliceous bedrock made of granite and gneiss (Imhof, 1965–1978; Landolt and Urbanska, 1989). Long-term climate data are available from the meteorological station at Grimsel Hospiz (1980 m), roughly 1 km north of Grimsel Pass. The mean annual temperature at Grimsel Hospiz was 1.61ºC during the period 1964–2004 (this and all subsequent values are derived from the on-line database of MeteoSwiss, Swiss National Meteorologi-cal Service). Northwesterly winds dominate. The mean annual precipitation total is 2130 mm, of which 62% falls as snow. The snow cover is thicker than 10 cm during 226 days of the year, and only 124 days were snow free on average. Maximum win-ter snow cover ranged between 1.9 m (1964) and 6.9 m (1970).

Fig. 1. Location of study sites: Gemmi Pass (calcareous bedrock) and Grimsel Pass (siliceous bed-rock). K606-01 © 2004 swisstopo.


The substrata of the Gemmi area consist of Mesozoic sedimentary rocks gener-ally rich in carbonates. Rendzinas and alpine Brown Earths have developed on the calcareous bedrock of this area (Imhof, 1965–1978; Landolt and Urbanska, 1989). The climate conditions at Grimsel Hospiz are also representative for the Gemmi area (Imhof, 1965–1978).

Table 1. Characteristics of the Alpine Plant Communities Studieda

Abbr. Community Area Alt. Exp. Slope Cover Characteristic species

thl Thlaspi repensscree communityThlaspietum rotundifolii

Ge 2370 NW 27 32 Cerastium latifolium, Hutchinsia alpina, Linaria alpina, Moehringia ciliata, Thlaspi repens

leo Leontodon montanusscree communityLeontodontetum montani

Ge 2280 S 23 11 Galium megalospermum, Leontodon montanus, Ranunculus parnassifolius, Trisetum distichophyllum, Viola cenisia

polG Polytrichum sexangularesnowbed communityPolytrichetum sexangularis

Gr 2182 – – 92 Agrostis schraderana, Gnaphalium supinum, Polytrichum sexangulare, Salix herbacea, Soldanella pusilla

sal Salix herbaceasnowbed communitySalicetum herbaceae

Ge 2460 – – 85 Alchemilla pentaphyllea, Cerastium cerastoides, Gnaphalium supinum, Salix herbacea, Taraxacum alpinum

salG Salix herbaceasnowbed communitySalicetum herbaceae

Gr 2180 – – 75 Arenaria biflora, Carex foetida, Gnaphalium supinum, Salix herbacea, Soldanella pusilla

ely Elyna myosuroidesturfElynetum myosuroides

Ge 2275 W 0–5 87 Agrostis alpina, Arenaria ciliata, Elyna myosuroides, Oxytropis campestris, Potentilla crantzii

fir Carex firmaturfCaricetum firmae

Ge 2255 NW 24 90 Androsace chamaejasme, Carex firma, Chamorchis alpina, Dryas octopetala, Helianthemum alpestre

ses Sesleria caeruleagrasslandSeslerio-Caricetumsempervirentis

Ge 2335 SE 27 64 Aster alpinus, Carduus defloratus, Carex sempervirens, Gentiana clusii, Sesleria caerulea

fer Carex ferrugineagrasslandCaricetum ferrugineae

Ge 1895 NW 27 100 Anemone narcissiflora, Astragalus frigidus, Carex ferruginea, Festuca pulchella, Pulsatilla alpina

fes Festuca violaceagrasslandFestuco-Trifolietum thalii

Ge 2130 – – 98 Alchemilla vulgaris, Crepis aurea, Festuca violacea, Poa alpina, Trifolium thalii

nar Nardus strictagrasslandSieversio-Nardetum strictae

Ge 2125 SE 22 97 Arnica montana, Campanula barbata, Gentiana acaulis, Geum montanum, Nardus stricta

curG Carex curvalaturfCaricetum curvulae

Gr 2240 S 0–22 82 Avenula versicolor, Carex curvula, Luzula lutea, Phyteuma hemisphaericum, Veronica bellidioides

loiG Loiseleuria procumbensheathLoiseleurio-Cetrarietum

Gr 2215 NE 17 80 Empetrum hermaphroditum, Hieracium alpinum, Loiseleuria procumbens, Vaccinium gaultherioides, Vaccinium vitis-idaea

cic Cicerbita alpinatall herb communityCicerbitetum alpinae

Ge 1765 SE 15 100 Achillea macrophylla, Adenostyles alliariae, Peucedanum ostruthium, Saxifraga rotundifolia, Viola biflora

aAbbr. = abbreviations of the community names used in the figures (syntaxon names according to Grabherr and Mucina, 1993); Area = study areas (Ge = Gemmi and Gr = Grimsel); Alt. = altitude in m a.s.l.; Exp. = exposure; Slope = slope angle in degrees; Cover = total vegetation cover in percent; Characteristic species = selection of species occurring in the communities under study.


