Tropospheric O3 moderates responses of temperate hardwood forests to elevated CO2: a synthesis of molecular to ecosystem results from the Aspen FACE project


†Author to whom correspondence should be addressed. E-mail:


  • 1The impacts of elevated atmospheric CO2 and/or O3 have been examined over 4 years using an open-air exposure system in an aggrading northern temperate forest containing two different functional groups (the indeterminate, pioneer, O3-sensitive species Trembling Aspen, Populus tremuloides and Paper Birch, Betula papyrifera, and the determinate, late successional, O3-tolerant species Sugar Maple, Acer saccharum).
  • 2The responses to these interacting greenhouse gases have been remarkably consistent in pure Aspen stands and in mixed Aspen/Birch and Aspen/Maple stands, from leaf to ecosystem level, for O3-tolerant as well as O3-sensitive genotypes and across various trophic levels. These two gases act in opposing ways, and even at low concentrations (1·5 × ambient, with ambient averaging 34–36 nL L−1 during the summer daylight hours), O3 offsets or moderates the responses induced by elevated CO2.
  • 3After 3 years of exposure to 560 µmol mol−1 CO2, the above-ground volume of Aspen stands was 40% above those grown at ambient CO2, and there was no indication of a diminishing growth trend. In contrast, O3 at 1·5 × ambient completely offset the growth enhancement by CO2, both for O3-sensitive and O3-tolerant clones. Implications of this finding for carbon sequestration, plantations to reduce excess CO2, and global models of forest productivity and climate change are presented.


Global atmospheric CO2 concentrations have risen by nearly 30% since pre-industrial times (Barnola et al. 1995; Stott et al. 2000; IPCC 2001). These increases are primarily due to fossil fuel emissions (Keeling et al. 1995). Similarly, emissions of oxidized nitrogen (NOx) and volatile organic compounds from fossil fuel combustion have increased background concentrations of O3 (Finlayson-Pitts & Pitts 1997; Fowler et al. 1998; Stevenson et al. 1998; Ryerson et al. 2001), which have risen some 36% over the same period (IPCC 2001). Fowler et al. (1999a, 1999b) suggest that nearly a quarter of the Earth's forests are currently at risk from tropospheric O3 where peak concentrations exceed 60 nL L−1. They further predict that half of the Earth's forests will be subjected to peak concentrations exceeding 60 nL L−1.

Thousands of studies have been conducted to examine the impacts of elevated CO2 (Ceulemans & Mousseau 1994; Saxe, Ellsworth & Heath 1998; Norby et al. 1999; Körner 2000) and O3 (Chappelka & Samuelson 1998; Matyssek & Innes 1999; Bortier, Ceulemans & Temmerman 2000) on plant growth and biomass accrual. Many of these studies have been confounded by the artificial greenhouse conditions inside the exposure chambers (Olszyk, Tibbitts & Hertzberg 1980; McLeod, Fackrell & Alexander 1985). They have been limited by the available space to include only single trees or a few young, immature trees. Thus there is a need for larger-scale and longer-term studies to examine the impact of these gases on ecosystem structure and function (Heck et al. 1998).

Elevated CO2 and O3 affect trees in opposite ways. Elevated CO2 stimulates photosynthesis (Tjoelker, Oleksyn & Reich 1998; Noormets et al. 2001a, 2001b) and growth above ground (Norby et al. 1999) and below ground (King et al. 2001; Kubiske & Godbold 2001), and delays autumnal foliar senescence (J.G. Isebrands, unpublished results). Trees grown under elevated CO2 generally have lower nitrogen concentrations in their foliage (Cotrufo, Ineson & Scott 1998), lower Rubisco concentrations (Moore et al. 1999), and altered concentrations of defence compounds (Lindroth, Kinney & Platz 1993; Lindroth et al. 1997) and of antioxidants and other secondary metabolites (Norby et al. 2001a; Wustman et al. 2001).

In contrast to the largely beneficial effects of CO2, O3 is generally detrimental to tree growth and forest productivity. Ozone induces foliar injury (Karnosky 1976), decreases foliar chlorophyll content (Gagnon et al. 1992), accelerates leaf senescence (Karnosky et al. 1996), decreases photosynthesis (Coleman et al. 1995a), alters carbon allocation (Coleman et al. 1995b) and epicuticular wax composition (Mankovska, Percy & Karnosky 1998; Karnosky et al. 1999, 2002a), predisposes trees to attack by pests (Stark et al. 1968; Karnosky et al. 2002a) and decreases growth (Wang, Karnosky & Borman 1986; Karnosky et al. 1992, 1996, 1998). Extrapolation of open-top chamber O3 exposures of Aspen to native Aspen stands suggests that 14–33% growth decreases could occur over 50% of its range in the eastern USA (Hogsett et al. 1997).

Current climate change scenarios predict further increases in atmospheric CO2 (Stott et al. 2000) and O3 concentrations (Stevenson et al. 1998; Fowler et al. 1999a, 1999b) over the next century. Little research has been done on the interactive impacts of these pollutants. Furthermore, conflicting results have been reported, even for a given species. For example, Volin & Reich (1996) and Volin, Reich & Givnish (1998) suggest that CO2 ameliorates the effects of O3 on trembling Aspen (Populus tremuloides Michx.) while Kull et al. (1996), McDonald et al. (2000, 2002), Sôber et al. (2002), Isebrands et al. (2001) and Wustman et al. (2001) suggest that CO2 does not ameliorate, but sometimes exacerbates the negative impacts of O3. Thus we do not fully understand how forest tree growth, or the composition and functioning of forests, will be influenced by interacting CO2 and O3.

