Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
Author for correspondence: Karina E. Clemmensen Tel: +45 3532 2298 Fax: +45 3532 2321 Email: firstname.lastname@example.org
• Shrub abundance is expected to increase with enhanced temperature and nutrient availability in the Arctic, and associated changes in abundance of ectomycorrhizal (EM) fungi could be a key link between plant responses and longer-term changes in soil organic matter storage. This study quantifies the response in EM fungal abundance to long-term warming and fertilization in two arctic ecosystems with contrasting responses of the EM shrub Betula nana.
• Ergosterol was used as a biomarker for living fungal biomass in roots and organic soil and ingrowth bags were used to estimate EM mycelial production. We measured 15N and 13C natural abundance to identify the EM–saprotrophic divide in fungal sporocarps and to validate the EM origin of mycelia in the ingrowth bags.
• Fungal biomass in soil and EM mycelial production increased with fertilization at both tundra sites, and with warming at one site. This was caused partly by increased dominance of EM plants and partly by stimulation of EM mycelial growth.
• We conclude that cycling of carbon and nitrogen through EM fungi will increase when strongly nutrient-limited arctic ecosystems are exposed to a warmer and more nutrient-rich environment. This has potential consequences for below-ground litter quality and quantity, and for accumulation of organic matter in arctic soils.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Annual mean temperature over Arctic land is expected to increase by 3–5°C within the next century if anthropogenic use of fossil fuel continues to increase concentrations of greenhouse gases in the atmosphere (IPCC, 2001; ACIA, 2005). A warmer climate is projected to shift vegetation zones northwards and promote expansion of forests into the arctic tundra, and tundra into the polar deserts (ACIA, 2005). In Alaskan tundra, increasing shrub abundance during the most recent decades has already been recorded by comparing historical and modern aerial photographs (Sturm et al., 2001) and by remote sensing (Stow et al., 2004). Long-term studies of local plots and field experiments with warming also suggest increased production and biomass, especially of woody shrubs, with increasing temperature (Chapin et al., 1995; Bret-Harte et al., 2001; van Wijk et al., 2004). Shrubs and trees decrease surface reflectivity because they are generally darker and more textured than tundra vegetation, and because they cover highly reflective snow and thus create a positive feedback to local and global warming (ACIA, 2005).
The shrubs most responsive to warming in arctic tundra, Salix spp. and Betula nana, form mutualistic symbioses with ectomycorrhizal (EM) fungi, which typically ensheath most of their root tips with a thick mantle of fungal hyphae and form an extensive external mycelium in the soil (Michelsen et al., 1996b; Smith & Read, 1997). The EM symbioses, depending on the specific EM fungal community, facilitate plant access to inorganic and organic nutrient pools, and are probably the main pathway of nutrient and water uptake for plants in natural ecosystems (Leake & Read, 1997; Smith & Read, 1997). Plant production in most arctic tundra ecosystems is limited by nitrogen availability, as most N is bound in recalcitrant forms in soil organic matter (SOM) accumulated over thousands of years because low temperatures and water saturation limit decomposition and mineralization rates (Chapin & Shaver, 1996; Nadelhoffer et al., 1997). Thus the observed increase in shrub abundance with enhanced temperature is probably closely linked to increased N availability and uptake by plants (Nadelhoffer et al., 1997; Shaver & Jonasson, 2001). However, despite their well established functional significance and their omnipresence at the key position at the plant–soil interface, mycorrhizal symbioses have been widely neglected in studies of arctic plant and ecosystem response to climatic change.
The external mycorrhizal mycelium probably represents the most dynamic component of the mycorrhizal symbiosis experiencing wide variation in environment, and probably responds directly to temperature (Staddon et al., 2002) and to nutrient availability (Wallenda & Kottke, 1998). However, only a few studies have focused on environmental effects on external EM mycelia in natural ecosystems (but cf. Nilsson & Wallander, 2003; Nilsson et al., 2005). This is probably because of a lack of methods to extract and distinguish EM mycelia from other mycelia in soils. Studies of C and N stable isotope abundance of sporocarp communities in forest ecosystems have identified a divide between EM and SAP fungi linked to their trophic status (Högberg et al., 1999; Henn & Chapela, 2001; Hobbie et al., 2001; Taylor et al., 2003; Trudell et al., 2004). EM fungi are generally depleted in 13C and enriched in 15N relative to SAP fungal sporocarps. A difference in δ13C of the principal C sources of the two groups of fungi is believed to cause the difference in their δ13C values (Högberg et al., 1999), whereas the 15N enrichment of EM relative to SAP fungal tissues is hypothesized to be related to fractionations of N isotopes within the EM mycelium, and preferential transfer of 15N depleted compounds to the plant (Högberg et al., 1996). Wallander et al. (2001) used this isotopic EM–SAP divide to verify the EM origin of mycelia extracted from sand-filled ingrowth bags after incubation in forest soil, and thus provided a method of estimating the production of EM fungi in soil.
