Introduction

Average global surface temperatures have increased by 0.6 ± 0.2 °C over the last century. The trend of increasing annual temperatures is expected to continue with predictions of a rise of global mean air temperatures of 1.8 to 4.0 °C by 2100 (Metz et al. 2007). Some natural ecosystems will be severely affected by climate change due to their limited adaptive capacity (Adger et al. 2007; Watson et al. 1997). In particular, alpine ecosystems are vulnerable to climate change because of their isolation and fragmentation and the existence of multiple stresses (Watson et al. 1997). Effects of climate warming on alpine vegetation have already been observed. For example, plant species have migrated upwards, increasing the number of species on high altitude summits (Grabherr et al. 1994; Pauli et al. 2001). Such changes in ecosystem composition, structure and diversity are apparent in high elevation ecosystems in Europe and have been attributed to warmer winter and summer temperatures, and longer growing seasons (Klanderud and Birks 2003; Kullman 2002; Moiseev and Shiyatov 2003; Virtanen 2003; Walther et al. 2005).

Vegetation changes in high altitude ecosystems as a response to climate change can also be associated with concomitant changes in soil processes as an altered temperature regime can affect the turnover of soil nutrients and plant community structures (Chapin et al. 1995). Soil enzymatic activities decrease with increasing altitude and decreasing temperature (Schinner 1982), and rates of N mineralization also decrease (Jacot et al. 2000b; Sveinbjornsson et al. 1995). Increased temperatures can therefore profoundly influence soil processes in cold environments, by increasing nutrient availability (Grogan and Chapin 2000; Michelsen et al. 1996) and enhancing rates of N mineralization (Hartley et al. 1999; Jonasson et al. 1999; Shaw and Harte 2001).

Although many studies have examined the effects of increasing temperature on nutrient cycling in arctic ecosystems (Grogan and Chapin 2000; Hobbie 1996; Jonasson et al. 2004; Jonasson et al. 1993; Schmidt et al. 2002; van Heerwaarden et al. 2003), only few have addressed nutrient transformations in high altitude ecosystems in-situ along natural temperature gradients (Jacot et al. 2000a; Morecroft et al. 1992; Sveinbjornsson et al. 1995) and their susceptibility to climate change. The aim of our study was to investigate the N dynamics of an alpine–nival ecotone on the basis of a relatively short altitudinal gradient (173 m) that was linked to a difference in mean annual temperatures of 2.4 °C. Due to the sensitivity of the prevalent vegetation to environmental conditions, the alpine–nival ecotone is ideally suited to study climate change effects (Gottfried et al. 1998) We hypothesized that lower air and soil temperatures due to increasing altitude will affect concentrations of plant available N in the soil and influence rates of soil N cycling. Natural abundance of the stable isotope 15N was used as a tool to study ecosystem processes as it acts as an integrator of the long-term ecosystem N balance (Högberg 1997). We expected that not only temperature but also other factors such as developmental stage and stability of the substrate influence N isotope signatures of soils and plants. In addition, roots were investigated for mycorrhizal colonisation and a complementary laboratory experiment was conducted to investigate the temperature sensitivity of two vital processes in the soil, N mineralization and nitrification.

Material and methods

Study area

The study was conducted at Mount Schrankogel, Stubaier Alps, Tyrol, Austria (11°05′58″E, 47°02′41″N; 3497 m) in August 1999 and 2000. The investigation area was located on a southwest slope between 2,900 m and 3,100 m a.s.l. The underlying parent material of the area around Mt. Schrankogel consists of siliceous material, mostly amphibolites (Patzelt 1976). Prevalent soil types under the alpine Carex curvula grassland and the patchy alpine–nival plant assemblages are alpine ranker, turf brown soils and alpine raw soils (leptosols and cambisols), respectively (Dullinger 1998). On the southwest ridge and slope, the closed alpine C. curvula grassland disintegrates into sward fragments at around 2,900 m and is finally replaced by nival rock and scree vegetation which is dominated by Androsace alpina, Poa laxa and Saxifraga bryoides (Gottfried et al. 1998; Pauli et al. 1999). This transition zone is referred to as alpine–nival ecotone (Gottfried et al. 1999). A vegetation classification by (Pauli et al. 1999) identified 14 vascular plant assemblages belonging to two vegetation types: alpine grassland (Caricion curvulae) and open, patchy nival plant assemblages (Androsacion alpinae). Closed alpine C. curvula grassland avoids unstable habitats with high cover of scree whereas nival vegetation establishes not only in rocky habitats but also on scree (Pauli et al. 1999).