The FACTS II (Aspen FACE) project was established in 1997 as the first open-air facility to examine the responses of forest trees to interacting CO2 and O3 (Dickson et al. 2000). Our objective was to examine how elevated atmospheric CO2 and O3 will affect the carbon and nitrogen cycles and ecological interactions of forests. Specifically, we are studying the impacts of these co-occurring greenhouse gases on aggrading northern forests in terms of carbon sequestration, physiological processes, growth and productivity, competitive interactions and stand dynamics, interactions with pests, and ecosystem processes such as foliar decomposition, mineral weathering and nutrient cycling.

This review (1) summarizes early results from Aspen FACE which show remarkable consistency from molecular through to ecosystem levels, in that relatively low O3 concentrations offset the responses of Aspen and Birch to elevated CO2; (2) places our findings in regard to forest productivity in the context of other CO2/O3 interaction studies; (3) draws implications in terms of global modelling; (4) summarizes effects on higher trophic levels; and (5) addresses research gaps and opportunities to better understand ecosystem responses to long-term exposures to interacting CO2 and O3.

summary of exposure methods and plant materials

The Aspen FACE project is a full factorial experiment with three replicate 30 m diameter rings of four treatments: control (ambient CO2, ambient O3); elevated CO2 (560 µmol mol−1 CO2 vs. ambient CO2 of ≈360 µmol mol−1; elevated O3 (1·5 × ambient); and elevated CO2 plus O3 (Fig. 1). The experiment was planted in July 1997. The two gases have been administered during daylight hours from budbreak to budset during 1998–2001 for total growing seasons of 166, 143, 139 and 143 days, respectively. Daytime CO2 concentrations in elevated CO2 treatments averaged 530, 548, 548 and 541 µmol mol−1 for the four growing seasons. The respective 1 min average CO2 concentrations were within 10% of the target for 78·5, 74·0, 67·3 and 71·2% of the time, and within 20% of the target for 94·0, 93·0, 91·9 and 92·7% of the time. O3 exposures, which are summarized in Figs 2 and 3, averaged 54·5, 51·1, 48·9 and 52·8 nL L−1 (12 h daytime mean during the growing season) compared to control ring O3, which averaged 34·6, 36·9, 36·0 and 36·6 nL L−1 for the same period. The growing season doses (SUM 00) for daylight hours were 97 900, 87 900, 78 800 and 90 700 nL L−1 h for O3 treatments, compared with control values at 59 100, 62 800, 58 200 and 66 100 nL L−1 h. The CO2 and O3 concentrations were chosen to represent the predicted atmospheric concentrations of these gases in the northern Great Lakes Region in the year 2050.

Figure 1.

Aerial view of the Aspen FACE project showing eight of 12 30 m diameter exposure rings (top). Each ring is divided into three sections as shown (bottom left). The gases are released through slots in the vertical vent pipes as shown in the bottom right photo.

Figure 2.

Seasonal mean O3 concentrations for ambient air (▴) and one elevated O3 ring (▪) in the Aspen FACE project.

Figure 3.

Hourly O3 concentrations from ambient air (black) and one elevated O3 ring (grey) in the Aspen FACE project during 1998–2001.

The rings were planted using 3–6-month-old potted plants in midsummer 1997. The eastern half of each ring was randomly planted at 1 × 1 m spacing in two tree plots of five Aspen clones differing in O3 tolerance (8L, 216 and 271 = relatively tolerant; 42E and 259 = relatively sensitive). The remaining half of each ring was further subdivided with a quarter ring being planted with alternating Aspen clone 216 and Birch (Betula papyrifera) and a quarter ring being planted with alternating Aspen clone 216 and Sugar Maple (Acer saccharum). Trembling Aspen is the most widely distributed tree species in North America. Aspen and Birch are the most important pulpwood species of the Great Lakes region, comprising over 70% of the round wood harvest (Piva 1996). According to the International Poplar Commission, the Aspen forest types make up more than 8·8 million ha in the USA and 17·8 million ha in Canada (Isebrands et al. 2001). Aspen–Birch–Maple stands are also important aesthetic components of northern forests. More details regarding the plant material, planting design, and the generation, dispensing and monitoring of CO2 and O3 are presented by Karnosky et al. (1999) and Dickson et al. (2000).

Synthesis of results

A summary of key results from the Aspen FACE project's establishment years, from time of plantation establishment in 1997 to crown closure in 2000, is shown in Table 1. The results are consistent across functional groups, from leaf biochemistry, gene expression and gas exchange through to ecosystem level, and across trophic levels in that elevated CO2 and O3 frequently exert opposite effects. When the two gases co-occur, low concentrations of ambient O3 offset or substantially moderate the responses attributable to elevated CO2.

Table 1.  Summary of responses of Trembling Aspen to elevated CO2 (+200 µmol mol−1), O3 (1·5 × ambient), or CO2 + O3 compared with control during 3 years of treatments at the Aspen FACE project
 CO2O3CO2 + O3Source
  • *

    Responses are shown as small but significant increases (↑), large and significant increases (↑↑), small but significant decreases (↓), large and significant decreases (↓↓), non-significant effects (n.s.) compared to trees grown in control rings with ambient CO2 and O3. Foliar analyses and leaf surface properties were largely determined from recently mature leaves of all three species during mid-season. Gas-exchange data were taken from all leaf ages and throughout the growing season.

  • RbcS = small subunit of Rubisco; PAL = phenylalanine ammonialyase; SOD = super oxide dismutase; ACC = 1-aminocyclopropane-1-carboxylic acid; C = carbon; N = nitrogen; Amax = maximum photosynthesis rate; LAI = leaf area index; NPP = net primary productivity.

