Author for correspondence: Matt Robson Tel: +1 435 7972555 Fax: +1 435 7973872 Email: email@example.com
• Tierra del Fuego is subject to increases in solar UV-B radiation in the austral spring and summer due to ozone depletion.
• Plastic films were used to filter solar UV-B radiation over peatland plots through six field seasons, resulting in near-ambient (c. 90%) and reduced (c. 17%) solar UV-B treatments.
• As in the first three field seasons of treatments, near-ambient UV-B caused reduced height growth but had no effect on biomass production of the moss Sphagnum magellanicum. It reduced leaf and rhizome growth of Tetroncium magellanicum. Height growth and morphology of Empetrum rubrum and Nothofagus antarctica were only affected by solar UV-B during the fourth to sixth field seasons. There was also a decrease in Tetroncium leaf nitrogen under near-ambient UV-B.
• Growth of Sphagnum was less affected than that of most emergent vascular plants. This enabled the Sphagnum mat to engulf more Nothofagus, and limit the escape of Empetrum under near-ambient UV-B. Yet, differences in the response of species to solar UV-B were not expressed as changes in plant community composition.
A stratospheric ‘ozone hole’ has formed above Antarctica during the austral spring (September–November) every year since the late 1970s (Farman et al., 1985). This depletion of the ozone layer often extends beyond Antarctica to Tierra del Fuego, along with the associated increased ultraviolet-B radiation (UV-B) (Frederick et al., 1994; Díaz et al., 2000).
In Antarctica, solar UV-B was shown to reduce the rate of vegetative growth and leaf expansion of a native grass, Deschampsia antarctica Desv., and a forb, Colobanthus quitensis (Kunth) Bartl., compared with plants under filters that blocked much of the UV-B (Ruhland & Day, 2000; Day et al., 2001; Xiong et al., 2002). In short experiments with mosses, solar UV radiation had no effect on Bryum argenteum Hedw. (Green et al., 2000) and Sanionia uncinata (Hedw.) Loeske (Montiel et al., 1999; Lud et al., 2003). However, significant changes in pigmentation due to short-term UV-B fluctuations have been detected in Antarctic mosses (Newsham et al., 2002).
Tierra del Fuego, situated at the most southerly tip of South America, experiences the greatest increase in UV-B radiation as a consequence of ozone depletion of any region outside Antarctica (Díaz et al., 1996; Cede et al., 2002). This is most pronounced during October, when the ‘ozone hole’ can sometimes pass directly over Tierra del Fuego. Consequently, short-term increases in UV-B radiation of up to 50% from day to day can occur at Ushuaia, Argentina (55° S) (Díaz et al., 2001). Also, the subsequent break-up of the Antarctic vortex (November–early December) allows pockets of ozone-depleted air to pass over this region later in the growing season.
High latitudes have historically received little UV-B radiation (Caldwell et al., 1980). Thus, current decreases in the thickness of the ozone layer cause large relative increases in UV-B flux, although compared to lower latitudes, especially the tropics, total UV-B flux remains small. Since organisms have evolved under low background UV-B flux, they may be sensitive to a large relative increase in UV-B radiation. The location of Tierra del Fuego presents an opportunity to study more diverse plant communities than occur in Antarctica and yet still experience the effects of increased solar UV-B radiation (Díaz et al., 2000, 2001).
Interspersed among mature Nothofagus forest stands, along stream and lake margins, peatlands form an important component of the landscape of southern Tierra del Fuego (Roig, 2000). The composition of the Tierra del Fuego peatlands is comparable to those of the Northern Hemisphere; Sphagnum magellanicum is a widespread peatland species (Daniels & Eddy, 1985), whilst Empetrum rubrum is a dominant vascular plant, similar to Empetrum hermaphroditum in the Northern Hemisphere.
The overall objective of the investigation was to follow changes in the peatland plant community through 6 years of solar UV-B manipulation. This paper concentrates on the last 3 years. After the first 3 years of UV-B manipulation in this peatland, Searles et al. (1999, 2002) observed a small reduction in Sphagnum height growth (c. 15%), and a decrease in size of the largest leaf of Tetroncium (c. 14% in November), but no growth responses of Empetrum or Nothofagus to the solar UV-B manipulations. Based on these findings, we continued with another 3 years of more intensive investigation into plant responses to solar UV-B treatments. We aimed to assess whether longer-term UV-B treatments would result in more conspicuous responses of the Sphagnum and emergent vascular plants to solar UV-B manipulations. Furthermore, we examined whether changes in growth and morphology of Sphagnum and the vascular plant species would be evident as shifts in species composition and density within the plant community.
