Effects of mineral and nutrient input on mire bio-geochemistry in two geographical regions

Authors

  • Luca Bragazza,

    Corresponding author
    1. Dipartimento delle Risorse Naturali e Culturali, Università di Ferrara, Corso Porta Mare 2, I-44100 Ferrara, Italy, and
      Luca Bragazza (fax +39 0532208561, e-mail luca.bragazza@unife.it).
    Search for more papers by this author
  • Renato Gerdol,

    1. Dipartimento delle Risorse Naturali e Culturali, Università di Ferrara, Corso Porta Mare 2, I-44100 Ferrara, Italy, and
    Search for more papers by this author
  • Håkan Rydin

    1. Department of Plant Ecology, Evolutionary Biology Centre, Uppsala University, Villavägen 14, SE-752 36 Uppsala, Sweden
    Search for more papers by this author

Luca Bragazza (fax +39 0532208561, e-mail luca.bragazza@unife.it).

Summary

1 We assessed the role of climatic conditions and the effects of different, long-term atmospheric depositions in controlling the mineral and nutrient contents in pore-water, surface peat and in living Sphagna at a boreo-nemoral mire in Sweden and an alpine mire in Italy.

2 The terrestrial contribution of Ca2+, Mg2+ and inline image in bulk precipitation was much greater at the Italian mire, in accordance with the different bedrock in the region and the higher level of atmospheric pollution.

3 At both mires, the contribution of bulk precipitation to the concentration of major ions in mire pore-water was much greater in the ombrotrophic than in the minerotrophic part, because of the raised morphology of the mires, which limited the inflow of mineral soil water to the margins. The only ions strongly depleted in mire pore-water compared with precipitation were K+, inline image and inline image and these were therefore limiting to plant growth.

4 Higher inline image concentration in pore-water at the Swedish mire, which experienced lower atmospheric inputs of sulphate, was probably caused by oxidative processes during a long dry period in the summer before sampling.

5 Higher rates of inline image, inline image, as well as inline image atmospheric inputs at the Italian mire were reflected in significantly higher N and, partly, S concentrations in ombrotrophic Sphagna. Higher inline image concentration in pore-water at the Italian mire was associated with a lower N retention coefficient of the ombrotrophic Sphagnum plants, suggesting a reduced nitrogen filtering ability of the moss layer.

Introduction

Ombrotrophic mires, or bogs, are wetland ecosystems characterized by a dense cover of Sphagnum mosses, sedges and ericaceous shrubs forming a variable microtopography of hummocks, lawns, carpets, hollows and pools. From a biogeochemical point of view, bogs are oligotrophic, i.e. their interstitial water has a low nutrient and mineral content associated with very low pH values, generally < 4.5 (Gorham et al. 1985; Bridgham et al. 1996). Such bio-geochemical conditions are the result of a ‘perched’ water table insulated from the surrounding mineral soil (Bridgham et al. 1996). As a consequence, the bog is exclusively fed by wet and dry atmospheric depositions and for this reason bogs are defined as ‘ombrotrophic’. The ‘lagg’ is a narrow fen in the border between bog and mineral soil. It receives water both from the bog and from the mineral soil and is minerotrophic in terms of pore-water and peat chemistry (Rydin et al. 1999). Ombrotrophic conditions favour plants adapted to acidic conditions and low-nutrient availability, such as Sphagnum mosses that form the bulk of the dead and living biomass in these ecosystems (Charman 2002).

The presence, development and differentiation of bogs are regulated by factors such as climate, hydrology and progressive peat accumulation (Charman 2002). In particular, moisture surplus and length of the growing season are important climatic parameters explaining the distribution and the corresponding surface landforms of bogs, which are therefore mainly found in humid cold-temperate climates. Ombrotrophic mires are widespread in North America, the British Isles, Fennoscandia, European Russia and central Siberia (Charman 2002), but they are very localized in central and southern Europe. In the Italian Alps, bogs are more frequently found in the upper subalpine belt (i.e. between 1600 and 2000 m a.s.l.) where moisture regime and length of the growing season favour peat accumulation (Bragazza & Gerdol 2003).

The nutrient dynamics in mires have been studied thoroughly (see e.g. Charman 2002), but very few papers have performed large scale comparisons of major nutrient dynamics in peat and pore-water under natural conditions (e.g. Verhoeven et al. 1996). In view of the changes in precipitation chemistry that have occurred, especially in recent years (see Vitousek 1994), we can assume that processes related to nutrient and mineral cycling can be modified in bogs subjected to different atmospheric depositions, with possible effects on the stability of ombrotrophic mires. Indeed, ombrotrophic bogs, receiving their mineral and nutrient inputs solely from atmospheric deposition, appear particularly sensitive to atmospheric changes with respect not only to climatic modifications, but also to alterations of precipitation chemistry (Gorham 1991; Bobbink et al. 1998). For example, several reports have shown that increased deposition rates of nitrogen (N) compounds can alter the competitive equilibria between mire plant species (Berendse et al. 2001), the decomposition rates of organic matter (Aerts et al. 2001) and the emission rates of carbon dioxide and methane (Aerts & de Caluwe 1999).