Foliar gene expression and biochemistry
Rubisco*↓↓Wustman et al. (2001); Noormets et al. (2001a)
RbcS transcripts↓↓Wustman et al. (2001)
Chalcone synthase transcriptsn.s.n.s.n.s.Wustman et al. (2001)
PAL transciptsWustman et al. (2001)
SODn.s.n.s.n.s.Wustman et al. (2001)
ACC oxidaseWustman et al. (2001)
Ascorbate peroxidasen.s.Wustman et al. (2001)
CatalaseWustman et al. (2001)
Glutathione reductaseWustman et al. (2001)
Phenolic glycosidesn.s.Lindroth et al. (2002); Kopper & Lindroth (2002)
Tanninsn.s.Lindroth et al. (2001); Kopper & Lindroth (2002)
Foliar nitrogenn.s.Lindroth et al. (2001); Kopper & Lindroth (2002)
C : N ratio of foliagen.s.↑↑Lindroth et al. (2001)
Starchn.s.Wustman et al. (2001)
Gas exchange
Amax lower canopyn.s.↓↓↑(young leaves) ↓(older leaves)Takeuchi et al. (2001); Noormets et al. (2001a)
Amax whole canopy↑↑↓↓n.s.Noormets et al. (2001b)
Carboxylation efficiencyn.s.↓↓Sôber et al. (2002)
Stomatal limitationn.s.Noormets et al. (2001a)
Stomatal conductance↓↑ôber et al. (2000), Noormets et al. (2001a)
Foliar respirationn.s.n.s.Takeuchi et al. (2001), Noormets (2001)
Soil respiration↑↑n.s.King et al. (2001)
Microbial respiration↑↑n.s.n.s.Phillips et al. (2002)
Stomatal densityn.s.n.s.n.s.Percy et al. (2002a)
Chlorophyll contentWustman et al. (2001)
Chloroplast structureOksanen et al. (2001), Takeuchi et al. (2001), Wustman et al. (2001)
O3 flux↑↑Noormets et al. (2001a)
Growth and productivity
Leaf thicknessn.s.n.s.Oksanen et al. (2001)
Leaf sizeWustman et al. (2001)
Leaf arean.s.Noormets et al. (2001b)
LAIn.s.Noormets (2001)
Height growthn.s.Isebrands et al. (2001)
Diameter growthn.s.Isebrands et al. (2001)
Volume growthn.s.Isebrands et al. (2001)
Leaf biomassn.s.McDonald & Isebrands, unpublished
Stem biomassMcDonald & Isebrands, unpublished
Coarse root biomassn.s.King & Pregitzer, unpublished
Fine root biomassn.s.King et al. (2001)
Fine root turnovern.s.n.s.King et al. (2001)
Spring budbreakn.s.Delayedn.s.Isebrands & Karnosky, unpublished
Autumn budsetDelayedEarlyn.s.Isebrands & Karnosky, unpublished
Foliar retention (autumn)↑↑↓↓n.s.Isebrands & Karnosky, unpublished
Wood chemical composition
Ligninn.s.n.s.Anttonen et al. (2001)
Cellulosen.s.n.s.n.s.Anttonen et al. (2001)
Hemicellulosen.s.n.s.n.s.Anttonen et al. (2001)
Extractivesn.s.n.s.n.s.Anttonen et al. (2001)
Leaf surfaces
Crystalline wax structure↓↓Karnosky et al. (1999); Karnosky et al. (2002a)
Stomatal occlusion↑↑Karnosky et al. (1999)
Wax amountn.s.Karnosky et al. (2002a); Percy et al. (2002a)
Wax chemical compositionn.s.Changen.s.Karnosky et al. (2002a)
Wax fatty acid de novo synthesis↑↑↑↑Percy et al. (2002a)
Wax hydrocarbon biosynthesis↑↑↑↑n.s.Percy et al. (2002a)
Wax carbon-chain length↑↑Karnosky et al. (2002a)
Wettabilityn.s.↑↑Karnosky et al. (2002a)
Trophic interactions
Melampsora leaf rustn.s.↑↑↑↑Karnosky et al. (2002a); Percy et al. (2002b)
Aspen aphidsn.s.n.s.n.s.Percy et al. (2002b)
Birch aphidsn.s.n.s.n.s.Awmack & Lindroth, unpublished
Aspen Blotch LeafminerKopper & Lindroth (2002)
Forest Tent Caterpillarn.s.n.s.Kopper (2001), Kopper & Lindroth (2002)
Oberea woodborern.s.Mattson, unpublished
Ecosystem level
NPPn.s.Kruger & McDonald, unpublished
Litter decomposition (k value)n.s.Parsons et al. (2000)
Nutrient mobilization↑↑n.s.King, unpublished
Water-use efficiencySober et al. (2000)
Soil moistureKing, unpublished
Competitive indices↓↓McDonald et al. (2000); McDonald et al. (2002)
Soil invertebrate diversityn.s.Loranger & Pregitzer, unpublished
Microbial enzymesn.s.n.s.Phillips et al. (2002); Larson et al. (2002)
Microbial biomassn.s.n.s.Philllips et al. (2002); Larson et al. (2002)

gene expression and biochemistry

Plants largely respond to stress by changes in assimilation of carbon and in the repartitioning of other resources. Elevated CO2 and O3 are sensed primarily by leaves (Dickson & Isebrands 1991) and result in dynamic and rapid changes in gene expression (Noormets, Podila & Karnosky 2000) and gas exchange (Hendrey et al. 1997). We have documented O3-induced stimulations of transcript production of several antioxidants, including ascorbate peroxidase, catalase and glutathione reductase (Wustman et al. 2001). Interestingly, these same antioxidants appear to be downregulated under elevated CO2, regardless of O3 exposure, as was phenylalanine ammonia-lyase (PAL), a key enzyme in the shikimic acid pathway. CO2- and O3-induced decreases in transcripts of the small subunit of Rubisco were closely linked to independently measured decreases in Rubisco concentrations (Noormets et al. 2001a). Decreases in chlorophyll content, as measured by Wustman et al. (2001), were consistent with the degradation of chloroplasts (Oksanen, Sober & Karnosky 2001) under elevated O3.