Materials and Methods
The study site is a peatland of approximately 1500 × 500 m. It is located in Tierra del Fuego National Park (54°51′ S 68°36′ W) to the north of a small lake (Laguna Negra) and stream, otherwise surrounded by mature Nothofagus forest. The peatland derives all its water from precipitation, and has a pH of 4.5–6.5. The groundwater level drops to c. 40-cm depth when the peatland is driest in late summer. This peatland was chosen because it has relatively high plant diversity, homogeneously distributed over its area.
The peatland receives 500–600 mm precipitation annually. Precipitation over the 5-month growing season (October–February) was approximately 260 mm (1999–2000), 312 mm (2000–2001) and 269 mm (2001–2002) (T. M. Robson, unpublished data). The mean annual temperature is 5–6.5°C, reflecting the oceanic climate.
The peatlands of Tierra del Fuego have been well characterized by Mark et al. (1995) and Roig (2000). S. magellanicum (Brid.) forms an almost continuous mat, with occasional hummocks and small pools bordered by Sphagnum fimbriatum Wils. Of the emergent vascular plants, Nothofagus antarctica (Forster f) Oersted (Nothofagaceae) (southern beech) is most abundant along the forest border. This species usually grows as a small tree, but is severely stunted in the peatland, typically reaching only a few cm above the Sphagnum mat. Empetrum rubrum Vahl ex Willd (Empetraceae) (crowberry) is a creeping evergreen woody perennial, which dominates on the drier tops and sides of the hummocks; T. magellanicum Willd (Juncaginaceae) (arrow-rush) grows in the wetter hollows and depressions between hummocks. In addition, Nanodea muscosa Banks ex C. F. Garertner, a hemiparasite of Sphagnum peatlands, Juncus scheuchzerioides Gavdich, and Pernettya pumila (cf.) Hooker are found at lower density throughout the site (Moore, 1983). Most of the vascular plant species are visible as ramets emerging from the Sphagnum mat. Nearby ramets are often part of the same genet divided beneath the peatland surface.
Ten pairs of 1.4 × 2-m experimental plots were selected during October of 1996 in level and homogeneous areas of the Sphagnum mat (Searles et al., 1999). Near-ambient and reduced UV-B treatments were randomly assigned within the pairs of plots. Plastic-film filters were suspended horizontally from frames (c. 40 cm) above the plots to create the UV-B treatments (see photo, Ballaréet al., 2001). A uniform pattern of slits was melted into the filters producing a matrix of small louvres (2 × 25 mm) distributed over the entire filter that enabled water from precipitation to penetrate evenly to the Sphagnum mat. Also, we did not wish to totally exclude UV-B in the reduced UV-B treatment. The reduced UV-B treatment was achieved using perforated polyester filters (100-m thick, optically equivalent to ‘Mylar-D’, Dupont Co., Wilmington, Delaware, USA) that attenuate c. 83% of the short wavelength UV-B radiation and transmit most of the UV-A radiation (Searles et al., 2002). Near-ambient UV-B plots were covered by perforated polyfluorine filters (‘Aclar’ type 22A, 38-m thick, Honeywell, Pottsville, Pennsylvania, USA) that block c. 10% of the UV-B (Searles et al., 2002; Zaller et al., 2002).
The precipitation passing through the two filter types was found to be approximately equal. Photosythetically active radiation (PAR, total quantum flux in the 400–700 nm waveband) was measured with a quantum sensor (LiCor, Lincoln, Nebraska, USA) at the study site in Tierra del Fuego National Park. This showed that 83–95% PAR radiation was transmitted through both filter types (data not shown).
Filters were installed in late September to coincide with snow melt in the peatland, and remained in place until late March in each of the three field seasons (1999–2002), following the same protocol as Searles et al. (1999, 2002). Broken filters were replaced within a day, and a complete replacement of filters was made after three months, in late December. Filter attenuation remained unchanged over this time period (Searles et al., 2002).
