Effects of global change on a sub-Arctic heath: effects of enhanced UV-B radiation and increased summer precipitation

Authors

  • Gareth K. Phoenix,

    Corresponding author
    1. Department of Animal and Plant Sciences, University of Sheffield, Western Bank, Sheffield S10 2TN, UK;
    Search for more papers by this author
  • Dylan Gwynn-Jones,

    1. Department of Animal and Plant Sciences, University of Sheffield, Western Bank, Sheffield S10 2TN, UK;
    2. Abisko Scientific Research Station, S 981 07 Abisko, Sweden; and
    Search for more papers by this author
  • Terry V. Callaghan,

    1. Department of Animal and Plant Sciences, University of Sheffield, Western Bank, Sheffield S10 2TN, UK;
    2. Abisko Scientific Research Station, S 981 07 Abisko, Sweden; and
    Search for more papers by this author
  • Darren Sleep,

    1. Institute of Terrestrial Ecology, Merlewood Research Station, Grange-over-Sands LA11 6JU, UK
    Search for more papers by this author
  • John A. Lee

    1. Department of Animal and Plant Sciences, University of Sheffield, Western Bank, Sheffield S10 2TN, UK;
    Search for more papers by this author

  • Present address: Institute of Biological Sciences, Edward Llwyd Building, University of Wales, Aberystwyth, Ceredigio SY23 3DA, UK.

Gareth K. Phoenix, Department of Animal and Plant Sciences, University of Sheffield, Sheffield S10 2TN, UK (fax 0114-222-0002, e-mailg.phoenix@sheffield.ac.uk).

Summary

  • 1The responses of sub-Arctic heathland vegetation to enhanced UV-B radiation and increased summer precipitation over 7 years were investigated in a field experiment in northern Sweden.
  • 2Growth, phenology and reproduction of the dominant dwarf shrubs Vaccinium myrtillus, V. uliginosum, V. vitis-idaea and Empetrum hermaphroditum were studied after 5–7 years of manipulation and retrospective analyses were used to assess growth responses in earlier years. Leaf tissue N and P and 13C natural abundances were determined for V. myrtillus and E. hermaphroditum. Growth responses were also assessed for the moss Hylocomium splendens.
  • 3The deciduous V. myrtillus showed reduced growth, increased leaf thickness and increased flowering and berry production under enhanced UV-B in some years. V. uliginosum, V. vitis-idaea, E. hermaphroditum and H. splendens were, in general, tolerant of UV-B.
  • 4Increased precipitation affected growth only in the evergreen species: stem length and branching were sometimes stimulated in E. hermaphroditum, whereas V. vitis-idaea showed reduced branching.
  • 5Precipitation also increased leaf thickness in V. uliginosum and reduced flowering and berry production in V. myrtillus.
  • 6In the interactions that occurred between enhanced UV-B radiation and increased summer precipitation, combining the two treatments often negated any effect that either may have had separately. The effect of concurrent increases on this ecosystem is therefore likely to be much less than if either occurred singly.
  • 7Enhanced UV-B and increased summer precipitation appeared not to effect dwarf shrub abundances during the first 5 years of the experiment, suggesting that overall this heath may be more tolerant of these environmental changes than previously thought.

Introduction

Recent reports have described continued losses of stratospheric ozone over Arctic regions (Müller et al. 1997; Rex et al. 1997; Madronich et al. 1998) and eventual ozone recovery may be delayed by global warming (Shindell et al. 1998) or the persistence of ozone-depleting compounds in the atmosphere (Montzka et al. 1999). Such ozone depletion results in an increase in the amount of harmful UV-B radiation (280–315 nm) reaching the biosphere. However, the naturally thick ozone layer and low solar angles have ensured that Arctic plant communities have, until recently, been exposed to only low levels of harmful UV-B (Caldwell et al. 1980; Robberecht et al. 1980). Also, photochemical damage induced by increased UV-B radiation probably remains rapid in the cool climate of the sub-Arctic but enzymatic repair processes will take place much more slowly at low temperatures (Björn et al. 1997). Arctic vegetation is susceptible to increases in incident UV-B with plant responses ranging from reduced growth in dwarf shrubs (Johanson et al. 1995b) to increased fruit production and species-specific changes in herbivory (Gwynn-Jones et al. 1997). Such effects could have long-term implications for the community as competitive balances between species and interactions between trophic levels may be altered (Barnes et al. 1988, 1990; Gehrke et al. 1995, 1996). To date, data are only available for the first 2–3 years of manipulation, although responses may take many years to develop in such slow-growing species.

Responses to increased UV-B are likely to be affected by concurrent changes in other environmental factors. Large increases in precipitation are, for instance, predicted to occur in Northern Europe and soil moisture will be increased further because more precipitation will fall as rain rather than snow and there is little change in evapotranspiration at low temperatures (Kattenberg et al. 1996). The effects on Arctic vegetation are expected to be large because rainfall is low (≈ 300 mm per year at the site used in this study), slow-growing, long-lived perennials which rarely establish from seed cannot react rapidly to change, and low temperature and nutrient stress make them more sensitive to disturbance (Jonasson et al. 1996; see Crawford 1997 for an overview).