gas exchange

The three tree species that we examined have differing photosynthetic responses to CO2, O3 and CO2 + O3. In the response of the upper crown to elevated CO2, the rapid-growing, early successional species showed significant increases in light-saturated CO2 assimilation rate (Amax): 20–33% for Aspen (Noormets et al. 2001a, 2001b; Sôber et al. 2003), and 50–72% for Birch (Takeuchi et al. 2001). These values were seen consistently from year 1, and no acclimation to CO2 has yet been seen in sun leaves of Aspen and Birch. The relative limitation imposed by stomatal conductance on Amax in Aspen declined under elevated CO2, indicating that upper canopy leaves operated closer to their CO2-saturated rate (Noormets et al. 2001a). Improvements in shade photosynthesis of Aspen and Birch under elevated CO2 were small, so that net gains in daily canopy C fixation were largely realized at the top of the canopy and were driven by increases in Amax (Takeuchi et al. 2001). We found minimal effects of CO2 or O3 on leaf dark respiration, with significant late-season increases in respiration only under elevated O3 (Noormets 2001). Thus elevated CO2 increased upper canopy Amax in Aspen by 33% and in Birch by 64% (Fig. 4), but not in Sugar Maple. Moreover, we found that O3 and CO2 + O3 reduced the carboxylation efficiency of Aspen and Birch, but did not reduce Amax (Sôber et al. 2003). The declines in capacity were sufficient to eliminate any increase in Amax due to elevated CO2 in Aspen and Birch (compare CO2 and CO2 + O3 treatments in Fig. 4). Elevated O3 did not reduce photosynthetic capacity or Amax in Sugar Maple.

Figure 4.

Light-saturated CO2 assimilation rates of Aspen, Birch and Maple growing under experimental atmospheric CO2 and O3 treatments. Data represent the mean and SE of three trees from each of three replicates for three to five measurement times over the 1999 and 2000 growing seasons.

In addition to species-specific responses, we also observed substantial variation in photosynthetic assimilation among Aspen genotypes. Elevated CO2 stimulated Amax to a similar degree in two Aspen genotypes of contrasting sensitivity to O3. However, elevated O3 reduced Amax slightly in an O3-tolerant genotype (clone 216) and more so in an O3-sensitive genotype (clone 259). Elevated CO2 counteracted this reduction in the O3-tolerant genotype, in that Amax was 35% greater than that of clone 216 plants in the controls (Noormets et al. 2001a). In the O3-sensitive genotype, however, Amax in the CO2 + O3 treatment was equivalent to that of clone 259 individuals from the controls. These responses occurred across leaves of different developmental stages and throughout the growing season.

Significant effects of elevated CO2 and O3 were also found for stomatal conductance (gs). Generally, the stomata of Maple were much more responsive than those of Aspen and Birch to changing environmental conditions (light, CO2 and relative humidity) across all treatments. Elevated CO2 tended to reduce gs for all three species, as expected (Noormets et al. 2001a; A. Sôber & P. Sharma, unpublished results). The largest decreases in gs for Maple were under the combined CO2 + O3 treatment (A. Sôber & P. Sharma, unpublished results).

Increased leaf Amax of Aspen and Birch under elevated CO2 was accompanied by significantly greater canopy leaf-area production (Fig. 5), a finding we believe is simply linked to the trees being larger in the elevated CO2 rings, as we have not found changes in allometry associated with treatments. The O3-induced decline in Amax of Aspen (Fig. 4) was also reflected in the decreased LAI of the Aspen canopies. Interestingly, Paper Birch had no O3-induced decline in Amax, and the Aspen–Birch canopies exhibited no significant decrease in LAI under elevated O3, reflecting the contribution of Birch to total canopy leaf area. The combination of elevated CO2 + O3 did not affect canopy LAI of either pure Aspen or Aspen–Birch.

Figure 5.

Estimates of leaf area index (LAI, m2 leaf area per m2 ground area) for Aspen and Aspen–Birch stands in mid-season 2000. A destructive subsample harvest in August 2000 was used to calibrate the relationship between total leaf area and stem basal diameter. Significant linear regressions of log–log-transformed leaf area and diameter data were developed (R2 = 0·80, P < 0·0001), with specific intercepts for each species. Whole-stand estimates of LAI were obtained by estimating individual tree leaf area from mid-season basal diameter measurements using the regression equations, summing the total leaf area of all trees in each stand, and dividing by the total ground area in each stand type (means ± 1 SE).