METHODS

Floristic Composition

In each plant community, the presence of vascular plant species was recorded for 10 randomly placed plots (1 m2) at the end of July 2002. The plots were spaced widely enough to minimize spatial autocorrelation within the visually homogenous surface of the community. The vegetation frequency data obtained in this manner were used for subsequent analyses.

Climatic Factors

Air temperature, relative air humidity, global radiation, and wind speed were measured in each plant community during the growing seasons of 2002 and 2003. Each climate station consisted of an anemometer (type f.555.1.18, Schiltknecht, Gossau, Switzerland, installed 100 cm above ground), an air-temperature and rela-tive-air-humidity sensor (TRH 100, Pace Scientific Inc., Mooresville, N.C., USA, installed 15 cm above ground), a data logger (XR 440, Pace Scientific Inc., Mooresville, N.C., USA), and an amorphous solar cell (constructed at the Institute of Geography of the University of Bern, Switzerland, installed 40 cm above ground). Solar cells for measuring global radiation were produced and calibrated by Fischer (2002) using a pyranometer under natural irradiation conditions. The sen-sors were attached to a 1.3 m aluminum mast, with the part of the mast above the solar cells painted black to avoid reflection. Solar cells were mounted in a level position on the south side of the mast. Daily mean UV-B radiation was calculated from the measured daily mean global radiation data at the Gemmi and Grimsel Pass via an empirical model (Vonlanthen et al., 2004). Measurements of relative air humidity were converted to vapor pressure deficit (VPD; Warnecke, 1997).

Time and data were recorded in CEST (Central European Summer Time = UTC [Universal Time Coordinated] + 2 hours) in 10-minute intervals. These values then served as a basis on which to calculate the daily means. Missing daily mean values were interpolated using a linear regression. Microclimatic stations were intercali-brated in April 2002 (i.e., prior to the first field season), April 2003, and October 2003 (i.e., following the second field season). Days with a minimum temperature below 0ºC (measured 15 cm above ground) during the growing season were defined as days with frost. Analyses are based on daily means of the period during which all climate stations were installed (July 12, 2002–September 19, 2002, and July 13, 2003–September 6, 2003).

Length of Growing-Season

Growing-season length was obtained for each community by determining the dates of snowmelt in spring and formation of a lasting snow cover in autumn, in 2002 and 2003. Plant communities were considered snow free as soon as snow had disappeared from 80% of the defined surface. Analyses are based on two-year mean values.


Pedological Factors

According to Körner (1994), periodic water shortage in the top 5 cm of soil may induce water stress in plant communities above the treeline. During periods of low precipitation, soil suction was therefore measured, using a Quick-Draw Tensiometer (Soil Moisture Equipment Corp., Model 2900F1, Santa Barbara, CA, USA) at a depth of 4 cm. Soil suction was measured once in July and once in August, both in 2002 and in 2003 (altogether four measurement campaigns). Each time, three randomly placed measurements were made within one day in every plant community. The Tensiometer was calibrated prior to each measurement. Analyses are based on mean values.

At the end of July 2003, five soil samples were randomly collected from the upper 5 cm of mineral horizons in each plant community. The samples were pooled to one composite sample per community. All soil samples for Ntot, NH4, and NO3concentration measurements were placed in an icebox (–20º C) for up to three hours after sampling. Concentrations of soil nutrients, including Ntot, NH4, NO3, P, Mg, Ca, and K, as well as pH, were analyzed by Schweizer Samen AG (Thun, Switzerland) using standard methods described in FAL (1996). Concentrations of water-soluble K, Mg, and Ca were determined by atomic-adsorption spectroscopy, Ntot by Kjeldahl distillation, and water-soluble NH4, NO3, and P by photometry. The content of Norg was calculated as Ntot – NH4 – NO3. Soil pH was measured by shaking 20 g of sieved soil with de-ionized water and measuring pH with a glass electrode in the solution extracted after 18 hours.

Geomorphological Factors

Frost heave was measured with an apparatus developed by the authors (Vonlanthen et al., 2004). In August 2002, in each plant community, a 1.4 m long steel rod was inserted 1 m into the ground. Two pipe clips were fixed onto the pro-truding part (40 cm) of the steel rod. The pipe clips held a transparent tube (28 cm) in a vertical position. A grey plastic tube (22 cm) passed vertically through the trans-parent tube, remaining free to move up and down with the heaving soil. A 3 cm long piece of foam rubber was placed inside the transparent tube on top of the grey tube. The foam rubber was able to move upwards but not downwards with the grey tube. Changes of position of the foam rubber piece were measured in August 2003. To protect the apparatus against freezing, it was covered by a 41 cm high plastic tube. This apparatus measures only the maximum extent of frost upheaval.

To measure lateral soil movement stones were labeled and aluminium stripes buried vertically to a depth of 40 cm in July 2002 in every plant community. The positions of the stones and the aluminium bands were measured in August 2002 and August 2003 using a Real-Time GPS (Leica GPS500, Heerbrugg, Switzerland, absolute precision 1.5 cm). Based on these data, changes in the position of the alu-minium bands and the labelled stones were calculated. In summer 2003, the verti-cally buried aluminium bands were dug out to detect maximum soil movement at a depth of 0 to 40 cm.


