SITE DESCRIPTION

The study was carried out during July 2000 on the Taymyr Peninsula. We sampled from two sites that represent (1) the southern tundra subzone and (2) the typical tundra subzone of the Russian Arctic, described by Chernov and Matveyeva (1997).

The sampling site in the typical tundra subzone was located near Lake Nyagamya (72°42′N, 88°26′E). The vegetation at the site may be classified as erect dwarf-shrub tundra (Walker, 2000) consisting of a moss layer ∼8 cm high and a dwarf-shrub layer as high as 35 cm. At this site, three subsites were studied along a soil-moisture gradient from the shoreline of the lake (wet site) to a well-drained area on the top of a ridge (mesic site). The third subsite was established on a poorly drained, late-melting place at the foot of a hill, where interhummock areas were completely waterlogged during the study period (waterlogged site).

The sampling site in the southern tundra subzone was situated in the vicinity of River Pyasina (70°17′N, 88°31′E). The vegetation at the site may be classified as low-shrub tundra. Here, in addition to a moss and dwarf-shrub layer, some shrubs of up to 1 m height were present. The actual sampling site in the southern tundra subzone was a mesic zonal site on a gentle slope (∼2°).

Both research sites were characterized by the presence of hummocks (Fig. 1). The earth hummocks ranged from 10 to 30 cm in height and from 30 to 50 cm in diameter. They accounted for between 40 and 60% of the soil surface. The main soil and vegetation characteristics of the research sites are summarized in Tables 1 and 2. Closed plant communities occurred on both hummocks and interhummock areas. The hummocks were dominated by grasses and sedges; Carex bigelowii and different Eriophorum species were most abundant. Dwarf shrubs, mostly Dryas octopetala and D. punctata, covered 20–30% of the surface of hummocks at the southern tundra site and at the mesic subsite of typical tundra, but were absent at the other two subsites of typical tundra. The vegetation of the interhummock areas was mainly composed of mosses, which formed a cover ranging from 60 to 100%. On drier sites, the mosses were associated with lichens, but the lichens formed a low overall cover (<5%). On the wetter plots of the typical tundra site and throughout the southern tundra site, deciduous shrubs, mainly Salix species, also rooted in the interhummock areas. Additionally, a range of forbs grew in both subzones of tundra, although most of them appeared with low frequency, and some were only locally abundant.

Gleyic Cryosols represented the major soil types at the southern tundra site, and Orthic Cryosols dominated at the typical tundra site. All soils had a thin organic layer of largely undecomposed plant material at the surface (1–3 cm thick). The soil showed no distinct horizons at the southern tundra site, but was divided into upper A-horizon (Ap in hummocks, Ah in interhummock areas) and lower Bg-horizon at the typical tundra site. The Ah-horizon averaged 23 cm thick (minimum, 17 cm), and the Ap horizon averaged 39 cm thick (minimum, 28 cm). The permafrost table was roughly parallel to (but slightly flattened in comparison with) the microtopography of the surface. The climate at each site is continental-Arctic, and the mean annual temperature of the Taymyr Peninsula is −14°C (Matveyeva, 1994).

SOIL SAMPLING

For each subsite in the typical tundra subzone and for the southern tundra site, five hummocks and five adjacent interhummock areas were sampled within an area of 300 × 300 m. After removing the layer of undecomposed organic material, soil samples were taken either to the permafrost table or to a depth of 15 cm, sieved (2 mm) and thoroughly mixed. Subsamples were oven-dried (at ∼60°C), and total N and C and soil-moisture content were determined. Extracts were prepared from 1-g aliquots of fresh soil by shaking them for 1 h in 10 mL of 1 M KCl or deionized water followed by filtration. After addition of a microbial inhibitor, 10 μM phenylmercuric acetate, the extracts were stored at field temperatures (0–5°C) until transport to laboratories in Vienna. The KCl extracts were used to measure exchangeable NH₄⁺ and amino acids, whereas NO₃⁻ was determined in water extracts. The active-layer depth (ALD) was measured by means of a metal rod. Soil temperature was logged at a soil depth of 5 cm every 30 min throughout the 4 weeks of the study.

