Persistent stimulation of photosynthesis by elevated CO2 in a sweetgum (Liquidambar styraciflua) forest stand


Author for correspondence: David Tissue Tel: +1 (806) 742 2786 Fax: +1 (806) 742 2963 Email:


  • • The photosynthetic response of trees to rising CO2 concentrations ([CO2]) can be affected by plant source–sink relations, in addition to seasonal changes in environmental conditions. Characterization of biochemical and morphological feedbacks is important for understanding ecosystem responses to elevated atmospheric [CO2].
  • • The seasonal responses of leaf gas exchange and related biochemical parameters were measured during 3 yrs of exposure on established plantation sweetgum (Liquidambar styraciflua) trees at a Free-Air CO2 Enrichment (FACE) facility in eastern Tennessee, USA.
  • • Net photosynthetic rates (Agrowth) of upper-canopy leaves were 44% higher in trees grown in elevated [CO2] compared with ambient [CO2] over the 3-yr period. There were no significant CO2 treatment effects on photosynthetic or biochemical capacity (i.e. no change in Amax, Vcmax or Jmax) of L. styraciflua leaves, despite increased area-based leaf sugar (10%) and starch content (27%), and reduced mass-based leaf nitrogen concentration (NM; 10%).
  • • These results suggest that established L. styraciflua trees in closed-canopy forests might exhibit a long-term positive response to elevated [CO2] without reductions in photosynthetic capacity.


Photosynthesis in trees is stimulated an average of 44–66% by growth in elevated CO2 concentrations ([CO2]) (see reviews: Gunderson & Wullschleger, 1994; Curtis & Wang, 1998; Saxe et al., 1998; Medlyn et al., 1999; Norby et al., 1999). The short-term response can be attributed largely to stimulation of ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) carboxylation. Rubisco is CO2 substrate-limited under current atmospheric conditions, but an increase in atmospheric [CO2] stimulates carboxylation and suppresses oxygenation, thereby reducing photorespiration and increasing net photosynthesis (Sharkey, 1988; Long, 1991). Long-term photosynthetic responses to elevated [CO2], however, may be affected by biochemical, morphological and physiological feedbacks that balance carbon assimilation with growth (i.e. sink) demand (Long & Drake, 1991; Stitt, 1991; Sage, 1994; Griffin & Seemann, 1996; Peterson et al., 1999; Rogers & Ellsworth, 2002; Körner, 2003). An understanding of these long-term, leaf-level responses may provide insight into ecosystem responses to elevated atmospheric [CO2].

Trees exposed to CO2 enrichment in field experiments generally show less evidence of the loss of photosynthetic stimulation after long-term exposure to elevated [CO2], termed downregulation or acclimation, than had been suggested by earlier experiments conducted using seedlings grown in pots in growth chambers. Curtis & Wang (1998) found in a meta-analysis of 59 tree species that downregulation was minimal and inconsistent, with the greatest downregulation occurring in trees grown in small pots (−36%). However, a meta-analysis of 15 field-based elevated [CO2] experiments on European forest tree species found a downregulation of photosynthesis of 10–20% when measured at a common [CO2], in spite of a 51% enhancement of the light-saturated rate of photosynthesis when measured at growth [CO2] (Medlyn et al., 1999). Although photosynthetic downregulation may reduce the positive response of photosynthesis to elevated [CO2], it rarely completely eliminates it. Therefore, continued photosynthetic enhancement might be expected in forest ecosystems exposed to long-term CO2 enrichment, as is occurring in nature.

The lack of consistency with regard to the photosynthetic acclimation of trees can be partially explained by looking at plant-level source-sink relationships. The degree to which trees exhibit photosynthetic downregulation when grown in elevated [CO2] may be related, on a physiological level, to resource allocation within the plant, especially nitrogen, which often limits the productivity of forest ecosystems (Sage, 1994; Barnes et al., 1998). Within plant photosynthetic tissue, resources are partitioned between light-harvesting and nonharvesting components. At a broader scale, resources must be allocated between those plant tissues that are net exporters of photosynthate (sources) and net importers (sinks), typically characterized as new roots, shoots and leaves or reproductive organs (Herold, 1980; Stitt, 1991). A reduction in photosynthetic capacity may occur in elevated [CO2] as biochemical adjustments cause reallocation of resources, particularly nitrogen, from nonlimiting processes to limiting processes (Sage, 1994; Sage & Reid, 1994).

Maintenance of active sinks is necessary for the effective stimulation of photosynthesis (Eamus & Jarvis, 1989; Farrar & Williams, 1991; Stitt, 1991). The increased production of carbohydrates induced by higher photosynthetic rates in elevated [CO2] can be sustained only with continued translocation of photosynthates to active sink material (Xu et al., 1994; Wolfe et al., 1998). An imbalance in the source-sink relationship can result in an accumulation of leaf carbohydrates, which may ultimately trigger a feedback mechanism that reduces photosynthetic capacity (Van Oosten & Besford, 1994; Webber et al., 1994). Accumulation of large amounts of carbohydrates may lead to reductions in gene transcription and production of Rubisco protein, thereby reducing Calvin cycle activity and further sugar production (Sheen, 1990; Furbank & Taylor, 1995; Krapp & Stitt, 1995; Cheng et al., 1998). Most commonly, both the maximum carboxylation rate of Rubisco (Vcmax) and maximum electron transport rate (Jmax) are reduced in plants exhibiting photosynthetic downregulation (see reviews: Wullschleger, 1993; Gunderson & Wullschleger, 1994; Sage, 1994; Medlyn et al., 1999; Norby et al., 1999).