The occurrence of avalanches in each community was estimated based on per-sonal observations as well as on interviews with 10 persons familiar with the region. Avalanche frequency was quantified using a four-step scale from 0 (no ava-lanches) to 3 (regular occurrence of avalanches). Denudation of soil by water, snow, or wind was visually estimated and quantified using a five-step scale from 0 (more accumulation than denudation) to 4 (extremely high level of denudation).

Additional Factors

The proportion of waterlogged surface in each community was estimated (in %) during or after periods of heavy rainfall. In three permanently marked quadrats (0.25 m2) per community, wind damage (proportion of above-ground biomass damaged by wind, estimated to the nearest 5%), and herbivory (proportion of above-ground biomass removed by animals, estimated to the nearest 5%) were determined every two weeks during the growing season.

Data Analysis

The relationship between vegetation frequency data and environmental data was explored through ordination analysis, using detrended correspondence analysis (DCA) and nonmetric multidimensional scaling (NMDS; Kent and Coker, 1992; Ter Braak, 1995; Legendre and Legendre, 1998). We used both methods because there is currently no consensus on the most appropriate indirect ordination method (Kent and Coker, 1992; Ter Braak, 1995). DCA assumes a unimodal species response curve, whereas NMDS makes no such assumption, deriving configuration scores only from the rank order of the dissimilarities between samples or species (Faith et al., 1987). Where both methods produced similar results, we felt more confident that the patterns represented an inherent structure in the data. Because of the good agreement between DCA and NMDS, we only report the results obtained from the NMDS, which is the more appropriate choice for our dataset.

The analysis with NDMS was carried out with PC ORD (version 4.10 for Windows, McCune and Mefford, 1999). To reduce the noise of rare species, species occurring in fewer than three plots were deleted prior to analysis. The appropriate dimensionality for the ordination was determined by first running NMDS in autopi-lot mode with 50 runs for each of six dimensionalities; the lowest dimensionality was chosen that (1) showed substantial stress reduction compared to the next low-est dimensionality and (2) captured new information in all dimensions. In this dataset, ordinations were best described by two axes (instability < 0.00005). To ensure that the ordination avoids a local stress minimum, the analysis was then run 50 times using random starting configurations. The NDMS was run with 50 runs of the real data along with 100 runs with randomized data for a Monte Carlo test of significance.

Environmental variables were transformed to improve normality. The log (x) transformation was chosen for soil suction, Mg, K, P, herbivory, and frost upheaval; global radiation was square-root transformed.


RESULTS

Relationships between Floristic Composition and Environmental Factors (NMDS)

For examining patterns of community composition with respect to sites and envi-ronmental characters, a two-dimensional ordination obtained from NMDS was chosen (Fig. 2). A third dimension did not provide substantial stress reduction and did not add further information on environmental correlations. The ordination yielded a final stress value of 11.49 (p < .01), and the two axes accounted for 66% of the original variation in the dataset.

The arrows within the ordination diagram (Fig. 2) display the principal direction of variation and strength of correlation for major environmental variables. Only met-ric environmental variables with r2 > .35, which influence every plant community investigated, are shown. Even though global radiation and Mg content also had r2 > .35, they are not shown in the ordination diagram because these correlations are mainly driven by one site (Cicerbita tall-herb community). Axis 1 of the NMDS ordi-nation is interpreted as a climate gradient (temperature/wind speed): the temperature increases and wind speed decreases along axis 1. Axis 2 is interpreted as a soil gra-dient (Ca/pH/soil suction): Ca, pH, and soil suction values increase along axis 2.

Fig. 2. NMDS: nonmetric multidimensional scaling of vegetation frequency data. Arrows indicate the direction and strength of correlation of variables with r2 > .35. The sample sites are coded according to the phytosociological classes: = Mulgedio-Aconitetea; = Salicetea herbaceae; = Loiseleurio-Vaccinietea; = Thlaspietea rotundifolii; = Seslerietea albicantis; = Carici rupestris-Kobresietea bellardii; = Caricetea curvulae. For abbreviations of the community names, see Table 1.


The plant communities of the Thlaspietea rotundifolii class (Thlaspi scree com-munity and Leontodon scree community) are clustered in the upper left portion of the ordination space (Fig. 2), which is correlated with high pH values, high soil suc-tion values, and high Ca contents. The Carex ferruginea grassland, the Carex firmaturf, the Festuca grassland, and the Sesleria grassland, which belong to the class Seslerietea albicantis, have the same loadings on the soil-gradient axis but spread widely along the wind-speed and temperature gradient, with wind speed decreas-ing and temperature increasing from the Carex firma turf to the Carex ferrugineagrassland. The Cicerbita tall-herb community (class Mulgedio-Aconitetea) is located on the right side of the diagram, corresponding to high temperature values and low mean wind speed. The Polytrichum snowbed and the two Salix snowbeds (class Salicetea herbaceae), the Loiseleuria heath (class Loiseleurio-Vaccinietea), and the Carex curvula turf (class Caricetea curvulae) occupy the lower left corner of the diagram, which corresponds to low pH values, low soil suction values, low Ca contents, and high mean wind speed. With the exception of the Salix snowbed at Gemmi Pass (situated on siliceous limestone; Döbeli, 1997), these plant communi-ties occur on siliceous bedrock. Although the Nardus grassland belongs to the same class (Caricetea curvulae) as the Carex curvula turf, this plant community is situated in the middle of the ordination diagram. Finally, the Elyna turf (class Carici rupestris-Kobresietea bellardii) is located in the upper left corner of the diagram, correspond-ing to high wind speed, high Ca content, and high pH values.