SOIL NITROGEN AND CARBON CONTENTS

Amino acids were analyzed colorimetrically as α-amino N by applying a modified ninhydrin assay of Amato and Ladd (1988). Briefly, aliquots of the extract (1 mL) were mixed with 10 mg MgO and shaken for 12 h in open vials at room temperature to remove NH₄⁺. After centrifugation, the ninhydrin reagent (Sigma-Aldrich, Vienna, Austria) was added to an aliquot of the supernatant (1:1), and the mixture was heated to 100°C in a water bath for 15 min. After cooling, ethanol (50%) was added (1.25:1), and the absorbance was measured at 570 nm. NH₄⁺ was analyzed by use of the salicylic acid assay as described by Kandeler and Gerber (1988). NO₃⁻ was determined by ion-chromatography and conductivity detection after chemical suppression (DX 500 and ASRS-Ultra, Dionex, Vienna, Austria) by using an anion-exchange column (AS11; 250 × 4 mm i.d., Dionex) and NaOH gradient elution (2–40 mM in 8 min). Total N and C were determined with an elemental analyzer (EA 1110, CE Instruments, Milan, Italy). Atropine was used as the elemental reference material.

GROSS NITROGEN-TRANSFORMATION

Rates of gross N transformation (gross N mineralization, N immobilization) were assessed by using the ¹⁵N-pool-dilution method (Myrold and Tiedje, 1986; Schimel et al., 1986). To two subsamples each of 2-g sieved soil, 400-μL aliquots of 0.25 mM (¹⁵NH₄)SO₄ were applied uniformly. The assays were mixed thoroughly and incubated at ambient soil temperature for periods of 4 and 24 h. Incubations were terminated by extracting the samples with 12 mL of 1 M KCl. Prior to analysis, ¹⁵NH₄ was recovered in acid traps after adding 100 mg of MgO over 5 days. The acid traps consisted of circular, ash-free filter papers containing 10 μL of 2.5 M KHSO₄ and were wrapped in Teflon-coated tapes. After drying, isotopic enrichment was measured by continuous-flow isotope-ratio mass spectrometry (IRMS) using an elemental analyzer coupled to a gas IRMS system (Delta^PLUS; Finnigan MAT, Bremen, Germany), as described in detail by Wania et al. (2002). Rates of gross N mineralization and N immobilization were calculated by using the equation provided by Barrett and Burke (2000). N immobilization was calculated with the assumption that gross N nitrification rates are negligible, because cold and wet conditions of Arctic soils are considered to largely inhibit nitrifiers (Nadelhoffer et al., 1992).

MICROBIAL-BIOMASS NITROGEN

Microbial biomass was measured by the fumigation-extraction method according to Amato and Ladd (1988) with the following modifications: a subsample of 2 g of fresh soil was fumigated with ethanol-free chloroform for 24 h and thereafter extracted with 15 mL of 1 M KCl. The amount of α-amino N released by chloroform fumigation was estimated as the difference between the ninhydrin-reactive N in the fumigated and the nonfumigated soil extracts. Soil microbial-biomass N was calculated by using the following equation: biomass N = 3.1 × ninhydrin-reactive N released from fumigated soil (Amato and Ladd, 1988).

STATISTICS

Data were analyzed by using Statgraphics 4.0 (Stats Incorporated). Significance of differences between means of the study sites and between hummocks and interhummock areas of the southern and typical tundra subzones were assessed by using one-way analysis of variance (ANOVA) and the least significant difference test (LSD). By including the three subsites of the typical tundra subzone in the analysis of differences between relief-related microsites, the combined effects of relief and subsite were examined. Means were considered significantly different at the α = 0.05 level.

SOIL CHARACTERISTICS

Soil temperature was significantly lower in interhummock areas than in hummocks (Table 1). This difference was also reflected by shallower active-layer depths (ALDs) in the depressions. Soil-moisture contents were significantly higher for interhummock soils than for hummocks. However, the difference was least pronounced for the waterlogged subsite of the typical tundra subzone (Table 1). Although the soils at the latter subsite were apparently waterlogged, this interpretation was not reflected in data on soil-moisture content. The content of clay (particles < 0.002 mm), known to hold more water than larger particles, was relatively low at this typical tundra subsite (Table 1), most likely leading to a reduced water-holding capacity of these soils. The pH values of water in the soils were between 5.7 and 6.7 and showed no marked differences between hummocks and interhummock areas of either site (Table 1).

The soils differed in organic C and total N levels (Table 1). The contents of C and N were higher at the southern tundra site than in typical tundra. The interhummock areas always exhibited higher C and N contents than hummocks, with the exception of the waterlogged subsite in the typical tundra, where the contents were similar at both microsites. No significant differences were observed in C/N ratios, although they tended to be higher in interhummock areas compared to hummocks and in soils of the southern tundra compared to the typical tundra subzone.