In contrast to most of the studies dealing with the effects of altered atmospheric depositions, which rely on short-term experimental manipulations, we compared the geo-chemical conditions of two geographically contrasting mires (a boreo-nemoral mire in Sweden and an alpine mire in Italy) subjected to natural, long-term differences in atmospheric inputs. Specifically, our objective was to assess how precipitation chemistry and climate can account for: (a) chemical composition of mire pore-water; (b) surface peat chemistry; and (c) mineral and nutrient contents in ombrotrophic Sphagnum plants.

Materials and methods

study areas

The study was carried out at a mire in central-eastern Sweden and a mire in the south-eastern Alps of Italy. The Swedish mire, Ryggmossen (60°3′ N, 17°20′ E), is situated 25 km north-west of Uppsala (province of Uppland) at an elevation of c. 60 m a.s.l. The mire is c. 60 ha and has a roughly circular shape. The open central part is tree-less with a concentric pattern of hummocks and hollows. The hummocks are mainly dominated by Sphagnum fuscum (Schimp.) Klinggr. together with Calluna vulgaris (L.) Hull and other ericaceous species, whereas the hollows are mainly dominated by Sphagnum balticum (Russ.) C. Jens. and, locally, Sphagnum tenellum (Brid.) Brid and Eriophorum vaginatum L. In the wettest hollows Sphagnum cuspidatum Hoffm., Scheuchzeria palustris L. and Carex limosa L. also are frequent. A detailed description of Ryggmossen can be found in Du Rietz & Nannfeldt (1925).

The Italian mire, Wölfl Moor (46°26′ N, 11°24′ E), is located in South Tyrol, province of Bozen (Bolzano), at an elevation of c. 1300 m a.s.l. and is c. 8 ha. The vegetation of Wölfl Moor consists of a complex pattern of hummocks, lawns and carpets. The hummocks are dominated by Sphagnum fuscum, Sphagnum capillifolium (Ehrh.) Hedw., Vaccinium spp. and Calluna vulgaris; the lawns and carpets are characterized by Sphagnum rubellum Wils., Sphagnum magellanicum Brid., Sphagnum papillosum Lindb., Eriophorum vaginatum and Trichophorum caespitosum (L.) Hartman. A detailed description of the mire can be found in Alber et al. (1996).

Both mires presented a convex profile and originated from the in-filling of a lake-basin, as demonstrated by the gradual increase in peat thickness from the margin towards the mire centre, with maximum peat depths of c. 4.5 m at Ryggmossen (Du Rietz & Nannfeldt 1925) and c. 6 m at Wölfl Moor (L. Bragazza, unpublished observations). The bedrock consists of igneous rocks at both sites.

field work

The field work was performed from June to September 1998 at Ryggmossen and from June to October 1999 at Wölfl Moor. At each mire two transects were established from the margin to the inner portion. Along each transect we located five sampling plots corresponding to the main habitat types. Each plot was uniform in terms of plant composition and surface peat structure. At each plot a perforated PVC pipe was inserted into the peat down to 0.5 m depth from the level of the Sphagnum capitula. Depth to the water table was measured twice in all pipes to represent wet and dry conditions during the study period. At the peak of the growing season the percentage cover of all plant species was estimated at each sampling plot in a 25 × 25 cm quadrat.

Pore-water samples were extracted from each pipe with a syringe once during wet and once during dry conditions. Water pH and electrical conductivity were measured in the field by portable instruments. Electrical conductivity was corrected for the effect of hydrogen ions (Sjörs 1952). An aliquot of pore-water was put into acid-washed polythene bottles and tubes for subsequent chemical analyses. The pore-water samples for cation analyses were acidified and stored in a refrigerator, whereas the pore-water samples for nitrate, sulphate and chlorine analyses were deep frozen until analysis. Ammonium and reactive phosphorus concentrations were determined within 2 days from collection. All water samples were filtered through a 0.45-µm mixed cellulose ester MFS membrane filter.

At the end of the studying period, a 10-cm long surface peat sample was collected at each plot using a stainless steel cylindric corer (inner diameter 6.6 cm). After removing the top of living plant material, the peat samples were air dried and then ground prior to analyses. Peat subsamples were dried at 105 °C for 24 h to allow conversion of elemental data to a standard 105 °C oven dry weight.