Variation of Environmental Factors between and within Phytosociological Classes

Our study shows (Table 2; Fig. 3) that with the exception of the Nardus grassland and Carex curvula turf, plant communities belonging to the same phytosociological class are exposed to very similar environmental factor combinations when compared with communities belonging to other phytosociological classes. For example, all investigated snowbeds (class Salicetea herbaceae) were characterized by rare occur-rence of avalanches (Fig. 3U), low pH (Fig. 3I), a low Ca content (Fig. 3L), short veg-etation period (Fig. 3G), waterlogging (Fig. 3Q), and low soil suction values (Fig. 3H).

Moreover, this study reveals that scree slopes (class Thlaspietea rotundifolii) with their considerable soil movement and snowbeds (class Salicetea herbaceae) with their short vegetation period are extreme habitats for plants. Moreover, most of the grasslands investigated at Gemmi Pass (Carex ferruginea grassland, Carex firma turf, Festuca grassland, and Sesleria grassland—all from class Seslerietea albicantis) appear in the middle of the ordination (Fig. 2), thus indicating average conditions (Fig. 3) for the communities that we investigated. Furthermore, having favorable cli-mate conditions and a good nutrient supply (Table 2; Fig. 3) the Cicerbita tall-herb community (class Mulgedio-Aconitetea) represents lush and ideal conditions for plant life near the treeline. Finally, even though the Elyna turf and Loiseleuria heath do not belong to the same phytosociological class, they occupy almost the same ecological niche, where plants have to deal with high wind speeds (Fig. 3E), a long snow-free period (Fig. 3G), and considerable frost heave (Fig. 3R). However, Loiseleuria heath is located on siliceous bedrock and Elyna turf on calcareous bed-rock; hence, they differed completely in terms of soil nutrient contents (Figs. 3I–3P).


Fig. 3. Mean values of all environmental factors and their standard error (only for n > 1) as measured in the 14 alpine plant communities listed in Table 1. (A) Global radiation (mean values from 2002 and 2003), (B) UV-B radiation (mean values from 2002 and 2003), (C) temperature (mean values from 2002 and 2003), (D) vapor pressure deficit (mean values from 2002 and 2003), (E) wind speed (mean values from 2002 and 2003), (F) days with frost (mean values from 2002 and 2003), (G) vegetation period (mean values from 2002 and 2003), (H) soil suction (mean values from 2002 and 2003), (I) pH, (J) P, (K) Mg, (L) Ca, (M) Ntot, (N) NH4, (O) NO3, (P) K, (Q) waterlogging, (R) frost heave (in the Sesleria grassland the apparatus was destroyed in winter 2002/2003), (S) soil movement recorded with the aid of alumin-ium strips, (T) soil movement recorded with the aid of marked stones, (U) avalanches, (V) denudation, (W) wind damage (mean values from 2002 and 2003), (X) herbivory (mean values from 2002 and 2003).


Even though plant communities belonging to the same phytosociological class are exposed to very similar environmental factor combinations, there is still consid-erable variation of the environmental factors remaining within phytosociological classes (e.g., lower wind speeds were measured in the Polytrichum snowbed than in the two Salix snowbeds).

DISCUSSION

Relationships between Floristic Composition and Environmental Factors (NMDS)

The measured metric environmental factors that were highly correlated with the floristic composition were soil suction, pH, Ca, temperature, and wind speed (Fig.

Fig. 3. For caption see facing page.


2). In the following, these factors will be discussed with respect to their linkage with plant community composition.

Temperature. Low temperatures affect biochemical reactions such as these involved in photosynthesis, thereby limiting plant productivity (Larcher, 1995; Brunold et al., 1996). Many experimental studies using open-top chambers to raise temperatures have shown that reproductive and growth processes in arctic and alpine species are limited by temperature (e.g., Henry and Molau, 1997). Apart from having direct effects on plant growth, higher temperatures most likely also lead to increased mineralization rates in soils (Haynes, 1986). Furthermore, temperature co-determines the length of the growing season, which is one of the frequently men-tioned factors controlling the distribution of alpine plant species (e.g., Komárková, 1993; Stanton et al., 1994).

Wind speed. High wind speeds (e.g., measured in the Loiseleuria heath) can increase the abrasive mechanical effects of wind and lead to increased transpiration rates (e.g., Van Gardingen et al., 1991). Indirect effects of wind on vegetation may be even more important: (1) wind is the main factor in the horizontal distribution of snow (Franz, 1979); (2) it transports nutrient-rich soil particles (Franz, 1979); and (3) it accentuates the harsh microclimatic conditions of wind-exposed habitats (Körner, 1999).