For the duration of the experiment, under both of the UV-B treatments, air, surface and below-surface temperatures, precipitation, and UV-B radiation were monitored (21× datalogger, Campbell Scientific, Logan, Utah, USA) and compared with ambient values (data not shown).
Growth measurements were continued on Sphagnum, Empetrum, Tetroncium and Nothofagus, the same plant species studied during the first 3 years of the experiment (Searles et al., 2002). For continuity and comparability, growth was measured in the same way as during the first 3 years of the project. Because Sphagnum peatlands are known to exhibit considerable inherent microsite variability (Gerdol, 1996; Mitchell et al., 2000; Phoenix et al., 2003), and responses to UV-B in the first 3 years of this project were subtle (Ballaréet al., 2001; Searles et al., 2002), we increased the sample size to include 12 ramets of each species per plot.
Ten new colour-coded cranked wires (Clymo, 1970) were inserted into each plot in February of 1999 to allow time for the Sphagnum to recover from any associated disturbance (Searles et al., 2002) before the first measurements in September 1999. The vertical growth of the Sphagnum mat up these wires (and two existing wires) was measured four times through each growing season to determine seasonal growth patterns. Existing wires were used to compare height growth of Sphagnum under the filters and in the open, to determine the influence of the plot microclimate on growth during the fourth and fifth field seasons. These measurements showed that Sphagnum height growth was greater under the near-ambient UV-B filter than in the open (data not shown).
At the beginning and end of the sixth field season, Sphagnum capitulum density was measured nondestructively. The number of Sphagnum capitula in a 120 × 120-mm area was counted, and six randomly assigned counts per plot were taken. Sphagnum capitulum mass was sampled at the same time that density was censused and height growth was measured.
Annual Sphagnum biomass production and volumetric density of capitula were calculated from height growth, density, and biomass data. To calculate biomass increase per stem, a capitulum correction factor was used following Gehrke (1998). A random sample of 14 Sphagnum capitula were removed from each plot and cut into two 5-mm lengths, 0–5 mm and 5–10 mm from the apex (for calculation of the capitulum correction factor). Samples were oven dried at 65°C for at least 72 h and subsequently weighed to obtain dry mass.
Tetroncium leaf length (second to fifth leaves produced), and total number of green and senescent leaves were measured during November, December and towards the end of the growing season every year. In March 2002, 12 ramets per plot were harvested for measurement of rhizome growth. Tetroncium grows sympodially from a rhizome, branching from the leaf axil. Each leaf produced leaves a scar around the rhizome's circumference allowing retrospective growth measurements to be taken. Tightly packed leaf scars correspond to late season growth, allowing annual rhizome growth to be determined as in other rhizomous plants (Duarte et al., 1994; Moen & Walton, 1996). Total rhizome elongation over each of the last 7 years was measured, as was the distance between each of the most recent 13 leaf scars, to reveal the seasonal pattern of elongation over the year. Frequency of rhizome division and root production from the rhizome were also noted.
Twelve leaves per plot were harvested in February of 2001, air-dried, pooled, ground, and sealed in aluminium foil for analysis of their carbon and nitrogen composition. This analysis was carried out at Utah State University Analytical Laboratories, using a CHN1000 analyser (LECO Corporation, St Joseph, Michigan, USA).
Annual stem elongation of Empetrum was measured as the distance from an annual terminal bud scar on the stem to the shoot apex (Searles et al., 1999, 2002). Growth measurements of the same plants were taken once every field season in March, as was distance from the shoot apex to the Sphagnum mat. These plants had 3–4 years of growth above the Sphagnum mat when first selected in September 1999.
Annual terminal bud scars were used to carry out a retrospective analysis of Empetrum growth. Twelve ramets of various sizes were harvested from each plot every field season for more detailed measurements of the number of leaves, branching frequency, and total annual stem and branch growth (up to seven branches) following Shevtsova et al. (1997). A comparison was made between the last three field seasons (1999–2002), and the first three field seasons (1996–1999, ramets harvested in February 2000). The harvest of a range of ramets allowed the treatment effects on ramets differing in their years of growth above the Sphagnum mat to be assessed independently.