Sphagnum plants were harvested in eight 10 × 10 cm quadrats in the ombrotrophic central part at the two mires. At Ryggmossen, two quadrats were dominated by Sphagnum fuscum and the others by Sphagnum balticum, whereas Sphagnum fuscum and Sphagnum capillifolium were the dominant species at Wölfl Moor. Immediately after collection, all Sphagnum plants were separated into the capitulum (top centimetre) and an adjacent 2-cm long segment. All Sphagnum segments were dried at 40 °C for 48 h and ground prior to analyses.

chemical analyses

The pore-water samples were analysed for alkalinity, cations and silica (SiO2), inorganic nutrients and acid anion concentrations. Cations were Na+, K+, Ca2+, Mg2+, Mn2+, Fe3+ and Al3+. Inorganic nutrients were nitrate (inline image), ammonium (inline image) and soluble reactive phosphorus (inline image), whereas the acid anions were chloride (Cl) and sulphate (inline image). Alkalinity was determined by Gran titration (Gran 1952). Cations and silica were analysed by atomic absorption spectrophotometry (Allen 1989). Chloride, sulphate and nitrate were analysed by high-performance liquid chromatography (HPLC). Ammonium was determined by the phenate method (Wagner 1969). Soluble reactive phosphorus was determined by the molybdenum blue method (Allen 1989).

Mineral elements in surface peat, i.e. Ca, Mg, Na, K, Al, Fe and Mn, were analysed by atomic absorption spectrophotometry after digesting the peat samples with 65% nitric acid in a microwave oven. Total phosphorus (P) was determined colourimetrically by the molybdenum blue method on the same solution used for the mineral cation analyses (Allen 1989). Total nitrogen (N) was determined by the Kjeldahl method (Allen 1989).

Mineral elements in Sphagnum segments, i.e. Ca, Mg, Na and K, were analysed by atomic absorption spectrophotometry after digesting the segments with 65% nitric acid in a microwave oven. Total P was determined colourimetrically by the molybdenum blue method on the same solution used for the mineral cation analyses (Allen 1989). Nitrogen and sulphur (S) concentrations were determined by an elemental analyser (EA 1110, Carlo Erba). The N and S data are the averages of the two Sphagnum segments.

data analyses

For the Italian site, both long-term climatic data (1981–98) and climatic data for 1999 for Deutschnofen (4 km SE Wölfl Moor) were provided by the Hydrographic Office (Hydrographisches Amt) of the Bozen Province. For the Swedish site, long-term climatic data (1961–90) for Uppsala were obtained from SMHI (1998), whereas climatic data for the 1998 growing season at Ryda Kungsgård (32 km SSW Ryggmossen) were provided by the Swedish Environmental Research Institute. The number of growing days was calculated as the sum of all days having mean daily temperature > 5 °C, based on the records obtained for the 3 years before sampling.

Data on precipitation chemistry were drawn from records obtained at the closest meteorological stations during the 3 years preceding the sampling summers (Tait & Thaler 2000, and data provided by the Swedish Environmental Research Institute-IVL).

The matrix containing the cover values of all plant species in the 10 sampling plots at each site was subjected to a detrended canonical correspondence analyses (DCCA) (ter Braak & Smilauer 1998). The environmental variables selected were: mean water table position, pH, K+, Ca2+, SiO2 and inline image in pore-water and Ca, N and P in surface peat.

Differences in pore-water and surface peat chemistry among habitats were tested by one-way anova and Scheffé post hoc tests. Differences in precipitation chemistry and elemental concentrations in Sphagnum plants between the Italian and the Swedish areas were tested by t-tests.

The presence of extra sources of major ions in precipitation, as well as in mire pore-water, beyond the contribution of seawater (marine) and bulk precipitation, respectively, was assessed: (a) assuming that all Cl in rainwater and in mire pore-water was of marine origin, and (b) using Cl as a conservative element in mire pore-water (Appelo & Postma 1994). Ion concentrations in seawater were drawn from Stumm & Morgan (1981).

Results

climate

Mean annual temperature was similar in the two areas (Table 1). The Swedish site had lower annual precipitation than the Italian one, but precipitation during the study year was much higher than the long-term average at Ryggmossen. The growing season during the study year was c. 3 weeks shorter in the Swedish mire compared with the Italian mire.

Table 1.  Topographic and climatic data for the two mires. Climatic values in the study year are in parentheses (1999 and 1998 for the Italian and the Swedish mire, respectively). Growing season is the number of days with mean temperature > 5 °C
 Elevation (m a.s.l.)Bog surface (ha)Mean annual precipitation (mm)Mean annual temperature (°C)Growing season (days)
Wölfl Moor (Italy)1290 8  808 (919)  + 6.3 (+ 6.4)(212)
Ryggmossen (Sweden)  6060  554 (744) + 5.6 (+ 6.2)(189)

precipitation chemistry

Precipitation chemistry at the Italian mire was characterized by higher concentrations of Ca2+, Mg2+, K+, inline image, inline image and inline image compared with the Swedish mire (Table 2). Yearly mean pH of bulk precipitation was significantly higher at the Italian mire (pH = 5.19, SD = 0.52) compared with the Swedish mire (pH = 4.64, SD = 0.12).

