Responses of fen and fen-meadow communities to ozone


  • S. A. Power,

    Corresponding author
    1. Department of Environmental Science and Technology, Imperial College, Silwood Park, Ascot, Berkshire SL5 7PY, UK;
      Author for correspondence: S. A. Power Tel: +44 20 75942318 Fax: +44 20 75942339 Email:
    Search for more papers by this author
  • M. R. Ashmore

    1. Department of Environmental Science and Technology, Imperial College, Silwood Park, Ascot, Berkshire SL5 7PY, UK;
    2. Current address: Department of Environmental Science, University of Bradford, West Yorkshire BD7 1DP, UK
    Search for more papers by this author

Author for correspondence: S. A. Power Tel: +44 20 75942318 Fax: +44 20 75942339 Email:


  • •   Species of fen and fen-meadow communities, well supplied with water and nutrients, are characterised by high rates of growth, stomatal conductance values and specific leaf areas, all factors which have been associated with high sensitivity to ozone. We therefore examined the effects of ozone on 12 characteristic fen and fen-meadow species.
  • •   Plants received either filtered air or ozone; AOT40 exposures ranged from 9200 to 14 300 ppb h. Eight of the 12 species exhibited foliar injury in response to ozone exposure, with the first signs of injury on Vicia cracca, following an AOT40 exposure of only 1950 pbb h.
  • •   Ozone exposure significantly reduced plant photosynthetic rate, stomatal conductance and biomass production in four species. Cirsium arvense exhibited the greatest biomass response to ozone (32% and 58% reduction in above- and below-ground weight, respectively). Species with higher levels of visible injury tended to show greater reductions in biomass. There was a significant positive association between stomatal conductance and the magnitude of ozone effects on root biomass.
  • •   The widespread occurrence of either visible injury or growth reductions amongst the species screened, and the magnitude of effects on the most sensitive species, indicate that species of fens and fen-meadows may be more sensitive to ozone than other seminatural ecosystems which have been the focus of recent ozone studies.


It has long been established that ozone is damaging to vegetation, with many studies reporting effects of this pollutant at ambient concentrations throughout Europe (Skarby et al., 1998; Bungener et al., 1999a; Mills et al., 2000). The development of a critical levels approach to establishing thresholds for response to ozone over the past decade has led to an intensification in research into effects on a wide range of plant species. Responses of economically important agricultural crops and forest trees have been studied in most detail with progress now being made towards more refined (level II) critical levels for these groups. However, relatively little research has focused on the impacts of ozone on species of natural and seminatural plant communities, despite the conservation and biodiversity implications of effects on species composition in these habitats (Davison & Barnes, 1998).

Fumigation studies, using realistic concentrations, have shown a range of responses to ozone, including visible injury, reductions in above- and below-ground biomass, altered patterns of assimilate allocation, accelerated senescence and reduced seed production in a wide variety of characteristic European grassland species (Reiling & Davison, 1992a; Ashmore et al., 1995, 1996; Bergmann et al., 1995; Warwick & Taylor, 1995; Whitfield et al., 1997; Franzaring et al., 1998; Bungener et al., 1999b). A wide variation in inter- and intraspecific sensitivity to ozone has been demonstrated amongst grassland species, with evidence for the evolution of resistance to ozone also emerging, for example amongst populations of Plantago major (Reiling & Davison, 1992b; Lyons et al., 1997; Whitfield et al., 1998). However, relatively few studies have examined the response of species from other seminatural plant communities, although Franzaring et al. (2000) did examine the response of wetland species.

Critical levels of ozone have been proposed for seminatural vegetation using the AOT40 index originally developed for crop and forests species (Fuhrer et al., 1997). Values for annual species are currently set at 3000 ppb h over a growing season of 3 months, based on reductions in shoot biomass and seed production in sensitive species; those for perennial species are set at 7000 ppb h over a 6 month growth period, based on reductions in shoot or root weight in sensitive species (UBA, 1996/2001). A critical level has also been set for development of visible injury, of 500 ppb h over 5 d when the mean daytime vapour pressure deficit is above 1.5 kPa (UBA, 1996/2001). However, these critical levels are based on data from grassland species and communities, and very few data exist to assess whether the current critical level is appropriate for other communities, such as wetlands. The reasons for the differences in sensitivity between wild species are uncertain, although a number of factors have been identified as being of possible significance. There is evidence that ozone sensitivity is associated with higher values of stomatal conductance (Nebel & Fuhrer, 1994; Bungener et al., 1999a), higher growth rates (Bungener et al., 1999b) or higher values of specific leaf area ratio, although there is considerable variation in the relationships when different studies in the literature are compared, with only a weak association overall (Franzaring et al., 1998, 1999).