The regulation of photosynthetic rate and capacity, Vcmax and Jmax, by source–sink relationships suggests that photosynthetic downregulation may be more prevalent in trees whose growth processes (i.e. sinks) are limited annually or seasonally by environmental factors. No single factor appears predominant in controlling photosynthetic downregulation (Gunderson & Wullschleger, 1994; Medlyn et al., 1999), but it is most common to observe photosynthetic downregulation at the end of the growing season, when sink strength is reduced and many environmental factors become more limiting. For example, in three deciduous trees, Castanea sativa (El Kohen & Mousseau, 1994), Fagus sylvatica (Epron et al., 1996) and Betula pendula (Rey & Jarvis, 1998), the degree of photosynthetic downregulation in response to elevated [CO2] increased as the growing season progressed and sink strength declined. Improved soil nitrogen conditions through fertilization delayed photosynthetic downregulation in Castanea sativa, presumably due to increased growth and sink strength (El Kohen & Mousseau, 1994). In many pine species, downregulation is only observed in older needles, but not younger needles (Jach & Ceulemans, 2000; Tissue et al., 2001; Rogers & Ellsworth, 2002), concurrent with a reduction in their sink strength. In a closed-canopy forest, leaf production is constrained, which may limit the continued development of carbon sinks in elevated [CO2].

The primary objective of this study was to characterize the effects of seasonal variation in plant source-sink relations on the photosynthetic capacity of upper-canopy leaves to respond to elevated [CO2] in a closed-canopy sweetgum forest. Sweetgum (Liquidambar styraciflua) is a canopy-dominant, deciduous tree with indeterminate growth (Bormann, 1953). Measurements of photosynthetic and biochemical responses of upper-canopy foliage were used to determine the degree of photosynthetic enhancement, and potential photosynthetic downregulation, to elevated [CO2] at different times during the growing season, which represent different levels of leaf and plant sink strength. We hypothesized that reductions in sink capacity would feed back on photosynthetic enzyme activity, thereby reducing the positive effects of elevated [CO2] on leaf carbon balance. A decline in the photosynthetic enhancement of elevated [CO2] would be associated with the following: (i) reduced Amax; (ii) a decline in Vcmax and Jmax; (iii) increased leaf mass per unit leaf area (LMA); (iv) decreased mass-based leaf nitrogen concentration (NM); and (v) an increase in leaf carbohydrate content, especially starch.

Measurements were conducted over three growing seasons at the Oak Ridge National Laboratory's (ORNL) Free-Air CO2 Enrichment (FACE) site to determine whether photosynthetic responses to elevated [CO2] were persistent or transient. We hypothesized that canopy closure in this stand (i.e. leaf area index (LAI) not increasing) would constrain growth responses to elevated [CO2], in contrast to the responses of individual trees in an expanding system (Norby et al., 1999). Net primary productivity in this stand has been about 21% higher in elevated CO2 and there has been no sustained enhancement in aboveground growth (Norby et al., 2002), in contrast to the much larger CO2 responses sometimes reported for trees undergoing exponential growth (Norby et al., 1999). Hence, sink limitation to photosynthetic responses might be more likely in this stand than has been reported in previous studies with CO2-enriched L. styraciflua (Herrick & Thomas, 2001).

Materials and Methods

Oak Ridge National Laboratory Free-Air CO2 Enrichment (FACE) site

Research was conducted at the ORNL FACE facility in Roane County, Tennessee (35°54′ N, 84°20′ W). The FACE site was constructed within a sweetgum (L. styraciflua) plantation established in the autumn of 1988, at which time 1-yr-old trees were planted at a spacing of 2.3 m × 1.2 m. The FACE facility became operational in April 1998, at which time the live crown of the closed canopy extended from 8 to 9 m above the forest floor to a height of 13–14 m. Crown depth remained approximately the same each season, with 1 m of crown height at the top being added each year, but 1 m of crown height at the bottom being lost as lower branches died. Tree heights reached 14–15 m in spring 1999 and 15–16 m in spring 2000.

The ORNL FACE site consists of five circular plots (rings), each 25 m in diameter enclosing 80–90 trees inside a 2.5-m wide buffer zone. Two rings were treated with air enriched with CO2 (‘elevated CO2’), while the remaining three rings (‘ambient CO2’) served as controls. Two of the three control plots contain blower systems like those present in the elevated-CO2 plots. Experimental treatment was applied 24 h d−1 throughout the 1998–2000 growing seasons. The treatment set point for 1998 was a constant 565 µmol mol−1 CO2, which was approximately 200 µmol mol−1 above the global average [CO2]. The treatment protocol was adjusted in 1999 to incorporate the natural diurnal variation in [CO2], with the set point for the night-time concentration increased to 645 µmol mol−1. Actual mid-day [CO2] in the elevated-CO2 plots averaged 548, 556 and 555 µmol mol−1 in 1998, 1999 and 2000, respectively, compared to mid-day concentrations in the ambient rings of 362, 365 and 366 µmol mol−1 in 1998, 1999 and 2000, respectively (Gunderson et al., 2002).