Table 2.  Mean concentration of major ions (SD in parenthesis) in bulk precipitation at Wölfl Moor (Italy) and Ryggmossen (Sweden) during the 3 years before sampling. Data from Tait & Thaler (2000) and Swedish Environmental Research Institute-IVL. Percentage of seawater contribution to the concentration of each ion in precipitation was calculated multiplying the Cl concentration in precipitation by the relative concentration of the ion to chloride in seawater
 Ca2+ (µeq L−1)Mg2+ (µeq L−1)Na+ (µeq L−1)K+ (µeq L−1)inline image (µeq L−1)inline image (µeq L−1)inline image (µeq L−1)Cl (µeq L−1)
Wölfl Moor (Italy)16.7 (2.5)7.0 (1.0)6.4 (1.1)9.6 (1.5)27.7 (4.0)38.8 (5.4)29.7 (4.9)8.0 (1.4)
Seawater contribution1.8%24%98% 1.5%0%0%2.8%100%
Ryggmossen (Sweden)10.8 (4.5)4.4 (1.5)12.4 (4.6)5.4 (3.1)5.8 (1.4)17.9 (5.0)10.8 (2.2)13.1 (7.1)
Seawater contribution4.5%63%90%4.4%0%0%12%100%

The Na+ concentration in precipitation at both Ryggmossen and Wölfl Moor was comparable to that expected assuming marine input alone (Table 2). Magnesium was primarily derived from seawater at Ryggmossen, whereas at Wölfl Moor most of the Mg2+ was from a terrestrial input (Table 2). Ca2+, K+ and inline image showed very small percentages of marine contribution at both mires, particularly Wölfl Moor, whereas no seawater contribution was expected for inline image and inline image, based on their concentrations in precipitation (Table 2).

On the basis of annual amount of precipitation over the last 3 years (data not shown) it was possible to detect significantly higher depositions of Ca2+, Mg2+, K+, inline image, inline image and inline image at the Italian mire compared with the Swedish mire (P < 0.01).

habitat definition

Two main groups of plots were obtained by canonical ordination (Fig. 1). The first ordination axis showed a strong positive correlation with pH, Ca2+ and SiO2 in pore-water and with Ca and N in surface peat (Table 3). Hence, the first ordination axis clearly separates the minerotrophic lagg fens from the ombrotrophic parts of the mires (Fig. 1). The second ordination axis was related to water table position (Table 3). Ombrotrophic plots are located over a wide range along the gradient in depth to the water table forming different microhabitats such as hummocks, lawns and carpets. The minerotrophic plots at the mire margins showed a lower microtopographical variability and relatively high water table (Fig. 1).

Figure 1.

Detrended canonical correspondence (DCCA) scores of sampling plots along transects A (A1–A5) and B (B1–B5), from the margin to the inner portion, at the Italian mire (Wölfl Moor) and the Swedish mire (Ryggmossen). Note the different symbols indicating the sampling plots from the two mires. The vectors represent the direction of variation of environmental variables in pore-water (w) and surface peat (p). WT: water table position.

Table 3.  Eigenvalues and interset correlations of pore-water and peat variables with the first three ordination axes as obtained by detrended canonical correspondence analyses (DCCA)
 Axis 1Axis 2Axis 3
Eigenvalue0.79  0.57   0.34
Pore water variables
 Water table position0.22−0.53−0.09
 pH0.90−0.16 0.02
 K+0.44 0.45 0.24
 Ca2+0.84 0.15−0.19
 SiO20.90 0.08 0.19
 inline image0.43−0.11 0.09
Peat variables
 Ca0.87 0.06−0.17
 N0.88 0.12−0.16
 P0.63 0.33−0.03

surface peat chemistry

The minerotrophic parts of both mires had higher concentrations of Ca, Al, Fe, N and P than the corresponding ombrotrophic parts, significantly so except for P at Ryggmossen (Table 4). In addition, the Ca/Mg quotients in the surface peat from the minerotrophic lagg fens were significantly higher than those in local precipitation (3.3 and 3.7 at Ryggmossen and Wölfl Moor, respectively), but not in the ombrotrophic parts (Table 4). The N/P quotient in surface peat had similar values in the fen and the bog at each mire, but was significantly higher in the ombrotrophic part of the Italian than of the Swedish mire (Table 4).

