We measured the short-term direct and long-term indirect effects of elevated CO2 on leaf dark respiration of loblolly pine (Pinus taeda) and sweetgum (Liquidambar styraciflua) in an intact forest ecosystem. Trees were exposed to ambient or ambient + 200 µmol mol−1 atmospheric CO2 using free-air carbon dioxide enrichment (FACE) technology. After correcting for measurement artefacts, a short-term 200 µmol mol−1 increase in CO2 reduced leaf respiration by 7–14% for sweetgum and had essentially no effect on loblolly pine. This direct suppression of respiration was independent of the CO2 concentration under which the trees were grown. Growth under elevated CO2 did not appear to have any long-term indirect effects on leaf maintenance respiration rates or the response of respiration to changes in temperature (Q10, R0). Also, we found no relationship between mass-based respiration rates and leaf total nitrogen concentrations. Leaf construction costs were unaffected by growth CO2 concentration, although leaf construction respiration decreased at elevated CO2 in both species for leaves at the top of the canopy. We conclude that elevated CO2 has little effect on leaf tissue respiration, and that the influence of elevated CO2 on plant respiratory carbon flux is primarily through increased biomass.
To predict the influence of elevated CO2 on plant metabolism and carbon budgets, one must understand the response of both photosynthesis and leaf respiration. Experiments using growth chambers, open-top chambers and free-air carbon dioxide enrichment (FACE) rings have shown that photosynthesis increases in response to elevated CO2 (Gunderson & Wullschleger 1994; Curtis & Wang 1998; Saxe, Ellsworth & Heath 1998; Norby et al. 1999); however, there is no consensus on the effects on leaf respiration (Norby et al. 1999). Complicating any interpretation of the response of leaf respiration is the possibility of both short-term direct and long-term indirect effects (acclimation) (Amthor 1991; Amthor 1997).
Elevated CO2 may also influence leaf respiration indirectly by altering growth rate, non-structural carbohydrate concentration and other changes in tissue composition. Recent meta-analyses have found that elevated CO2 reduces leaf respiration rates by an overall 18% when expressed on a mass basis (µmol CO2 g leaf tissue−1 s−1) (Curtis & Wang 1998; Wang & Curtis 2000). Nevertheless, in individual studies, leaf respiration is often unresponsive or even increases with elevated CO2 (Amthor 1997; Norby et al. 1999; Tjoelker et al. 1999a). Thus, it is impossible to predict either direct or indirect effects of elevated CO2 on leaf respiration at the individual or the ecosystem level at present.
One approach to scaling leaf respiration to the ecosystem level is developing predictive relationships between respiration and tissue nitrogen concentration similar to those for photosynthesis (Peterson et al. 1999). In this analysis, respiration is conceptually partitioned into separate components for growth and maintenance (McCree 1970; Thornley 1970; Amthor 2000b). Growth respiration is the amount of CO2 respired to produce new tissue, and maintenance respiration is the amount of CO2 respired to maintain extant tissue. A strong relationship between maintenance respiration and tissue nitrogen content has been found when several species are considered together (Ryan 1995; Reich et al. 1998; Tjoelker, Reich & Oleksyn 1999b), leading to use of such relationships in scaling respiration to the ecosystem level (Ryan 1991a; Ryan 1991b). Nevertheless, many studies have found either a very weak relationship or no relationship at all between maintenance respiration and leaf nitrogen for individual species (Roberntz & Stockfors 1998; Mitchell, Bolstad & Vose 1999).
The purpose of this study was to quantify the short- and long-term effects of elevated atmospheric CO2 on leaf dark respiration of trees growing in an intact forest ecosystem. For long-term effects, we partitioned leaf respiration into growth and maintenance components to see whether or not elevated CO2 affected these components differently. We compared one evergreen (Pinus taeda) and one deciduous (Liquidambar styraciflua) species native to the south-eastern United States growing under ambient and elevated atmospheric CO2 in a FACE experiment.
Materials and methods
Site and species
The study site was the forest atmosphere carbon dioxide transfer and storage (FACTS-1) research facility, a 17-year-old loblolly pine (Pinus taeda L.) plantation in the Piedmont region of North Carolina (35°58′N, 79°05′W). In this plantation, an understorey of deciduous trees has established naturally, with sweetgum (Liquidambar styraciflua L.) occasionally reaching the canopy. Three 30-m-diameter FACE rings are fumigated continuously with CO2 to raise the atmospheric concentration 200 µmol mol−1 above the current ambient concentrations; three fully instrumented rings serve as controls (Hendrey et al. 1999). For our measurements, we selected three individuals each of codominant loblolly pine and canopy-emergent sweetgum from each ring. Fumigation with elevated CO2 started in August 1996, and as loblolly pine needles live an average of 18 months, needles from treatment plots measured in this study developed under elevated CO2 concentrations.