Gas-exchange measurements

The responses of net photosynthesis (A) to intercellular [CO2] (Ci) (A-Ci curves) were measured during each field season, which for these L. styraciflua trees extended from leaf-out in mid April to leaf fall in late October (Norby et al., 2003). All A-Ci curves (five to eight curves per ring per sampling period) were collected using fully expanded (mature), upper-canopy leaves in at least four trees, which were accessible from a stationary hydraulic lift located near the centre of each plot. A portable photosynthesis system (LI-6400, LiCor, Inc., Lincoln, NE, USA) was used to measure the rate of leaf assimilation at each of 11 external [CO2] (Ca), ranging from 1500 to 0 µmol mol−1 CO2. An automated programme was utilized to log assimilation readings at each Ca set point when photosynthesis rates had equilibrated, which was typically 1–2 min after a stable Ca set point was reached. All gas-exchange measurements were made at saturating light levels (1800 µmol m−2 s−1) provided by a blue-red light source mounted above the 6 cm2 leaf cuvette. Chamber temperatures were set throughout the season to approximate the average mid-afternoon temperatures that occurred during the week that measurements were taken. Temperatures within the leaf cuvette typically ranged from 27 to 36°C. All gas-exchange measurements were taken between 10:00 and 16:00 h EST. A-Ci curves were collected in July and September 1998, monthly in May until September 1999, and in May, July and September 2000.

A-Ci curves were analysed using Photosyn Assistant (Dundee Scientific, Dundee, Scotland, UK). For each curve, this software uses equations that relate the effect of changing [CO2] to photosynthetic rate. The equations estimate potential limitations to CO2 assimilation due to the maximum Rubisco carboxylation rate (Vcmax), and the rate of ribulose bisphosphate (RuBP) regeneration via electron transport (Jmax). The program uses the Farquhar et al. (1980) model, as modified by von Caemmerer & Farquhar (1981), Sharkey (1985), Harley & Sharkey (1991), and Harley et al. (1992). Temperature-response functions were included in modelling these photosynthetic parameters to account for the wide range of mean leaf temperatures encountered throughout this study (Bernacchi et al., 2001). The maximum CO2- and light-saturated rate of photosynthesis (Amax) and net rate of photosynthesis at the respective growth [CO2] (either Ca = 360 or 560 µmol m−2 s−1, Agrowth) were also determined from A-Ci curves.

Leaf properties

All leaves used for A-Ci measurements were harvested for mensuration and biochemical analysis. The thickness of individual fresh leaves was measured in the field using digital calipers at the time of leaf collection beginning in July 1999. Leaves were then harvested and placed on ice until they could be returned to the laboratory for processing. Leaf area was determined for each individual leaf using a leaf-area meter (LI-3100). Following the measurement of individual leaf areas, all leaves were dried in a 70°C oven for at least 3 d, and then each leaf was weighed individually. These data were used to calculate leaf mass per unit area (LMA). Estimates of leaf volume and density were also calculated, because leaf function is affected by the interaction of leaf morphology (area and thickness) and composition (dry matter, liquid content, and fractional air space) (Niinemets, 1999; Roderick et al., 1999a,b).

Leaf soluble sugar and starch contents were determined colorimetrically by a phenol-sulfuric acid technique (Tissue & Wright, 1995). Total nonstructural carbohydrate (TNC) was calculated as the sum of soluble sugar and starch. Total chlorophyll content was determined by immersing a 1.33 cm2 leaf punch from each leaf in 10 ml of 95% ethanol for 2 d in a dark environment, followed by measurement of the absorbance of the extract at 665, 649 and 470 nm (Lichtenthaler & Wellburn, 1983). Total leaf nitrogen was measured in the dried leaf samples using a CN Analyzer (NCS 2500, Carlo Erba Inc., Milan, Italy).

Statistical analyses

Treatment effects were calculated using the ring means of gas-exchange and leaf- property parameters for each sampling date. A two-way analysis of variance (anova) was used to compare the effects of CO2 treatment (n = 3 ambient [CO2]; n = 2 elevated [CO2]) and sampling date (n = 10) and the interaction of [CO2] and sampling date on all measured and modelled parameters. Parameters were considered significantly different when P < 0.05. Least squares linear regression analysis was performed to analyse the relationship of Vcmax to Jmax, leaf nitrogen per unit area (NA) to the various photosynthetic parameters measured and modelled in this experiment, and total chlorophyll per unit area (ChlA) to Jmax. In those instances where statistically significant regression results were noted for both ambient- and elevated-CO2 treatments, homogeneity of slopes and intercepts was tested using analysis of covariance (ancova). All statistical analyses were performed using systat (Version 8.0, SPSS Inc. Chicago, IL, USA, 1998).