Table 4.  Mean values (SD in parenthesis) of surface peat variables in the minerotrophic and ombrotrophic parts of the Italian and Swedish mire (see Fig. 1). The means followed by the same letter within a row are not significantly different at P < 0.05 based on anova and Scheffé tests. Element concentrations in peat are as mg kg−1 dry mass. (n is the number of sampling plots)
 Italian minerotrophic part (n = 4)Swedish minerotrophic part (n = 2)Italian ombrotrophic part (n = 6)Swedish ombrotrophic part (n = 8)
Ca 4698a (1641) 5416a (208)1952b (577)1308b (555)
Mg  685a (190)  821a (146) 542a (127) 461a (162)
Na  158a (40)  182a (111) 221a (140) 190a (88)
Al 1706b (511) 5069a (952) 708c (23) 226d (91)
Fe 3595a (1992) 4277a (447)1209b (356) 513b (90)
Mn   31a (21)   96a (65)  23a (30)  68a (53)
N17350a (4361)15750a (3890)5583b (2472)5200b (1475)
P  702a (305)  708a (151) 233b (122) 396ab (119)
K  586ab (280) 1332a (375) 562b (279) 680ab (322)
Ca/Mg    6.8a (1.2)    6.7a (1.5)   3.6b (0.4)   3.1b (1.5)
N/P   27a (10)   22ab (1)  25a (6)  14b (5)

mire pore-water chemistry

In both mires, the minerotrophic parts had significantly higher values of alkalinity, pH, and Ca2+, Mg2+, Na+, Al3+, Fe3+, Mn2+ and SiO2 concentrations in pore-water than the corresponding ombrotrophic parts (Table 5). At Wölfl Moor the minerotrophic part also had higher inline image concentrations in pore-water compared with the ombrotrophic part (Table 5).

Table 5.  Mean values (SD in parenthesis) of pore-water variables in the minerotrophic and ombrotrophic parts of the Italian and Swedish mire. Means followed by the same letter within a row are not significantly different at P < 0.05 based on anova and Scheffé tests. Electrical conductivity (El. Cond.) is expressed as µS cm−1, alkalinity as meq L−1, silica as mg L−1, and charged ions as µeq L−1. (n is the number of sampling plots)
 Italian minerotrophic part (n = 4)Swedish minerotrophic part (n = 2)Italian ombrotrophic part (n = 6)Swedish ombrotrophic part (n = 8)
El. Cond. 48.2b (27.5) 76.0a (21.6) 37.9b (8.3) 42.1b (24.3)
Alkalinity  0.48a (0.21)  0.67a (0.61)  0.04b (0.02)  0.0b (0.0)
pH  5.45a (0.46)  5.04ab (0.28)  4.64b (0.41)  3.73c (0.15)
Ca2+195a (78)255a (1.8) 50b (44) 70b (34)
Mg2+ 66a (16) 82a (8) 39b (21) 18c (7)
Na+112a (82)112ab (12) 56bc (39) 66c (9)
K+ 17a (9)  7b (2) 15a (11)  7b (3)
Al3+ 88ab (66)154a (33) 33c (11) 55bc (44)
Fe3+ 53b (86)210a (130) 23b (10) 34b (32)
Mn2+  0.7a (0.3)  0.8a (0.1)  0.2b (0.1)  0.3b (0.2)
SiO2 12.4a (2.8) 15.4a (2.4)  7.1b (2.4)  5.8b (1.9)
inline image132b (25)273a (86)108b (2)300a (139)
Cl 35a (12) 36a (8) 31a (17) 34a (3)
inline image  1.3a (0.5)  0.4b (0.2)  1.0a (0.3)  0.4b (0.2)
inline image  3.2a (1.6)  3.5a (2.2)  1.2b (1.0)  1.9ab (1.1)
inline image  0.29a (0.32)  0.10a (0.10)  0.32a (0.13)  0.29a (0.64)

Pore-water in the minerotrophic part at the Italian mire differed significantly from that in the minerotrophic part at the Swedish mire by having lower inline image concentrations and higher inline image and K+ concentrations. On the other hand, pore-water in the ombrotrophic part at the Italian mire had higher pH, Mg2+, K+ and inline image concentrations, but lower inline image concentrations compared with the ombrotrophic part at the Swedish mire (Table 5).

Concentrations of major ions in mire pore-water showed a seawater contribution decreasing in the order: Na+ > Mg2+ > K+ > inline image > Ca2+ > inline image = inline image (Table 6). The marine contribution was much more effective in pore-water from the ombrotrophic parts than the minerotrophic parts, with the exception of K+ and inline image, which showed the same percentage of seawater contribution for both parts of both mires (Table 6).