Plants of fen and fen-meadow habitats, with high availability of both nutrients and water, are characterised by high growth rates, high values of stomatal conductance and high values of specific leaf area ratio, all characteristics which have been associated with high ozone sensitivity. Furthermore, the high water availability of such communities in the field is likely to mean that the ozone flux to such communities, unlike that to many grassland systems, is likely to remain high even in the hot dry summers associated with high ozone exposures. It is therefore hypothesised that species of fen and fen-meadow plant communities will be sensitive to ozone. There is little empirical evidence of the responses of species from any wetland communities to ozone, although Franzaring et al. (2000) reported significant effects of ozone inducing premature leaf senescence in several wetland species. In this paper, we report the results of a study of the effects of summertime ozone exposure on 12 species characteristic of UK fen and fen-meadow communities. We examined a range of response variables, including visible injury, shoot and root biomass, flower development and photosynthetic rate, in order to assess which variables might be the best indicators of ozone effects, and the best basis for critical levels for species of these communities.

Materials and Methods

Plant material and growth conditions

Seeds of species characteristic of UK fen and fen-meadow communities were obtained from commercial seed suppliers (Emorsgate Seeds, Norfolk, UK, and John Chambers, Northants, UK). Species were selected to represent those typical of communities classified by the National Vegetation Classification (NVC –Rodwell, 1991, 1994) as S24 (Phragmites australis – Peucedanum palustre tall herb fen), S25 (Phragmites australis – Eupatorium cannabinum tall herb fen), S26 (Phragmites australis – Urtica dioica tall herb fen), S27 (Carex rostrata – Potentilla palustris tall herb fen), M22 (Juncus subnodulosus – Cirsium palustre fen-meadow) and M24 (Molinea caerulea – Cirsium disectum fen-meadow). Seeds were germinated in John Innes commercial seed compost and then transplanted into a soil mix comprising 18 parts peat: 3 parts vermiculite: 1 part perlite: 1 part sand, with additional slow release fertiliser pellets (14 : 13 : 13 NPK) at 2 g l−1 soil. Plants were kept fully watered whilst in the glasshouse and later, during the fumigation period, twice daily irrigation using an automatic watering system ensured that all plants were well supplied with water throughout the experiment. Irrigation periods were of 45 minute duration, at 5 am and 10 pm daily, bringing all pots up to field capacity. In addition, all plants received ambient rainfall. Final species choice was determined by both the availability of a seed supply and germination success.

During two consecutive summers, 1997 and 1998, a total of 12 plant species were exposed to ozone in two separate 2-month long experiments. In 1997, plants of five species were used (Symphytum officinale, Valeriana officinalis, Lythrum salicaria, Iris pseudacorus and Mentha aquatica), with the experiment lasting from June 17th until August 27th. In 1998, seven different species were fumigated with ozone (Vicia cracca, Lathyrus pratensis, Rumex acetosa, Cirsium arvense, Lychnis flos-cuculi, Lotus uliginosus and Filipendula ulmaria) during the period July 11th to August 24th. Valeriana officinalis was also included, as an 8th species, in the second year, to compare results with those from 1997 in order to assess interannual variability in species response. Three plants were transplanted into each of two replicate three litre pots per species, per chamber. All plants were subsequently acclimatised to chamber environments for 5 d before the first day of ozone exposure.

Climatic conditions were similar during the experimental periods in 1997 and 1998. In 1997, mean daily maximum and minimum temperatures at Silwood Park were 23.0°C and 12.3°C, respectively, and average daily rainfall was 2.3 mm. In 1998, the mean daily temperatures were 22.3°C (maximum) and 11.2°C (minimum), while average daily rainfall was 1.0 mm.