The net photosynthetic rates (Agrowth) of L. styraciflua in elevated [CO2] were significantly higher than those of trees in ambient [CO2] (Table 1, Fig. 1), with increases ranging from 33% to 93% (Fig. 2a). The average net photosynthetic rate was 44% higher with CO2 enrichment (Fig. 1). Sample date also significantly affected the net rate of photosynthesis (Table 1). There were seasonal and annual differences in Agrowth response with the highest rates occurring in mid-season and lower rates in early and late-season (Fig. 2a). There was no significant CO2 effect on Amax over the course of the experiment (Table 1, Fig. 2b). There was a significant effect of sample date on Amax (Table 1). There were seasonal and annual differences in Amax response with the highest rates occurring in mid-season and lower rates in early and late-season (Fig. 2b).

Table 1. anova results for Liquidambar styraciflua leaf photosynthetic and biochemical parameters
Parameter[CO2]Sample date[CO2]* Sample date
  1. Mature upper-canopy leaves of L. styraciflua grown in ambient (∼364 µmol m−2 s−1) and elevated (∼553 µmol m−2 s−1) CO2 were sampled in July and September 1998, monthly from May through September 1999, and in May, July and September 2000. Significance of treatment effect, as indicated by the P-value, is reported for CO2 concentration, sample date, and the interaction between CO2 treatment and date of sample collection ([CO2]* Sample date).

Agrowth (µmol m−2 s−1)< 0.0001< 0.00010.9397
Amax (µmol m−2 s−1)  0.2139  0.00010.7585
Vcmax (µmol m−2 s−1)  0.8580  0.00030.9603
Jmax (µmol m−2 s−1)  0.2508  0.00040.9182
Jmax/Vcmax  0.5324  0.00370.0921
PNUE (µmol g−1 N s−1)< 0.0001< 0.00010.4657
LMA (g m−2)  0.0002< 0.00010.9052
Leaf Density (g cm−3)  0.5449< 0.00010.1443
NM (mg g−1)< 0.0001< 0.00010.5545
NA (g m−2)  0.2576  0.00840.7560
NV (mg cm−3)  0.0463< 0.00010.6167
ChlM (mg g−1)  0.0291  0.00040.6827
ChlA (g m−2)  0.4813  0.06750.2674
ChlM (mg g−1)/NM(mg g−1)  0.8970< 0.00010.7370
Sugar (g m−2)  0.0010< 0.00010.1610
Starch (g m−2)< 0.0001< 0.00010.0006
TNC (g m−2)< 0.0001< 0.00010.0247
Figure 1.

Percent change of photosynthetic and leaf characteristic parameters measured in mature upper-canopy leaves of Liquidambar styraciflua exposed to ambient (∼364 µmol mol−1) or elevated (∼553 µmol mol−1) CO2. Values were calculated as ([Elevated CO2– Ambient CO2]/Ambient CO2× 100), with ring means serving as the experimental unit in each of 10 sampling periods (n = 30 ambient [CO2]; n = 20 elevated [CO2]) measured over the course of the project. The presence of asterisks denotes a significant CO2 treatment effect (*, P < 0.05; **, P < 0.01; ***, P < 0.001) in the overall [CO2] * Sampling Date anova.

Figure 2.

Photosynthetic response of mature upper-canopy leaves of Liquidambar styraciflua exposed to ambient (∼364 µmol mol−1, n = 3, open circles) and elevated (∼553 µmol mol−1, n = 2, closed circles) CO2. Photosynthesis at (a) growth [CO2] (Agrowth) and (b) CO2 and light saturation (Amax) were determined using A-Ci curves collected between May and September in 1998, 1999, and 2000 using a LiCor 6400 photosynthesis system. Values are presented as means ± sem.

Growth in elevated [CO2] did not have a significant effect on Vcmax, Jmax or the ratio of Jmax to Vcmax (Table 1, Fig. 1). Sampling date had a significant effect on both Vcmax and Jmax (Table 1), with rates generally higher in mid-season than early and late-season (Fig. 3a,b). Date of sampling also significantly affected the ratio of Jmax to Vcmax (Table 1). This ratio was generally lower in mid-season when Agrowth and Amax tended to exhibit their highest rates (Fig. 3c). In addition, there was a positive linear relationship between Jmax and Vcmax for leaves grown in both ambient and elevated [CO2] (R2 = 0.87 ambient; R2 = 0.78 elevated) (Fig. 4). The regressions for both treatments were statistically significant (P < 0.0001 ambient, P < 0.0001 elevated), with similar slopes and intercepts.

Figure 3.

Response of modelled photosynthetic parameters of mature upper-canopy leaves of Liquidambar styraciflua exposed to ambient (∼364 µmol mol−1, n = 3, open circles) or elevated (∼553 µmol mol−1, n = 2, closed circles) CO2. A-Ci curves were analysed using a C3 model of photosynthetic biochemistry to determine: (a) Vcmax, the maximum carboxylation rate of Rubisco, and (b) Jmax, the RuBP regeneration capacity mediated by electron transport rate. (c) The ratio of Jmax/Vcmax was also calculated. Values are presented as means ± sem.

Figure 4.