The differences in Amax and LAI that we observed indicate that CO2 and O3 will alter carbon assimilation into terrestrial ecosystems in a manner consistent with the physiological response of the dominant vegetation. This key observation supports our hypothesis that ecosystem C assimilation, allocation and cycling are strongly influenced by the life-history traits of the dominant plant taxa, coupled with the manner in which plants increase or decrease amounts of C and N allocated to plant growth, storage and defence.

growth and productivity

The photosynthetic responses of the species and genotypes in our experiment have influenced growth and litter production above and below ground (Parsons, Bockheim & Lindroth 2000; Isebrands et al. 2001; King et al. 2001). Enhanced rates of photosynthesis under elevated CO2 contributed to increased above-ground growth of Aspen and Birch (Fig. 6). In contrast, decreases in photosynthesis by O3 depressed above-ground growth in Aspen, but not Birch. Interacting CO2 and O3 resulted in intermediate responses in Aspen and Birch, such that these treatments did not generally differ significantly from controls. Although Maple photosynthesis exhibited little response to any of the treatments, we did observe an overall negative above-ground growth response to all treatments. Nevertheless, variability in above-ground growth was extremely high in the smaller Maple saplings (CV = 250%), and it is premature to conclude that the overall growth of Maple will decline as CO2 and O3 accumulate in the atmosphere.

Figure 6.

Volume growth index (diameter2 × height) of Aspen (a), Birch (b) and Maple (c) grown for 3 years in full factorial combinations of elevated atmospheric CO2 (solid lines) and elevated O3 (solid symbols).

As with photosynthesis, above-ground growth responses varied widely in magnitude among Aspen genotypes. For example, the response to elevated CO2 was least in genotype 42E and greatest in genotype 271 (Isebrands et al. 2001). Similarly, O3-sensitive genotype 259 had the greatest reduction in above-ground biomass under elevated O3, consistent with its reduced Amax. In contrast, mean above-ground biomass of genotype 8L was not influenced by elevated O3 to any extent. Given the photosynthetic and above-ground growth responses of these genotypes, we expect those genotypes that are most responsive to CO2 and least responsive to O3 will eventually dominate our aggrading stands. By observing the growth and allometry of individual genotypes over the next several years, our intensive measurements of each plant in the experiment will enable us to document mortality and dominance at the genotype level over time, as well as to study interspecific interactions (McDonald et al. 2002).

wood chemistry

Concentration of total lignin (= gravimetric + acid-soluble lignin) in stem wood of Aspen increased by 2·5% under elevated O3 as compared to the control, while elevated CO2 had no effect. This result agrees with a previous study where no changes in lignin concentration caused by elevated CO2 were found (Hättenschwiler, Schweingruber & Körner 1996). Increases in total lignin concentration under elevated O3 are interesting, as there was previously no evidence for O3 effects on wood chemical composition. Increase in lignin may indicate changes in carbon allocation leading to enhanced activity of the phenylpropanoid biosynthetic pathway, which is consistent with the finding that PAL transcripts increased under O3 exposure (Wustman et al. 2001). The concentrations of the other main structural components of cell wall, α-cellulose and hemicellulose, were not affected by any treatment, nor were the minor nonstructural components (acetone-soluble extractives). A CO2 effect on N dilution was also evident in Aspen branch wood, although there was a significant CO2–O3 interaction, with low-CO2/high-O3 branches containing 30% more N per unit dry mass (W.J. Mattson & R. Julkunen-Tiitto, unpublished results). In the case of Aspen branch wood, the sum total of phenolics (total per g dry weight) tended to increase under high CO2, but the differences were not statistically significant. Out of 14 molecular species of phenolics, the concentrations of only four increased significantly, whereas the others did not. Likewise, neither starch nor fibre content of branch wood changed in response to CO2 and O3 (W.J. Mattson & R. Julkunen-Tiitto, unpublished results).

Apart from the environmental control, genetic factors also have a strong impact on wood properties (Costa e Silva et al. 1998; King et al. 1998; Denne, Calahan & Aebischer 1999; Hylen 1999). In Aspen, genotype significantly affected total lignin and acetone-soluble extractives (Anttonen et al. 2001). The O3-sensitive clone 259 had 3% higher total lignin concentration than the O3-tolerant clone 216. The O3-sensitive clone 259 also differed from the other clones in having 26% lower concentration of acetone-soluble extractives. Genotype did not have an effect on α-cellulose and hemicellulose. Wood chemical composition and fibre properties may be different in the juvenile phase from that in mature trees (Zobel & van Buijtenen 1989; Hatton & Hunt 1992); further studies are needed as these trees reach maturity to predict the effects of future climate on wood chemical composition.

foliar chemistry, surface properties and heterotrophic interactions

Our hypothesis is that changes in the quantity and chemistry of plant tissues, elicited by CO2 and O3, cascade through terrestrial ecosystems and alter the performance of heterotrophic organisms, and thus potentially the entire community structure. Elevated CO2 and O3 can alter foliar chemistry and surface properties. Foliar N concentrations in Aspen and Birch declined by 16–21% in response to enriched CO2, but declined only marginally in response to elevated O3 (Lindroth et al. 2001). Decreased N concentrations in foliage under elevated CO2, as has been commonly reported in CO2 studies with other tree species (Cotrufo, Ineson & Scott 1998; Norby et al. 2000), increased C : N ratios in Aspen and Birch foliage, particularly under the combination of elevated CO2 and O3 (Lindroth et al. 2001). For example, as N declined in late-season Birch leaves, starch concentrations increased threefold under elevated CO2 (W.J. Mattson & R. Julkunen-Tiitto, unpublished results). Differences in C : N ratios among the treatments were maintained through leaf senescence and litterfall (Lindroth et al. 2001; Parsons, Bockheim & Lindroth 2000). CO2 enrichment, regardless of whether it was combined with O3 or not, increased litter C : N by 39% in Aspen, and by 24% in Birch, relative to the controls.