NITROGEN POOLS

NH₄⁺ dominated the pool of soluble N in all habitats (2–20 μg N g⁻¹ dry weight [DW]), followed by amino acids (0.4–8 μg N g⁻¹ DW) and NO₃⁻ (0.5–1.7 μg N g⁻¹ DW; Fig. 2). Levels of microbial-biomass N ranged from 3.5 to 150 μg g⁻¹ DW (Fig. 3) and were, thus, relatively high compared to levels of KCl-extractable N. Microbial-biomass N represented on average 130% of the concentration of available N in typical tundra and 440% at the southern tundra site and accounted for 0.3% and 1.4% of total N at the sites, respectively. Soils at the southern tundra site exhibited significantly greater N-pool sizes than soils of typical tundra. The difference was most pronounced in microbial-biomass N, with a 10-fold higher N concentration at the southern tundra site compared to the typical tundra site.

Considerable variations in N concentrations were also observed between soils of hummocks and interhummock areas. All sites showed a strong tendency toward higher N concentrations in the interhummock soils, with statistically significant differences for NO₃⁻ and NH₄⁺ at the wet and the mesic subsites of the typical tundra site (P < 0.05, ANOVA, LSD) and for all N pools of the southern tundra site (P < 0.01). Only the waterlogged subsite of the typical tundra site, where generally the lowest N availabilities were detected, did not show any differences in N pools. Microbial-biomass N was also higher in interhummock soils compared to hummock soils at the southern tundra site (P < 0.05). Total N, extractable N, and microbial-biomass N were similar in hummock soils of the waterlogged, wet, and mesic subsite. Inorganic N decreased with increasing soil-moisture content in soils of interhummock areas.

NITROGEN-TRANSFORMATION PROCESSES

The general patterns of gross N mineralization were similar to that of N pools. The highest rates of gross mineralization were observed in interhummock soils of the southern tundra site, and the lowest were seen in soils at the waterlogged subsite of the typical tundra site (Fig. 4). Again, no differences between the hummocks of waterlogged, wet, and mesic sites in typical tundra could be found, whereas in interhummock areas, the lowest N-mineralization rates were observed in the waterlogged soils. Mean rates of gross mineralization were fivefold higher at the southern tundra site than in typical tundra. Except for the typical-tundra waterlogged subsite, where no differences between hummocks and interhummock areas were observed, rates of gross mineralization were on average threefold higher in interhummock areas. However, these higher rates were only significant for soils of the southern tundra (P < 0.01). No differences between hummocks and interhummock areas were found for N immobilization (Fig. 5). The rates were higher in southern tundra compared to typical tundra.

EFFECTS OF MICRORELIEF AND SOIL CONDITIONS

In hummock tundra, a lateral flow of water from elevated hummocks to lower-situated interhummock areas occurs (Quinton, 2000). Because higher N pools were found in interhummock soils compared to hummock soils, it is likely that dissolved organic and inorganic N forms accumulate in the depressions of the interhummock areas as a result of their lower microtopographic position in the relief. Both surface as well as subsurface runoff along the permafrost table, which was roughly similarly shaped as the surface, may contribute to the accumulation of nutrients in interhummock areas. A similar flow of nutrients has frequently been demonstrated along several topographic sequences (Giblin et al., 1991; Högberg, 2001; Kummerow et al., 1987). In tussock tundra—a system that is structured similarly to hummock tundra—higher concentrations of mineral nutrients have also been found in intertussock compared to tussock soils (Cheng et al., 1998).

The relief may not only promote flow of water and nutrients away from the hummocks, but may also promote litter redistribution by wind and snow from exposed to lower-situated parts of the relief (Fahnestock et al., 2000). At the sites studied here, the litter of the majority of the species accumulated in interhummock areas, providing organic matter input to the soil (data not shown). This redistribution may have led to the considerably higher amounts of organic N and C observed in soils of interhummock areas (Table 1). Higher organic N and C, in turn, may have favored higher N mineralization in interhummock soils, as evidenced by faster N-mineralization rates (Fig. 4). Soil C and N have been shown to be primary drivers for gross N-mineralization rates (Booth et al., 2003). In our study, both higher organic matter contents and greater N availabilities due to water inflow possibly contributed to faster N-mineralization rates of interhummock areas, although the relative contribution of the two processes to the observed pattern cannot be determined by our data. However, the differences in microbial-mineralization rates and soil-nutrient status were accompanied by differences in microbial biomass, a pattern that was also found in tussock tundra (Cheng et al., 1998). N-immobilization rates were lower than gross N-mineralization rates, indicating net N mineralization in Arctic soils. The rates were similar between hummocks and interhummock areas, which is consistent with the higher N availability in interhummock soils.