Study plots

Within the study area, five plots of 3 × 3 m were selected along an altitudinal transect (2,906 to 3,079 m a.s.l.). Vascular plant assemblages at the chosen altitudes are shown in Table 1. The study plots on Mt. Schrankogel were located in the immediate vicinity of permanent plots as they had been established and described previously (Gottfried et al. 1998). Since 1997, temperature loggers (Tinytag logger, Gemini Data Loggers, UK) have measured the air or snow temperature at 1–3 cm above the soil surface which in terms of alpine plant life represents a key microclimatic zone (Gottfried et al. 1998). Mean annual temperatures for a period of three years are presented in Table 2 for each study plot. Temperature data clearly show that the altitudinal gradient is linked to a temperature gradient. (Gottfried et al. 1998) found on average an annual mean temperature lapse rate of 0.95°C per 100 m altitudinal increase between 2,900 and 3,500 m at Mt. Schrankogel.

Table 1 Altitude (meter above sea level) and description of plant assemblages of the study plots on Mt. Schrankogel, Tyrol, Austria
Full size table
Table 2 Characteristics (pH, C–N ratios, total N, total C and δ 13C) of top soils (0–10 cm) of the five study plots on Mt. Schrankogel, Tyrol, Austria
Full size table

Soil and plant sampling

Soils were collected in August 1999 and 2000, while plants were sampled in August 2000. Five replicate soil samples were taken from the upper 10 cm of the soil profile of each 3 × 3 m study plot. The samples were put into cold storage (4 °C) immediately after collection and transported to Vienna. Soils were sieved (<2 mm) and after the removal of fine roots stored at 4 °C. Extraction with 1 M KCl for solute analysis was conducted within a week after collection. Aliquots of soil samples were oven-dried at 60 °C for 2 days before isotope analysis. Above- and belowground plant parts were collected from 19 plant species (Table 3). Plant species were chosen according to their abundance at this altitudinal transect, focusing on the major species and on those that occurred at least in two different altitudinal plots at the study site. Samples were collected from within the 3 × 3 m plots where soil samples have been taken. Three replicate plants were sampled at each site as only a few individuals of each species were present in the plots. Plants were separated into roots and shoots and oven-dried at 60 °C.

Table 3 Plant species collected on Mt. Schrankogel, Tyrol, Austria for analysis of natural abundance of 15N
Full size table

Soil pH, \({\text{NO}}_3^ - \), \({\text{NH}}_4^ + \), dissolved organic nitrogen

Soil pH (0.01 M CaCl2) was determined from sieved soil samples (<2 mm) using a soil:solution ratio of 1:2.5. Nitrate was analysed in soil water extracts (1:10, w/v) by chemically suppressed ion chromatography (DX 500; Dionex, Vienna, Austria). Anions were separated on an anion exchange column (AS11; 250 × 4 mm i. d., Dionex) using a linear KOH gradient (2 to 40 mM in 7 min.). Ammonium and α-amino-N concentrations were determined in aliquots of 1 M KCl extracts (1:10, w/v) of the soils using colorimetric techniques in a microtiterplate reader photometer (Tecan Rainbow, Austria). Quantification of ammonium was carried out using the indophenol method (Kandeler and Gerber 1988) and of α-amino-N using the ninhydrin reaction (Amato and Ladd 1988). For the latter method, any interfering ammonium was removed by increasing the pH of extracts with MgO resulting in gaseous loss of ammonium as ammonia. From hereon we refer to the ninhydrin-reactive N as dissolved organic nitrogen (DON).