Diurnal and seasonal patterns of drought stress and UV-B maxima are similar (Bornman & Teramura 1993) and some studies have shown that plant tolerance can actually increase when the stresses are combined (Murali & Teramura 1986; Balakumar et al. 1993; Drilias et al. 1997). Mechanisms include increases in antioxidant capacity, increased UV-B absorbing compounds and leaf anatomical changes. Precipitation is, however, predicted to increase and nothing is known of how this may effect areas in which naturally low levels of rainfall may at present provide some tolerance to UV-B.

We established a long-term field experiment to investigate the effects and interactions of enhanced UV-B radiation and increased summer precipitation on a natural sub-Arctic heath community. We hypothesized that:

  • (1) UV-B damage would lead to reductions in growth, rate of development and reproduction

  • (2) continued exposure throughout the longer life of leaves would make evergreen species more susceptible to UV-B (Sullivan & Teramura 1992; Johanson et al. 1995b)

  • (3) species-specific responses would lead to changes in competitive balances and thus to changes in species abundances

  • and

  • (4) increased summer precipitation would increase sensitivity to UV-B.

Methods

Experimental design

The first and only study combining enhanced UV-B radiation and increased summer precipitation in a natural ecosystem was established on a sub-Arctic heathland community at the Abisko Scientific Research Station, Abisko, Sweden (68°35′ N, 18°82′ E) in 1993. It has run during every subsequent growing season from 1993 to 1999 inclusive, and hence also forms one of the longest running UV-B enhancement experiments to date.

The enhanced UV-B treatment, which simulated a 15% reduction in stratospheric ozone under clear sky conditions at Abisko, and the watering treatment, which increased average summer precipitation from 121 mm to ≈ 220 mm, were applied in a two-way factorial design with four replicates of each treatment (total of 16 plots).

The heathland community corresponds to the Empetrum–V. myrtillus variant described by Sonesson & Lundberg (1974), and consists of an open birch canopy (Betula pubescens Ehrh. ssp. czerepanovii) with a dense dwarf shrub understorey dominated by E. hermaphroditum Hagerup, V. vitis-idaea L. (both evergreen), V. myrtillus L. and V. uliginosum L. (both deciduous). Below this is a ground layer of predominantly mosses and lichens.

Six parallel fluorescent tube lamps (Q-PANEL UVB-313, Cleveland, OH, USA) were mounted 1.5 m above each plot (2.5 × 1.3 m) on metal frames, the lamps being 0.5 m apart. The middle 70 cm of the two central lamps was covered in aluminium foil to allow even irradiance throughout the plot at canopy height (Johanson et al. 1995a). Fluorescent lamps were preburned for 100 h before use to achieve stable output and new lamps were used each season. Individual filters were positioned directly below each lamp and not between lamps so that plots received ambient solar UV and precipitation. Shading was also minimal. In control (ambient UV-B) plots, window glass filters were used to exclude UV-B and UV-C, whereas in enhanced UV-B plots, UV-transparent Plexiglas (Röhm GmbH, Darmsradt, Germany) carrying presolarized cellulose diacetate foil (0.13 mm, Courtaulds, Derby, UK) excluded only ecologically irrelevant UV-C radiation (< 280 nm). Diacetate filters were replaced after every 50 h of field irradiation to maintain stable transmittance.

The extra UV-B exposure required to simulate 15% ozone depletion was calculated from a computer model (Björn & Murphy 1985; Björn & Teramura 1993) for clear sky conditions at Abisko and weighted using a modified generalized plant action spectrum from Caldwell (1971) normalized at 300 nm. Lamp output was centred around noon by timers which controlled three of the six lamps per frame simultaneously to gain a step-wise ‘square wave’ increase and decrease in daily UV-B enhancement. The timers were adjusted every second week to simulate seasonal variation in natural UV-B levels. Maximum biologically effective UV-B doses were calculated to be 5.8 kJ m−2 d−1 for UV-B-enhanced plots and 4.6 kJ m−2 d−1 for control plots. Spectral irradiances within the plots were measured using an Optronics 724 spectroradiometer (Orlando, FL, USA). Typical radiation environments under treatment and control frames are shown in Fig. 1.

Figure 1.

Ultraviolet radiation (280–400 nm) at dwarf shrub canopy height under enhanced UV-B and control (ambient UV-B) frames. Irradiance includes incident solar ultraviolet. Data from midday on 8 June 1999.

Plots were watered every 1 or 2 days (9 L m−2 added per week, total equivalent to 100 mm additional precipitation per season). Treatments were applied from late May to early September from 1993 to 1999 and the long-term effects were determined in 1997, 1998 and 1999 (years 5, 6 and 7 when ambient summer precipitation was 88, 135 and 179 mm, respectively).

Stem growth and branching

Annual stem length increments and annual branching in E. hermaphroditum and V. myrtillus from 1993 to 1997 were determined in late August 1997. The same measurements were also taken for V. vitis-idaea for 1996 and 1997 only (the number of pre-1996 stems was statistically insufficient). The length between the scars left by each year’s bursting bud along the main axis stem (Johanson et al. 1995a) was measured in 10 randomly selected individuals of each dwarf shrub species in each plot and the number of stems produced (including the main stem itself) by the main axis in each year was counted.

Two additional independent sets of data were collected for V. myrtillus, V. vitis-idaea and E. hermaphroditum in 1998 and 1999 using the current year’s growth in a new set of plants selected at random. Retrospective measurements of stem growth and branching in V. uliginosum were taken in 1998 (readings taken back to 1995) and an additional set of independent measurements was taken in 1999.

For the most abundant shrub species, E. hermaphroditum, the growth of the main apical stems of 10 individuals was measured at weekly intervals throughout the 1997 growing season.