Ozone fumigation system and experimental design

Plants were exposed in a system of eight open top chambers (1.5 m height and diameter), at Imperial College's field station in Ascot, Berkshire, UK. Details of chamber design are given by Ashmore et al. (1989). Half of the chambers were ventilated with charcoal filtered air, while the other half received charcoal filtered air with additional ozone. Filtered air chambers received a daytime (09.00–18.00) mean ozone concentration of 17.2 nl l−1 in 1997 and 18.1 nl l−1 in 1998; AOT40 values for filtered air chambers were zero, in both years. Charcoal filtration removed almost 100% of ambient ozone; mean values therefore represent concentrations resulting from ingress of ambient air through the top of fumigation chambers during periods of turbulence, and when chamber doors were open for plant measurements.

Ozone was generated by passing dried, compressed air through an ozone generator (Peak Scientific, Paisley, Scotland), with gas being bubbled through a water trap to remove N2O5 before delivery (Brown & Roberts, 1988). Gas concentrations were monitored using an ozone monitor (Monitor Laboratories, Model 8810; Enviro Technology Ltd., Stroud, UK), which was calibrated at the beginning of each experimental season. A computer controlled solenoid system was used to deliver ozone to half of the chambers between 09.00 and 18.00 daily, with target concentrations of 80 nl l−1. This was increased to 90 nl l−1 in the latter half of 1998 (when fewer days were appropriate for ozone fumigation) in order to reach cumulative AOT40 values which were comparable to those in 1997. Plant pot position was randomised within each chamber and pots were rotated between replicate chambers at regular intervals throughout the fumigation period. Ozone exposure was restricted to warm, dry, still days, where maximum temperatures were forecast to exceed 17°C. In total, 24 plants were exposed to ozone and an additional 24 received filtered air, for each species. Values from replicate plants within pots were meaned to give a pot value; pot means were then averaged to provide a chamber mean value per species, for each measured parameter. Data from the four replicate chambers per treatment were used in all subsequent analyses.

Plant measurements

Initial growth measurements (leaf number and plant height) were carried out immediately before the start of fumigation and used as covariates in subsequent data analysis. Plants were monitored daily for the first signs of visible injury, with injury assessments (number of leaves with light, moderate or severe ozone injury) being carried out at regular intervals following the onset of symptoms. Since the total number of leaves became very large for some species, only the number of injured, and not the total number of leaves, was determined at each injury assessment. Leaves exhibiting ozone injury symptoms were assigned to one of three categories, corresponding to increasing extent and severity of symptoms: light (category 1; < 10% of leaf surface area affected), moderate (category 2; 10–30% of leaf area affected) or severe (category 3; > 30% leaf area affected) injury. In the results section, visible leaf injury is expressed as the sum of moderately and severely injured leaves.

Measurements of both stomatal conductance and photosynthetic rate were made in both summers, using a portable Infra Red Gas Analyser (LCA 2; ADC Ltd, Hertfordshire, UK). The youngest fully expanded leaf from two plants per pot was tagged on the first measurement occasion and the same leaf re-measured a further two (1998) or three (1997) times during the fumigation period. Measurement intervals were chosen to cover the experimental period, and avoided periods of wet weather. All measurements were carried out on bright, sunny days, and all plants were well watered beforehand. Due to small leaf size, these measurements were not carried out for V. cracca or L. pratensis. Reproductive development was monitored at weekly intervals throughout the experiment. On each occasion, counts were made of the number of buds/flowers at different developmental stages. Stages differed for the different species, but were categorised as unexpanded buds, expanded buds, open flowers, petal drop and ovary development. Destructive harvests of above- and below-ground biomass were carried out at the end of each experiment. Above-ground plant material was separated into live and dead, flowering and vegetative parts, before drying (72 h at 80°C) and weighing. Leaf numbers were also counted at the final harvest but are not included in the results as effects mirrored those seen for biomass. Roots were thoroughly washed in sequential bowls of clean water, by hand. Care was taken to ensure that all soil was removed and that all root material was retained. The latter was then dried for 72 h, at 80°C, and weighed. Plant species were harvested sequentially, giving slightly different AOT40 exposures for each species (Table 1).