The main objective of the present work was to determine whether predicted increased summer temperature and nutrient availability altered EM fungal abundance in plant roots and soil. We used ergosterol as a biomarker for living fungal biomass; EM mycelial production was measured by ingrowth bags in the field with EM fungi identified by 15N and 13C natural abundance. A second objective was to examine the extent to which these effects of climate change on mycorrhizas were direct or mediated through (proportional to) responses in the abundance of host plants. We hypothesized that both summer warming and increased nutrient availability altered EM fungal abundance in roots and soil, and that this, at least partly, was mediated through altered host plant density.
Materials and Methods
This study took place during the summers of 2002 and 2003 in two arctic tundra ecosystems, a heath tundra located near Abisko in subarctic northern Sweden close to the forest line at 450 m above sea level (68°38′ N, 20°51′ E); and a tussock tundra located well above the forest line at the south side of Toolik Lake in the northern foothills of Brooks Range in arctic Alaska (68°38′ N, 149°34′ W, 760 m above sea level). The Toolik Lake site is part of the US Long-Term Ecological Research network. Previous work has shown contrasting treatment responses of growth and tissue chemistry of the EM shrub Betula nana L. between the sites (Bret-Harte et al., 2001; Graglia et al., 2001; van Wijk et al., 2004).
The heath tundra site is situated on a west-north-west-facing slope and has a well drained, compact organic layer 10–15 cm deep and a pH of 7.1 (Schmidt et al., 2002). The organic layer rests on rocky mineral soil without permafrost. The vegetation is dominated by ericoid dwarf shrubs and mosses, with deciduous shrubs and dwarf shrubs as subdominants (Havström et al., 1993; Jonasson et al., 1993, 1999). The ectomycorrhizal shrub B. nana forms part of the up to 40 cm tall overstorey and makes up approx. 6% of the above-ground vascular plant biomass, while the other EM shrubs, Salix spp. and Dryas octopetala L., and the EM herb Polygonum viviparum L., make up approx. 3% of the above-ground biomass. The tussock tundra site is situated on a gentle north-facing slope, has a moist, loose 20–30-cm deep organic soil horizon with a pH of 5.0 (Schmidt et al., 2002), and rests on mineral soil with continuous permafrost. The site is dominated equally by graminoids (mainly the tussock-forming Eriophorum vaginatum L.), evergreen and deciduous shrubs and dwarf shrubs, and mosses (Shaver & Chapin, 1991; Chapin et al., 1995; Chapin & Shaver, 1996; Boelman et al., 2005; Hobbie et al., 2005b). Betula nana makes up approx. 20% of the above-ground vascular plant biomass, and other EM species (Salix spp.) less than 1%. The mean annual precipitation is approx. 300 mm at the heath tundra and 350 mm at the tussock tundra, of which about one-third and half, respectively, falls as summer precipitation at the two sites. For more details on both sites see Schmidt et al. (2002).
At both sites, field experiments were started in 1989, 14 growing seasons before sampling, and have been maintained since. Our study included four treatments: warming, fertilization, warming combined with fertilization, and untreated control plots, all replicated across six blocks at the heath and four blocks at the tussock tundra, and randomized within each block. Each year from early June to late August or early September, passive ecosystem warming was accomplished by raising plastic greenhouses over the vegetation. The greenhouses at the heath were 50-cm-high, dome-shaped constructions with a basal area of 1.2 × 1.2 m. They were made of spectrally neutral 0.05 mm thick polyethylene supported by PVC tubes, and had 40 × 40-cm air vents in the top and a 5–10-cm gap above the ground on two sides to allow free passage of precipitation and water vapour, and to reduce temperature extremes. Mean air temperature increased by 2.8°C, from 10.9 to 13.7°C; and soil temperature at 4 cm depth by 0.4–0.6°C, from 6.7–7.1 to 7.1–7.7°C (Jonasson et al., 1993; Michelsen et al., 1996a). At the tussock tundra, greenhouses were made of 0.15-mm-thick polyethylene fixed on permanent wooden frames with a 2.5 × 5-m surface area and up to 1.5 m in height. Uneven microtopography allowed air circulation beneath the bases of the greenhouses, similarly to those at the heath tundra. Similar greenhouses nearby increased the growing season mean air temperature by 3.5°C, from 11.2 to 14.7°C; and soil temperature by 2.2°C, from 3.6 to 5.8°C, at 10 cm depth (Chapin et al., 1995). Measurements in August, 9 yr after initiation of the experiment, showed treatment effects of increased depth of soil thaw in warmed plots and decreased thaw depth after fertilizer addition. In general, the thaw reached into the mineral soil in all treatments, allowing plant roots to penetrate the entire horizon of organic soil (Bret-Harte et al., 2001). The greenhouses reduced photosynthetically active radiation by approx. 10% at the heath and 20% at the tussock tundra, whereas soil-water content at both sites was largely unaffected after 2–5 yr treatment (Chapin et al., 1995; Michelsen et al., 1996a). The reduction of light could possibly have affected plant C assimilation negatively, but at both sites earlier shading experiments showed that plant biomass and production are not very responsive to 50–64% reductions in light input (Chapin et al., 1995; Jonasson et al., 1999; van Wijk et al., 2004).