Growth of the moss H. splendens was also assessed. The distance along the main axis between the branching points of each new year’s growth to the apex of each branch (for 1996–99) was measured on 10 randomly selected individuals within each plot (Økland 1995) and annual segments were then separated and weighed after oven drying at 60 °C for 48 h.

Percentage cover survey

Each plot was surveyed for cover of all species present in early August (at peak biomass) in both 1993 and 1997. The percentage cover of all species present was estimated in each of 64 5 × 5 cm squares in the centre of each plot. While this method is biased towards canopy-forming species, subdivision of the 40 × 40 cm survey area improves accuracy as percentage cover estimations are far easier on such small units.

Leaf thickness

Individuals of V. myrtillus, V. uliginosum and V. vitis-idaea (n = 10 per species per plot) were chosen at random in mid-August 1997, 1998 and 1999 and the thickness of the second from top leaf of each plant measured using outside callipers (0.001 mm: 0–25 mm, Radio Spares). This was not undertaken on E. hermaphroditum as it has small cylindrical leaves.

Vegetative bud phenology

Randomly selected individuals of each of the four dwarf shrub species (n = 10 per species per plot) were labelled in May 1998 then re-visited at weekly intervals. The phenological stage of the apical bud was recorded (1 = dormant bud, 2 = first signs of bud burst, 3 = bud burst with the first leaf fully expanded) until all individuals had reached stage 3. Tagged plants of the deciduous species (V. myrtillus and V. uliginosum) were surveyed again in late August for signs of senescence in the second upper leaf on the main stem (0 = no senescence, 1 = leaf tips or fringes brown, 2 = at least half brown, 3 = fully necrotic, as standing dead, 4 = fallen leaf).

Flowering and berry production

Flowering buds were counted in the first week of July, for V. myrtillus in 1997 and for all three Vaccinium species in 1998 and 1999 (E. hermaphroditum flowers are small and inconspicuous and could not be surveyed accurately). Berries were harvested in mid-August, counted and weighed fresh. Again, only V. myrtillus was included in 1997 but berries of all four dwarf shrub species were harvested in 1998 and 1999. Total flower counts and berry numbers and weights were adjusted for the cover of species within the plots (based on the survey taken in 1997).

leaf nutrients and δ13C signatures

Leaves of V. myrtillus and E. hermaphroditum were harvested in 1994, 1997 and 1998 and dried at 60 °C for 48 h. Samples were then sent for 13C natural abundance analysis (δ13C) at the NERC stable isotope facility (Institute of Terrestrial Ecology, Merlewood, Cumbria, UK), 1994 samples were stored dry and analysed along with 1997 samples. δ13C results are reported as ‰ where δ13C = 1000 × (Rsample − Rstandard)/Rstandard (‰). R = 13C/12C. Plant δ13C is controlled by the carbon isotope ratio of the surrounding air and the discrimination between the two carbon isotopes during carbon assimilation. Ribulose bisphosphate carboxylase preferentially assimilates 12CO2 except when stomata are closed; almost all intercellular CO2 is then used regardless of its isotopic composition and as this carbon is then integrated into plant tissue, leaf δ13C represents a leaf life-time integrated measure of the relationship between carbon assimilation and stomatal conductance (Griffiths 1991; Lajtha & Marshall 1994; Michelsen et al. 1996).

Ground material (0.03–0.05 g) was analysed for ammonium-N and phosphate-P by flow injection analysis using the indophenol blue and molybdenum blue methods, respectively (S. E. Allen 1989).

Data analysis

Main factor effects and interactions were determined using a two-factor General Linear Model in minitab 11.12 for Windows (Minitab Inc., PA, USA). Retrospective growth analyses were analysed with two-way repeated measures anovas using SPSS for Windows 6.0 (SPSS Inc., Chicago, IL, USA) (stem increments for each year were taken from the same plants and hence were not independent). Phenological development of growing buds was analysed by repeated measures anovas following rank transformation of developmental stage values (Conover & Iman 1981; SAS Institute Inc. 1988). Data were transformed where necessary. Where measurements were taken from several plants in each plot (i.e. stem lengths, branching, leaf thickness, phenology), mean values per plot were used in the analyses.

Results

Effects of enhanced uv-b radiation

The deciduous V. myrtillus was by far the most responsive of the four dwarf shrubs to UV-B radiation with annual stem length increments reduced by an average of 12% between 1993 and 1997 (Table 1), whereas leaf thickness (+7% 1997; Table 2), flowering (+133% in 1999), and both the number (+153%) and total fresh weight (+342%) of berries (Table 3) all increased significantly. With the exception of flower numbers in 1998, UV-B always increased reproductive measures in this species.

Table 1.  Main factor effects and interaction of enhanced UV-B radiation and increased summer precipitation (W) on annual stem length increments (mm) of four dwarf shrubs and a moss. The values for main factor effects are the difference between the means of plots that received a treatment and those that did not. The values for interaction effects are the difference between the means for those plots in which treatments were applied in combination and the expected values calculated from the effects of those treatments applied separately (Snedecor & Cochrane 1973). Figures in bold indicate significant effects
  Main factor effects and interactionsMean values from individual treatment combinations (± SE) n = 4
SpeciesYear +UVB +WInteraction−UVB − W+UVB − W−UVB + W+UVB + W
  1. Statistical significances of main factor effects: *P < 0.05; **P < 0.01. †In retrospective analyses means from individual treatment combinations calculated from all years combined and compared using repeated measures anova.