Table 1.  Cumulative ozone exposures for each plant species
 AOT40 (ppb h)Number of fumigation daysAvg. hours fumigation per day
  1. Lower exposures in 1998 due to a smaller number of fumigation days associated with a shorter experimental period.

 Symphytum officinale11 600467.2
 Valeriana officinalis(1)13 200537.5
 Iris pseudacorus13 200547.4
 Lythrum salicaria13 900567.5
 Mentha aquatica14 300577.5
 Vicia cracca  9 200257.5
 Lathyrus pratensis  9 200257.5
 Rumex acetosa  9 200257.5
 Lychnis flos-cuculi  9 300267.7
 Cirsium arvense10 000237.9
 Lotus uliginosus10 100268.0
 Valeriana officinalis(2)10 200248.2
  Filipendula ulmaria10 600218.8

Statistical analysis

The effects of ozone on plant parameters were analysed using one-way ANOVA in the GLIM statistical package (Crawley, 1993). Regression analysis, also carried out in GLIM, was used to investigate the relationship between stomatal conductance and growth parameters. Root : shoot ratio data were arc-sine transformed prior to analysis, and back-transformed for presentation.


Ozone exposure

Cumulative ozone exposures and the fumigation periods for each of the species are summarised in Table 1. The total number of days of ozone exposure varied between species as a result of either staggered harvest dates (to enable the large quantities of plant material to be processed appropriately) or delayed starts for a given species. The latter resulted from slower germination in 1998, and thus slow seedling growth. Despite a later start in 1998, AOT40 exposures were similar between the two years, as a result of slightly increased numbers of fumigation hours per day and slightly higher target ozone concentrations (as outlined above). The number of hours during which plants received ozone in a given day varied largely as a result of climate; the addition of ozone was suspended during periods of rainfall, and until plant leaves had re-dried. In practice, there were fewer episodes of rain between 09.00 and 18.00 during the experimental period in 1998 than in 1997.

Visible injury

Recognisable ozone injury was observed on eight of the 12 species, with the signs of (moderate or severe) injury first recorded on V. cracca following an ozone exposure of only 1950 ppb h (day six of the experiment). For most species (V. cracca, L. pratensis, L. uliguinosus, V. officinale, F. ulmaria), injury symptoms typically began with the development of small necrotic stippling, in the interveinal areas, particularly in the older leaves. In S. officinale, small necrotic flecks rapidly developed into larger areas of necrosis and were frequently accompanied by bronzing of the upper leaf surface. In M. aquatica, injury was manifest as red mottle, developing into necrosis in the most damaged leaves. Species with small leaves (e.g. V. cracca, L. pratensis) shed damaged leaves, while those with larger leaves (e.g. S. officinale, C. arvense) retained their leaves in a damaged state.

The development of visible symptoms was also rapid for L. uliginosus and C. arvense; moderate/severe injury symptoms were visible on the 6th (L. uliginosus) and 9th (C. arvense) day of exposure, corresponding to AOT40 values of 2050 and 3000 ppb h, respectively. These three species appear to be relatively sensitive to ozone, in terms of early development of visible foliar injury. Injury was seen on L. pratensis following an ozone exposure of approximately 4500 ppb h, while signs of injury were not apparent in the remaining species until after an exposure in excess of 6000 ppb h (F. ulmaria, V. officinalis, M. aquatica and S. officinale). The development of visible injury in relation to ozone exposure is illustrated in Fig. 1. Whilst S. officinale was relatively slow to develop injury symptoms, it was nevertheless the species with the greatest proportion of visibly injured leaves at final harvest in 1997 (20% of leaves in injury categories 2–3). C. arvense exhibited the greatest injury of all affected species in 1998, with 9% of leaves moderately or severely injured at the final harvest.

Figure 1.

Development of visible injury in relation to ozone exposure.

Flower development

Two species (L. salicaria and M. aquatica) flowered during the course of the experiment. Bud formation began earlier and total number of whorls of buds was higher in plants exposed to ozone for both species (Fig. 2). Although the experiment ended before flowering was finished in either species, no effects of ozone were recorded for any other stage of flower development, nor on the total numbers of flowers (of any developmental stage).