PH. It is unlikely that pH per se affects plants, but there is evidence that mobili-zation and availability of nutrients increase with increasing pH (Sjörs, 1952;

Table 2. Characteristics of Plant Communities Belonging tothe Same Phytosociological Class

Phytosociological classesMain environmental similarities within

phytosociological class

Main environmental dissimilarities within

phytosociological class

Scree slopes(class Thlaspietea rotundifolii)

Highest pH, highest soil suction values, regular occurrence of avalanches, highest denudation level, highest lateral soil movement

Climatic conditions, nutrient supply

Base-rich grasslands(class Seslerietea albicantis)

Moderate ecological conditions, an exception is the high herbivory level


Base-poor grasslands(class Caricetea curvulae)

Moderate herbivory level, low denudation level, low Ca content, moderate soil suction values


Snowbeds(class Salicetea herbaceae)

Shortest vegetation period, low pH, low soil suction values, low Ca content, rare occurrence of avalanches, low denudation level, waterlogging


Tall-herb community(class Mulgedio-Aconitetea)

Highest mean temperature, lowest mean global radiation, lowest mean wind speed, high contents of Mg and Ca

Loiseleuria heath (class Loiseleurio-Vaccinietea) and Elyna turf (class Carici rupestris-Kobresietea bellardii)

High wind speed means, longest vegetation period, considerable frost upheaval

Nutrient supply


Larcher, 1995). On the other hand, a high pH like that measured in the scree slopes lessens the availability of both micronutrient elements and the macro-element P for plants (Scheffer and Schachtschabel, 1998). Moreover, pH has an impact on the cat-ion exchange capacity and is related to the base saturation. Recent investigations in the arctic tundra treat pH as a variable of interest in ordinations, often finding that it statistically explains a significant amount of variation among communities and species assemblages (van Raamsdonk, 1988; Timoney et al., 1993; Walker et al., 1994; Heikkinen, 1996; Heikkinen and Neuvonen, 1997).

Ca. The Ca content of soils mainly depends on soil pH and the geological sub-strate. Thus, it is no surprise that Ca content is also highly correlated with floristic composition. Many studies have shown that alpine plant distribution is strongly influenced by the abundance of calcium (e.g., Gigon, 1971; Körner, 1999; Michalet et al., 2002). This is corroborated in this study, in which plant communities on sili-ceous bedrock (Loiseleuria heath, Carex curvula turf, Polytrichum snowbed, and Salix snowbed on Grimsel Pass) and communities on calcareous bedrock form two well separated groups in the ordination (Fig. 2). Many species are found mainly on Ca-rich substrates (calcicole species) and others mainly on Ca-poor substrates (calcifuge species).

Soil suction. On the scree slopes, where the driest soils occurred, water seeps away quickly but may be retained in deeper layers. The lack of water at the surface may be fatal during the juvenile stages of plants. Moreover, desiccation of the top soil layer can block mineral cycling (Körner, 1995). By contrast, in the snowbeds, with the wettest soils, waterlogging leads to anaerobic conditions during rainfall that can influence the floristic composition.

In summary, we found that temperature, pH, soil suction, wind speed, and Ca are likely to influence a community’s floristic composition directly or via their effects on other environmental variables, which, in turn, affect plant species occur-rences directly. Temperature, soil suction, and pH (as an indicator of soil nutrient availability) are three of the indicator values identified by Ellenberg (e.g., Ellenberg et al., 1992). This study corroborates the importance of these factors in determining the ecological environment of plants. Other factors such as waterlogging, lateral soil movement, wind damage, and herbivory did not seem to play a primary role in controlling the floristic composition in most of the communities we studied.

Variation of Environmental Factors between and within Phytosociological Classes

Although, the Nardus grassland and Carex curvula turf plant communities belong to the same phytosociological class (Caricetea curvulae), these two commu-nities differ considerably in site conditions and floristic composition (only 11% of all species are common to both plant communities). This difference is related to geologic substrate: our Carex curvula turf was located on siliceous bedrock, whereas the Nardus grassland was situated on calcareous bedrock. Therefore, the Nardus grassland harbors an unusually high proportion of calcicole species and shows affinities with the class Seslerietea albicantis. This might be a local deviation from what appears to be the general rule that plant communities that belong to the same phytosociological class are closely correlated with a specific combination of


environmental factors (Table 2; Fig. 3). Variation of environmental conditions within the classes are reflected in the floristic composition (e.g., the Thlaspi scree community and Leontodon scree community have only 13% of common species).

CONCLUSIONS

We found two environmental gradients to be most relevant to the linkage between the floristic composition of alpine plant communities and environmental conditions in the Swiss Alps: (1) the temperature/wind-speed gradient; and (2) the soil-suction/pH/Ca content gradient. The inverse relationship between wind speed and temperature along the first axis suggests that global change scenarios that only provide projections of temperature could also be used for alpine plant communities where wind is the dominant factor.