V. myrtillus1993–97−3.4*  +3.1  +2.8*29.0 (± 2.1)22.8 (± 1.6)29.3 (± 2.4)28.7 (± 1.9)
 1998 +4.1  +1.2  +1.731.4 (± 4.9)33.8 (± 2.1)30.9 (± 4.1)36.6 (± 3.4)
 1999 +6.3  +2.9 +14.431.7 (± 4.1)30.8 (± 1.5)29.7 (± 2.2)34.6 (± 3.0)
V. uliginosum1995–98 +2.2  +1.6 +10.8**46.1 (± 1.9)37.5 (± 1.0)36.9 (± 0.6)49.9 (± 2.2)
 1999−2.3−2.8 +13.434.5 (± 5.6)29.9 (± 6.3)29.7 (± 2.8)33.8 (± 0.0)
V. vitis-idaea1996–97−1.5  +4.1  +4.8*30.2 (± 1.0)23.8 (± 1.4)29.4 (± 2.0)32.7 (± 3.1)
 1998−4.0−2.7  +6.629.8 (± 6.0)19.3 (± 1.3)20.6 (± 1.3)23.2 (± 2.5)
 1999−7.9−13.8  −1.025.1 (± 0.6)21.1 (± 0.6)19.7 (± 1.2)20.2 (± 3.0)
E. hermaphroditum1993–97 +1.3  +4.6  −2.333.5 (± 1.4)37.0 (± 1.0)40.3 (± 4.2)39.3 (± 2.1)
 1998 +0.5  +5.4  +3.929.7 (± 2.6)26.3 (± 1.5)31.2 (± 1.8)35.6 (± 6.8)
 1999−4.5−6.8  +9.934.2 (± 3.1)28.2 (± 2.8)27.5 (± 2.5)30.6 (± 4.1)
H. splendens1996–99 lengths +4.2*  +3.4*  +3.5*23.9 (± 1.0)24.5 (± 1.2)23.7 (± 1.2)31.5 (± 2.4)
 1996–99 weights +1.4*−0.6  +1.3* 8.17 (± 0.41) 8.26 (± 0.34) 6.35 (± 0.59) 8.97 (± 0.66)
Table 2.  Main factor effects and interaction of enhanced UV-B radiation and increased summer precipitation (W) on the leaf thickness (µm) of three Vaccinium species
  Main factor effects and interactionsMean values from individual treatment combinations (± SE) n = 4
SpeciesYear+UVB+WInteraction−UVB − W+UVB − W−UVB + W+UVB + W
  • *

    P < 0.05;

  • **

    P < 0.01;

  • P = 0.051.

V. myrtillus1997+10.3** +1.8 −4.5152 (± 3)167 (± 5)158 (± 2)164 (± 2)
 1998 +0.0 +4.4 +1.9162 (± 8)160 (± 3)164 (± 8)166 (± 7)
 1999 +6.4 −2.7−10.3159 (± 3)175 (± 15)166 (± 10)163 (± 8)
V. uliginosum1997 +0.4+25.4 −5.9241 (± 9)247 (± 7)272 (± 0)267 (± 6)
 1998 +3.5 +6.6 −8.8243 (± 7)255 (± 10)258 (± 13)253 (± 4)
 1999 +3.6+23.1**−11.8*237 (± 2)253 (± 2)272 (± 2)264 (± 9)
V. vitis-idaea1997 +0.46 +1.06 −1.81339 (± 12)362 (± 18)368 (± 15)354 (± 6)
 1998 +0.6+10.9−30.4**327 (± 8)358 (± 6)369 (± 13)339 (± 6)
 1999 −4.1 +9.4−22.7*337 (± 12)356 (± 8)369 (± 10)342 (± 6)
Table 3.  Main factor effects and interaction of enhanced UV-B radiation and increased summer precipitation (W) on flowering and berry production of four dwarf shrubs. All counts and weights are per m2 of species cover
  Main factor effects and interactionsMean values from individual treatment combinations (± SE) n = 4
SpeciesResponse+UVB+WInteraction−UVB − W+UVB − W−UVB + W+UVB + W
  • *

    P < 0.05.

  • Flowering and berry production of V. uliginosum occurred in too few plots to allow statistical analysis but numbers are shown for completeness.