Linear relationship between the maximum rate of electron transport (Jmax) and the maximum rate of Rubisco carboxylation (Vcmax) of mature upper-canopy leaves of Liquidambar styraciflua. Vcmax and Jmax were derived as best-fit estimates from individual A-Ci curves. A single regression line is presented because the regressions generated for each treatment were not statistically different. The regression for the combined data was statistically significant (P < 0.0001, n = 45). Open circles, ambient [CO2]; closed circles, elevated [CO2].

The photosynthetic nitrogen-use efficiency (PNUE), defined as carbon assimilation per unit of leaf nitrogen (Lambers et al., 1998), of L. styraciflua leaves was significantly greater in elevated [CO2] (Table 1, Fig. 1). Sample date was shown to significantly affect PNUE, with the greatest CO2 effect occurring in mid-season (Table 1, Fig. 5).

Figure 5.

Response of photosynthetic nitrogen-use efficiency (PNUE) of mature upper-canopy leaves of Liquidambar styraciflua exposed to ambient (∼364 µmol mol−1, n = 3, open circles) or elevated (∼553 µmol mol−1, n = 2, closed circles) CO2. PNUE was calculated as the rate of photosynthesis at growth CO2 (Agrowth) per unit leaf nitrogen concentration (NA). Values are presented as means ± sem.

Leaf properties

Leaf mass per unit leaf area (LMA) was consistently higher in leaves from trees grown in elevated [CO2] compared with ambient [CO2], with an average 9% increase over the 3-yr study period (Table 1, Fig. 1). The absolute increase in LMA due to elevated [CO2] ranged from 2.9 to 16.7 g m−2, with the greatest differences occurring in mid-season, while the relative increase ranged from 3% to 19% (Fig. 6a). There were no differences in LMA on an annual basis. There was no CO2 treatment effect on leaf density (g cm−3), but date of sampling did have a significant effect (Table 1, Fig. 1) with greater mean leaf densities noted late in the growing season in 2000 (Fig. 6b).

Figure 6.

Response of growth characteristics of mature upper-canopy leaves of Liquidambar styraciflua exposed to ambient (∼364 µmol mol−1, n = 3, open circles) or elevated (∼553 µmol mol−1, n = 2, closed circles) CO2. Growth characteristics were determined for the leaves harvested following the collection of A-Ci curves. (a) Leaf mass per unit leaf area (LMA) measurements were collected in all 10 sample periods, while (b) leaf density measurements were limited to the latter half of the 1999 season and all of the 2000 season. Values are presented as means ± sem.

Leaf biochemistry

There was no significant CO2 treatment effect on nitrogen content when expressed on a leaf-area (NA, g m−2) basis (Table 1). Elevated [CO2] significantly decreased leaf nitrogen when expressed on a dry-mass (10%; NM, mg g−1) and leaf-volume (9%; NV, mg cm−3) basis (Table 1). The relative reduction in NM due to elevated [CO2] ranged from 4% to 21%. Sampling date significantly affected NA, NM and NV (Table 1). Foliar nitrogen was generally highest in 1999 and lowest in 2000 (Fig. 7).

Figure 7.

Response of foliar nitrogen concentration of mature upper-canopy leaves of Liquidambar styraciflua exposed to ambient (∼364 µmol mol−1, n = 3, open circles) or elevated (∼553 µmol mol−1, n = 2, closed circles) CO2. Biochemical analyses were performed following the harvest of the mature upper-canopy leaves upon which photosynthesis measurements were taken. Leaf nitrogen results are reported (a) per unit leaf mass (NM) and (b) per unit leaf area (NA). Values are presented as means ± sem.

Foliar total chlorophyll concentrations were significantly decreased by growth in enriched [CO2] when expressed on a mass (ChlM), but not area (ChlA) basis (Table 1, Fig. 1). ChlM was also significantly affected by sample date (Table 1). The relative decline in ChlM due to elevated [CO2] ranged from 2% to 25%, with the greatest reduction occurring in mid-season (Fig. 8a). ChlM did not vary annually. There was no significant effect of sample date on ChlA (Table 1, Fig. 8b). The investment of plant nitrogen in chlorophyll was examined by calculating the ratio of ChlM per unit NM. Elevated [CO2] had no effect on this ratio, but there was a significant effect of sampling date (Table 1).

Figure 8.

Response of foliar chlorophyll concentration of mature upper-canopy leaves of Liquidambar styraciflua exposed to ambient (∼364 µmol mol−1, n = 3, open circles) or elevated (∼553 µmol mol−1, n = 2, closed circles) CO2. Biochemical analyses were performed following the harvest of the mature upper-canopy leaves upon which photosynthesis measurements were taken. Leaf chlorophyll results are reported (a) per unit leaf mass (ChlM) and (b) per unit leaf area (ChlA). Values are presented as means ± sem.

Area-based sugar, starch and TNC content of L. styraciflua leaves were significantly increased in elevated [CO2] (Table 1, Fig. 1). Sugar was increased an average of 10%, starch 27%, and TNC 17% over the 3-yr period of the study. The relative change in sugar ranged from −20% to 36% (Fig. 9a), while the relative change in starch ranged from −3% to 94% (Fig. 9b). Consistent seasonal CO2 effects on sugar content were not as pronounced as those for starch, where the greatest CO2 effects occurred in mid-season (Fig. 9). On a mass-basis, elevated CO2 increased leaf starch (17%) and TNC (18%), but did not significantly affect leaf sugar concentration.