To access the long-term indirect effect of exposure to elevated CO2 on leaf dark respiration, we measured respiration rates of leaves at average night-time CO2 concentrations in the ambient (400 µmol mol−1) and elevated (600µmol mol−1) plots. Average night-time CO2 concentrations in the rings between 1 June 1999 and 30 September 1999 were 404 µmol mol−1[standard deviation (SD) = 41·2] in the ambient plots and 595 µmol mol−1 (SD = 60·7) in the elevated plots. We selected one fully expanded leaf (sweetgum) or 8–12 fascicles (loblolly pine) from the top and bottom of the canopy for each individual. Respiration measured on these fully expanded leaves represents maintenance respiration. Dark respiration was measured at 5 week intervals (mid-June, average night-time temperature 19·7 °C; late July, average night-time temperature 23·5 °C; early September, average night-time temperature 20·8 °C) on detached leaves using an open gas-exchange system with a conifer cuvette (LI 6400; LiCor Inc., Lincoln, NE, USA) and flow rates of 100 µmol sec−1. Actual leaf area in the cuvette was 17–40 cm2 for sweetgum and 22–47 cm2 for pine. Leaves were stored in humidified plastic bags and all measurements were made within 20 min of detaching. Prior to this experiment, we measured respiration rates of leaves in growth chambers and in the field before and after detaching and found that respiration rates were unaffected for several hours. Night-time measurements of respiration were not comparable to measurements made during the daytime on darkened leaves, even when corrected for temperature differences. Therefore, all measurements were made between 2100 and 0400 h (EST). We checked for possible changes in respiration rate at a given temperature over the course of a night and found none.
To eliminate leaks and minimize diffusion of CO2 through the gas exchange cuvette, all gaskets in the cuvette were covered with a thick layer of silicone putty (Permagum; Virginia KMP Corp., Dallas, TX, USA) and the entire cuvette covered with plastic film (Saran wrap; S.C. Johnson, Racine, WI, USA) (McDermitt et al. 2001). To account for any residual diffusion, the respiration rate of each leaf or group of needles was recalculated using correction factors, measured by sealing an empty cuvette and measuring the difference between CO2 in the sample and reference chambers for the reference CO2 concentration used (400 or 600 µmol mol−1). Correction factors used were an average of three independent runs of an empty sealed cuvette. After respiration measurement, leaf tissue was sampled from the sweetgum for determination of specific leaf area (SLA); leaf tissue was dried at 70 °C and weighed. Pine needles were trimmed to the actual amount of tissue in the cuvette, dried and weighed. Projected leaf area of pine needles was calculated according to Naidu et al. (1998). Dried leaf tissue was analysed for carbon and nitrogen using a micro-Dumas CHN analyser (model NA1500; Carlo Erba Strumentazione, Milan, Italy).
To measure the effect of short-term exposure to elevated CO2 on leaf respiration, we measured respiration rates at four different CO2 concentrations (200, 400, 600, 800 µmol mol−1) in early September on leaves from the top and the bottom of the canopy. After changing the measurement CO2 concentration, respiration stabilized within 5–10 min. All measurements were corrected for residual diffusion as above.
The temperature response of leaf respiration (Q10) in both species was measured in mid-June for leaves from the top of thecanopy. Respiration was measured at four temperatures, from about 22 °C to 32 °C, and was described as Rt = R0[exp(t × lnQ10)/10)], where Rt is the respiration rate at temperature t and R0 is the respiration rate at 0 °C (Johnson & Thornley 1985; Ryan 1991a). We started the measurements at ambient temperature (∼ 26 °C), then at approximately 4 °C below and two temperatures above ambient. There was no evidence of hysteresis in the measurements.