Figure 2.

Bud development for Lythrum salicaria and Mentha aquatica during 1997. */*** represent significant treatment effects (P < 0.05/0.001). (*) represents a difference between ozone and filtered air at the P < 0.10 level of probability. Squares, ozone; diamonds, filtered air.

Stomatal conductance

Statistically significant effects of ozone on stomatal conductance were recorded on at least one occasion for five species. In all cases, conductance was lower in plants receiving ozone (Table 2). V. officinalis and I. pseudacorus had significantly lower conductance values in ozone at the end of the experiment (F = 38.47, P < 0.001; F = 16.09, P < 0.01, respectively). S. officinale, M. aquatica and C. arvense experienced significant transient effects on stomatal conductance earlier in the season, although these were no longer significant by the end of the experiment. No significant ozone–time interactions were found for any species.

Table 2.  Stomatal conductance (mol m−2 s−1) and photosynthetic rate (µmol m−2 s−1) in filtered air (F/A) or ozone, on each measurement occasion
  Time interval 1Time interval 2Time interval 3Time interval 4
  1. Species marked with a # were measured in 1997, all others were investigated in 1998. Measurements were taken on different dates for each species in each year, but were spaced more or less evenly throughout the experimental period in all cases. Time interval 1 = 2nd week of July in 1997, 4th week of July in 1998; Time interval 2 = 3rd week of July in 1997, 1st week of August in 1998; Time interval 3 = 1st week of August in 1997, 3rd week of August in 1998; Time interval 4 = 3rd week of August in 1997 only; no. 4th measurement occasion in 1998. V. officinalis (1) = 1997, V. officinalis (2) = 1998. Degrees of freedom = 1,6. */** represent significant treatment F ratios at the P < 0.05/0.01 level of probability.

S. officinale # Gs  0.338  0.319  0.342  0.31  0.296**0.369  0.351  0.497
A  4.65  4.60  2.84  0.13 −2.69 −2.30 −3.08 −5.45
V. officinalis # (1)Gs  0.266  0.284  0.190  0.223  0.345  0.418  0.141***  0.279
A  6.03  6.59  6.54***  7.49  3.82  5.37  1.30*  4.21
L. salicaria # Gs  0.245  0.236  0.190  0.188  0.339  0.340  0.195  0.165
A  8.69  8.06  6.12  5.82  7.16  6.98  5.06  4.02
I. pseudacorus # Gs  0.586  0.582  0.436  0.474  0.402**  0.587  0.233**  0.360
A10.6310.44  9.2511.18  3.93*  6.69  2.53**  6.78
M. aquatica # Gs  0.385  0.224  0.151  0.156  0.282*  0.419  0.130*  0.249
A  7.76  5.29  4.94  3.73  2.93  3.93  1.34  2.21
R. acetosaGs  0.442  0.447  0.283  0.259  0.335  0.275  
C. arvenseGs  0.347*  0.680  0.370  0.623  0.440  0.465  
A  7.02*10.83  4.64**10.41  4.92*  9.50  
L. flos-cuculiGs  0.246  0.242  0.180  0.154  0.186  0.193  
A  9.47  8.43  7.60  6.15  8.20  8.02  
L. uliginosusGs  0.428  0.343  0.307  0.372  0.199  0.235  
A  6.20  6.86  5.28*  8.38  4.78*  6.25  
V. officinalis(2)Gs  0.264  0.299  0.155  0.198  0.119  0.160  
A  6.99  9.35  3.68  5.12  4.60*  8.18  
F. ulmariaGs  0.221  0.197  0.159  0.224  0.240  0.283  
A  6.49  5.72  5.56*  9.19  5.23**  9.62  

Photosynthetic rate

Exposure to ozone resulted in large and significant reductions in photosynthetic rate for the same five species as above (Table 2). For three species (V. officinalis, I. pseudacorus and C. arvense) significant reductions in photosynthesis accompanied reductions in stomatal conductance on at least one occasion. The largest reduction in the rate of photosynthesis was measured for V. officinalis in 1997 (69% lower in ozone on the final measurement occasion), with the magnitude of reduction having steadily increased with increasing exposure to ozone. A significant ozone–time interaction was found for F. ulmaria (F = 4.84, P < 0.05); the magnitude of photosynthetic reduction in response to ozone increased as the experiment progressed for this species.