Nitrogen and carbon isotope composition

Dried soil and plant samples were ground to a fine powder in a ball mill (Retsch MM2; Retsch, Vienna, Austria), weighed into tin capsules and analysed by continuous-flow gas isotope ratio mass spectrometry (IRMS). The IRMS system consisted of an elemental analyser (EA 1110, CE Instruments, Milan, Italy) coupled via a ConFlo II interface (Finnigan MAT, Bremen, Germany) to the gas isotope ratio mass spectrometer (DeltaPLUS, Finnigan MAT). Reference gas (high purity CO2 and N2 gas; AGA, Vienna, Austria) was calibrated to the Vienna-Pee Dee Belemnite (V-PDB) standard using IAEA-CH-6 and IAEA-CH-7, and to the atmospheric N2 standard (at-air) using IAEA-NO-3, IAEA-N-1 and -2 reference material (International Atomic Energy Agency, Vienna, Austria). The abundance of C and N isotopes of each sample was determined in a single run, and N2 and CO2 reference gases were run with each analysis. The natural abundance of 15N and 13C was calculated as follows:

where R is the ratio of 15N/14N for nitrogen and 13C/12C for carbon isotope abundance. The precision of isotope measurements of a laboratory standard was 0.15‰ for δ 15N and 0.10‰ for δ 13C.

Gross mineralization and gross nitrification

Gross mineralization and gross nitrification were determined in the replicated soil samples that were collected from the 3 × 3 m plots at three altitudes (2,906 m, 2,944 m and 3,030 m) in 1999 and 2000, and incubation temperature was varied between 0 and 24 °C. Rates of gross nitrogen transformations were determined by 15N pool-dilution technique. Duplicate samples (2 g) of sieved soil (<2 mm) were transferred into polyethylene vials and mixed with 0.5 ml of a 0.25 mM (15NH4)2SO4 or 0.5 ml of a 0.5 mM 15KNO3 solution for the mineralization and nitrification assays, respectively. Mineralization and nitrification assays were stopped after 4 h in half of the samples and in the second half after 20 h by extracting with 14 ml of 1 M KCl for 4 h. In order to remove interfering organic nitrogen from the nitrification assays, samples were treated with 30% H2O2 and concentrated H2SO4 in a UV digester (60 min). In this reaction, organic nitrogen was decomposed and the ammonium released was removed by increasing the pH >9.5 through addition of 32% NaOH and shaking on a horizontal shaker for 2 days. By adding Devarda reagent, nitrate was then converted to ammonium which was collected by microdiffusion in acid traps at 25 °C for 48 h (Khan et al. 1998). Acid traps were made of discs of ash-free filter paper amended with 15 μl 2.5 M KHSO4 and coated in strips of teflon tape. Mineralization samples were treated with MgO to increase the solution pH to >9.5 and the ammonia released into the headspace was collected as above. Thereafter, acid traps were dried over concentrated H2SO4, the filter papers were removed from the Teflon coating, transferred into tin capsules and the amount of 15N and total N measured by IRMS.

Rates of gross mineralization and gross nitrification were calculated using the equations described in (Barrett and Burke 2000):

$$\begin{array}{*{20}c}{m = \frac{{M_0 - M}}{t} \times \frac{{{{\log H_0 M} \mathord{\left/{\vphantom {{\log H_0 M} {HM_0 }}} \right.\kern-\nulldelimiterspace} {HM_0 }}}}{{{{\log M_0 } \mathord{\left/{\vphantom {{\log M_0 } M}} \right.\kern-\nulldelimiterspace} M}}}} \\{n = \frac{{M_0 - M}}{t} \times \frac{{{{\log H_0 M} \mathord{\left/{\vphantom {{\log H_0 M} {HM_0 }}} \right.\kern-\nulldelimiterspace} {HM_0 }}}}{{{{\log M_0 } \mathord{\left/{\vphantom {{\log M_0 } M}} \right.\kern-\nulldelimiterspace} M}}}} \\\end{array} $$
m :

gross mineralization in mg kg−1 d−1

n :

gross nitrification in mg kg−1 d−1

H0, H:

15N g soil−1 at time t = 0 and t = t respectively (t = 20 h)

M0, M:

total N (14N +15N g soil−1) at time t = 0 and t = t respectively (t = 20 h)

Determination of mycorrhizal root colonisation

For assessment of root colonisation by arbuscular mycorrhizal (AM) and dark septate (DS) fungi, roots were cleared in 10% KOH at 90 °C for 30 min and thereafter stained employing the ink/vinegar method (Vierheilig et al. 2005). Randomly selected root pieces (three per plant) from species collected in all five study plots were mounted on microscope slides and the percentage of root length colonised by the respective fungi was assessed under a microscope at 100 × magnification (McGonigle et al. 1990). Colonisation data are expressed as the percentage of total root intersects examined that contain fungal structures (= % root length colonised).

Radiocarbon determination

Radiocarbon measurements of soil organic matter were made by accelerator mass spectrometry (AMS) at the Vienna Environmental Research Accelerator (VERA #2502−#2504). For this analysis, the humic acid fraction was used. The 14C data (corrected for isotopic fractionation) are presented as pMC (percent Modern Carbon) and, where pMC values were <100, as radiocarbon age (years BP) (Stuiver and Polach 1977).

Statistical analysis

Statistical analysis was carried out by polynomial regression analysis based on first and second order models using Statgraphics Centurion XV software (Statistical Graphics Corp.).

Results

Temperature, soil pH, soil N concentration and C isotope composition

With increasing altitude mean annual air temperature on Mt. Schrankogel, measured at 1–3 cm above soil surface, decreased from 1.7 °C at the lowest elevation (2,906 m) to −0.7 °C at the highest elevation (3,079 m) (Table 2). The temperature difference between the lowest and highest altitude was therefore 2.4 °C for the years 1997–2000. Soil pH increased from 4.6 to 5.5 with altitude (Table 2). Soil C–N ratio, total N and total C decreased along the altitudinal gradient (Table 2), whereby the lowest site (2,906 m) was always markedly different from the sites at higher altitudes. Soil δ 13C did not change with increasing elevation (Table 2).

15N natural abundance of soil and plant material

The δ 15N signatures of soil (polynomial regression analysis, second order model; P < 0.001) and plant material (P < 0.001) decreased significantly with elevation and were more positive at the lowest altitude (Fig. 1). The biggest difference of soil δ 15N was between the site at 2,906 m and at 3,030 m, where δ 15N values decreased by 5.5‰ (from +2.15‰ to −3.34‰.). Soil pH and δ 15N of soil exhibited a significant negative linear relationship (Fig. 2) where δ 15N of soils decreased (R 2 = 0.57, P < 0.001) with decreasing soil acidity. δ 15N values of soils were always more positive than those of plant tissues whereas δ 15N signatures of leaves and roots were similar (Fig. 1). The shift in δ 15N values along the altitudinal gradient was more pronounced in soils (5.5‰) than in plants where δ 15N values only decreased by about 2‰. The four plant species that were collected across the whole altitudinal transect showed similar, significant negative relationships between δ 15N and altitude. The trend was more pronounced for Leucanthemopsis alpina (P < 0.001) and Luzula spicata (P < 0.001) than for P. laxa (P = 0.023) and S. bryoides (P = 0.065) (3.0 to 3.5‰ versus 1.5 to 1.0‰).