Leaf tissue for measurements of construction cost and construction respiration was collected in late July. Samples were dried and ground to a fine powder and ash-free heat of combustion was determined using a microbomb calorimeter (Gentry Instruments, Aiken, NC, USA). Ash fraction was obtained by combusting samples in a muffle furnace at 500 °C. Construction cost (in g glucose g tissue−1) was calculated according to Williams et al. (1987) and converted to CO2 units. Construction respiration (an estimate of g carbon respired g tissue−1) was calculated from construction cost by subtracting the carbon content of the tissue (Nobel, Alm & Cavelier 1992; Carey et al. 1996; Carey, Callaway & DeLucia 1997).
Statistical comparisons of the indirect effects of CO2 were performed with analysis of variance (anova) (Proc Mixed version 6·12; SAS Institute, Cary, NC, USA) using a repeated measures mixed model design with treatment, species, canopy position and day as fixed factors, block as a random factor and leaf nitrogen as a covariate. Respiration rates calculated on a leaf area and leaf mass basis were analysed separately. For the analysis of direct effects, we performed a regression of respiration versus CO2 for each leaf (Proc Reg version 6·12, SAS Institute), and then analysed the regression slopes using anova. The model for construction respiration and construction cost used treatment, species and canopy position as fixed factors, with block as a random factor. The Q10 model used treatment and species as fixed factors and block as a random factor. Using block as a random factor is more conservative than a fixed-factor analysis, but it allows for generalization to the forest as a whole. Respiration rates and leaf nitrogen were transformed to meet the assumptions of the tests. In all analyses, we started with a model including all main effects and two-way interactions. The final model was determined by sequentially removing terms from the starting model (starting with the largest P-values first) until only significant factors remained.
Direct effects of CO2 on leaf respiration
For loblolly pine, there was essentially no direct effect of CO2 on leaf respiration rate (Fig. 1c). After correcting the raw rates of CO2 production (Fig. 1a) for chamber diffusion (Fig. 1b), the slope of respiration versus reference CO2 concentration was 6·224 × 10−5µmol CO2 m−2 s−1(µmol mol−1 CO2)−1, producing a 1·5–2·5% stimulation in respiration rate for a 200 µmol mol−1 increase in CO2. This slope was statistically significant (d.f.= 5; t=−4·70; P= 0·0053), but should be regarded with caution because an effect of this magnitude is beyond the accuracy of the instrument.
For sweetgum, there was a small residual suppression of leaf respiration rate by CO2 (Fig. 1f). The corrected slope of respiration versus reference CO2 concentration was −1·328 × 10−4µmol CO2 m−2 s−1(µmol mol−1 CO2)−1 (d.f. = 5; t= 10·02; P= 0·0002); a 200 µmol mol−1 increase in CO2 decreased respiration rates by 7–14% depending on measurement date and canopy position. For both loblolly pine and sweetgum, the effect of CO2 on respiration was independent of the CO2 concentrations under which trees were grown (F= 0·72, P= 0·48).
Indirect effects of CO2 on leaf respiration
There was no evidence that trees grown in elevated CO2 had different rates of mass-based or area-based leaf respiration than those grown in ambient CO2(Table 1, Fig. 2). For both species, respiration was significantly higher in leaves from the top of the canopy than in those from the bottom (Table 1, Fig. 2).
Table 1. Final statistical model for the indirect effect of elevated atmospheric CO2 on leaf respiration measured on an area or mass basis. Three trees of each species were sampled in each ring
NDF, numerator degrees of freedom; DDF, denominator degrees of freedom.
Square root nitrogen
Canopy position ×day
Both leaf nitrogen concentration (g N g leaf−1) and content (g N m2 leaf−1) changed through time, peaking in July (Table 2). Leaf nitrogen content was higher for leaves at the top of the canopy for both species (sweetgum P= 0·0001; pine P= 0·0001), but nitrogen concentration did not vary with canopy position. Averaged across the three measurement dates, there was a small reduction in total leaf nitrogen concentration under elevated CO2 for sweetgum (0·0175 versus 0·0159 g N g leaf−1; P= 0·001) but not for loblolly pine (0·0124 versus 0·0113 g N g leaf−1; P= 0·29). Leaf nitrogen content increased under elevated CO2 for sweetgum (0·788 versus 0·989 g N m2 leaf−1; P= 0·0001) but was unchanged in pine (2·97 versus 2·79 g N g leaf−1; P= 0·6017). For sweetgum, leaf nitrogen content correlated positively with area-based respiration rates (Table 1, Fig. 3), with nearly identical relationships between ambient and elevated treatments. There were no significant relationships between leaf nitrogen concentration and mass-based respiration rates.