The pattern of response to ozone differed between species, and can be broadly categorised as either; a stimulation, followed by a reduction in photosynthesis (M. aquatica); a fairly constant reduction in photosynthesis throughout the experiment (C. arvense, L. uliginosus); or an increasing magnitude of reduction as the experiment progressed (V. officinalis, F. ulmaria, I. pseudacorus). Figure 3 illustrates temporal changes in plant photosynthetic response in relation to ozone exposure, for selected species.

Figure 3.

Relationship between AOT40 exposure and effects on plant photosynthetic rate for (a) Valeriana officinalis (b) Mentha aquatica and (c) Cirsium arvense. */**/*** represent significant treatment F-ratios at the P < 0.05/0.01/0.001 levels of probability. The solid line represents ozone exposure (AOT40) at different time intervals for each species (values correspond to 2nd Y axis). Light grey columns, ozone; dark grey columns, filtered air.

Plant biomass

A general trend was observed towards a reduction in both above- and below-ground biomass in response to ozone exposure (Table 3), with M. aquatica being a notable exception. Significant reductions in above-ground biomass were found for S. officinale, V. officinalis (1997) and F. ulmaria. The effect of ozone on above-ground weight of C. arvense was not quite statistically significant, although a 42% reduction was found. V. officinalis was also affected below-ground, with a highly significant reduction in root weight. In general, below-ground biomass was more affected by ozone than above-ground, as is reflected in the general trend for lower root : shoot ratios; three of the species had significantly lower root : shoot ratios in ozone, with a further three species having root : shoot ratios more than 10% lower in ozone compared with filtered air, although the latter were not statistically significant. The greater weights recorded for the first five species listed reflects the longer experimental period in 1997 compared with 1998.

Table 3.  Above- and below-ground biomass (g dry weight) and root:shoot ratios for all species
 Live (leaf + stem) weightRoot weightRoot : shoot ratio
 O3F/AF ratioO3F/AF ratioO3F/AF ratio
  1. Live (leaf + stem) weights exclude senescent/dead leaves (and flower stalks for L. salicaria and M. aquatica). */**/*** represent significant treatment effects at the P < 0.05/0.01/0.001 level of probability. Degrees of freedom = 1,6. (1) and (2) refer to 1997 and 1998 data for V. officinalis, respectively.

S. officinale18.323.7  9.82*42.754.2  3.561.711.85  0.24
V. officinalis(1)14.318.2  6.50*25.036.554.6***1.581.9510.02*
L. salicaria48.352.4  0.9645.852.1  3.410.890.92  0.25
I. pseudacorus  7.2  7.4  0.06  7.8  8.8  4.411.041.16  3.91
M. aquatica13.4  8.511.12*  3.9  3.4  4.48
R. acetosa  6.8  7.1  0.36  4.3  4.4  0.080.620.61  0.01
V. cracca  2.0  2.5  1.69  1.1  1.5  3.090.580.63  2.12
L. pratensis  1.4  1.3  0.43  0.9  1.1  7.22*0.660.8314.58**
C. arvense  2.7  4.0  5.48  0.9  2.2  7.12*0.330.5217.15**
L. flos-cuculi  4.2  3.6  3.78  1.2  1.2  0.480.300.30  0.13
L. uliginosus  4.3  3.8  1.90  1.1  1.2  2.15
V. officinalis(2)  1.8  2.1  2.99  1.2  1.4  0.380.68075  0.44
F. ulmaria  1.1  1.4  6.13*  0.8  0.8  4.820.710.65  6.57*

The weight of dead leaves was significantly increased in ozone for R. acetosa (F1,6 = 8.041, P < 0.05) and V. officinale (1998) (F1,6 = 19.74, P < 0.01). For all other species, the dead leaf biomass was not affected by ozone.