Figure 9.

Response of foliar carbohydrate content of mature upper-canopy leaves of Liquidambar styraciflua exposed to ambient (∼364 µmol mol−1, n = 3, open circles) or elevated (∼553 µmol mol−1, n = 2, closed circles) CO2. Foliar (a) sugar (b) starch and (c) total nonstructural carbohydrate (TNC) content was determined for leaves harvested following the collection of A-Ci curves. Values are presented as means ± sem.


Net photosynthetic rates (Agrowth) of leaves in the upper canopy of the L. styraciflua forest stand varied seasonally, but they were consistently higher (44–54%) in elevated [CO2] compared with ambient [CO2] throughout the three seasons of this study. Similarly, Gunderson et al. (2002) reported a sustained 46% photosynthetic enhancement in mid- and upper-canopy foliage of L. styraciflua grown in this same stand. These values were similar to, or slightly lower than the levels of enhancement reported for other deciduous tree species exposed to a doubling of current atmospheric [CO2] (Ceulemans & Mousseau, 1994; Saxe et al., 1998; Norby et al., 1999). Individual L. styraciflua trees that were expanding into the overstorey of a Pinus taeda stand in the Duke Forest FACE experiment were more responsive to CO2 enrichment: Herrick & Thomas (2001) reported that sun leaves in elevated [CO2] had net photosynthetic rates that were stimulated an average of 63% over the course of their 3-yr study. (Note that ‘sun leaves’ of the L. styraciflua trees in the mixed pine-sweetgum canopy of the Duke FACE experiment had morphological characteristics of the mid-canopy leaves in the ORNL FACE, and the upper-canopy leaves at the ORNL FACE, which were fully exposed to the sun, had higher LMA and NA.) Although the response to CO2 enrichment is somewhat lower in the closed-canopy ORNL stand compared to that of individual emergent trees in the Duke FACE experiment, the photosynthetic response of L. styraciflua is strong and persistent, even in closed-canopy conditions.

Photosynthetic capacity (Amax) of L. styraciflua was not significantly affected by 3 yrs of exposure to elevated [CO2]. During this period, net primary productivity was enhanced 21% by CO2 enrichment (Norby et al., 2002), indicating a sustained increase in carbon sinks in elevated [CO2], even though growth responses were constrained by the nonexpanding canopy. Analyses of more than 350 A-Ci curves collected in this experiment revealed no decrease in the rate of Rubisco carboxylation capacity (Vcmax) or RuBP regeneration mediated by electron transport (Jmax), further indicating that photosynthetic acclimation to elevated [CO2] did not occur during the experimental period in this forest stand. Similarly, Herrick & Thomas (2001) found that photosynthetic downregulation did not occur for individual L. styraciflua trees after 3 yrs of [CO2] enrichment at the Duke Forest FACE site. The ratio of Jmax to Vcmax, which represents the balance between the rate of regeneration of RuBP by electron transport and the rate of Rubisco carboxylation, has been theorized to increase in plants grown in elevated [CO2] as leaf N is reallocated to light-harvesting processes from carboxylation processes (Sage, 1994; Medlyn, 1996). However, this was not observed in L. styraciflua in our experiment, indicating that the relationship between carboxylation and light-harvesting processes, which varied seasonally, was not affected by CO2 treatment. Similarly, no shift in the Jmax to Vcmax ratio was observed in a meta-analysis of field-based experiments on European tree species grown in elevated [CO2] (Medlyn et al., 1999), in Populus tremuloides and Acer saccharum (Kubiske et al., 2002) or in L. styraciflua at the Duke FACE site (Herrick & Thomas, 2001).

Photosynthetic nitrogen-use efficiency (PNUE) was higher in L. styraciflua grown in elevated [CO2]. Centritto & Jarvis (1999) noted a similar increase in PNUE in elevated-CO2 plants, along with a reduction in Vcmax per unit leaf nitrogen. Significant positive correlations between Vcmax and Jmax with leaf NA have been shown for many plants (Evans, 1989; Sage et al., 1990; McMurtrie et al., 1992; Leuning et al., 1995; Kellomäki & Wang, 1997), including L. styraciflua (Herrick & Thomas, 2001). Changes in these relationships would be suggestive of reallocation of nitrogen within the trees associated with improved nitrogen-use efficiency (Kellomäki & Wang, 1997). However, no such correlation between Vcmax and Jmax with NA was found for L. styraciflua in this study, which provides no further indication of N reallocation. Apparently, increased N demand in elevated [CO2] was met by increased N uptake (Johnson et al., 2003).

Although changes in leaf characteristics were observed in L. styraciflua grown in elevated [CO2], this did not affect photosynthetic (Amax) and biochemical capacity (Vcmax and Jmax). In many studies, a decline in NM in elevated [CO2] is associated with photosynthetic downregulation (Medlyn et al., 1999; Peterson et al., 1999). In L. styraciflua, NM was generally lower in leaves grown in elevated [CO2] in all sample periods. The average decrease in NM over the 10 sample periods was 10%, compared to the 15% reduction reported by Medlyn et al. (1999) for 12 woody species exposed to elevated [CO2] and 16% for many different species (Curtis & Wang, 1998). Despite the reduction due to growth in elevated [CO2], NM of leaves of L. styraciflua ranged from 15 to 20 mg g−1, which is generally higher than the average NM of forest-grown L. styraciflua (Blinn & Buckner, 1989). Therefore, the lack of photosynthetic downregulation in L. styraciflua may be partially explained by the lack of foliar N deficiency.