Nothofagus ramets with at least 3 years of stem growth visible above the Sphagnum mat were selected in September 1999. Stem growth and leaf production were measured twice during each field season. Expansion of the first and third true leaves produced was followed through each field season.
Terminal bud scars were used to determine the years of Nothofagus stem growth above the Sphagnum mat, retrospective annual stem extension, and branching frequency, in the same manner used with Empetrum. The vertical height of each stem above the Sphagnum mat was measured at the end of each season, and compared to the previous field season to calculate the rate of engulfment by Sphagnum.
Community composition was assessed by placing a quadrat over a 120 × 60-cm area in the centre of the plot. The area was divided into 50 equal squares. All the Nothofagus ramets and at least two-fifths of the more abundant Empetrum ramets in the quadrat were censused. The number of years of growth of each ramet above the Sphagnum mat was assessed using annual terminal bud scars. Ramets were recruited into ‘apparent age’ classes dependent on whether they were outpacing or being engulfed by the Sphagnum mat. This allowed changes in the ‘apparent age’ structure of the Nothofagus and Empetrum populations to be assessed. The number of Tetroncium individuals and the less abundant N. muscosa, J. scheuchzerioides and P. pumila were counted at the end of the fifth and sixth field seasons of the project.
To allow for hydrological gradients and heterogeneity of vegetation across the site, each pair of plots was considered as a block. However, where between-block variability was low, block was omitted from model.
The effect of UV-B treatment on cumulative stem growth, distance between leaf scars on the Tetroncium rhizome, morphological characteristics of Sphagnum and Empetrum, and Tetroncium leaf carbon and nitrogen, was assessed using an anova in a one-way factorial, randomized complete block design. The effect of UV-B treatment on Tetroncium leaf expansion was assessed using an anova in a one-way factorial, completely randomized design. The effects of UV-B treatment and year on annual stem and rhizome growth, Tetroncium and Nothofagus morphological characteristics, change in height growth, and plant population size were assessed using an anova in a two-way factorial, blocked split-plot-in-time design. Year and UV-B treatment were fixed-effects factors, and blocks (where used) were a random-effects factor. The effects of UV-B treatment and apparent age on annual stem growth of Empetrum were likewise assessed using an anova in a two-way factorial, blocked split-plot-in-time design. Tetroncium rhizome elongation data were Loge or square-root transformed as appropriate before analysis to better meet assumptions of normality and homogeneity of variance. Back-transformed means and standard errors are presented.
We tested various covariance matrix structures: based on the assumption that growth in the peatland at any sampling time was most strongly correlated with growth at the next sampling time, together with Schwarz Bayesian information statistics, we determined that the first-order autoregressive structure was most appropriate for analysis of growing plants; and that the compound symmetry structure was most appropriate for repeatedly harvested samples. Growth responses of the Tetroncium rhizome, and Sphagnum capitulum were regressed against temperature, precipitation, and daily UV-B radiation data.
All computations were performed in SAS version 8.2 (SAS Institute Inc., Cary, North Carolina, USA). P-values, where given in the text or with figures in anova tables, are accompanied by numerator and denominator degrees of freedom, and F-values.
Sphagnum height growth was consistently slightly less (c. 9%) under near-ambient than under reduced UV-B throughout the experiment (Fig. 1). Measurements throughout the growing season showed that most height growth occurred from October to January when the groundwater level is relatively high (data not shown), although this also coincides with maximum ozone depletion.
Capitulum density was slightly increased (c. 5%) under near-ambient compared with under reduced UV-B, and together with a tendency (P = 0.062) for increased capitulum mass (c. 8%), contributed to an increase (c. 32%) in volumetric density under near-ambient UV-B (Table 1). This offset the reduction in Sphagnum height growth seen under near-ambient UV-B (Fig. 1), and resulted in no change in Sphagnum biomass production between UV-B treatments (Table 1).
Samples taken during field season 2001–2002. Mean of 10 plots per UV-B treatment (±1 SE).
1.38 (± 0.06)
13 169 (± 396)
8.06 (± 0.75)
199.2 (± 10.9)
1.27 (± 0.05)
12 588 (± 380)
6.10 (± 0.69)
191.7 (± 14.5)
Annual rhizome elongation was reduced (c. 13%) under near-ambient UV-B (Fig. 1). There was no difference (F1,18 = 0.44, P = 0.516) in the magnitude of UV-B effect over the course of the experiment (Fig. 2a).