Elevated CO2 and O3 altered concentrations of C-based secondary metabolites in Aspen and Birch (Lindroth et al. 2001; Lindroth, Wood & Kopper 2002); but the direction and magnitude of response differed among particular metabolites and between Aspen clones. Common secondary compounds, such as tannins and the phenolic glycosides of Aspen, were responsive to these two gases (Kopper & Lindroth 2002; Lindroth, Wood & Kopper 2002). Eight of 11 Birch leaf phenolic compounds increased 15–30% under high CO2 (W.J. Mattson & R. Julkunen-Tiitto, unpublished). As was the case with C : N ratios, secondary metabolite concentrations were highest in litter originating in the CO2-enriched rings and lowest in the rings exposed to high O3 (W.F.J. Parsons, R.L. Lindroth & J.G. Bockheim, unpublished results).

Elevated CO2 and O3 also altered rates of leaf epicuticular wax biosynthesis with increases or decreases depending on clone and treatment, modified amounts of carbon allocated to various wax forms, and changed chemical composition, with O3 increasing the ratio of long-chain alkane compounds (Karnosky et al. 1999, 2002a). O3 modified wax structure from crystalline to amorphous masses (Mankovska et al. 1998) in Aspen and Birch. These changes in leaf surface properties may have contributed to the threefold to fivefold increased incidence of the Aspen leaf rust Melampsora medusa Thuem. f.sp. Tremuloidae in the O3 and CO2 + O3 treatments (Karnosky et al. 2002a; Percy et al. 2002b) by altering the wettability of the leaf surface and providing a leaf surface more conducive to spore germination.

Elevated CO2 and O3 may alter the performance of insects through changes in bottom-up (plant) and top-down (natural enemy) factors. Colonization of Aspen trees by the Aspen Blotch Leafminer (Phyllonorycter tremuloidiella) declined under enriched CO2 and O3, a response probably caused by changes in epicuticular waxes easing insect attachment and leaf surface penetration (Kopper & Lindroth 2002). Tent Caterpillar (Malacosoma disstria) pupal weights improved under elevated O3, but this was negated under enriched CO2, responses consistent with changes in concentrations of foliar phenolic glycosides. Such effects were not always consistent, however, across Aspen clones. In contrast to well established findings with potted Aspen trees in greenhouses (Lindroth, Kinney & Platz 1993; Bezemer & Jones 1998), CO2 fumigation did not increase leaf foliar consumption by insects. On the other hand, first and late-instar Gypsy Moth larvae consistently increased consumption (>15%) of CO2-fumigated Birch leaves, though growing no better than control larvae (W.J. Mattson & T. Trier, unpublished results). The beetle Oberea schaumii, which bores into the stems and young branches of Aspen trees, responded to CO2 and O3 by being least abundant on trees under O3 fumigation, but most abundant on trees under the CO2 + O3 fumigation. Its populations were intermediate in abundance under elevated CO2, but still exceeded the controls. The Bark Scale, Chionaspis, and the Fly Gall-maker, Hexomyza, were most abundant on trees growing under CO2 fumigation, showing no other treatment responses (W.J. Mattson, unpublished results).

Elevated CO2 and O3 also have the potential to alter insect community composition. Population censuses of aphids and natural enemies on Aspen revealed that while aphid abundance was unaffected by CO2 or O3, natural enemy abundance increased at elevated CO2 and declined at elevated O3 (Percy et al. 2002b). Awmack & Harrington (2000) showed that the damage caused by aphids increased at elevated CO2 and negated the large growth stimulations otherwise expected of Bean plants (Vicia faba) under elevated CO2. Therefore increases in pest numbers may have a considerable impact on forest health and productivity in the future. The importance of insect pest and disease eruptions in altering carbon fluxes from ecosystems has been highlighted by Kurz & Apps (1999), who detected a decade-long shift from carbon sink to source in the boreal forests of Canada, attributable to an increase in disturbances by pests and fire.

ecosystem responses

Most studies dealing with forest trees and elevated CO2 and O3 have been conducted in chambers in which it is not possible to address long-term, large-scale, ecosystem-level questions (Heck et al. 1998; Hendrey et al. 1999; McLeod & Long 1999; Karnosky et al. 2001). After 4 years’ research at Aspen FACE, we are beginning to see indications that the physiological and genetic responses, which we detected early in this experiment, are cascading through the ecosystem and resulting in significant ecosystem-level responses to elevated CO2 and O3. It should be noted that we have found slight differences in soil fertility across our site, but we accommodated these differences in our experimental design by using a randomized complete block design with the blocking being done by maintaining rings of common fertility within each replicate (Dickson et al. 2000). Our soil fertility levels are in the range of natural Aspen forest soil fertility, but sufficient to ensure that our tree growth responses to elevated CO2 have been unconstrained by nutrient limitations.

litter production, chemistry and decomposition

Several pieces of evidence collectively suggest that greater rates of carbon assimilation and growth under elevated CO2 directly influence the amount of organic substrate entering the soil for microbial metabolism, and that this response has been dampened by elevated O3. Elevated atmospheric CO2 increased leaf litter production by 36% in Aspen, and Birch leaf litter production doubled in the Aspen–Birch community. Elevated O3 did not substantially alter leaf litter production relative to the control, but elevated O3 decreased the CO2-induced production of litter in Aspen and Birch. Similarly, elevated CO2 significantly increased the mass of dead fine roots by 140% beneath Aspen, and by 340% beneath Aspen–Birch (King et al. 2001). Although elevated O3 did not influence dead fine-root mass (relative to the control), it did nullify the increase in fine-root mass caused by elevated CO2. We also found that elevated CO2 increased live fine root biomass by 113% beneath Aspen and by 83% beneath Aspen–Birch; elevated O3 did not significantly alter live fine root biomass (King et al. 2001). We also observed no change in the lignin content of live and dead fine roots (J.S. King, unpublished results), a result supported by greenhouse studies (W.F.J. Parsons, B.J. Kopper & R.L. Lindroth, unpublished results). Taken together, our results suggest that CO2 substantially increased above-ground and below-ground litter production beneath Aspen and Aspen–Birch, but this response was almost eliminated by elevated O3.