Table 6.  Percentage of contribution from seawater (Seawater contr.) and bulk precipitation (Precipitation contr.) to the concentrations of major ions in mire pore-water at the Italian mire and the Swedish mire. Contributions from seawater and bulk precipitation were obtained multiplying the Cl concentration in mire pore-water by the relative concentration of the ion to Cl in seawater and precipitation, respectively. Figures > 100% were due to ionic concentrations in mire pore-water lower than the expected contribution from precipitation, indicating a depletion of the ion compared with precipitation
 Italian minerotrophic part (n = 4)Swedish minerotrophic part (n = 2)Italian ombrotrophic part (n = 6)Swedish ombrotrophic part (n = 8)
Ca2+Seawater contr.    0.7%  0.5%    2%   2%
Precipitation contr.   37%  12%  129%  40%
Mg2+Seawater contr.   11%   9%   17%  40%
Precipitation contr.   46%  15%   69%  64%
Na+Seawater contr.   26%  27%   47%  34%
Precipitation contr.   25%  30%   44%  37%
K+Seawater contr.    4%   9%    4%   9%
Precipitation contr.  247% 212%  248% 194%
inline imageSeawater contr.    3%   1%    3%   1%
Precipitation contr.   98%  10%  106%   9%
inline imageSeawater contr.    0%   0%    0%   0%
Precipitation contr. 9322%3984%10850%3740%
inline imageSeawater contr.    0%   0%    0%   0%
Precipitation contr.53050%1400%12400%2505%

Contribution of bulk precipitation to the concentration of major ions in mire pore-water showed a more complex pattern than seawater contribution. The contribution of bulk precipitation to Mg2+ and Na+ concentrations in mire pore-water was generally greater for the ombrotrophic than the minerotrophic parts at both mires (Table 6). The contribution of bulk precipitation to Ca2+ concentration in mire pore-water was greater for the ombrotrophic part only at the Swedish mire, as Ca2+ at ombrotrophic parts of Wölfl Moor was depleted compared with bulk precipitation (Table 6). K+, inline image and inline image were strongly depleted in mire pore-water relative to precipitation at both mires (Table 6). The contribution of bulk precipitation to inline image concentrations in mire water at Wölfl Moor was much greater than at Ryggmossen, indicating a major role of extra sources at the Swedish mire (Table 6).

mineral and nutrient contents in ombrotrophicsphagnumplants

Sphagnum plants from the ombrotrophic parts exhibited significantly higher concentrations of Ca, Mg, K and N at the Italian than at the Swedish mire (Table 7). S and P concentrations did not differ significantly between the two sites. The N/P quotient in ombrotrophic Sphagnum plants was significantly higher at the Italian mire (Table 7).

Table 7.  Mean concentrations (SD in parenthesis) of selected elements in Sphagnum plants from the ombrotrophic part at the Italian mire and the Swedish mire. Retention coefficient (Ret. coeff.) for each element was calculated as the ratio between the concentration of the element in Sphagnum plant and in precipitation. Significance of differences for element concentrations between the two mires were assessed by t-test. (n is the number of Sphagnum sampling plots)
 Italian ombrotrophic Sphagna (n = 8)Swedish ombrotrophic Sphagna (n = 8)t-valueP
Ca (mg g−1)  1.89 (0.26) 0.74 (0.5)   5.600.00
 Ret. coeff.  5.7 3.4  
Mg (mg g−1)  0.54 (0.10) 0.36 (0.03)   4.650.00
 Ret. coeff.  6.7 6.7  
Na (mg g−1)  0.46 (0.15) 0.54 (0.31)−0.720.48
 Ret. coeff.  3.1 1.9  
K (mg g−1)  4.26 (0.62) 3.49 (0.81)   2.120.05
 Ret. coeff. 11.516.6  
S (mg g−1)  0.52 (0.20) 0.37 (0.11)   1.620.12
 Ret. coeff.  0.4 0.7  
N (mg g−1) 10.0 (2.2) 6.9 (1.9)   2.90.01
 Ret. coeff.  4.210.1  
P (mg g−1)  0.33 (0.08) 0.35 (0.06)   0.770.45
 Ret. coeff.n.a.n.a.  
N/P 30 (3.7)20 (3.8)   3.920.00

Retention coefficients, calculated as ratio between the Sphagnum content of the element and the corresponding concentration in precipitation, decreased in the order K > Mg > Ca > N > Na > S at the Italian mire and in the order K > N > Mg > Ca > Na > S at the Swedish mire (Table 7).

Discussion

The separation between the ombrotrophic and minerotrophic parts of the two mires suggested by ordination analysis is mirrored by significant differences in pore-water and surface peat chemical variables. We found pH, alkalinity, Ca2+, Mg2+, Al3+ and SiO2 concentrations in pore-water to be significantly higher in the minerotrophic parts (Shotyk 1988). In addition, there were clear differences in surface peat content of elements such as Ca, Al, Fe, N, P and the Ca/Mg quotient (see for example Shotyk 1988, 1996). Wheeler & Proctor (2000) have argued that the primary distinction along the poor–rich gradient is between rich fen (high pH) and poor fen (more similar to bog in pH and Sphagnum dominance), rather than between minerotrophy and ombrotrophy. As our transects did not include any part with high pH, our data have no bearing on this discussion. Therefore, we retain the terms ‘fen’ and ‘bog’ for indicating minerotrophic and ombrotrophic conditions, respectively (Sjörs & Gunnarsson 2002).