V. myrtillusFlowers no. m−2 1997+400−328−362714 (± 121)1476 (± 314)748 (± 176) 787 (± 177)
 Flowers no. m−2 1998 −44 −93  −1404 (± 97) 360 (± 81)311 (± 126) 268 (± 125)
 Flowers no. m−2 1999+132*−128* −38144 (± 41) 314 (± 57) 54 (± 24) 148 (± 76)
 Berry no. m−2 1997 +75 +32 +60178 (± 59) 193 (± 59)150 (± 62) 285 (± 95)
 Berry no. m−2 1998+100 +20 +30212 (± 67) 282 (± 74)202 (± 152) 332 (± 137)
 Berry no. m−2 1999+125*−136* −39119 (± 32) 284 (± 88) 22 (± 9) 109 (± 60)
 Berry mass m−2 1997 +14.2 +12.4 +14.8 32.9 (± 13.5)  32.3 (± 10.7) 30.5 (± 13.8)  59.4 (± 20.1)
 Berry mass m−2 1998 +15.2  +7.1  −5.1 37.0 (± 13.1)  57.3 (± 17.7) 49.2 (± 37.7)  59.3 (± 31.0)
 Berry mass m−2 1999 +16.6* −15.3*  −9.4  7.8 (± 3.4)  33.8 (± 11.1)  1.9 (± 1.2)   9.1 (± 5.4)
V. uliginosumFlowers no. m−2 1998  −6−282 −42322 (± 116) 358 81 (± 10)  34 (± 34)
 Flowers no. m−2 1999 −19−115 +31146 (± 81)  96 (± 96)  0 (± 0)  11 (± 11)
 Berry no. m−2 1998 +82 −91−100  9 (± 9) 191 18 (± 18)   0 (± 0)
 Berry no. m−2 1999 +12.0 −12.0 −12.0  0 (± 0)  24 (± 24)  0 (± 0)   0 (± 0)
 Berry mass m−2 1998 +12.3 −13.0 −18.1  0.6 (± 0.6)  31.0  5.8 (± 5.8)   0.0 (± 0.0)
 Berry mass m−2 1999  +1.4  −1.4  −1.4  0.0 (± 0.0)   2.8 (± 2.8)  0.0 (± 0.0)   0.0 (± 0.0)
V. vitis-idaeaFlowers no. m−2 1998−171 +91 +85405 (± 170) 150 (± 87)411 (± 91) 325 (± 65)
 Flowers no. m−2 1999 +26+128 −94 62 (± 22) 182 (± 80)284 (± 135) 216 (± 27)
 Berry no. m−2 1998 +27+416 +26564 (± 258) 565 (± 277)954 (± 453)1006 (± 211)
 Berry no. m−2 1999 +65 +27 −12  0 (± 0)  77 (± 61) 39 (± 24)  92 (± 64)
 Berry mass m−2 1998   −4.6 +44.6 −17.6 36.8 (± 16.2)  49.8 (± 26.3) 99.0 (± 52.7)  76.8 (± 21.9)
 Berry mass m−2 1999  +2.9  +1.7  −0.4  0.0 (± 0.0)   3.3 (± 2.8)  2.1 (± 1.2)   4.6 (± 2.7)
E. hermaphroditumBerry no. m−2 1998 +40 −82−149 58 (± 58) 246 (± 143)124 (± 67)  15 (± 15)
 Berry no. m−2 1999 −18−213 −37266 (± 231) 285 (± 226) 90 (± 44)  36 (± 28)
 Berry mass m−2 1998 +15.9 −14.3 −41.5*  0.0 (± 0.0)  57.4 (± 31.9) 27.2 (± 16.8)   1.5 (± 1.5)
 Berry mass m−2 1999  +8.8 −33.2 −15.8 27.6 (± 23.0)  52.2 (± 43.7) 10.2 (± 4.4)   3.2 (± 2.4)

UV-B radiation did not have significant effects on growth, reproduction and phenology of any of the other dwarf shrub species. Closer inspection of the apparent 10% greater total nitrogen in leaves of E. hermaphroditum in 1998 (Table 4) showed that it occurred only with increased precipitation. Growth (in terms of both length and weight) of the moss H. splendens was enhanced between 1996 and 1999 (Table 1).

Table 4.  Main factor effects and interaction of enhanced UV-B radiation and increased summer precipitation (W) on leaf total nitrogen (mg N g−1) and phosphorus (mg P g−1), and 13C natural abundances (δ13C) of Vaccinium myrtillus and Empetrum hermaphroditum
  Main factor effects and interactionsMean values from individual treatment combinations (± SE) n = 4
SpeciesResponse; Year +UVB +WInteraction−UVB − W+UVB − W −UVB + W +UVB + W
  • *

    P < 0.05;

  • **

    P < 0.01.

V. myrtillusNitrogen; 1994 −1.30 +1.02 +4.15  31.8 (± 2.5) 26.3 (± 1.6)  28.6 (± 3.1)  31.5 (± 1.5)
 Nitrogen; 1997−0.19 +1.94 +1.83  22.4 (± 1.6) 20.4 (± 0.7)  22.5 (± 1.6)  24.2 (± 2.0)
 Nitrogen; 1998 −1.45 +0.09 +2.96  28.4 (± 1.4) 24.0 (± 1.2)  25.5 (± 2.4)  27.0 (± 3.0)
 Phosphorus; 1994 +0.26 +0.35* +0.17   2.77 (± 0.13)  2.89 (± 0.12)   2.98 (± 0.19)   3.37 (± 0.18)
 Phosphorus; 1997 +0.11 +0.11 +0.04   2.71 (± 0.10)  2.78 (± 0.01)   2.78 (± 0.07)   2.93 (± 0.05)
 Phosphorus; 1998 +0.23 +0.18 +0.22   1.86 (± 0.11)  1.89 (± 0.04)   1.82 (± 0.16)   2.28 (± 0.22)
 δ13C; 1994−0.08 +0.04 +0.03 −31.1 (± 0.2)−31.2 (± 0.2) −31.1 (± 0.2) −31.1 (± 0.3)
 δ13C; 1997−0.29−0.20−0.06 −31.5 (± 0.2)−31.8 (± 0.1) −31.8 (± 0.1) −32.0 (± 0.2)
 δ13C; 1998−0.21−0.00−0.01 −31.7 (± 0.1)−31.9 (± 0.1) −31.7 (± 0.3) −31.9 (± 0.2)
E. hermaphroditumNitrogen; 1994 +0.06 −1.07 +1.57  21.6 (± 0.7) 20.0 (± 0.9)  18.9 (± 0.5)  20.5 (± 1.5)
 Nitrogen; 1997 −1.29−0.61 +1.25  22.5 (± 1.1) 19.9 (± 1.3)  20.6 (± 1.1)  20.6 (± 1.3)
 Nitrogen; 1998 +1.96*−0.97 +1.96*  21.4 (± 0.6) 21.4 (± 0.9)  18.4 (± 0.3)  22.4 (± 0.9)
 Phosphorus; 1994 +0.10−0.11 +0.09   2.50 (± 0.10)  2.51 (± 0.19)   2.30 (± 0.11)   2.50 (± 0.18)
 Phosphorus; 1997 +0.04−0.00+0.09   2.57 (± 0.09)  2.52 (± 0.17)   2.48 (± 0.50)   2.61 (± 0.07)
 Phosphorus; 1998 +0.07−0.11 +0.14   1.42 (± 0.09)  1.34 (± 0.07)   1.17 (± 0.02)   1.38 (± 0.09)
 δ13C; 1994−0.040.41*−0.64**−25.8 (± 0.2)−25.2 (± 0.1)−25.6 (± 0.1)−26.2 (± 0.2)
 δ13C; 1997−0.19−0.52−0.15−26.2 (± 0.3)−26.3 (± 0.2)−26.6 (± 0.2)−26.9 (± 0.3)
 δ13C; 1998 +0.04 +0.10−0.59−27.3 (± 0.1)−26.7 (± 0.1)−26.6 (± 0.1)−27.2 (± 0.3)