The 1 year decay rate (k value) of Birch litter was significantly reduced by elevated CO2 regardless of O3 (Parsons, Bockheim & Lindroth 2000). Aspen decay showed similar trends, although differences among treatments were not significant. Initial differences in foliar quality among the treatments were sustained throughout Aspen and Birch decomposition, and these distinctive chemical signatures probably contributed to controlling mass loss from the decomposing litter, whether it was returned to its ring of origin or transplanted into another treatment ring. From reciprocal litter transplant experiments we observed slight moderation of litter quality effects: a weak substrate–environment interaction (Parsons, Bockheim & Lindroth 2000).

soil and microbial respiration

Elevated CO2 substantially increased soil respiration rates beneath Aspen (by 30%) and Aspen–Birch (by 60%); however, soil respiration increased much less (10%) beneath Aspen–Maple. Elevated O3 had a relatively minor influence on mean soil respiration at both atmospheric CO2 concentrations except for late in each growing season, possibly due to increased fine root senescence caused by O3. This pattern of soil respiration was reflected in differences in soil pCO2 among our experimental treatments (King et al. 2001), wherein elevated CO2 increased soil pCO2 by 27% (averaged over three depths from 15 to 125 cm and two growing seasons), and O3 had little effect. We believe this result is important because higher soil pCO2 could lead to more carbonic and organic acids in the soil, leading to more rapid mineral weathering, nutrient leaching, and the export of dissolved inorganic C. Like soil respiration, microbial respiration under elevated CO2 was significantly increased beneath Aspen (33%) and Aspen–Birch (55%), but the increase beneath Aspen–Maple was small (1%) and not significant (Phillips et al. 2002). We also observed that elevated O3 did not significantly decrease microbial respiration relative to the control, but elevated O3 did reduce rates of microbial respiration under elevated CO2 (in the CO2 + O3 treatment). Greater rates of microbial respiration indicate that increased litter inputs under elevated CO2 are being metabolized by a soil microbial population that is larger, more active, or both. We observed a nonsignificant increase in soil microbial biomass under elevated CO2 (28%) across two growing seasons (Larson et al. 2002), while microbial biomass in the O3 and CO2 + O3 treatments were equivalent to that of the control (data not shown). We have not yet determined if microbial respiration changes are in any way related to changes in the biochemical composition of the plant litter.

microbial community function

We believe that changes in the quantity of organic substrate entering the soil from our experimental treatments have altered the metabolism of soil microbial communities. Again, we do not yet know if litter quality changes are also having an effect. Elevated CO2 increased the activity of enzymes involved in plant and fungal cell-wall degradation at O3 concentrations. We detected CO2-induced increases in cellobiohydrolase, an enzyme catalysing the release of cellobiose during cellulose degradation, and N-acetylglucosaminidase, which catalyses the release of N-acetylglucosamine during chitin degradation (we did not find significant interactions between CO2 and species or O3 treatments; Larson et al. 2002). These results suggest that greater inputs of plant and fungal cell-wall substrates (cellulose and chitin) under elevated CO2 have altered the metabolism of these plant-derived compounds in soil and the transport of C through the soil food web. This response was confirmed by increased recovery of 13CO2 that was respired from labelled cellobiose and N-acetylglucosamine added to soil in a preliminary soil incubation experiment (Phillips et al. 2002). It is possible that these responses are driven largely by greater inputs of dead fine roots and associated mycorrhizal fungi, and we intend to explore this possibility.

soil n cycling

Results from a short-term 15N tracer experiment suggest that changes in microbial metabolism among experimental treatments have altered rates and patterns of soil N cycling. We followed the flow of inline image and 15inline image in soil collected from each FACE ring during the 1999 field season. O3 had no effect on the recovery of 15N in microbial biomass or soil organic matter. In contrast, elevated CO2 (main effect) significantly increased the amount of 15inline image recovered in microbial biomass and soil organic matter. Recovery of 15inline image in these pools was greater under elevated CO2, but this increase was not significant. These results indicate that larger amounts of N are forming soil organic matter under elevated CO2. Because the C : N ratio of soil organic matter is relatively constant, this finding suggests that greater amounts of C also are forming soil organic matter in our experiment. This indicates that more C may be stored in soil as the atmospheric CO2 concentration increases, a finding not strongly supported in the Duke FACE study (Schlesinger & Lichter 2001).

water balance

There has been much speculation that ecosystem water balances will be altered under elevated CO2 (Curtis 1996; Curtis & Wang 1998). However, this type of response has been impossible to test in greenhouse chamber or open-top chamber studies. We have found two independent lines of evidence suggesting possible water balance changes attributable to elevated CO2 and O3. First, water-use efficiency calculations by Sôber et al. (2000) suggest that water-use efficiency was highest for Aspen clones exposed to elevated CO2 and lowest in those exposed to O3. Trees exposed to elevated CO2 + O3 were intermediate between treatments, but still greater than controls in water-use efficiency. Second, during a relatively dry year (1999) but not in relatively wet year (2000), changes in volumetric soil moisture, as determined by time domain reflectometry, were detectable throughout the growing season under elevated CO2, O3 and CO2 + O3 (J.S. King & K.S. Pregitzer, unpublished results).