Our study reveals that with the exception of the Nardus grassland and Carex curvula turf, plant communities belonging to the same phytosociological class are exposed to very similar environmental factor combinations. Moreover, our study shows that the variation between phytosociological classes is much larger than within classes. Still, the variation of environmental factors within individual classes also leads to a further differentiation of the floristic composition. Overall, our study confirms the phytosociological classification, which assumes that differences in site conditions should translate into differences in plant communities in a predictable way.

Although this study generally supports existing phytosociological systems, it is highly recommended for future studies that existing databases should be expanded to include replicates of plant communities at different locations. This would allow a more detailed assessment of the extent to which each plant species responds to environmental factors independently and as a member of a community—that is, through interactions such as competition and facilitation and as affected by the existing species pool.

The plant communities investigated are all located within the central Swiss Alps and do not necessarily represent compositional patterns across the entire European Alps. It would be desirable to conduct similar studies in other alpine areas to improve the geographic scope of our scientific understanding of alpine plant com-munities.

Acknowledgment: This study was supported by the Stiftung zur Förderung der wissenschaftlichen Forschung der Universität Bern, the Hochschulstiftung der Burgergemeinde, the Centre of Research and Developement of LLB Bern, and SEVA-Lotteriefonds (Kanton Bern). We are very grateful to all persons without whose contributions this study would not have been possible, in particular T. Reist (Bern), C. Schöb (Bern), and L. Vonlanthen (Freiburg).

REFERENCES

Brockmann-Jerosch, H. (1907) Die Flora des Puschlav (Bezirk Bernina, Kanton Graubünden) und ihre Pflanzengesellschaften. Leipzig, Germany: Engelmann.

Brunold, C., Rüegsegger, A., and Brändle, R. (1996) Stress bei Pflanzen: Ökologie, Physiologie, Biochemie, Molekularbiologie. Bern, Switzerland: Haupt.


Cernusca, A. (1976) Bestandesstruktur, Bioklima und Energiehaushalt von alpinen Zwergstrauchbeständen. Oecologia Plantarum, Vol. 11, No. 1, 71–102.

Cernusca, A. (1989) Ecosystem research in the Austrian central Alps (Hohe Tauern). In A. Cernusca, ed., Struktur und Funktion von Graslandökosystemen im Nationalpark Hohe Tauern. Innsbruck, Austria: Wagner, Veröffentlichungen des österreichischen MAB-Programmes 13, 587–625.

Cernusca, A. and Seeber, M. C. (1989a) Mesoklimatische Hinweise und Beschreibungvon Witterungsablauf und Phänologie während der Ökosystemstudie “Höhentransekt” in den Hohen Tauern. In A. Cernusca, ed., Struktur und Funktionvon Graslandökosystemen im Nationalpark Hohe Tauern. Innsbruck, Austria: Wagner, Veröffentlichungen des österreichischen MAB-Programmes 13, 311–330.

Cernusca, A. and Seeber, M. C. (1989b) Phytomasse, Bestandesstruktur und Mikroklima von Grasland-Ökosystemen zwischen 1612 und 2300 m in den Alpen. In A. Cernusca, ed., Struktur und Funktion von Graslandökosystemen im Nationalpark Hohe Tauern. Innsbruck, Austria: Wagner, Veröffentlichungen des österreichischen MAB-Programmes 13, 419–461.

Dierschke, H. (1994) Pflanzensoziologie. Stuttgart, Germany: Ulmer Verlag.Döbeli, C. (1997) Das hochalpine Geoökosystem der Gemmi (Walliser Alpen).

Unpublished dissertation thesis, Department of Geography, University of Basel, Basel, Switzerland.

Dullinger, S., Dirnböck, T., Greimler, J., and Grabherr, G. (2003) A resampling approach for evaluating effects of pasture abandonment on subalpine plant spe-cies diversity. Journal of Vegetation Science, Vol. 14, No. 2, 243 –252.

Ellenberg, H. (1996) Vegetation Mitteleuropas mit den Alpen (5th edition). Stuttgart, Germany: Ulmer.

Ellenberg, H., Weber, H. E., Düll, R., Wirth, V., Werner, W., and Paulissen, D. (1992) Zeigerwerte von Pflanzen in Mitteleuropa. Scripta Geobotanica, Vol. 18, 1–258.

Erschbamer, B. (1989) Vegetation on avalanche paths in the Alps. Vegetatio, Vol. 80, 139–146.

Faith, D., Minchin, P., and Belbin, L. (1987) Compositional dissimilarity as a robust measure of ecological distance. Vegetatio, Vol. 69, 57–68.

FAL (1996) Schweizerische Referenzmethoden der Eidg. Landwirtschaftlichen Forschungsanstalten. Volume 1. Zürich, Switzerland: Reckenholz.

Favarger, C. and Robert, P.-A. (1995) Flore et végétation des Alpes—étage alpin (3rd edition). Lausanne, Switzerland: Delachaux et Niestlé S.A.

Fischer, E. (2002) Globalstrahlungsmessung mit Solarzellen. Unpublished seminar work, Department of Geography, University of Bern, Bern, Switzerland. (Copies can be obtained from the corresponding author.)