Fig. 1
figure 1

Relationship between soil concentrations of dissolved organic N (a), ammonium (b), nitrate (c) and total plant-available N (d) and altitude at Mt. Schrankogel, Austria. Error bars represent standard errors (n = 5–15). Results of a polynomial regression analysis (second order model) are given

Full size image
Fig. 2
figure 2

Relationship between assay temperature and gross mineralization in top soils collected from an alpine–nival ecotone at Mt. Schrankogel, Austria, (a) in 1999 and (b) in 2000. Error bars represent standard errors (n = 9). Results of a polynomial regression analysis (second order model) are given

Full size image

Mycorrhizal root colonisation

L. alpina from all five sites was heavily colonised by arbuscular mycorrhizal (AM) fungi (up to 80%), and only to a much lesser extent by dark septate (DS) fungi (between 0 to 10%). L. spicata roots contained no AM fungal structures, while root colonisation with DS fungi varied between 0 and 100% of total root length. P. laxa showed low levels of AM (0 to 15%) and highly variable levels of colonisation with DS fungi (0 to 65%) along the altitudinal transect. Root colonisation of S. bryoides ranged from 0 to 45% by AM, and <5 to 100% by DS fungi. For all species, neither AM nor DS fungal root colonisation was correlated with altitude (P > 0.1).

Plant available forms of N in soils

The sum of plant available N (i.e., ammonium, nitrate and DON) in soil extracts decreased with increasing altitude and concentrations at 2,906 m differed significantly from those at 2,987 m and 3,079 m (Fig. 3, polynomial regression analysis, second order model; P < 0.036). A closer view on the fractions of available N in the soil extracts showed that the concentrations of ammonium (P = 0.061) and of DON (P = 0.014) decreased with altitude, whereas soil nitrate concentration did not change significantly with altitude (P = 0.464). Consequently, the relative contribution of each form of N to the sum of available nitrogen changed with increasing altitude. At 2,906 m, ammonium, DON and nitrate made up for 43%, 46% and 11% of soluble N in the soil, respectively. At 3,079 m, the contribution of ammonium to plant available N in the soil was similar (46%), however, that of DON decreased to 22% and the contribution of nitrate increased to 32%.

Fig. 3
figure 3

Modelled rates of gross mineralization for a growing season (May to August) at different altitudes at Mt. Schrankogel, Austria. Cumulative rates of \({\text{NH}}_4^ + \)production were calculated using the temperature-mineralisation relationships derived from Fig. 2 and mean monthly temperatures during May to August in 1999 and 2000

Full size image

Rates of gross mineralization and nitrification – measured and calculated

We found no significant differences in the temperature sensitivity of gross transformation rates between high and low altitude soils and therefore combined the data for Fig. 4. Rates of gross mineralization increased significantly from 0 °C to 24 °C as presented in Fig. 4a,b (polynomial regression analysis, second order model; P = 0.019 and 0.007 for 1999 and 200). The amounts of ammonium produced increased from 1.03 mg N kg−1 DW d−1 at 0°C to 2.56 mg N kg−1 DW d−1 at 24 °C in 1999 and from 0.98 to 2.30 mg N kg−1 DW d−1 in 2000. Rates of gross nitrification are not shown as they were low (<0.4 mg N kg−1 DW) and since the data did not indicate any temperature dependency of nitrate production. We used the results of the incubation experiment to model cumulative nitrogen release by gross mineralization in 1999 and 2000 (Fig. 5). Total nitrogen release during the growing season (May to August) was calculated by relating the rates of gross mineralization from the laboratory experiment to mean monthly air temperatures (measured at 1–3 cm above ground) in 1999 and 2000. The data showed that with increasing altitude cumulative seasonal N production decreased from 124 (144) mg N kg−1 DW season−1 to 119 (123) mg N kg−1 DW season−1 in the year 1999 (2000). Calculated in this way, seasonal N production was higher in 2000 than in 1999 due to a steeper gradient of mean annual temperatures between the lowest and highest altitude (slope k of quadratic regression was −0.2 in 1999 and −0.8 in 2000).

Fig. 4
figure 4

Relationship between soil pH and δ 15N of soils from different altitudes between 2,906 m and 3,079 m at Mt. Schrankogel, Austria (n = 25). Results of a polynomial regression analysis (second order model) are given

Full size image
Fig. 5
figure 5

Relationship between δ 15N signatures of top soils (0–10 cm), δ 15N of leaves and roots of plant species and altitude at Mt. Schrankogel (see Table 3 for species list). Mean vegetation δ 15N values of above- and belowground plant material were calculated from single plant species after weighing by species cover on the respective plots. Error bars represent standard errors (n = 5–36). Results of a polynomial regression analysis (second order model) are given

Full size image

Radiocarbon analysis of humic fraction

14C-dating of the humic acid fraction of soils from three different altitudes showed that soils from the lowest altitude were 665 ± 30 years old (92.1 ± 0.4 pMC) whereas humic acids from higher altitudes were less then 50 years old (105.2 ± 0.5 and 101.6 ± 0.6 pMC for site 2 and 4, respectively).