Enhanced UV-B radiation had no significant effects on bud phenology or leaf senescence (data not shown) or on leaf δ13C signatures (Table 4), branching (Table 5) and species percentage cover (Table 6).

Table 5.  Main factor effects and interaction of enhanced UV-B radiation and increased summer precipitation (W) on branching of four dwarf shrubs. Branching is calculated as the number of stems growing from the previous years main stem
  Main factor effects and interactionsMean values from individual treatment combinations (± SE) n = 4
SpeciesYear+UVB+WInteraction−UVB − W+UVB − W−UVB + W+UVB + W
  1. * P < 0.05. †In retrospective analyses means from individual treatment combinations calculated from all years combined and compared using repeated measures anova.

V. myrtillus1993–97−0.09+0.06+0.071.93 (± 0.06)1.84 (± 0.05)1.98 (± 0.13)1.90 (± 0.10)
 1998−0.20+0.10−0.43*2.45 (± 0.12)2.68 (± 0.13)2.98 (± 0.25)2.35 (± 0.16)
 1999+0.06−0.26−0.062.40 (± 0.14)2.53 (± 0.09)2.20 (± 0.19)2.20 (± 0.17)
V. uliginosum1995–98−0.02−0.25−0.142.17 (± 0.16)2.29 (± 0.05)2.06 (± 0.03)1.90 (± 0.14)
 1999−0.14−0.01+0.192.10 (± 0.21)1.78 (± 0.08)1.90 (± 0.10)1.95 (± 0.25)
V. vitis-idaea1996–97−0.08−0.10*−0.031.15 (± 0.06)1.10 (± 0.03)1.08 (± 0.03)0.98 (± 0.04)
 1998+0.03+0.05+0.031.00 (± 0.00)1.00 (± 0.00)1.03 (± 0.03)1.08 (± 0.05)
 1999+0.038+0.038+0.0131.05 (± 0.05)1.08 (± 0.05)1.08 (± 0.05)1.12 (± 0.05)
E. hermaphroditum1993–97+0.10+0.22*−0.131.77 (± 0.05)1.99 (± 0.14)2.11 (± 0.16)2.08 (± 0.08)
 1998+0.15+0.33−0.202.48 (± 0.32)2.83 (± 0.23)3.00 (± 0.16)2.95 (± 0.23)
 1999+0.11+0.04+0.413.18 (± 0.31)2.88 (± 0.19)2.80 (± 0.27)3.33 (± 0.27)
Table 6.  Main factor effects and interaction of enhanced UV-B radiation and increased summer precipitation (W) on the change in percentage cover of species between 1993 and 1997
 Main factor effects and interactionsMean values from individual treatment combinations (± SE) n = 4
Species+UVB+WInteraction−UVB − W+UVB − W−UVB + W+UVB + W
  • *

    P < 0.05;

  • **

    P < 0.01.

V. myrtillus+2.52  +0.22+1.90−0.62 (± 0.92)  0.05 (± 1.61)−2.30 (± 0.60) 2.13 (± 2.11)
V. uliginosum−0.75  −2.27+0.292.67 (± 2.19)1.64 (± 1.64)0.12 (± 0.35)−0.34 (± 0.54)
V. vitis-idaea−0.95  +1.11+2.77*3.05 (± 1.31)−0.67 (± 1.86)1.39 (± 0.65)3.21 (± 0.92)
E. hermaphroditum−0.38 −2.74−1.07−4.93 (± 2.47)−4.24 (± 3.62)−6.60 (± 3.84)−8.05 (± 4.29)
H. splendens−0.07 +11.6**+2.63−7.20 (± 2.11)−9.91 (± 4.40)1.84 (± 4.07)4.40 (± 2.71)
Litter−0.81  −8.54*−3.11  7.51 (± 3.17)  9.81 (± 2.15)  2.08 (± 3.09)−1.84 (± 3.35)

Effects of increased summer precipitation

Although retrospective analysis of stem length increments from 1993 to 1997 did not show any main effects of precipitation overall, watering significantly increased stem length in the evergreen E. hermaphroditum throughout the 1997 growing season (Fig. 2). Branching was significantly effected in the evergreen species but in the opposite direction (Table 5). The decreased δ13C signatures of E. hermaphroditum leaves (Table 4) proved to occur only at enhanced levels of UV-B.