The most striking aspect of the experimental results was the number of species (nine of the 12 studied) that were negatively affected by ozone in terms of either visible injury or detrimental effects on growth or physiology. Initial species selection did not involve any prior knowledge about relative sensitivities to ozone and was largely constrained by seed supply and germination success. However, it should be noted that genotypic differences amongst native plant species may result in large intraspecific variation in sensitivity to ozone (Davison & Barnes, 1998), with the results from this experiment therefore representing only the sensitivities of the genotypes used. Hence any conclusions about the relative sensitivity of the 12 species need to be made with caution. Nevertheless, the fact that significant adverse effects were found on 75% of the species at ozone exposures comparable to those experienced over many areas of Europe supports the original hypothesis that species of fen and fen-meadow habitats are very sensitive to ozone.

In terms of visible injury, three groups of species appeared: those showing no visible injury, those only showing visible injury after substantial cumulative ozone exposure (4500 ppb h or more), and those showing visible injury after an exposure of 1500–2500 ppb h. By contrast, Franzaring et al. (2000) found visible injury on only one of 10 wetland species subjected to comparable or greater ozone exposures. Although differences in experimental procedures or climatic conditions may account for this difference, it more likely reflects real differences in sensitivity between the species used, as the only two species common to both our study and that of Franzaring et al. (2000) –Lychnis flos-cuculi and Lythrum salicaria– showed no visible injury symptoms in either study. Visible injury was reported for L. flos-cuculi grown in open top chambers by Bungener et al. (1999a), although in the latter study, symptoms did not appear until after an ozone exposure of more than 15 000 ppb h, and then only a relatively small proportion of leaves were affected. This, together with a lack of effect on above-ground biomass, suggests that L. flos-cuculi is not a particularly sensitive species.

Skelly et al. (1999) showed that widespread occurrence of foliar injury symptoms in the field was linked to ozone sensitivity of native vegetation in southern Europe, suggesting that a significant proportion of naturally occurring plant communities may be experiencing phytotoxic concentrations of ozone. The fact that some of our species showed visible injury after exposure to 1500–2500 ppb h over a few days suggests that such injury may be commonly found in the field and that the short-term critical level of 500 ppb h is appropriate for these wetland species. Visible foliar injury is, however, not generally regarded as a good indicator of overall plant sensitivity to ozone, since injury does not necessarily lead to a reduction in biomass or reproductive success (Pleijel & Danielsson, 1997). Furthermore, biomass reductions have been frequently reported in the absence of any signs of visible injury (Mortensen & Nilsen, 1992; Reiling & Davison, 1992a). In this study, five of the eight species showing injury symptoms also had statistically significant reductions in biomass. Conversely none of the species which failed to exhibit visible signs of injury experienced significant reductions in live weight. Hence, our study suggests that, for these species, visible injury is a reasonable indicator of ozone sensitivity. Rumex acetosa was the only species which had an increased number (and biomass) of dead leaves, in the absence of either ozone injury or effects on live biomass. This, in combination with the fact that only one of the species which were visibly injured had a significant increase in dead leaves suggests that, under these experimental conditions, injury does not translate into increased leaf death, but rather reduced growth.

The inclusion of V. officinalis in both summer-long experiments allows an assessment of interannual variability in species response to ozone. Physiological responses and the development of visible injury in this species were very similar between years. Despite much lower biomass production of all species in 1998, the proportional above-ground weight reduction for V. officinalis, whilst not quite statistically significant, was of a similar magnitude to that found in 1997. Below-ground effects were however, much smaller in the second year and may reflect the shorter experimental season and lower ozone exposures in 1998, compared with 1997.

Three of the species in this study showed significant reductions in above-ground biomass, while three also showed significant reductions in root biomass and root : shoot ratio. By contrast, one species (Mentha aquatica) showed a significant stimulation in above-ground biomass. Stimulations of above-ground growth of certain species in response to moderate ozone exposures have been reported in other studies (Franzaring et al., 2000Molinia caerulea; Bungener et al., 1999aCentaurea jacea, Crepis biennis, Taraxacum officinalis), although the mechanisms of this increased growth are unclear.