Discussion

The N dynamics in the alpine–nival ecotone on Mt. Schrankogel clearly changed as mean annual air temperatures declined with increasing altitude. As hypothesized, lower temperatures reduced rates of ecosystem N turnover and N accumulation. This resulted in lower amounts of available and total N in soils at our higher altitude study plots (Table 2 and Fig. 1). A decline of soil organic matter with increasing altitude was also reflected by decreasing concentrations of total C in soil that led to a narrowing of C:N ratios with increasing altitude (Table 2). In vivo measurements of gross ammonium production at various temperatures (0–24 °C) supported our findings showing a reduction of rates of mineralization with decreasing temperature (Fig. 2). Interpolation of these numbers to overall seasonal production of N yet again demonstrated lower rates of mineralization at higher altitudes where temperatures are lower (Fig. 3). Our data are in accordance with numerous studies in arctic and alpine ecosystems that reported a strong relationship between temperature and rates of N mineralization (e.g. (Hartley et al. 1999; Hobbie 1996; Jonasson et al. 1993; Nadelhoffer et al. 1991; Peterjohn et al. 1994).

Changes in rates of soil N turnover along the altitudinal gradient on Mt. Schrankogel influenced the availability of the different forms of N in the surface soil (Fig. 1). The concentrations of ammonium and DON in soils decreased with increasing altitude reflecting decreasing rates of N mineralization with altitude. In contrast, the concentrations of nitrate increased slightly with increasing altitude, possibly due to the insensitivity of nitrification to temperature shown in this study when compared with N mineralization.

In this context, it is important to acknowledge that temperature also influences plant assemblages and cover which in turn affect soil N cycling through litter quality and quantity (Bowman et al. 2004; Steltzer and Bowman 1998). The influence of plants on N dynamics potentially has important consequences for soil processes if climate warming alters the structure and composition of plant communities (Hobbie 1996). In addition, in such a scenario temperature may influence plant N uptake directly via increased productivity (Atkin and Cummins 1994; Garnett and Smethurst 1999; Volder et al. 2000). Moreover, soil microbial community composition (which itself is tightly interconnected with plant species composition) may play a crucial role in driving the cycling of N.

Although natural abundance of 15N does not allow a clear assessment of mechanisms or a quantification of process rates or process identity (Handley and Raven 1992) it provides a general understanding of underlying ecosystem processes. In our study, soil δ 15N and pH were strongly correlated showing that soils with more acidic soil pH (as found in lower altitudes) had more positive δ 15N values. Older soils usually have more positive δ 15N signatures due to longer microbial processing (Kramer et al. 2003) and greater cumulative losses of 15N depleted N such as nitrate and nitrous oxide over an extended period of time (Brenner et al. 2001; Nadelhoffer and Fry 1994). Further, more developed soils usually have more acidic pH values as soil pH decreases with age (Cox and Whelan 2000; de Kovel et al. 2000; Munroe and Bockheim 2001) and soil development (Berendse et al. 1998). This relationship and results from the 14C-dating of the humic acid fraction of soils from three different altitudes suggest that not only temperature influenced N cycling in the alpine–nival ecotone but also soil age and developmental stage. Indeed, soils under closed alpine grassland (alpine ranker and turf brown soils) are of a different type than soils under patchy nival plant assemblages (alpine raw soils) (Dullinger 1998), supporting the findings that soils at higher altitudes are younger and less well developed.