In L. styraciflua, there was no significant CO2 effect on NA. These results are in agreement with many studies that have shown significant reductions in NM, but not NA, because of reciprocal changes in NM and LMA (Curtis, 1996; Curtis & Wang, 1998; Medlyn et al., 1999; Norby et al., 1999). Similarly, ChlM, but not ChlA, was significantly reduced in elevated [CO2] in this study due to reciprocal changes in LMA. The reductions in NM may be explained by the presence of increases in nonstructural, and potentially also structural, carbohydrates (i.e. increased LMA in elevated [CO2], which dilutes NM; Epron et al., 1996). The enhanced LMA of L. styraciflua that was noted with growth in elevated [CO2] could produce a secondary effect of increasing Agrowth if it were partially due to the presence of additional photosynthetic tissue (Thomas & Harvey, 1983; Peterson et al., 1999). However, linear regression analysis revealed that for L. styraciflua grown in elevated [CO2] there was no significant relationship between Agrowth and LMA (results not shown).

The photosynthetic response of L. styraciflua varied through the growing season, as has been observed in many other trees (Curtis et al., 1995; Tissue et al., 1997; Rey & Jarvis, 1998; Tissue et al., 1999; Rogers & Ellsworth, 2002), due in part to varying environmental conditions. In this study, Agrowth in both CO2 treatments generally declined about 50% from a high in early season to a low at the end of our measurement period in September; however, the relative enhancement of Agrowth due to elevated [CO2] was maintained throughout this period. In a related study on L. styraciflua conducted during the same 3-yr period at the ORNL FACE site, light-saturated rates of photosynthesis and stomatal conductance were tightly correlated (Gunderson et al., 2002). Reductions in photosynthesis and stomatal conductance observed during mid- and late-season (August to September) in 1998 and 1999 were primarily associated with drought conditions (Gunderson et al., 2002). During these drought conditions, the most significant variable reducing photosynthesis was increased vapour pressure deficit (VPD), followed by reduced soil water potential (Gunderson et al., 2002). Under extreme conditions of dry soil and high VPD, as experienced in late 1998, stomatal conductance and photosynthesis were reduced sharply in both ambient- and elevated-CO2 trees, and differences between the CO2 treatments were small. At other times, CO2 stimulation of photosynthesis was sustained despite late-season reductions in gas exchange. Likewise, in the Duke Forest, L. styraciflua exhibited a pattern of seasonal decline in Agrowth, which was attributed to a reduction in stomatal conductance; similarly, a significant elevated-CO2 effect on photosynthesis was maintained until leaf senescence in early November (Herrick & Thomas, 2003).

In some deciduous tree species, this seasonal variation in photosynthetic response may also be linked to seasonal variations in source activity and sink capacity. For example, Epron et al. (1996) noted that tree species, such as Fagus sylvatica, that continue to produce new sinks beyond the first flush of leaves in spring in response to increased carbohydrate availability, do not show a decrease in early season Amax, in contrast to the decline in Amax observed in species with determinate growth, such as Quercus petraea (Epron et al., 1994). However, when aboveground growth slows dramatically, thus limiting sink capacity, trees show a late-season decline in photosynthesis (Epron et al., 1996). In the herbaceous species, Lolium perenne, growing in the Swiss FACE site, the ratio of carbon sink activity to source activity was greatly increased by removing 89% of the canopy using a defoliant (Rogers et al., 1998). Photosynthetic acclimation in L. perenne was not observed until much of the canopy had regrown and the ratio of sink to source activity had greatly declined (Rogers et al., 1998).

In L. styraciflua, which exhibits indeterminate growth (Kuers & Steinbeck, 1998), general bud set marks a major decline in leaf production and branch extension. Bud set was noted in this experiment (over three growing seasons) to occur around 20 July, which is also the time of peak LAI (Norby et al., 2003). Thus, the marked decline in aboveground sink capacity of these trees was accompanied by a decrease in Amax in the second half of the growing season, which was more pronounced in leaves grown in elevated [CO2] as compared to those grown in ambient [CO2]. The monthly data collected in 1999 for Amax, Vcmax and Jmax illustrate this decline in photosynthetic capacity in elevated [CO2] at the time of bud set, and hence decreased sink capacity, in July. These data suggest that even transient reductions in sink capacity may lead to reduced photosynthetic capacity, thereby reducing, but not eliminating, the positive effects of elevated [CO2] on leaf photosynthesis.