Several physicochemical parameters such as water content and temperature may also vary with microrelief. Within this generally wet ecosystem (Christensen et al., 1998), interhummock areas of each subzone and subsite were always wetter than hummocks, if not waterlogged, and mean temperatures were lower. Both factors—low temperatures and high water contents—generally restrict microbial activity (Christensen et al., 1998; Flanagan and Veum, 1974; Johnson et al., 1996). The inhibiting effect of waterlogging on microbial activity was shown in the typical tundra subsites by the lower inorganic N availability and N-mineralization rates of waterlogged soils in the interhummock areas. This effect even overruled the effect of temperature, because the waterlogged site was the warmest of the three subsites of typical tundra. Nevertheless, compared to hummocks, we found higher N-mineralization rates and microbial N pools at interhummock areas, which were wetter and cooler than the hummocks. This result suggests that both effects—soil-moisture content and temperature—were overruled by other factors at the small spatial scale considered here.

Although many studies, especially those utilizing experimental approaches, have demonstrated that temperature (Hartley et al., 1999; Hobbie, 1996; Oberbauer et al., 1991) and soil-moisture content (Johnson et al., 1996) are major determinants on nutrient-cycling rates of Arctic soils, increasing evidence also indicates that other topographic patterns are controlling nutrient turnover. For example, an incubation study with Arctic soils at different sites along a toposequence revealed that the overall effect of quality of soil organic matter was more important in controlling C- and N-mineralization rates than temperature (Nadelhoffer et al., 1991). Environmental factors such as substrate quantity (Fahnestock et al., 2000; Giblin et al., 1991) or quality (Johnson and Damman, 1991; Moore, 1991) often appear to overrule effects of temperature or soil humidity in the field. As a result of greater water saturation of soils, Fisk et al. (1998) suggested that plant-mediated factors may have greater influence on nutrient-mineralization rates in Arctic (as compared to, e.g., alpine) ecosystems than soil climate. This situation may lead to an adaptation of Arctic microorganisms to anaerobic soil conditions (Schmidt, 1999; Schmidt and Bolter, 2002). Our data add support to this possibility, because the observed differences in N pools and N-mineralization rates between hummocks and interhummock areas were not related to differences in temperature and soil-moisture content, but may be attributed to differences in dissolved nutrients and substrate input in the form of litter. Thus, our primary hypothesis was not confirmed by our data. However, whenever plant composition and structure are similar between Arctic sites with different climatic regimes, the temperatures and soil-moisture contents may control nutrient dynamics. In our study, this was evident by (1) higher N pools and turnover rates at the lower-latitude site (southern tundra subzone) compared to the site in the typical tundra subzone and (2) the negative effect of waterlogging on N mineralization in interhummock areas along the moisture gradient in the typical tundra subzone.

EFFECTS OF PLANT FUNCTIONAL TYPES

As mentioned already, plant species and, thus, substrate quality and quantity may also be a crucial determinant of biogeochemical cycles in Arctic soils (Hobbie, 1995; Shaver et al., 1998). In hummock tundra the dominant plant species of hummocks are sedges and dwarf shrubs, whereas the interhummock areas are predominantly colonized by mosses (Table 2). Mosses are generally thought to slow down nutrient turnover and decomposition, because of their recalcitrant litter and abundant antimicrobial substances (Aerts et al., 1999; Banerjee and Sen, 1979; Basile et al., 1999; Turetsky, 2003; Malmer et al., 2003). However, we found significantly higher N-mineralization rates in interhummock areas under mosses. A closer look into the literature revealed that higher rates of N mineralization may occur together with lower decomposition rates in Arctic and boreal regions and temperate bogs and fens (Hobbie, 1996; Aerts et al., 1999; Verhoeven et al., 1990; Scheffer and Aerts, 2000; Fisk et al., 2003; Wardle et al., 2003). Possible reasons for these observations include a low N-immobilization potential of microbes at moss-dominated sites (Hobbie, 1996; Verhoeven et al., 1990) or different microbial communities (Scheffer and Aerts, 2000). This possibility suggests that under such circumstances, decomposition is more limited by carbon than by nitrogen, leading to a net N mineralization (Hobbie, 1996; Schimel and Weintraub, 2003).

It is interesting to note that we also found higher microbial biomass under mosses, and thus, another mechanism may have been of importance. Mosses access nutrients mainly from wet and dry deposition, with a low uptake of nutrients from the soil (Bates, 2000; Marion et al., 1982). Hence, in interhummock soils, microorganisms may experience only low competition from plants for nutrients in soil and litter, thus promoting microbial biomass and high net mineralization rates. In contrast to mosses, sedges and grasses that dominate hummocks are characterized by dense and deep root systems (Eissenstat et al., 2000; Gebauer et al., 1995), thereby effectively acquiring both inorganic and organic N. Hence, vascular plants may efficiently compete with microorganisms for nutrients, thereby controlling microbial biomass and activity (Wang and Bakken, 1997; Marion et al., 1982; Schimel and Bennett, 2004). However, patterns of competition between microorganisms and plants may change during the course of the season (Bardgett et al., 2002; Jaeger et al., 1999).