At both sites, fertilizer was added once each year just after snowmelt. At the heath tundra, fertilizer solution of NH4NO3, KH2PO4 and KCl was added at a rate of 10.0, 2.6 and 9.0 g m−2 yr−1 of N, P and K, respectively (in 1989 4.9, 1.3 and 6.0 g m−2 yr−1 of N, P and K, respectively), except in 1993, 1998 and 1999, when sampling took place. At the tussock tundra, fertilizer was added as pellets of NH4NO3 (10 g N m−2 yr−1) and P2O5 (5 g P m−2 yr−1) dissolving over days to weeks at the soil surface. At the tussock tundra we included a shorter-term fertilizer treatment started in 1996, 6 yr before sampling, with the same fertilizer addition rate.
Sampling, preparation and analyses
In August 2002 we randomly harvested three 3.8-cm-diameter cores from each of the replicate plots and divided them into horizons: litter; O1 (0–5 cm depth); O2 (5 to approx. 10 cm depth); and O3 (a very dense organic horizon mixed with mineral soil, approx. >10 cm depth). The O3 horizon was not consistently present in the tussock tundra, and the horizon from 5 cm depth to the mineral soil is presented as O2. Roots of ectomycorrhizal plant species (mainly B. nana and, at the heath tundra, Salix spp.) were carefully separated from the soil and split into fine roots (FR, <0.5 mm diameter) and coarser roots. FR represented the nonwoody part of the root system. The three samples of root free soil or roots from the same horizon within a plot were pooled, frozen at −20°C and later freeze-dried and weighed. Soil moisture was determined gravimetrically, and soil organic matter content was measured as loss on ignition after 6 h at 550°C. We collected fungal sporocarps throughout July and August 2002 at both sites, and throughout August 2003 at the heath tundra site. Half of each sporocarp was dried at 40°C and used for identification (see Appendix Table A1 for full species names). Of the other half, the cap centre in fleshy species and all parts in smaller species were rinsed, dried at 40°C and ground with mortar and pestle.
Fungal ingrowth bags (8 × 4 × 1.5 cm) were constructed of nylon mesh (50 µm mesh size; Sintab Produkt AB, Sweden) and filled with 80 g acid-washed sea sand (Wallander et al., 2001). The mesh bags were placed at a depth of 4–8 cm at an angle of 45° and incubated from 16 June to 23 August 2002 (68 d) at the heath tundra and 2 July to 17 August 2002 (46 d) at the tussock tundra. Another set of mesh bags were incubated for a year and recovered on 20 June 2003 at the heath and 1 July 2003 at the tussock tundra. An aliquot of sand was frozen at −20°C and freeze-dried, whereas the main part of the sand was extracted in water, and mycelium floating in the water phase was collected on a nylon mesh and freeze-dried (Wallander et al., 2001). Sand and organic particles contaminating samples under bag recovery were carefully removed under a dissection microscope.
Ergosterol was used as a biomarker for living fungal biomass in roots and soils, and was determined on HPLC as by Nylund & Wallander (1992). Dry sand (5 g) or ground soil (0.2–0.5 g) or FR (20–30 mg) were saponified in 4 ml 10% KOH in methanol, sonicated for 15 min, and heated for 90 min at 70°C. After cooling, 1 ml H2O and 3 ml cyclohexane were added, and samples were vortex-mixed for 30 s. Samples were centrifuged for 5 min at 1000 g (for sand, 3 min at 250 g) and the upper cyclohexane phase, containing neutral lipids, was collected. Then 2 ml cyclohexane was added, samples were mixed and centrifuged once more, and the upper phase was collected and combined with previous wash-out. Samples were dried under air flow at 40°C, dissolved in methanol and filtered through a 0.45 µm Teflon syringe filter (Millex LCR-4, Millipore, Billerica, USA) and injected into an HPLC (Waters 600, Waters Corporation, Milford, MA, USA). The chromatographic system consisted of a C18 reverse-phase column (Nova-Pak, 3.9 × 150 mm, Waters) preceded by a C18 reverse-phase guard column (Nova-Pak, 3.9 × 22 mm). Extracts were eluted with methanol at a flow rate of 1 ml min−1 and ergosterol was quantified with a UV detector (Waters 996) at 282 nm. A conversion factor of 3 µg ergosterol per mg fungal biomass was used (Salmanowicz & Nylund, 1988). To convert fungal mass estimated per g sand to g m−2, we used the volumetric mass of the sand used (1.54 g cm−3) to calculate fungal mass per cm3, then extrapolated this amount to the upper 10 cm of the organic soil (Wallander et al., 2004).