sources of ions in precipitation and mire pore-water

The chemical composition of precipitation showed the seawater contribution to Ca2+, Mg2+, inline image and K+ concentrations to be much lower at the Italian mire than at the Swedish mire. The excess of Ca2+ and Mg2+ at Wölfl Moor can be accounted for by a higher role played by external terrestrial inputs, primarily related to (i) the geographical position, which probably receives more alkaline dust from the surrounding carbonate mountains, and (ii) the inflow of air masses from the Sahara rich in alkaline dusts (Prodi & Fea 1979; Tait & Thaler 2000). The effect of alkaline dust was also mirrored by the higher rainfall pH at Wölfl Moor. The higher excess of inline image in precipitation over the Italian mire relative to seawater must be related to higher levels of pollution at Wölfl Moor. Indeed, the Italian study area is exposed to air masses coming from the Po plain (Camuffo et al. 1991; Della Lucia et al. 1996), the most densely populated and industrialized area in Italy. The contribution of seawater to K+ concentration in precipitation was rather low both at the Swedish and the Italian mire, and the K+ present was probably mainly due to local inputs, probably agricultural activities (Proctor 1992).

All major ions showed substantial excess in mire pore-water compared with the concentrations expected from marine input, indicating a very low or absent sea-spray influence at both mires. When mire pore-water concentrations are compared with the relative composition of local precipitation it is possible to detect the contribution of precipitation to ion concentration in mire pore-water. An increase or decrease in ion concentration relative to precipitation can be due to: (a) terrestrial dust washed down by rain or wind-borne particles not detected by the funnels used for collecting bulk precipitation; (b) soil groundwater inflows rich in mineral solutes; or (c) release or adsorption by peat. The lower contribution of precipitation to Ca2+, Mg2+ and Na+ concentrations in mire pore-water in the minerotrophic parts compared with the ombrotrophic parts can be explained as a primary effect of mineral soil water inflows along the mire margins. Bulk precipitation did not, in fact, totally explain the cation contents in pore-water on the ombrotrophic parts, which are probably due to the mire surface trapping droplets and solid particles more effectively than the funnels used for collecting bulk precipitation (see Oldfield et al. 1979; Schauffer et al. 1996). Calcium was depleted in ombrotrophic pore-water at Wölfl Moor compared with precipitation very probably because of the greater cation exchange activity of living Sphagna, as indicated by the higher Ca concentration in ombrotrophic Sphagnum plants at the Italian mire (see below).

K+, inline image and inline image were the only ions showing a strong depletion in mire pore-water compared with precipitation at both mires, as a consequence of the rapid absorption of these nutrients by living plants (Proctor 1994; Vitt et al. 1995; Bragazza et al. 1998).

Lower sulphate concentrations in pore-water were found at the Italian mire compared with the Swedish mire, even though the Italian mire experienced significantly higher sulphate deposition rates and precipitation contributed most of inline image concentration into mire pore-water. Generally, high levels of sulphate depositions result in higher inline image concentration in bog waters (Gorham et al. 1985; Proctor 1992). The role of air pollution in controlling sulphate concentrations in bog water is not, however, always detectable (Adamson et al. 2001), as the sulphate supplied by rainfall is rapidly taken up by plants (Urban & Bayley 1986), reduced by soil microbes (Bottrell & Novak 1997) and incorporated into organic and inorganic compounds (Urban et al. 1989; Steinmann & Shotyk 1997). Short-term, extreme climatic events may have temporarily masked the effect of atmospheric deposition on pore-water chemistry. Sulphate is particularly enriched in water during dry periods when a low water table favours the oxidation of the H2S formed under anaerobic conditions (Bayley et al. 1986; Urban et al. 1989; Proctor 1994; Mandernack et al. 2000). Such acid pulses can result in high sulphate and H+ concentrations in pore-water for several months before concentrations return to the previous values (De Vito & Hill 1999; Adamson et al. 2001). At Ryggmossen, the year before sampling (1997) was considerably drier than 1998 (annual precipitation of 443 vs. 744 mm, Swedish Environmental Research Institute). Therefore, it is likely that the inline image formed during the dry summer of 1997 was then washed out, yielding a higher inline image concentration in pore-water during the sampling period (1998). Lower pore-water pH at Ryggmossen can be explained by the higher sulphate concentration (cf. Proctor 1994), whereas higher pore-water pH at Wölfl Moor could also be due to local alkaline dust inputs exerting a buffering effect on the water pH, as happens for local precipitation.

effect of climate and atmospheric depositions on element concentrations in ombrotrophic living sphagna

The ombrotrophic part, being fed exclusively by atmospheric deposition, appears particularly suitable to investigate the effects of precipitation chemistry on the mineral and nutrient content in ombrotrophic living Sphagna.