Each ramet measured produced 5–6 leaves per growing season. By measuring elongation between individual leaf scars, we determined that the treatment effect was highly seasonal, most pronounced during periods of rapid growth and not apparent late in the season (Fig. 2b). No effect on annual root production or the frequency of clonal division was found from rhizome or ramet analysis (data not shown). Tetroncium growth response to the near-ambient UV-B treatment was significantly negatively correlated with rainfall (R2 = 0.894), but not with other climatic variables (temperature and net radiation) (data not shown).
All measured leaves expanded more slowly under near-ambient UV-B, and reached a smaller final size (c. 13%) than under reduced UV-B (only Leaf 3 is shown; Fig. 3). The number of leaves produced and the rate of leaf senescence were not affected by UV-B treatment (data not shown). There was no change in leaf carbon content (c. 43% carbon, both treatments; F1,9 = 0.06, P = 0.813), but there was significantly (F1,9 = 8.51, P = 0.017) less leaf nitrogen under near-ambient (2.6 ± 0.1%, mean ±1 se) than under reduced UV-B (3.0 ± 0.1%).
Annual stem growth of a repeatedly measured cohort of ramets was consistently reduced from 1999 to 2002 under near-ambient UV-B (Fig. 1). This trend was more pronounced in harvested ramets, representing a range of sizes, over the same period (c. 10%; Table 2). There was also a slight tendency for less frequent branching and less branch growth under near-ambient UV-B from 1999 to 2002 (Table 2). Empetrum stem growth outpaced height growth of the Sphagnum mat under reduced UV-B (Fig. 4). However, the increase in Empetrum height above the growing Sphagnum mat was c. 53% less under near-ambient than under reduced UV-B (Fig. 4).
Table 2. Retrospective analysis of annual stem and leaf growth and morphology of Empetrum ramets
Retrospective analysis of harvested plants (1999–2002)
New growth of main stem (mm)
New growth of stem and branches (mm)
Branching frequency (Branches year−1 growth)
Number of leaves (year−1 growth)
12.4 (± 0.4)
29.0 (± 1.8)
1.2 (± 0.2)
25.8 (± 0.7)
13.7 (± 0.5)
34.5 (± 2.2)
1.5 (± 0.1)
28.3 (± 1.3)
Retrospective analysis of harvested plants (1996–1999)
New growth of main stem (mm)
New growth of stem and branches (mm)
Branching frequency (Branches year−1 growth)
Number of leaves(year−1 growth)
1999–2002 ramets harvested in February of 2002. 1996–1999 ramets harvested in February of 2000. Data for three field seasons are pooled. Mean of 10 plots per UV-B treatment (±1 SE).
19.0 (± 1.0)
26.1 (± 1.4)
0.7 (± 0.1)
31.1 (± 1.4)
19.4 (± 0.9)
27.9 (± 1.9)
0.8 (± 0.1)
34.9 (± 1.9)
The apparent age structure of the ramet population above the Sphagnum mat also underwent some change due to the UV-B treatments. Ramets with ≥ 4 years of growth above the Sphagnum exhibited much slower growth than those with fewer years of growth above the mat. The net result of slow growth of the ‘older’ ramets, and the consistently depressed growth of ‘younger’ ramets under near-ambient UV-B (data not shown), resulted in 20% of these ‘older’ ramets being partially engulfed by the Sphagnum between the 2001 and 2002 growing seasons (data not shown).
Nothofagus stem growth was highly variable among shoots, and there was no statistically significant effect of UV-B treatment on stem growth (Fig. 1). When measured in the sixth field season, there was a decrease in branching frequency (c. 35%) under near-ambient compared to reduced UV-B (Table 3).
Table 3. Leaf production and branching frequency of Nothofagus ramets
Number of leaves on new stem growth (Nov)
Number of leaves on new stem growth (Jan)
Leaf Area – Nov (mm2)
Leaf Area – Dec (mm2)
Branching frequency (Branches year−1 growth)
Split-plot anova on data for three field seasons, except branching frequency (2001–2002 only; one-way anova). Mean of 10 plots per UV-B treatment (±1 SE).