Total C and N, and 13C/12C and 15N/14N ratios, in mycelia and sporocarps were used to identify the EM–SAP fungal divide and analysed on an Isoprime isotope ratio mass spectrometer (Micromass-GV Instruments, Manchester, UK) coupled to a Eurovector CN elemental analyser (Milan, Italy) using continuous flow, at the Institute of Biology, University of Copenhagen. Natural abundance of isotopes is expressed in the δ-notation relative to international standards: Vienna Pee Dee Belemnite for C; atmospheric N2 for N: δXsample (‰) = 1000 ×[(Rsample/Rstandard) − 1] where R is the molar ratio of heavyX : lightX. Samples were analysed with reference gas calibrated against international standards IAEA C5, CH6, CH7, N1, N2 and USGS 25, 26, 32, and drift corrected using internal standards of leaf material calibrated with these standards. The standard deviation of isotopic measurements of the standards used was ±0.2‰ for δ15N and ±0.1‰ for δ13C.
Statistical analyses were performed with the SAS 8.0 system (2001). anovas were used with type III sums of squares in the GLM procedure and α = 0.05. Tendencies towards significance (0.05 < P < 0.10) are also reported. For each site, treatment effects were analysed by four-factor anovas with warming (T), fertilization (F), horizon (H) and block as main factors, including the H × T, H × F, T × F and H × T × F interactions. In analyses of ergosterol in ingrowth bags, incubation time (Time) was used as the main factor instead of H. For tussock tundra data, the 6-yr fertilization treatment was not included in the statistical analyses. Treatment effects within each horizon were analysed using three-way anovas with T, F and block as factors, and including the T × F interaction. Significant anova results were evaluated using Tukey's HSD multiple comparisons of means test. Differences between sites were tested using t-tests for control plots. Before analyses, variables were tested for homogeneity of variances (Levene's test), and necessary log or root transformations were made. Pearson correlations were used to evaluate linear relations between response variables across treatments.
Fungal biomass in soil
The ergosterol concentration in control plots at the heath was about twice as high as at the tussock tundra (P = 0.0220, t-test; Fig. 1). At both sites the concentration of fungal biomass decreased strongly down the organic profile, and fertilization increased the concentration. This increase took place mainly in the litter and upper organic layer (O1), as seen by the significant interaction between horizon and fertilization in the anovas. At the heath, this led to increased total fungal biomass per unit area in fertilized soils, while there was no effect of fertilization on total fungal biomass at the tussock tundra because of a nonsignificant decrease in total SOM content in fertilized soils (Fig. 2). Warming increased both ergosterol concentration and total fungal biomass at the tussock tundra, but not at the heath (Fig. 1, 2).
At the tussock tundra, litter layer mass was significantly higher in fertilized than in unfertilized plots (P < 0.0001); litter mass was c. 6 and 20 times higher in 6-yr (286 ± 77 g m−2) and 14-yr (945 ± 126 g m−2) fertilized plots, and 20 times higher in fertilized plus warmed plots (982 ± 174 g m−2), than in control plots (48 ± 19 g m−2). Warming doubled litter layer mass at the heath (P = 0.0444, all warmed plots, 454 ± 114 g m−2; all ambient temperature plots, 226 ± 63 g m−2). At the time of sampling, the moisture content of the O1 soil horizon in control plots was higher at the tussock (492 ± 73% of DW) than at the heath tundra (110 ± 8% of DW, P = 0.0029). At the tussock tundra, warming significantly decreased the moisture content of the litter layer (P = 0.0034, all warmed, 150 ± 35%; all ambient temperature, 285 ± 40% of DW), whereas no treatment effects were found in deeper soil layers. At the heath tundra, the moisture content of the litter layer was approx. 90% of DW in all treatments, whereas warming led to drier O1 (P = 0.0415, all warmed, 72 ± 7%; all ambient temperature, 92 ± 7% of DW) and O2 (P = 0.0417, all warmed, 136 ± 10%; all ambient temperature, 162 ± 5% of DW) horizons.
EM–SAP fungal divide and production of external EM mycelium
At both sites, the dual isotope plot clearly separated EM and SAP sporocarps, and mean δ13C and δ15N values of EM and SAP sporocarps differed significantly (Fig. 3; Table 1). At both sites, δ13C in mycelia extracted from ingrowth bags was similar to δ13C in EM, but differed strongly and significantly from that of SAP sporocarps (Table 1), strongly indicating that the mycelia were ectomycorrhizal (Wallander et al., 2001). δ15N in mycelia at the tussock tundra was lower than EM, and similar to SAP sporocarp δ15N (Table 1). In all three types of fungal material at both sites, mean C concentration was 38–42%. Mean N concentration of SAP sporocarps, however, was two and three times higher than that of mycelia and EM sporocarps at the tussock and heath tundra, respectively, leading to significantly lower C : N ratios of SAP sporocarps (Table 1). When these comparisons were made using sporocarp species means as replicates, the same picture emerged, except that δ15N values of the three types of fungal material were not significantly different at the heath.
Table 1. Comparison of δ13C, δ15N, %N and C : N ratio of ectomycorrhizal (EM) and saprotrophic (SAP) sporocarps and of mycelia extracted from ingrowth bags at heath tundra in subarctic Sweden and at tussock tundra in arctic Alaska
Values are means (± 1 SE).
Replicates (n) represent single sporocarp or mycelial collections, the latter across treatments. Results of anova and Tukey's HSD test are presented. In each row, means that do not share letters are significantly different (α= 0.05).