Significantly higher inline image concentrations in pore-water at Wölfl Moor could be accounted for by higher rates of atmospheric nitrogen deposition at the Italian site associated with lower nitrogen retention by the moss cover. In pristine bogs, N is a growth limiting element that is rapidly assimilated and effectively retained by plants and microorganisms (Urban & Bayley 1986; Urban et al. 1988). However, under increasing atmospheric nitrogen deposition Sphagnum plants tend to become more and more nitrogen saturated and to lose their ability to bind airborne nitrogen (Lamers et al. 2000). In light of the preferential absorption of nitrate by Sphagnum species compared with vascular plants (Lee & Stewart 1978), more nitrate is expected to pass through the Sphagnum carpet, thus increasing nitrate availability in pore-water, as nitrate deposition increases (Press et al. 1986; Bobbink et al. 1998; Heijmans et al. 2001). The proposed critical load for atmospheric N deposition in bogs is c. 10 kg ha−1 year−1, with a predicted strong increase of N availability to the rhizosphere for depositional levels > 20 kg ha−1 year−1 (Bobbink & Roelofs 1995; Lamers et al. 2000). The Italian mire, which is exposed to air masses coming from the Po plain, is subjected to an annual N input above the critical load. On the other hand, the Swedish mire receives an annual N input of c. 4 kg ha−1 year−1, i.e. intermediate between the depositional values in south Sweden, c. 7–9 kg ha−1 year−1 and those in north Sweden, c. 0.6–2 kg ha−1 year−1 (Malmer 1988), but in every case lower than the proposed critical load. If the main reason for the observed differences in pore-water nitrate concentrations is the lowered filtering ability of the ombrotrophic Sphagnum cover, we would expect a high N content in Sphagnum plants at the Italian bog. Accordingly, the Sphagnum plants at Wölfl Moor presented a total N concentration of c. 10 mg g−1, i.e. above the proposed concentration threshold for N-saturated Sphagna (c. 9 mg g−1, Lamers et al. 2000). Moreover, the retention coefficient of Sphagnum plants at the Italian mire is much lower than at the Swedish mire. Hence, it appears that the Sphagnum cover at the Italian mire has reached, or is close to, N saturation, causing a leaching through the moss carpet, whereas at the Swedish mire the Sphagnum plants are no longer N limited, as indicated by a N/P quotient < 14 (Aerts et al. 1992), but the Sphagnum layer does not yet appear N saturated (Press et al. 1986; Berendse et al. 2001).

Although the Italian mire is subjected to higher atmospheric depositions of K, Sphagnum plants there did not present higher K concentrations than at the Swedish mire, probably because of the active translocation of this cation to the growing capitula (Malmer 1988).

The input of alkaline dust from the surrounding Dolomites is probably the main reason for higher Ca and Mg concentrations in Sphagnum plants at Wölfl Moor, in accordance with the strong correlation between Ca and Mg depositions and the corresponding Ca and Mg concentrations in Sphagnum plants reported by different authors (e.g. Pakarinen 1981; Malmer 1988; Malmer et al. 1992). Contrary to K, Ca and Mg are mainly accumulated on cation exchange sites in mosses (Malmer 1988) and their accumulation pattern mirrors that of the cumulative growth (Gerdol 1990). The linear growth of Sphagnum fuscum at Ryggmossen during the study year was c. 6 mm year−1 (Bragazza, unpublished). No growth data are available for Wölfl Moor, but Sphagnum capillifolium (a hummock-forming species taxonomically similar to S. fuscum) at a bog close to Wölfl Moor showed a mean annual elongation of c. 23–24 mm (Gerdol 1995). Linear growth is generally positively correlated with dry mass production (Gerdol 1995) and a higher Sphagnum elongation at the Italian mire is consistent with a longer growing season. Therefore, under higher productivity and with a higher external supply of these bivalent cations, the accumulation capacity of the Sphagnum plants is expected to lead to higher Ca and Mg concentrations in moss tissues (Gerdol 1990; Malmer et al. 1992).

Higher atmospheric sulphate depositions are supposed to be reflected in the S content of ombrotrophic Sphagnum plants (Malmer 1988; Novak et al. 2001). As expected given the higher levels of sulphate depositions, ombrotrophic Sphagnum plants at the Italian mire presented greater S content compared with the Swedish mire, reflecting, once more, the different chemistry of precipitation over the two mires.

Acknowledgements

We thank B. Schipperges and I. Targa for field assistance. Some of the chemical analyses were performed at the Biology Laboratory of the Province of Bozen, at the Department of Chemistry of the University of Ferrara, and at the Department of Limnology of Uppsala University. For assistance in chemical analyses, thanks are due to R. Marchesini, R. Alber and A. Marchi. A. Svensson, at the Swedish Environmental Research Institute, kindly provided data on precipitation chemistry. The research carried out in Sweden was supported by a grant to L.B. by the University of Padua (Italy). Financial support was also obtained from the Swedish Natural Science Research Council and Ferrara University. Finally, we thank M.C.F. Proctor for helpful suggestions.

Ancillary

Advertisement