3.3 (± 0.2)
4.2 (± 0.2)
1913 (± 123)
3135 (± 123)
1.3 (± 1.1)
3.5 (± 0.2)
4.7 (± 0.3)
2313 (± 236)
3543 (± 186)
1.6 (± 1.1)
UV-B × Year
Annual measurements of vertical height show that Nothofagus ramets were outpaced by the Sphagnum mat. Under near-ambient UV-B Nothofagus ramets were engulfed by the Sphagnum mat c. 57% more rapidly than under reduced UV-B (Fig. 4).
The UV-B treatments had no effect on Nothofagus leaf size or number in November, but in December when fully expanded, the first leaf produced annually was smaller (c. 15%) under near-ambient than under reduced UV-B(Table 3). There was also a tendency (P = 0.057) for fewer leaves (c. 10%) under near-ambient UV-B (Table 3).
Like Empetrum, the growth of Nothofagus ramets declined with years above the Sphagnum mat (Fig. 1). When broken down into apparent age classes, many Nothofagus ramets (c. 50%) had only one year of growth above the Sphagnum mat. No changes in the apparent age structure of Nothofagus due to solar UV-B treatment occurred during the last 2 years of the study (data not shown).
Sphagnum annual height growth was greater than that of Nothofagus (Fig. 4) and Tetroncium (Fig. 1), but less than that of Empetrum in both treatments (Fig. 4). Irrespective of the UV-B treatments, Empetrum increased (c. 8%), whilst Tetroncium (c. 23%) and Nothofagus (c. 13%) decreased in ramet density from the fifth to the sixth field season (Table 4). However, there were no overall changes in the species composition of the plant community due to the UV-B treatments at the end of the sixth field season (Table 4).
Table 4. Annual population of peatland plant species, as total number of ramets above the Sphagnum mat per plot over the final two field seasons
Species UV-B treatment
Mean of 10 plots per UV-B treatment (±1 SE).
124 (± 20)
925 (± 175)
434 (± 87)
15 (± 5)
12 (± 4)
29 (± 14)
106 (± 26)
936 (± 142)
443 (± 120)
14 (± 7)
16 (± 3)
17 (± 5)
108 (± 20)
972 (± 179)
350 (± 85)
7 (± 2)
17 (± 6)
19 (± 8)
93 (± 14)
1059 (± 154)
327 (± 97)
10 (± 3)
16 (± 3)
9 (± 3)
UV-B × Year
Overall, growth of Sphagnum and the vascular plant species was reduced by near-ambient UV-B over the last 3 years of our study. The effects, although small, were generally consistent over time. Many of the trends in measured parameters were not significant when taken in any one year, but were when repeatedly measured over several years.
A reduction in Sphagnum height growth and increase in volumetric density to due solar UV-B occurred during the fourth to sixth field seasons and is comparable to that reported during the first three field seasons (Searles et al., 2002). There was no evidence of an incrementally greater treatment effect accumulating with time. A similar change in the growth and morphology occurred in Sphagnum fuscum growing under supplemental UV-B lamp treatments (c. 30% greater than ambient) in a sub-Arctic (68° N) peatland (Gehrke, 1998). In a peatland microcosm study under similarly enhanced UV-B, there was no effect of the supplemental UV-B on Sphagnum angustifolium, and, if anything, a slight decrease in Sphagnum balticum, capitulum mass (Niemi et al., 2002a, 2002b).
The most responsive species to UV-B radiation in our study was Tetroncium, which exhibited reduced rhizome growth and leaf expansion during the 6 years of treatments (also see Searles et al., 2002). In other experiments in Tierra del Fuego, near-ambient UV-B similarly reduced summer root-growth in a Carex fen (Zaller et al., 2002), and leaf expansion of perennial herb Gunnera magellanica, and fern Blechnum penna-marina (Rousseaux et al., 1998, 2001). On the Palmer Peninsula, Antarctica, solar UV-B also reduced leaf expansion of Deschampsia antarctica (Day et al., 2001) and Colobanthus quitensis (Xiong et al., 2002).