As in many previous studies, growth of L. styraciflua in elevated [CO2] resulted in enhanced leaf TNC concentration (Xu et al., 1994; Van Oosten & Besford, 1996; Saxe et al., 1998; Stitt & Krapp, 1999; Curtis et al., 2000). Starch accumulation in leaves grown in elevated [CO2] was especially pronounced, with an average enhancement of 27% ± 10. There were no consistent seasonal trends as starch was highest in mid-season in 1998, late-season in 1999 and early- to mid-season in 2000. However, in all three growing seasons the greatest effect of elevated [CO2] on starch accumulation occurred in July, during the period of bud set (i.e. reduced sink strength), suggesting a negative relationship between sink strength and carbohydrate accumulation. Nonetheless, starch accumulation in elevated [CO2] did not affect photosynthetic capacity or biochemical capacity.

The lack of photosynthetic acclimation in canopy-dominant L. styraciflua in our experiment in the ORNL FACE site over a 3-yr period is consistent with observations for L. styraciflua during the first 3 yrs of exposure to elevated [CO2] at the Duke Forest FACE site (Herrick & Thomas, 2001). However, these results were not consistent across all species sampled at the Duke FACE site. For example, photosynthetic acclimation (i.e. decrease in Rubisco activity) was observed in 1-yr-old, but not current-year needles of canopy-dominant Pinus taeda (loblolly pine) trees following 2.5 yrs of exposure to elevated [CO2] (Rogers & Ellsworth, 2002). It had been previously shown that P. taeda growing in an adjacent FACE ring in Duke Forest displayed both photosynthetic enhancement (32%) and downregulation (15–20%) following 3 yrs of exposure to elevated [CO2] (Ellsworth et al., 1998). Photosynthetic downregulation has also been observed in open-top chambers after multiple-year exposure to elevated [CO2] in P. taeda (Lewis et al., 1996; Tissue et al., 1996, 1997), P. ponderosa (Tissue et al., 1999) and P. radiata (Turnbull et al., 1998; Griffin et al., 2000).

A possible explanation for the differing responses of L. styraciflua and the various pine species to long-term exposure to elevated [CO2] involves differences in the phenology of deciduous versus evergreen trees. Evergreen gymnosperms are typically characterized by their maintenance of photosynthesizing green leaves (i.e. needles) throughout the entire year. Evergreen leaves generally have lower rates of light-saturated photosynthesis and experience seasonal changes in photosynthetic capacity more gradually than do leaves of deciduous species. In addition, evergreen species have the capacity to maintain needles for more than one growing season, which results in trees having different-aged needles with different physiological characteristics (e.g. source-sink relations). As a result, differences have been noted in CO2 effects on current-year and 1-yr-old needles in evergreen species. Photosynthetic acclimation has been shown to be limited to 1-yr-old needles in P. taeda (Rogers & Ellsworth, 2002) and P. radiata (Turnbull et al., 1998; Griffin et al., 2000). Current-year needles of P. radiata grown in elevated [CO2] have exhibited greater increases in net photosynthesis than 1-yr-old needles, and had enhanced photosynthetic capacity, while 1-yr-old needles showed no CO2 effect (Tissue et al., 2001). It should be noted, however, that deciduous trees exposed to long-term elevated [CO2] may also exhibit photosynthetic downregulation, including Betula pendula (Rey & Jarvis, 1998; Juurola, 2003), Populus tremuloides and Acer saccharum (Kubiske et al., 2002).

In rapidly growing deciduous angiosperms, such as L. styraciflua, the rate of photosynthesis typically accelerates rapidly in the spring as trees refoliate, remains high during the summer, and declines rapidly in late summer as leaves senesce before abscising (Kozlowski & Pallardy, 1997). As a consequence, deciduous trees, which have only current-year leaves, accomplish all of their growth before the dormant season, whereas evergreens are capable of continuing to accumulate dry matter during the dormant period. These differences in phenology result in significant differences in the overall source-sink relations of trees, which reduce the likelihood of downregulation of fast-growing deciduous species such as L. styraciflua.

In conclusion, photosynthetic enhancement (44–54%) of L. styraciflua was maintained for 3 yrs at the ORNL FACE site. There were no significant CO2 treatment effects on photosynthetic (Amax) or biochemical (Vcmax and Jmax) capacity, despite increased sugar and starch accumulation and reduced NM in elevated [CO2], indicating that photosynthetic downregulation did not occur. Elevated [CO2] also increased PNUE in leaves that are not apparently limited by N in this forest. These results suggest that the enhancement of photosynthesis in L. styraciflua in elevated [CO2] is likely to continue for many years, thereby providing more carbon for growth in this closed-canopy deciduous forest.


We gratefully acknowledge the assistance of Sarah Schweitzer, Carol Benner, Zeke Duke, Don Yee, Rich Sobiecki, Carrie Pendley, Sally Hileman and Emily Greer in collecting gas-exchange measurements, Brent Wehner for his tireless efforts performing biochemical analyses, and Jeff Riggs and Don Todd for keeping the lifts and other FACE facilities operational. This research was supported by the NSF/DOE/NASA/USDA/EPA/NOAA Interagency Program on Terrestrial Ecology and Global Change (TECO) by the U.S. Department of Energy, Office of Science, Biological and Environmental Research Program. Support was also provided by the Graduate Program of the Texas Tech University Department of Biology. The research was performed at the Oak Ridge National Laboratory, which is managed by UT-Battelle, LLC for the U.S. Department of Energy under contract DE-AC05–00OR22725.