Franz, H. (1979) Oekologie der Hochgebirge. Stuttgart, Germany: Ulmer.Fries, M. (1985) Bodenkundliche Studien unter einem Caricetum firmae auf dem

Munt la Schera im Schweizerischen Nationalpark. In Physische Geographie 18. Zürich, Switzerland: Geographical Institute University of Zürich, Switzerland.

Galland, P. (1982) Etude de la Végétation des Pelouses Alpines au Parc National Suisse. Unpublished Ph.D. thesis, Departement of Biology, University of Neuchâtel, Neuchâtel, Switzerland.

http://www.ingentaconnect.com/content/external-references?article=0042-3106()69L.57%5Baid=7000816%5D


Gigon, A. (1971) Vergleich alpiner Rasen auf Silikat- und Karbonatboden. In Veröffentlichungen Geobotanisches Institut ETH, Stiftung Rübel Zürich 48. Zürich, Switzerland: Stiftung Rübel.

Grabherr, G. and Mucina, L. (1993) Die Pflanzengesellschaften Österreichs. Teil 2. Natürliche waldfreie Vegetation. Jena, Germany: Fischer.

Haynes, R. J. (1986) The decomposition process: Mineralization, immobilization, humus formation, and degradation. In R. J. Haynes, ed., Mineral nitrogen in the plant-soil system. Orlando, FL: Academic Press, 52–126.

Heikkinen, R. K. (1996) Predicting patterns of vascular plant species richness with composite variables: A meso-scale study in Finnish Lapland. Vegetatio, Vol. 126, 151–165.

Heikkinen, R. K. and Neuvonen, S. (1997) Species richness of vascular plants in the subarctic landscape of northern Finland: Modelling relationships to the environ-ment. Biodiversity and Conservation, Vol. 6, 1181–1201.

Henry, G. H. R. and Molau, U. (1997) Tundra plants and climate change: The Inter-national Tundra Experiment (ITEX). Global Change Biology, Vol. 3, Suppl. 1, 1–9.

Imhof, E., ed. (1965–1978) Atlas der Schweiz. Bern, Switzerland: Bundesamt für Landestopographie.

Johnson, P. L. and Billings, W. D. (1962) The alpine vegetation of the Beartooth Plateau in relation to cryopedogenic processes and patterns. Ecological Mono-graphs, Vol. 32, No. 3, 105–135.

Kammer, P. M. and Möhl, A. (2002) Factors controlling species richness in alpine plant communities. An assessment of the importance of stress and disturbance. Arctic, Antarctic, and Alpine research, Vol. 34, No. 4, 398–407.

Kent, M. and Coker, P. (1992) Vegetation Description and Analysis—A Practical Approach. London, UK: Belhaven Press.

Klug-Pümpel, B. (1982) Effects of microrelief on species distribution and phytomass variations in a Caricetum curvulae stand. Vegetatio, Vol. 48, 249–254.

Komárková, V. (1993) Vegetation type hierarchies and landform disturbance in arc-tic Alaska and alpine Colorado with emphasis on snowpatches. Vegetatio, Vol. 106, 155–181.

Körner, C. (1994) Impact of atmospheric changes on high mountain vegetation. In M. Beniston, ed., Mountain Environments in Changing Climates. London, UK, and New York, NY: Routledge, 155–166.

Körner, C. (1995) The impact of atmospheric changes on alpine vegetation: The eco-physiological perspective. In A. Guisan, J. I. Holgen, R. Spichiger, and L. Tessiers, eds., Potential Impacts of Climate Change in the Alps and Fennoscandian Moun-tains. Genève, Switzerland: Edition des Conservatoires des Jardins botaniques, 113–120.

Körner, C. (1999) Alpine Plant Life. Berlin, Germany: Springer.Körner, C., Wieser, G., and Cernusca, A. (1989) Der Wasserhaushalt waldfreier

Gebiete in den Oesterreichischen Alpen zwischen 600 und 2600 m Höhe. In A. Cernusca, ed., Struktur und Funktion von Graslandökosystemen im Nationalpark Hohe Tauern. Innsbruck, Austria: Wagner, Veröffentlichungen des österreichischenMAB-Programmes 13, 119–153.






Körner, C., Wieser, G., and Guggenberger, H. (1980) Der Wasserhaushalt eines alpinen Rasens in den Zentralalpen. In H. Franz, ed., Untersuchungen an Alpinen Böden in den Hohen Tauern 1974–1978: Stoffdynamik und Wasser-haushalt. Innsbruck, Austria: Wagner, Veröffentlichungen des österreichischen MAB-Programmes 3, 243–264.

Kudo, G. (1991) Effects of snow-free period on the phenology of alpine plants inhabiting snow patches. Arctic and Alpine Research, Vol. 23, No. 4, 436–443.

Landolt, E. and Urbanska, K. M. (1989) Our Alpine Flora. Chur, Switzerland: Swiss Alpine Club.

Larcher, W. (1995) Physiological Plant Ecology (3rd edition). Berlin, Germany: Springer.