Substrate stability and duration of snow cover most likely also contribute to differences in soil development within the alpine–nival ecotone. Species in the nival zone can tolerate unstable scree whereas alpine species such as C. curvula preferably grow on more stable substrate (Pauli et al. 1999). Unstable habitats are subject to greater erosive forces leading to greater ecosystem N losses. Plant species in the nival zone on Mt. Schrankogel experience an even shorter vegetation period than species in the alpine zone owing to longer-lasting snow cover until the summer months (Gottfried et al. 2002). This also slows N cycling at higher altitudes.

Natural abundance of 15N in plant species decreased with increasing altitude (Fig. 6) in a similar way as soil δ 15N which is expected as soil N and plant N nutrition are tightly connected. Various factors control plant δ 15N, such as the isotopic composition of N sources and 15N discrimination during plant N acquisition, both of which are affected by microbial colonization of roots, by growth N demand and competition, and by nitrogen form. Internal N cycling in plants did not affect plant δ 15N values as leaves and roots had similar δ 15N values (Fig. 5). Symbiotic associations with microorganisms that have access to N sources not available for plants can affect plant N isotope composition; however, all plant species investigated were non-N2 fixing species. Further, level and type of mycorrhizal association have been shown to affect plant δ 15N (Michelsen et al. 1998). In this study mycorrhizal colonisation by dark septate and arbuscular mycorrhizal fungi did not change with increasing altitude and were quite variable, although others have found fungal colonisation of roots to decrease with altitude in alpine regions (Haselwandter and Read 1980).

Fig. 6
figure 6

Relationship between foliar δ 15N of four alpine plant species and altitude at Mt. Schrankogel. Data points represent all individuals (n = 15) of four selected species, Leucanthemopsis alpina, Luzula spicata, Poa laxa and Saxifraga bryoides. These species were sampled from all five study plots on Mt. Schrankogel between 2,906 and 3,079 m. Results of a polynomial regression analysis (second order model) are given

Full size image

Competition and nutrient limitations may also influence the δ 15N of plant material. On Mt. Schrankogel plant–plant competition was visibly less at the open nival vegetation (higher altitude) than at the alpine grassland site (lower altitude). This is further corroborated since the effect of neighbouring plants on each other changes with altitude from competition to facilitation in alpine communities (Brooker et al. 2005; Callaway et al. 2002; Choler et al. 2001; Kikvidze et al. 2005). Limited N supply to plants – as may be the case in our lowest altitude plot – has reportedly resulted in no or only little N isotope fractionation (Evans et al. 1996; McKee et al. 2002; Nadelhoffer and Fry 1994). Consequently, more efficient uptake of N by plant roots would lead to less depleted foliar δ 15N values which were not evident in this study, comparing the divergence between plant and soil δ 15N with decreasing altitude.

Temperton et al. (2007) found that plant δ 15N signatures in a biodiversity experiment decreased with increasing diversity of the plant assemblages. This suggested that changes in mineralization and or nitrification rates in more diverse versus less diverse plant assemblages could produce different δ 15N signals in plants. We did not find a similar but an inverse relationship between plant δ 15N and plant diversity, however, plant and soil δ 15N values were also likely controlled by soil microbial processes that result in 15N depletion of soil inorganic N and consequently of plant N (Amundson et al. 2003).

It can therefore be speculated that the different ecosystems across the alpine–nival ecotone may differently respond to climate and atmospheric change. At the nival zone, where N limitation is not likely, an increase in atmospheric N deposition may not affect plant primary production. However, warmer temperatures may directly lead to increased plant growth and altered plant composition (Pauli et al. 2007) and may also increase rates of N mineralization. In contrast, at the alpine sites, where plant growth may be N limited (Callaway et al. 2002) due to stronger competition, it may be speculated that increased atmospheric deposition of N could markedly affect plant productivity and lead to changes in community structure.

In summary, we could show in this study that ecosystem nitrogen dynamics changed remarkably along a short altitudinal gradient (173 m) at the alpine–nival ecotone on Mt. Schrankogel. Our data from the field and from laboratory experiments suggest that temperature, soil age and development and plant competition all influence ecosystem N cycling, although to a varying extent at the different ecosystems.