Figure 2.

Lengths of Empetrum hermaphroditum apical growing stems during 1997 under enhanced UV-B and increased summer precipitation (W).

Reproduction was affected only in the deciduous V. myrtillus and only in 1999 (Table 3). This species also showed an effect on leaf phosphorus in 1994 (Table 4). The other deciduous species V. uliginosum showed increased leaf thickness in 2 years (Table 2). The watering treatment increased the growth of H. splendens but only at enhanced levels of UV-B radiation (Table 1).

The percentage cover of the dwarf shrubs remained unaffected, but H. splendens increased significantly and litter cover decreased with increased precipitation (Table 6). Bud phenology and leaf senescence were not effected (data not shown).

Interactions between enhanced uv-b radiation and increased summer precipitation

The various interactions between enhanced UV-B radiation and increased summer precipitation can be divided into three types. (i) The commonest pattern was for the combined treatment to induce a (small) opposite response to that predicted from combining the effects of the separate treatments. In both V. uliginosum and V. vitis-idaea, both UV-B and increased precipitation decreased stem growth (measured retrospectively) but increased it when in combination (Table 1). One treatment effected each of E. hermaphroditum leaf nitrogen (1998) and δ13C (1994), but in each case the combined treatment more than counteracted this effect (Table 4). V. myrtillus branching (Table 5), and changes in V. vitis-idaea percentage cover (Table 6) and stem weights of H. splendens (Table 1) also showed such interactions. (ii) The combined treatment sometimes reduced the effect due to either treatment alone. Thus UV-B reduced V. myrtillus stem growth more markedly under ambient than increased precipitation (Table 1). Similarly, the development of V. myrtillus buds appeared to be delayed by enhanced UV-B for 1 week  after initial bud burst but only under ambient precipitation (Fig. 3). Increases in leaf thickness (V. uliginosum and V. vitis-idaea) showed this type of interaction in some years (Table 2), as did berry weight in E. hermaphroditum (Table 3). (iii) In the case of H. splendens main axis stem lengths, separate treatments had little effect but the combination caused a large increase (Table 1).

Figure 3.

Development of apical growing buds of Vaccinium myrtillus in 1998 under enhanced UV-B radiation and increased summer precipitation (W).

Discussion

Effects of enhanced uv-b radiation

We hypothesized that enhanced UV-B would be detrimental to the dwarf shrubs studied and the deciduous V. myrtillus was indeed sensitive. Its reduced stem growth, also observed by Johanson et al. (1995b) in the first 2 years under enhanced UV-B, may lead to reduced competitive fitness. Although the mechanism is unknown, possible causes include damage-induced reduction of photosynthesis (for reviews see Bornman & Teramura 1993; Baker et al. 1997), photodestruction of the plant growth regulator indol-3-acetic acid, which facilitates cell expansion (Ros & Tevini 1995), and true morphological preadaptation allowing plants to avoid full exposure to UV-B (Ballaréet al. 1995). Increased flowering and berry production (apparent in this species in most years, even when not statistically significant) may also be disadvantageous; reproduction from seed is rarely successful in this community (Callaghan & Emanuelsson 1985; Callaghan et al. 1992; Phoenix et al. 2000) and may therefore represent a poor use of resources (Carlsson & Callaghan 1990; Karlsson et al. 1990). Shorter term responses of berry production (but not flowering) to UV-B have also been reported within the first 3 years of a similar experiment (Gwynn-Jones et al. 1997). Such changes may have profound effects for vertebrate herbivores as well as the human populations for which it is a food source (M. Anderson 1985).

The increase in leaf thickness observed in 1997 may represent a protective mechanism that places sensitive photosynthetic apparatus deeper within the leaf (Cen & Bornman 1990; Day 1993). The tendency of UV-B to increase leaf thickness in all three Vaccinium species in each of treatment years 5, 6 and 7 (Table 2) contrasts with decreases in V. myrtillus and V. uliginosum in the first 2 years of a similar study (Johanson et al. 1995b) and suggests that such responses can change over time.

We predicted that UV-B would have a greater effect on evergreen than deciduous species, but E. hermaphroditum and V. vitis-idaea showed few responses. Although the leaves of evergreens are exposed for longer periods during the growing season and there is the potential for damage to accumulate over the years (Sullivan & Teramura 1992; Johanson et al. 1995b), their bud burst is later (by 20 days in V. vitis-idaea compared with V. myrtillus and V. uliginosum, Phoenix 2000) and they thus avoid much of the highest radiation period (early to mid growing season, see Johanson et al. 1995a; Fig. 3). UV-B levels can be reduced by > 50% in late season, lessening the advantage of leaf loss in deciduous species. Furthermore, the previous year’s leaves may be increasingly protected by new growth above and the effects of any damage reduced as they contribute less to the plants total photosynthetic capacity (Karlsson 1985).