Table 2. Leaf characteristics of loblolly pine and sweetgum growing in the FACTS-1 research facility
Projected SLA (cm2 g−1)
Projected g N m2 leaf−1
g N g leaf−1
Temperature response of leaf respiration
The Q10 of loblolly pine and sweetgum leaves averaged 2·71 and 2·05, respectively, with no differences between CO2 treatments (Table 3). Basal respiration rates (Ro) were higher for sweetgum than for loblolly pine when expressed on either a mass or area basis. For sweetgum, area-based Ro was significantly higher for trees grown in ambient CO2.
Table 3. Temperature response of leaf maintenance respiration (Q10) and basal leaf respiration rate (Ro). Each value is an average of nine independent measurements
Ro (µmol m−2 s−1)
Ro (µmol g−1 s−1)
Mean values between treatments that were not statistically different (P > 0·05).
There were no differences in leaf construction costs between trees in ambient or elevated CO2 for either loblolly pine or sweetgum (Table 4). Also, there were no differences between the top and bottom of the canopy for loblolly pine, although there was an effect of canopy height on construction cost for sweetgum. Sweetgum leaves at the top of the canopy showed a trend for higher fraction of total carbon in elevated compared with ambient plots (elevated 0·49 ± 0·01 versus ambient 0·44 ± 0·02 g C g leaf−1), but leaves at the bottom of the canopy were similar (elevated 0·45 ± 0·02 versus ambient 0·47 ± 0·01 g C g leaf−1). For loblolly pine, carbon concentrations of leaves at the top (elevated 0·49 ± 0·01 g C g leaf−1; ambient 0·46 ± 0·03 g C g leaf−1) and bottom of the canopy (elevated 0·50 ± 0·02 g C g leaf−1; ambient 0·47 ± 0·03 g C g leaf−1) had similar differences between ambient and elevated plots. Differences in carbon fraction between treatments were not due to differences in ash content because there was no effect of elevated CO2 on ash content (data not shown). Our estimates of leaf construction respiration were significantly lower in elevated compared with ambient plots for both species at the top of the canopy but were not different between treatments at the bottom of the canopy (Table 4).
Table 4. Growth component of respiration in late July measured by calorimetry. Construction cost includes the carbon atoms incorporated into the tissue; construction respiration includes only carbon respired to construct tissue
Heat of combustion (kJ g−1)
Construction cost (mol CO2 kg tissue−1)
Construction respiration (mol CO2 kg tissue−1)
Letters indicate significant differences (P < 0·05) between columns (capital letters) or rows (lower-case letters) within a species.
Direct effects of CO2 on leaf dark respiration
Short-term elevation of CO2 suppressed leaf dark respiration by about 10% for a 200 µmol mol−1 increase in CO2 in sweetgum and had essentially no effect on loblolly pine. If we had not accounted for the diffusion of CO2, we would have concluded erroneously that this effect was much larger. In an extensive study, Amthor (2000a) concluded that direct effects of CO2 on leaf respiration are small (averaging 1·5% decrease for a 400 µmol mol−1 increase in CO2). Our study supports this conclusion for loblolly pine, but sweetgum showed a larger suppression. The existence of a direct effect of CO2 on leaf respiration may have implications on daily plant carbon budgets, as ambient CO2 concentrations change between day and night. For example, as CO2 concentrations increase at night, total night-time respiratory losses of sweetgum could be reduced as leaf respiration is suppressed. However, a direct effect of CO2 on leaf respiration will probably not alter plant carbon budgets as global atmospheric CO2 increases because there appears to be no long-term effect of CO2 on leaf respiration (Fig. 2).
Indirect effects of CO2 on leaf dark respiration
Leaf maintenance respiration did not appear to be altered by long-term exposure to elevated CO2 (Table 1, Fig. 2). Inconsistency in the literature does not allow confidence in predictions of long-term responses of leaf respiration to elevated CO2, but the generalization is that indirect effects are related to changes in leaf tissue composition (Amthor 1997; Saxe et al. 1998; Norby et al. 1999). That is, lower respiration rates are often linked to reduced nitrogen concentrations resulting from accumulation of carbohydrates. In this study, sweetgum showed reduced concentrations of total leaf nitrogen under elevated CO2, but this did not translate into a difference in leaf maintenance respiration rate. Also, there did not appear to be a relationship between leaf respiration rate and total leaf nitrogen concentration.