Tetroncium favours the wettest areas of the peatland, and in high rainfall years the reduction in rhizome growth under near-ambient UV-B was less pronounced. The UV-B effect was apparently dampened by the greater annual growth under both treatments during wetter years. Tetroncium leaf nitrogen concentration was lower under near-ambient than under reduced UV-B. It is possible that these two effects were symptomatic of reduced nutrient availability in low rainfall years. Nutrient limitation is often considered to be the most important stress restricting growth of vascular plants in peatland competition (Aerts et al., 1999; Bridgham, 2002). Even under long-term nitrogen deposition, S. magellanicum monoliths were able to capture and retain most nitrogen in the capitulum, and exhibited reduced height growth and increased density. This, in turn, limited nitrogen availability for the emergent vascular plants (Heijmans et al., 2002a, 2002b).
In our study, both Empetrum and Nothofagus were less competitive against the Sphagnum mat under near-ambient UV-B. Empetrum was less successful at escaping from the Sphagnum mat, and Nothofagus was engulfed by Sphagnum to a greater extent under near-ambient than under reduced UV-B. These trends can only partially be explained by reduced stem growth due to solar UV-B, and it is likely that the concurrent reduction in branching frequency and growth were also contributing factors. Indeed, changes in the production and orientation of branches are known to affect competition between Sphagnum and emergent peatland vascular plants (Svensson, 1995). In a comparable study of the effects of solar UV-B on a sub-Arctic heath community, a reduction in branching frequency was the only significant treatment effect on E. hermaphroditum, perhaps due to high within-treatment growth variation (Phoenix et al., 2003).
After six field seasons of treatments, changes in growth and competition between Sphagnum and vascular plant species were still not reflected at the community level. This is similar to the results of other solar UV-B manipulation experiments of shorter duration (Day et al., 2001; Phoenix et al., 2003). The only population-level change was the partial engulfment by Sphagnum of ‘older’Empetrum ramets subject to near-ambient UV-B for five and six field seasons of treatments. This may indicate that six field seasons is insufficient time for growth effects to be expressed at the community level in such a slow growing system, particularly when many other confounding environmental variables affect growth and competition (Shevtsova et al., 1995; Press et al., 1998).
Our reduced-UV-B treatment is lower than solar UV-B levels in this region before ozone depletion occurred (Searles et al., 2002); therefore, our differences in UV-B treatments were substantially larger than the difference between normal solar UV-B before and with ozone depletion. Yet growth responses of Sphagnum and vascular plants to the solar UV-B manipulations were still rather subtle. If ozone depletion in this region substantially ameliorates within the next few decades, it is unlikely that significant changes in the structure of the Tierra del Fuego peatland plant community will occur. In our experiment, Sphagnum gained a competitive advantage over emergent vascular plants under near-ambient UV-B, but this was not substantial enough to alter the community composition between the fourth and sixth field seasons of this study. Biomass production of Sphagnum was unaffected by UV-B treatments, but hypothetically the change in Sphagnum morphology could alter carbon storage by affecting peatland hydrology (van Breemen, 1995), since the denser capitulum layer could retain more water, reducing decomposition in the acrotelm (Malmer et al., 1994). To become meaningful in a broader context, the long-term effects of UV-B radiation should be studied together with other climatic changes and this has received little attention (Björn et al., 1999; Phoenix et al., 2001; Sonesson et al., 2002).
Funding was provided by the United States National Science Foundation Terrestrial Ecology and Global Change grants (95-24144 and 98-14357). We are particularly grateful to the Argentinean National Park Service (Administración de Parques Nacionales) for allowing us to use Parque Nacional de Tierra del Fuego to perform our experiments. We gratefully acknowledge the vital collaboration with CADIC-CONICET (Centro Austral de Investigaciones Científicas), Ushuaia (Dir. Eduardo Olivero), including access to the NSF UV monitoring station data (Susana Díaz), and weather data for Ushuaia (Rodolfo Iturraspe). Technical support and field assistance in Tierra del Fuego was provided by Nicolás Garibaldi, Ricardo Saenz-Samaniego, and Florencia Díaz. We also appreciate Peter Searles, Carla Giordano, Renato Gerdol and Hans Zaller for research advice, and Susan Durham's substantial statistical input. Two anonymous reviewers gave valuable suggestions on an earlier version of this manuscript.