Our presetting of lamp output results in UV-B overdosing on cloudy days when ambient solar UV is low. Detailed monitoring at Abisko in 1994 by Björn & Holmgren (1996) showed that this would lead to a simulation of 19% rather than 15% depletion of the ozone layer. Such an ozone depletion is still realistic, and while it should be taken into account, this does not invalidate our findings. Indeed, the lack of effects on most species, suggests that Arctic vegetation is less sensitive to enhanced UV-B than previously supposed.

Effects of increased summer precipitation

Some effect of increased precipitation was observed in the evergreen dwarf shrub E. hermaphroditum which increased stem length and branching. This may increase its competitive fitness although increased stem length could prove disadvantageous if buds were then exposed above the insulating snow layer during winter months (Sonesson & Callaghan 1991). The suggestion that water availability may limit growth is supported by the coincidence of the greatest responses in a relatively dry year (1997). Additional precipitation had no effect on stem growth and branching in similar vegetation at a nearby site in the first 2 years of manipulation (Parsons et al. 1994) and a lack of an early effect can also be seen in our study. Increased soil moisture may have led to increased mineralization thus accounting for the greater long-term growth (J. M. Anderson 1991). Alternatively, greater water availability may have increased cell turgor and thus promoted cell expansion or reduced stomatal resistance to gas exchange and thus increased photosynthesis (D. J. Allen et al. 1999). The latter is consistent with the significantly more negative δ13C of E. hermaphroditum leaves under the precipitation treatment and the former with trends towards increased leaf thickness in Vaccinium species.

The significant increase in percentage cover of H. splendens is unlikely to be due to hydration of existing tissue as the first survey was taken after the precipitation treatment had started. Such increases in moss cover could have profound implications for associated plants and their roots as well as the soil system through effects on temperature, humidity, water storage and the alteration of the nutrient content of through flow (see Longton 1997 for overview). Potter et al. (1995) observed an increase in the dry weight of growing shoots but not cover of H. splendens after 3 years of additional summer precipitation on a similar community. It is possible that changes in cover take longer to become apparent or that our greater frequency of watering maintained better hydration (and thus allowed greater growth) of the moss mats.

The reduction in litter cover may be due to either greater rates of decomposition at increased moisture levels, to H. splendens growing over old litter or to watering washing litter from the plots.

Interactions between enhanced uv-b radiation and increased summer precipitation

Our prediction that precipitation would reduce the potential for water stress to alleviate the effects of enhanced UV-B was not generally supported. Rather than exacerbating, combining the two treatments often negated any effect that either treatment had separately. In many of the cases in which interactions occurred, the separate treatments induced the same direction of response. Non-additive effects where therefore unexpected.

We conclude that the effects of future combined increases in UV-B radiation and summer precipitation cannot be predicted from studies of the effects of either perturbation separately. Furthermore, if these environmental changes do occur concurrently, then the effects on this sub-Arctic heathland will be much less than effects resulting from either change separately. This emphasizes that experiments must continue to involve multifactorial manipulations.

The only example of a synergistic interaction in increasing H. splendens stem lengths is somewhat surprising as the leaves of many mosses are one cell thick and have no protective layer and are therefore expected to be particularly susceptible to UV-B damage. It has been suggested by Gehrke et al. (1996) that stimulation by UV-B might rapidly lead to exhaustion of reserves, making long-term positive effects unlikely, but long-term positive effects of UV-B (either alone or in combination) have been observed in our study (e.g. H. splendens stem growth; V. myrtillus flowering and berry production). Enhanced UV-B may not necessarily be an environmental stress, and positive effects should not be considered as transient anomalies.

Cloud cover may increase with summer precipitation and the consequent reduction of incident UV-B might mitigate the effects of ozone depletion. Although such interactions are uncertain, increased precipitation may therefore prove to be more crucial than enhanced UV-B.

Community-level effects

We hypothesized that species-specific responses to UV-B and precipitation would affect interspecific competitive balances. Although changes in H. splendens cover show that responses can be observed at the community level, none of the plant-level changes in dwarf shrubs are detectable as changes in species abundances. It is harder to detect differences in canopy cover than differences in stem and leaf growth and, as much of the biomass was present prior to manipulation, field experiments may need to run over many years before changes are large enough to become apparent (Parsons et al. 1994; Press et al. 1998; Phoenix et al. 2000).

The lack of community change, however, probably indicates tolerance of this vegetation to the simulated environmental changes. Responses to enhanced UV-B radiation and increased summer precipitation do effect growth and reproduction within this sub-Arctic community but they are species- and often season specific. If such heaths are indeed more tolerant of the simulated environmental changes than previously thought, particular concern for their future under increased UV-B and summer precipitation may be misplaced.

Acknowledgements

This research was undertaken as part of the UVECOS consortium and was funded by the CEC (Contract EV5V-CT910032). The Royal Swedish Academy of Sciences provided additional financial support. GP was supported by the Natural Environment Research Council (UK). NERC also funded the carbon stable isotope analysis. We would like to thank Professors M. Sonesson and L. O. Björn, Drs U. Johanson, C. Gehrke, A. Hartley, A. Kyparisses, N. Ekelund (irradiance measurements) and C. Quarmby without whom the work would not have been possible. The field assistance of N. Money and O. Khitun is gratefully acknowledged. Finally, we would like to thank the Abisko Scientific Research Station (Abisko, N. Sweden) for allowing us to conduct these experiments and for providing such excellent technical and logistic support.

Received 11 June 2000 revision received 17 October 2000

Ancillary