C : N ratio
C : N ratio
Based on the assumption that the mycelia from ingrowth bags were ectomycorrhizal, fertilization increased growing-season (June–August) production of external EM mycelium by 80% at the heath and by 110 and 180% at the tussock tundra after 6 and 14 yr, respectively (Fig. 4). Warming increased production of mycelium by approx. 50% at the tussock tundra, but had no effect at the heath (Fig. 4). The growth rates of mycelium into the bags during the growing season were similar at the two sites (approx. 0.35 ng ergosterol per g sand per day). At both sites, mycelial mass in the ingrowth bags incubated for one growing season or 1 yr was similar, although at the heath the amount of mycelium in ingrowth bags tended to be lower after incubation for 1 yr.
Fine root characteristics of ectomycorrhizal plants
The fine root biomass of EM plants was not significantly affected by the treatments at the heath, whereas at the tussock tundra, fertilization increased FR biomass (Table 2). Also, at the tussock tundra fertilization increased EM plant leaf and total above-ground biomass more than FR biomass, resulting in a significant decrease in FR : leaf and FR : above-ground biomass ratios (Table 2). At the heath, warming decreased the FR ratios, but only in plots with simultaneous fertilization. Nitrogen concentration in fine roots increased with fertilization at the tussock tundra, and with warming at the heath tundra (Table 2). At the heath tundra, EM colonization of fine roots (µg ergosterol g−1 FR DW) and EM fungal biomass in fine roots (mg ergosterol m−2) was higher in O1 than in O2 (P < 0.05, data not shown), and warming tended to increase EM colonization (Table 2). At the tussock tundra there were no significant effects of treatments on EM colonization, whereas EM fungal biomass in fine roots increased with fertilization and tended to increase with warming (Table 2). Overall, although we found some treatment effects on EM colonization of fine roots at the heath tundra, FR biomass was the major control on EM fungal biomass in fine roots, as also seen by the strong correlation between FR biomass and FR fungal biomass at both sites (Table 3). In untreated plots, N concentration and EM fungal biomass in fine roots were higher at the tussock than at the heath tundra (P = 0.0045 and P = 0.0173, respectively), and FR biomass and EM colonization tended to be higher at the tussock than at the heath tundra (P < 0.10, t-tests). The ratio between EM mycelial production and EM fungal biomass in fine roots tended to be higher at the heath than at the tussock tundra (P = 0.0594), and the ratio tended to increase with simultaneous warming and fertilization at the heath (Table 2).
Table 2. Ectomycorrhizal (EM) plant fine root (FR, <0.5 mm diameter) characteristics in control (c), warmed (t), fertilized (f) and combined f and t plots after 14 yr treatment and in plots fertilized for 6 yr (f6) at heath tundra in subarctic Sweden and at tussock tundra in arctic Alaska
Significant main factor and interaction effects in anovas (T, temperature enhancement; F, fertilization) are indicated: ***, P < 0.001; **, P < 0.01; *, P < 0.05; †, P < 0.10.
EM plant total above-ground and leaf biomass data are from biomass harvests after 10 yr treatment at the heath tundra (S.J. and A.M., unpublished data) and after 6 and 14 yr treatment at the tussock tundra (G.R.S. and J. Laundre, unpublished data, http://ecosystems.mbl.edu/arc).
Table 3. Ectomycorrhizal (EM) fungal variables correlated with soil and EM plant variables at heath tundra in subarctic Sweden (H, n = 24) and at tussock tundra in arctic Alaska (T, n = 20)
EM mycelial production
FR EM colonization
FR EM fungal biomass
EM production : FR EM biomass
, P < 0.001;
, P < 0.01;
, P < 0.05;
, P < 0.10; ns, not significant; +, positive correlation; –, negative correlation; FR, fine roots.
EM plant total above-ground and leaf biomass data from biomass harvests after 10 yr treatment at heath tundra (S.J. and A.M., unpublished data) and after 6 and 14 yr treatment at tussock tundra (G.R.S. and J. Laundre, unpublished data, http://ecosystems.mbl.edu/arc).
Correlations with EM fungal parameters across treatments
At the tussock tundra, EM mycelial production was highly correlated with total above-ground and leaf biomass of EM plant species with r2 values of 0.72 and 0.61, respectively (Table 3). EM plant above-ground biomass also correlated with EM fungal biomass in fine roots because of the strong control of FR biomass on FR fungal biomass. At the heath tundra, we found no direct correlations between EM fungal and EM plant biomass parameters. However, with increasing EM plant above-ground biomass, the ratio of EM mycelial production to EM biomass in fine roots tended to increase: EM mycelial production tended to increase more than FR EM biomass (Table 3). In fine roots at both sites, N concentration was positively correlated with ectomycorrhizal colonization (r2 = 0.58 and r2 = 0.39 at the heath and tussock tundra, respectively; Table 3). Soil moisture did not correlate significantly with any of the EM fungal parameters.