A positive correlation between respiration and leaf nitrogen concentration has been found for tree seedlings (Griffin, Ball & Strain 1996a; Tjoelker et al. 1999b) and when several species are used in the same analysis to obtain wide variation in nitrogen (Ryan 1995; Reich et al. 1998); but for mature trees within a species, there is often no relationship (Ryan 1995; Roberntz & Stockfors 1998; Mitchell et al. 1999). The absence of a correlation between maintenance respiration and leaf nitrogen concentration may be explained, in part, by amounts of respiratory enzymes generally being in excess of those required for observed rates of respiration (Amthor 1991). In addition, leaf maintenance respiration may be related directly to protein turnover instead of just the total amount of protein (Amthor 2000b). Thus, leaf nitrogen concentration may not predict leaf maintenance respiration in forest stands, and should be used only if it is first shown to be predictive.
When we expressed leaf nitrogen on a leaf area basis (g N cm2 leaf−1) instead of a concentration basis (g N g leaf−1), we found a significant relationship with respiration for sweetgum (Fig. 3). This reflected a relationship with SLA and area-based respiration rates instead of an underlying physiological relationship between respiration and leaf nitrogen.
We found no evidence that elevated CO2 altered leaf construction costs for loblolly pine or sweetgum (Table 4). At higher concentrations of elevated CO2 than we used in this experiment (300–350 µmol mol−1 increase), changes in non-structural carbohydrates sometimes reduce construction costs of leaves (Griffin et al. 1993; Griffin et al. 1996b; Wullschleger et al. 1997). Even in these cases, however, the reduction is only about 3%.
We found that growth under elevated CO2 reduced construction respiration by 21% for loblolly pine and by 39% for sweetgum for leaves at the top of the canopy. We know of no other estimates of the effects of elevated CO2 on leaf construction respiration using calorimetric methods. However, estimates of the construction component of respiration (called growth respiration) from chemical composition (Poorter et al. 1997) or from regression techniques (Wullschleger & Norby 1992; Wullschleger, Norby & Gunderson 1992) have found reductions in leaf construction respiration under elevated CO2 of 10–20%. This reduction was found to be primarily a result of reduced protein concentration, although accumulation of total non-structural carbohydrates (TNC) also played a role. Reduced protein concentrations can lower growth respiration because synthesis of proteins are accompanied by large CO2 production, whereas increased concentrations of TNC can reduce growth respiration because compounds such as starch can be formed with little CO2 production (Poorter et al. 1997).
In our case, lower construction respiration for sweetgum leaves at the top of the canopy under elevated CO2 can be explained partly by lower nitrogen concentrations in elevated compared with ambient plots, and partly by higher fractions of total carbon in elevated compared with ambient plots for leaves at the top of the canopy but not for those at the bottom. For loblolly pine, we found a trend for reduced nitrogen and higher-fraction total carbon for leaves at both the top and bottom of the canopy under elevated CO2. These differences were enough to allow for statistically significant reduced construction respiration at the top of the canopy but not at the bottom.
Increasing atmospheric CO2 concentration by 200 µmol mol−1 did not appear to have a large impact on leaf dark respiration of mature loblolly pine or sweetgum trees growing in an intact forest ecosystem. There appeared to be a short-term direct suppression of respiration by elevated CO2 in sweetgum, but not in loblolly pine. We found no evidence for an indirect long-term effect on leaf maintenance respiration. Although elevated CO2 reduced leaf construction respiration for leaves at the top of the canopy, construction respiration is a small fraction of maintenance respiration on a yearly basis (Ryan et al. 1996). Leaf nitrogen concentration did not predict leaf respiration and should be used with caution for scaling to the stand level. Because elevated atmospheric CO2 did not appear to influence leaf-tissue-specific respiration rates, the effects of elevated CO2 on plant respiratory carbon flux are primarily at the whole-plant level through increased biomass.
We thank Eliza Maclean, Ann Singsaas, Eric Singsaas and Carl Bernacchi for assistance with the field measurements. We also thank Susanne Aref for providing statistical advice, and the Stable Isotope/Soil Biology Laboratory of the University of Georgia Institute of Ecology for C/N analyses. Comments from Richard Norby and two anonymous reviewers greatly improved this manuscript. This work was supported by grant number DE-FG02–95ER62127 from the Department of Energy, Office of Biological and Environmental Research. Additional support from NASA is gratefully acknowledged.
Received 31 October 2000;received inrevised form 25 April 2001;accepted for publication 4 May 2001