There was an overall trend in our data for reduced root : shoot ratios following ozone exposure, with three species having a significantly reduced proportion of biomass below-ground, and a further three species exhibiting a nonsignificant reduction of at least 10%. By contrast, Warwick & Taylor (1995) found variable root responses of calcareous grassland species under ozone stress. These differences may reflect different priorities in resource allocation in species adapted to different levels of nutrient and water availability. Ozone has also been shown to induce changes in partitioning between vegetative and reproductive growth in different directions in individual species (Bergmann et al., 1995). Our data show a stimulation of flower bud production by ozone. By contrast, Franzaring et al. (2000) found no effect of ozone on flower production of wetland species but a stimulation of flower biomass.

Whilst this study was not designed to explicitly investigate mechanisms responsible for plant sensitivity to ozone, the data were investigated to test if a relationship exists between stomatal conductance and the magnitude of growth response to ozone. The relationship between stomatal conductance and reductions in above-ground weight was not significant (r2 = 0.081), but there was a significant positive relationship (Fig. 4) between stomatal conductance and the reduction in ozone for both root biomass (F1,8 = 6.136, P < 0.05; r2 = 0.434) and root : shoot ratio (F1,8 = 5.427, P < 0.05; r2 = 0.404). These results provide support for the hypothesis that stomatal conductance, through its role in controlling ozone flux into leaves, might be a good predictor of species sensitivity to ozone. Many studies of ozone impacts on seminatural vegetation have not reported effects on below-ground biomass, and the fact that earlier analyses of relationships between growth and physiological characteristics and responses to ozone have used effects on shoot or leaf biomass (Franzaring et al., 1998) may be an important constraint to their interpretation. Furthermore, current critical levels are based on a criterion of reduced shoot biomass, while our data suggest that further consideration should be given to whether root biomass may be a more sensitive criterion of ecological relevance.

Figure 4.

Relationship between stomatal conductance and ozone-induced reductions in (a) root weight and (b) root : shoot ratio.

Those species which did not experience significant reductions in photosynthetic rate were also unaffected in terms of biomass (L. salicaria, R. acetosa, L. flos-cuculi) whereas most of those which consistently had lower rates of photosynthesis in ozone experienced the greatest effects on growth (V. officinalis, C. arvense, F. ulmaria). Responses to ozone for M. aquatica were interesting, with both stomatal conductance and photosynthetic rate being approx. 50% higher in ozone on the first measurement occasion (AOT40 of 3200 ppb h), then becoming progressively reduced in response to increased ozone exposure. This early stimulation of photosynthesis may explain the significant increase in above-ground biomass recorded at the final harvest. However, if the experiment had been of a greater duration, perhaps even for more than a single growth season, adverse effects on biomass may have developed. By contrast, Franzaring et al. (2000) reported that the negative effects of ozone after 28 days exposure of wetland species were reduced in scale at the end of the season, possibly because of a greater tolerance of the later-season leaves.

Since this was not designed as a dose–response experiment, it is not possible to identify thresholds for effects which might be used to identify critical levels. However, the magnitude of biomass reductions observed for the most affected species (as high as 58% for C. arvense roots and 30% for V. officinalis roots) implies, assuming linear relationships with AOT40 (Fuhrer et al., 1997), that a 10% reduction in root biomass would occur at 2000–4000 ppb h, a value comparable to the critical level of 3000 ppb h set to prevent 10% loss of shoot biomass in annual species. Hence, based on the genotypes screened, certain species of fen and fen-meadow communities may be amongst the most sensitive species to ozone. Furthermore, if factors which increase ozone flux into leaves (e.g. low soil moisture deficit and vapour pressure deficit) result in a greater plant response (Bungener et al., 1999a; Emberson et al., 1999, 2000), fen and fen-meadow communities, experiencing little soil moisture deficit in the field, might be expected to be among the most sensitive seminatural communities. For example, in the UK, the highest summer concentrations of ozone are found in the warmer, drier southern and eastern parts of the country, where soil moisture availability may be a more important determinant of plant community sensitivity. Many fen and fen-meadow communities occur in these areas and can therefore be expected to experience relatively high ozone fluxes during the growing season. Further research should therefore be directed at these (and similar) habitats to quantify the medium to long-term impacts of ambient ozone concentrations and to establish appropriate critical levels for the protection of the most sensitive European plant communities.


This study was supported by funding from the former UK Department of the Environment, Transport and the Regions.