Fungal biomass in arctic soils
The present study is the first attempt to estimate total living fungal biomass together with production of EM mycelia in organogenic arctic soils, and responses to long-term fertilization and warming. Fungal biomass, estimated with ergosterol as a biomarker, at our sites mainly includes SAP, EM and ericoid mycorrhizal fungi, as the ergosterol content of arbuscular mycorrhizal fungi is very low (Olsson et al., 2003). Fungal biomass in the complete soil column in control plots was 550 and 200 g m−2 at the heath and tussock tundra, respectively (Fig. 2), which is lower than fungal biomass estimated down to 70 cm depth in temperate forest soils (700–950 g m−2; Wallander et al., 2004). However, the concentration of fungal biomass in the upper organic horizon at our sites (Fig. 1) is very similar to that reported in the humus layer of forest soils (50–70 mg fungal biomass per g SOM; Wallander et al., 2004), and the main cause of higher fungal mass in forest soils is the large accumulated biomass in the mineral soil. The low fungal biomass in tussock tundra is caused mainly by lower biomass in the deepest horizon compared with the heath (Figs 1,2), which is probably linked to impeded drainage, high water content and low temperatures caused by the underlying permafrost. At the heath, the organic horizon overlays better drained, rocky mineral soil.
Our estimate of fungal biomass in the upper 10 cm of SOM in control plots (17 ± 3 and 21 ± 2 mg C g−1 SOM at the tussock and heath tundra, respectively, assuming 40% C in fungal biomass) is slightly lower (tussock tundra) or 2.3 times higher (heath tundra) than total microbial biomass C in August, estimated earlier by the chloroform fumigation method at the same two sites (tussock tundra, 19; heath, 9 mg C g−1 SOM; Schmidt et al., 2002). At the tussock tundra, the C : N ratio of the microbial biomass suggested fungal dominance of microbial biomass (Schmidt et al., 2002), which agrees with our study. At the heath tundra, the apparent discrepancy in biomass estimates could be caused by differences in extractability of microbial C with the fumigation–extraction method in different soils (Joergensen, 1996), or by uncertain conversion from ergosterol to fungal biomass. While mycelia extracted from ingrowth bags at the heath were dark brown and rigid (melanized), the mycelia from the tussock tundra were light coloured and fragile, indicating that EM mycelia at the heath were more recalcitrant than those at the tussock tundra to chloroform–fumigation. Fumigation appears to underestimate the fungal contribution to microbial biomass (Ingham et al., 1991). On the other hand, fungal ergosterol concentration varies with fungal species, and mean ergosterol concentrations in fungal populations in the soil may be different from the concentration used, obtained from cultured EM fungi (3 µg mg−1, Salmanowicz & Nylund, 1988). For instance, Montgomery et al. (2000) determined an average ergosterol concentration of 4 µg mg−1 mycelial biomass in six SAP fungal species isolated from soil, and Padgett & Posey (1993) reported 2.3 µg mg−1 in an isolate from the genus Hymemoscyphus, also including ericoid mycorrhizal species.
Fungal δ15N and δ13C patterns and production of external EM mycelium
Our study confirms for the first time that the general pattern of δ15N and δ13C in sporocarp communities emerging from studies in forest ecosystems (Högberg et al., 1999; Henn & Chapela, 2001; Hobbie et al., 2001; Taylor et al., 2003; Trudell et al., 2004) also applies to treeless tundra ecosystems. We identified a clear divide between EM and SAP sporocarps at both sites, caused by lower δ13C and higher δ15N in EM than in SAP sporocarps (Fig. 3; Table 1). The EM origin of mycelia in the sand-filled ingrowth bags was confirmed by the similar δ13C values of mycelia and EM sporocarps (Wallander et al., 2001). At both tundra sites, the fungal pattern of N concentration coincided with that of δ13C (Table 1). The lower N concentration in EM fungi may arise because they transfer part of the absorbed nutrients to their host in return for C, whereas SAP fungi may accumulate nutrients because their growth is primarily limited by available C. This difference in N accumulation, caused by different trophic strategies, could be accentuated in ecosystems with an inherently slow decomposition rate and nutrient cycling. The mean δ15N of extracted mycelia at the tussock tundra, however, was lower than EM sporocarp d15N. This is probably because EM fungal taxa differ in δ15N (Taylor et al., 2003) and different EM fungal species dominated mycelial and sporocarp communities. However, several other explanations are possible. As discussed by Hobbie et al. (2005a), sporocarps may generally be enriched in 15N relative to vegetative mycelia, because 15N-enriched protein is transported to the sporocarps, while relatively 15N-depleted chitin is retained in mycelia. Also, EM mycelia extracted from different depths in a soil profile may have different δ15N values, although they have the same δ13C (Wallander et al., 2004). This is probably because all EM mycelia have the same C source (symbiotic plants), but may exploit different N sources in the soil.