Legendre, P. and Legendre, L. (1998) Numerical Ecology. Amsterdam, The Nether-lands: Elsevier.

McCune, B. and Mefford, M. J. (1999) Multivariate Analysis of Ecological Data, Ver-sion 4.10.—MjM Software Design. Gleneden Beach, OR.

Mertz, P. (2000) Pflanzengesellschaften Mitteleuropas und der Alpen. Landsberg, Germany: Ecomed Verlagsgesellschaft AG and Co.

Michalet, R., Gandoy, C., Joud, D., Pagès, J.-P., and Choler, P. (2002) Plant commu-nity composition and biomass on calcareous and siliceous substrates in the northern French Alps: Comparative effects of soil chemistry and water status. Arctic, Antarctic, and Alpine Research, Vol. 34, No. 1, 102–113.

Mosimann, T. (1985) Untersuchungen zur Funktion subarktischer und alpiner Geoökosysteme Finnmark (Norwegen) und Schweizer Alpen. In Physiogeo-graphica 7. Geographical Institute, University of Basel, Switzerland.

Rehder, H. (1970) Zur Ökologie insbesondere Stickstoffversorgung subalpiner und alpiner Pflanzengesellschaften im Naturschutzgebiet Schachen (Wettersteinge-birge). In Dissertationes Botanicae 6. Stuttgart, Germany: Cramer.

Rehder, H. (1976) Nutrient turnover studies in alpine ecosystems. Oecologia, Vol. 22, 411–423.

Rehder, H. and Schäfer, A. (1978) Nutrient turnover studies in alpine ecosystems. Oecologia, Vol. 34, 309–327.

Rübel, E. (1911–1912) Pflanzengeographische Monographie des Bernina-Gebietes. Botanische Jahrbücher für Systematik, Pflanzengeschichte und Pflanzengeographie, Vol. 47, 1–616.

Scheffer, F. and Schachtschabel, P. (1998) Lehrbuch der Bodenkunde (14th edition). Stuttgart, Germany: F. Enke.

Sjörs, H. (1952) On the relation between vegetation and electrolytes in Swedish mire waters. Oikos, Vol. 2, 241–258.

Stanton, M. L., Rejmánek, M., and Galen, C. (1994) Changes in vegetation and soil fertility along predictable snowmelt gradient in the Mosquito Range, Colorado, U.S.A. Arctic and Alpine Research, Vol. 26, No. 4, 363–374.

Stebler, F. G. and Schröter, C. (1892) Beiträge zur Kenntnis der Matten und Weiden der Schweiz X: Versuch einer Übersicht über die Wiesentypen der Schweiz. Landwirtschaftliches Jahrbuch der Schweiz, Vol. 6, 95–212.





Ter Braak, C. J. F. (1995) Ordination. In R. H. G. Jongman, C. J. F. ter Braak, and O. F. R. van Tongeren, eds., Data Analysis in Community and Landscape Ecology. Cambridge, UK: Cambridge University Press, 119–153.

Timoney, K. P., La Roi, G. H., Zoltai, S. C., and Robinson, A. L. (1993) Vegetation communities and plant distributions and their relationships with parent materials in the forest-tundra of northwestern Canada. Ecography, Vol. 16, 174–188.

Van Gardingen, P. R., Grace, J. B., and Jeffree, C. E. (1991) Abrasive damage by wind to needle surfaces of Picea sitchensis (Bong.) Carr. and L. Plant, Cell and Environment, Vol. 14, 185–193.

Van Raamsdonk, L. W. D. (1988) Principal component analysis of ecological indi-cator values of Swiss alpine flora. Botanica Helvetica, Vol. 98, No. 2, 195–205.

Vonlanthen, C., Bühler, A., Veit, H., Kammer, P., and Eugster, W. (2004) Charak-terisierung ökologischer Standortfaktoren in alpinen Pflanzengemeinschaften. Naturforschende Gesellschaft Bern, Vol. 61, 49–77.

Walker, M. D., Walker, D. A., and Auerbach, N. A. (1994) Plant communities of a tussock tundra landscape in the Brooks Range foothills, Alaska. Journal of Vege-tation Science, Vol. 5, 843–866.

Wallossek, C. (1990) Vegetationskundlich-ökologische Untersuchungen in der alpinen Stufe am SW-Rand der Dolomiten (Prov. Bozen und Trient). In Disserta-tiones Botanicae 154. Stuttgart, Germany: Cramer.

Warnecke, G. (1997) Meteorologie und Umwelt—eine Einführung (2nd edition). Berlin, Heidelberg, Germany: Springer.

Wieser, G., Neuwinger, I., and Cernusca, A. (1989) Der Bodenwasserhaushalt alpiner Graslandgesellschaften zwischen 1600 und 2600 m Höhe. In A. Cernusca, ed., Struktur und Funktion von Graslandökosystemen im Nationalpark Hohe Tauern. Innsbruck, Austria: Wagner, Veröffentlichungen des österreichischenMAB-Programmes 13, 95–118.



afg?hi%gfahj%.kllmh?j?i>n%a%>j a j ?>j?. a f%a>>i>>lihj%ko...

Documents