While elevated CO2 (Vasseur & Potvin 1998) and O3 (Berrang et al. 1986; Barbo et al. 1998) have been implicated in altering plant communities and biodiversity, few studies have investigated the impacts of the interacting effects of CO2 and O3 on community composition. We have evidence that the composition of forest communities may be altered under these two greenhouse gases. First, in a study of competitive interactions between Aspen clones differing in O3 sensitivity, McDonald et al. (2000, 2002) found that elevated CO2 exacerbated growth reductions as elevated O3 decreased growth 35% in elevated CO2 compared with 24% in ambient CO2. Clone-competitive interactions were significant, suggesting that interacting CO2 and O3 could exacerbate clonal competition for fitness, as previously described under elevated O3 by Karnosky et al. 1999, 2002b). We expect these competitive interactions to increase over the next few years as interactions among crowns and among root systems intensify, and as seed production and dispersal occur in the pioneer species.

implications for carbon sequestration and net primary productivity models

Our results suggest that elevated O3 at relatively low concentrations can significantly reduce the growth enhancement by elevated CO2. Our results follow similar trends found for many agricultural crops, other hardwood trees and a few conifers (Fig. 7). Together, these studies on plants of different genetic backgrounds, growth characteristics and life histories suggest that O3 can seriously alter the capacity of vegetation to grow under elevated CO2 and to sequester carbon. For example, the projected co-occurrence of elevated O3 (as predicted by Fowler et al. 1999a) over a large portion of the natural range of the circumpolar Leuce (Aspen) section of Poplar (Fig. 8) may mean that worldwide growth stimulations will not be as great as predicted from previous studies of elevated CO2. It is important to bring an understanding of O3 as a moderator of CO2 responses to global models of terrestrial net primary productivity. It is also important to expose forests for their entire rotation or life cycle to understand the practicality of using forests and forest plantations to sequester carbon and to offset anthropogenic CO2 emissions.

Figure 7.

Relative effects of controlled exposure to elevated CO2 on normalized plant growth under CO2 alone (striped bars, 500–713 µmol mol−1 CO2) and elevated CO2 plus ozone (dotted bars). (Modified and expanded from Barnes & Wellburn 1998.) Data presented for wheat (Triticum aestivum) from Barnes, Ollerenshaw & Whitfield (1995); Rudorff et al. (1996); McKee, Bullimore & Long (1997); Bender, Herstein & Black (1999); Hudak et al. (1999); soybean (Glycine max) from Heagle, Miller & Pursley (1998) and Miller, Heagle & Pursley (1998); tomato (Lycopersicon esculentum) from Olszyk & Wise (1997) and Hao et al. (2000); rice (Oryza sativa) from Olszyk & Wise (1997); potato (Solanum tuberosum) from Donnelly et al. (2001) and Lawson et al. (2001); corn (Zea mays) from Rudorff et al. (1996); hardwood trees including hybrid poplars (Populus hybrids) from Dickson et al. (1998); Trembling Aspen (Populus tremuloides) from Volin & Reich (1996); Volin et al. (1998); Isebrands et al. (2001); oak (Quercus petrea) from Broadmeadow & Jackson (2000); conifers including Ponderosa Pine from David Olszyk (personal communication); Scots Pine (Pinus sylvestris) from Broadmeadow & Jackson (2000); Utriainen et al. (2000). Each pair of bars represents one species.

Figure 8.

Worldwide distribution of Aspen and projected areas with elevated O3 in the year 2100. Ozone map from Fowler et al. (1999a).

Conclusions and research needs

The suite of responses to elevated CO2 and/or O3 at the Aspen FACE project have been remarkably consistent across functional groups, species and genotypes differing in O3 tolerance, and from molecular to ecosystem levels. However, it must be noted that these responses are being found in a young, aggrading forest. We have not yet detected a diminishing of CO2 growth enhancement, as has been reported for Loblolly Pine (Oren et al. 2001) and Sweetgum (Norby et al. 2001b). However, our forest is in a much younger stage of development than these other two sites. Comparisons among FACE sites are difficult because several factors differ, including soils, climate, species and treatments. Several new areas of study at our site – such as the occurrence and abundance of mycorrhizal fungi; the diversity and quantity of understorey vegetation; and canopy-level sap flow – are still premature to discuss at this point, but should be an integral part of our project's future research efforts. Finally, we continue to use our results to parameterize and test various growth models such as ECOPHYS (Martin et al. 2001) to broaden the inferences from our results.


This research was partially supported by the US Department of Energy's Office of Biological and Environmental Research (BER) (DE-FG02-95ER62125; DE-FG02-93ER6166; DE-FG02-98ER62680), USDA Forest Service Northern Global Change Program, the National Science Foundation (DBI-9601942; IBN-9652675; DEB-9707263; DBI-972395), the USDA NRI Competitive Grants Programs (000-2982; 001-00796; 001-01193), the National Council of the Paper Industry for Air and Stream Improvement (NCASI), Michigan Technological University, the Slovakian Forest Research Institute, the Praxair Foundation, the University of Wisconsin Foundation, the McIntire–Stennis Program, the Brookhaven National Laboratories/US Department of Energy (725-079), the Academy of Finland (Projects 39935; 42702; 43168), the Ministry of Agriculture and Forestry in Finland, Natural Resources Canada – Canadian Forest Service, the British Ecological Society, and the Natural Environmental Research Council of the United Kingdom.