EM mycelial production in the upper 10 cm of control plots was estimated to be close to 1 g m−2 yr−1 at both sites (Fig. 4): approx. 0.5% of total fungal biomass in the soil. In the humus layer of two forest soils, standing EM fungal biomass was estimated to make up approx. 80% of total fungal biomass, and EM production was 10% of EM biomass; that is, it had an estimated turnover time of 10 yr (Wallander et al., 2004). Although EM mycelia probably contributed less to total fungal biomass in our soils because of lower EM plant dominance, the low production estimates suggest that EM mycelia had very long turnover times (>50 yr). However, the method may generally underestimate production of mycelia (Wallander et al., 2004), especially in arctic soils where fungal populations have limited access to mineral substrates and are adapted to exploit a highly organic soil matrix. The similar level of fungal mass in the mesh bags after one growing season and after 1 yr indicates that there had been little dieback of mycelium during winter. Ectomycorrhizal fungi have been shown to survive subzero temperatures, but the persistence of basidiomycete mycelium through periods of cold remains unknown (Robinson, 2001; Tibbett et al., 2002). Because the growing season at high latitudes is short, a perennial mycelium would be highly advantageous for perennial plants, and our results support the hypothesis that EM fungal mycelia in arctic ecosystems are long-lived (Tibbett et al., 2002).
Direct and plant-mediated responses of soil fungi
A number of recent studies in forest ecosystems have documented decreased fungal biomass and production of EM mycelium with increased N availability (Högberg et al., 2003; Nilsson & Wallander, 2003; Nilsson et al., 2005). Also in alpine ecosystems, N fertilization decreased soil microbial biomass (Schmidt et al., 2004). In contrast, we demonstrated increased soil fungal biomass and EM mycelial production after 14 yr fertilization in both of the sites studied and after 14 yr warming at the tussock tundra (Figs 1, 2, 4). However, although fungal biomass and production levels, and their treatment responses, were similar at the two sites, the mechanisms behind them were not the same, seen in very different magnitudes of EM plant responses at the two sites.
At the tussock tundra across treatments, EM mycelial production was strongly correlated with EM plant above-ground and leaf biomass (Table 3). This suggests that increased EM mycelial production with fertilization and warming was largely, if not entirely, a result of increased biomass production of EM plants. After 14 yr fertilization, above-ground biomass and production of EM plants had increased eight- and sevenfold, respectively, as B. nana outcompeted most of the other plant species (Hobbie et al., 2005b; G.R.S. and J. Laundre, unpublished data). This magnitude of change in host plant biomass and production has not been observed in any of the forest ecosystems studied. For instance, in a natural nutrient-availability gradient studied by Högberg et al. (2003) and Nilsson et al. (2005), production of trees increased threefold while production of mycelia decreased to one-third. Likewise, N fertilization of mature forests decreased mycelial production to approx. 50% of that in control areas (Nilsson & Wallander, 2003). Thus, although EM mycelial production was strongly correlated with EM plant above-ground biomass, the tripling in mycelial production that we observed with fertilization (Fig. 4) represents a status quo or even a decline relative to EM plant above-ground production. Also, above-ground biomass increased relatively more than fine root biomass and EM biomass in fine roots with fertilization (Table 2), consistent with the general pattern of decreased C allocation to roots and mycorrhizal fungi relative to shoots when plants are released from nutrient limitation (Högberg et al., 2003).
At the heath tundra, in contrast, EM mycelial production was not correlated with EM plant biomass (Fig. 4; Table 3), which suggests a more direct stimulation of mycelial growth in the fertilized treatment. We observed a doubling in EM mycelial production with fertilization, whereas there was no significant increase in above-ground biomass of EM plants at the heath after 10 yr fertilization (van Wijk et al., 2004; S.J. and A.M., unpublished data). Soil inorganic N availability in the fertilization treatment changed much less at the heath than at the tussock tundra, probably partly because of better drainage at the heath (Jonasson et al., 1999). This, together with increased vegetation biomass and plant nutrient pools, indicates that vegetation and associated mycorrhizal fungi still constituted a strong nutrient sink at the heath, whereas large accumulation of inorganic N in the soil at the tussock tundra suggests that this site was becoming nutrient saturated (Shaver & Jonasson, 1999). Decreased mycelial growth after fertilization in forest soils is believed to be caused by fungal intolerance to high levels of inorganic N in their surroundings (Wallenda & Kottke, 1998). Thus the difference in relative EM mycelial responses at the two sites was probably because nutrient limitation persisted at the heath tundra, whereas inorganic nutrient levels were critically high for EM mycelial growth at the tussock tundra. Changes in the composition of EM fungal communities may also have played a role in the different relative responses at the two sites. Fertilizer-induced reductions in mycelial growth in forest ecosystems are partly explained by increased abundance of EM fungal species with short or no external mycelia radiating from root tips at the expense of species forming extensive mycelia (Wallenda & Kottke, 1998; Nilsson & Wallander, 2003). Changes in EM fungal communities in response to fertilization have also been found in arctic ecosystems (Urcelay et al., 2003; Clemmensen & Michelsen, 2006). Such changes can affect nutrient and C turnover, because fungal species differ in their abilities to attack different organic substrates (Leake & Read, 1997). We are currently investigating possible site and treatment effects on EM fungal communities associated with B. nana. Despite the lack of correlation between EM mycelial production and EM plant biomass at the heath, the positive fertilization effect on EM mycelial production may yet be partly plant-mediated. Growth of mycorrhizal mycelium depends on current photosynthate from the host plant (Smith & Read, 1997; Olsrud et al., 2004), and high production of mycelia could be supported by an increased rate of C fixation after fertilization, although not manifested in increased plant biomass production.
In comparison with the effects of fertilization, responses to warming were more subtle and, at the tussock tundra, coincided with increased EM plant biomass, similarly to responses to fertilization. At the heath, warming did not change fungal biomass or EM mycelial production. However, warming decreased fine root biomass relative to above-ground biomass in ectomycorrhizal plants and tended to increase EM colonization of fine roots and, in combination with fertilization, the ratio of EM mycelial production to fine root biomass (Table 2). This suggests a trade-off between C allocation to fine root and EM fungal biomass production, and that warming shifted C allocation in favour of EM fungi.
Consequences of increased EM fungal abundance
Fungal material probably is a precursor of stable soil organic matter and recalcitrant N (He et al., 1988; Wedin et al., 1995; Högberg et al., 1996; Aber et al., 1998; Langley & Hungate, 2003). Our study suggests that an increased amount of N is cycled through and accumulated in EM fungal biomass as a consequence of long-term fertilization at both arctic tundra sites, and of warming at the tussock tundra. We hypothesize that this has led to increased input of fungal derived N and C with relatively long turnover time to the SOM pool, and thus has counteracted possible positive feedback to CO2 concentration in the atmosphere, observed by Mack et al. (2004). They found that long-term nutrient fertilization reduced bulk soil C storage in arctic tussock tundra with vegetation responses similar to those observed here, and only approx. 1 km away, on the East shore of Toolik Lake. This result is consistent with the nonsignificant decrease in SOM mass we observed at the tussock tundra site (Fig. 2). A study by Neff et al. (2002), however, showed that N additions accelerated decomposition of light soil C fractions while further stabilizing soil C compounds in heavier fractions. Thus C fractions other than those derived from fungi may have disappeared in the fertilized tussock tundra soils. Also, we found large increases in litter layer mass with fertilization at the tussock tundra, and with warming at both sites. This suggests that the main pools of soil C and N were changing, and that soils in future shrub-dominated arctic ecosystems may become more similar to forest floors with a continuous litter layer underlain by an organic horizon, where more complex organic molecules form the major long-term reservoir of soil C and N.
As concluded by Leake et al. (2002), the overall effect of EM fungi on the N cycle is to increase the supply of N to EM plants, while reducing the supply of both N and C to saprotrophs and to potential competitor plants lacking EM symbioses. Thus increased EM fungal abundance probably interacted with long-term vegetation responses as a consequence of improved competitive abilities of EM plants (mainly B. nana and Salix spp.). At the tussock tundra, where B. nana dominated vegetation and litter production after fertilization, the high fungal biomass in the litter layer (approx. 8% of litter mass) and the similar responses to fertilization of fungal biomass and EM mycelial growth, together with the strong correlation between FR EM biomass and N concentration, point to increased internal cycling of N between EM plants and fungal partners after fertilization. Such a positive feedback loop between increased B. nana C fixation and litter production, and increased EM fungal growth and return of N to B. nana, may have facilitated increasing dominance of B. nana at the tussock tundra.
This is the first study focusing on long-term responses of soil fungal biomass and EM mycelial production to fertilization and warming in ecosystems, without trees as the dominant EM plant type. Our key findings were as follows. (i) Fungal biomass in soil and EM mycelial production increased with fertilization in both arctic tundra ecosystems, in contrast to responses reported from forest ecosystems. This was partly caused by increased biomass of EM plants and partly by stimulation of fungal growth by the treatment. (ii) The ratio of EM mycelial production to EM biomass was low, suggesting that the mycelia are perennial and long-lived. (iii) Our study supports the generality of δ15N and δ13C sporocarp patterns, and confirms the presence of an EM–SAP divide in two arctic tundra ecosystems. At both sites, N concentration was higher in SAP sporocarps than in EM sporocarps and mycelia. (iv) Future work should focus on functional aspects of external mycelia of mycorrhizal fungi, such as exo-enzyme production and decomposability, in order to unravel their impact on SOM accumulation in arctic ecosystems.
We are much indebted to the staff at the Abisko Scientific Research Station (ANS) and at Toolik Field Station for logistical support. We thank H. Wallander for introduction to the ergosterol assay, H. Knudsen, J. Vesterholt and T. Borgen for identification of fungal sporocarps, and L. Christiansen, E. Stieve, Y. Yano, E. Stuckenbrock, R. Clemmensen and K. Føns for their assistance in the field and lab. Financial support for this research was provided by grants from ANS, The Fiedler Foundation, the Danish Natural Sciences Research Council, and the University of Copenhagen. The experimental plots at Toolik Lake are maintained by the NSF Arctic Long-term Ecological Research project (NSF grants DEB-9810222, DEB-0423385 and related grants).
Table A1. Elemental concentration of C and N, C : N ratio and natural abundance of 13C and 15N isotopes of fungal sporocarps collected at a heath tundra in subarctic Sweden and a tussock tundra in arctic Alaska