Dr Mark G. Tjoelker University of Minnesota, Department of Forest Resources, 115 Green Hall, 1530 Cleveland Ave. N., St. Paul, MN 55108 USA. Fax: 612–625–5212; e-mail: firstname.lastname@example.org
We tested the hypothesis that acclimation of foliar dark respiration to CO2 concentration and temperature is associated with adjustments in leaf structure and chemistry. Populus tremuloides Michx., Betula papyrifera Marsh., Larix laricina (Du Roi) K. Koch, Pinus banksiana Lamb., and Picea mariana (Mill.) B.S.P. were grown from seed in combined CO2 (370 or 580 μmol mol–1) and temperature treatments (18/12, 24/18, or 30/24 °C). Temperature and CO2 effects were predominately independent. Specific respiration rates partially acclimated to warmer thermal environments through downward adjustment in the intercept, but not Q10 of the temperature–response functions. Temperature acclimation of respiration was larger for conifers than broad-leaved species and was associated with pronounced reductions in leaf nitrogen concentrations in conifers at higher growth temperatures. Short-term increases in CO2 concentration did not inhibit respiration. Growth in the elevated CO2 concentration reduced leaf nitrogen and increased non-structural carbohydrate concentrations. However, for a given nitrogen concentration, respiration was higher in leaves grown in the elevated CO2 concentration, as rates increased with increasing carbohydrates. Across species and treatments, respiration rates were a function of both leaf nitrogen and carbohydrate concentrations (R2 = 0·71, P < 0·0001). Long-term acclimation of foliar dark respiration to temperature and CO2 concentration is largely associated with changes in nitrogen and carbohydrate concentrations.
Increasing atmospheric concentrations of carbon dioxide (CO2) and potential climate warming may alter the net carbon exchange of forest trees through changes in rates of respiration. In the short term, rates of dark respiration in plants are highly responsive to ambient temperatures. Some studies have shown that dark respiration may be reversibly inhibited by short-term increases in CO2 concentrations (Wullschleger, Ziska & Bunce 1994; Drake, Gonzàlez-Meler & Long 1997; but see Ryle, Powell & Tewson 1992; Ziska & Bunce 1994; Mitchell et al. 1995). However, acclimation of dark respiration in plants to temperature (Larigauderie & Körner 1995) and possibly CO2 concentration (Poorter et al. 1992; Curtis 1996) calls into question long-term predictions of respiratory losses of carbon based on short-term response functions. An understanding of the nature of acclimation and species differences in this regard is critical in order to predict plant response to potential climate change (Körner 1995).
It is well known that rates of dark respiration in plants acclimate to changes in ambient temperature in both field studies (e.g. Pereira et al. 1986) and in controlled environments in periods as short as several days (Rook 1969; Pearcy 1977; Tranquillini, Havranek & Ecker 1986). In the short term (minutes to hours), rates of dark respiration in plants typically increase exponentially with increases in temperature. A downward shift of the entire short-term temperature response function is commonly observed in plants acclimated to warmer temperatures (Rook 1969; Pearcy 1977; Tranquillini et al. 1986). As a result, comparisons of dark respiration at common temperatures reveal a lower rate of dark respiration in plants acclimated to higher temperatures. Furthermore, acclimation responses to temperature have been shown to vary widely among alpine and lowland herbaceous species (Larigauderie & Körner 1995; Arnone & Körner 1997). Little is known concerning the leaf traits and factors underlying temperature acclimation of respiration either within or among species.
Leaves grown in an elevated CO2 concentration frequently exhibit decreased N and specific leaf area (SLA, projected leaf area per unit leaf mass), and increased concentrations of starch and total non-structural carbohydrates (TNC) compared with controls (Curtis 1996; Poorter et al. 1997). Although respiration rates are considered to be primarily controlled by demand for respiratory products (Farrar 1985; Farrar & Williams 1991), increased rates of respiration have been associated with increased carbohydrate concentrations in leaves (Azcón-Bieto & Osmond 1983), especially in response to CO2 enrichment (Hrubec, Robinson & Donaldson 1985; Thomas et al. 1993; Thomas & Griffin 1994). Sugars serve as substrates for glycolysis and mitochondrial respiration. Thus, in response to CO2 enrichment, declining leaf N and increasing TNC may affect rates of dark respiration in opposite directions. In this study we examine leaf respiration and relationships with SLA, leaf N, and TNC both within and among seedlings of five boreal tree species. General models of dark respiration based on leaf traits would enable better predictions of tree and canopy-scale respiration.
The broad-leaved species, Populus tremuloides Michx. and Betula papyrifera Marsh., the deciduous conifer, Larix laricina (Du Roi) K. Koch, and evergreen conifers Pinus banksiana Lamb. and Picea mariana (Mill.) B.S.P., comprise the major tree species of the North American boreal forest and were selected for study in a controlled-environment experiment on the potential interactive effects of CO2 concentration and temperature on growth and physiology (Tjoelker, Oleksyn & Reich 1998a,b). These five tree species provided a test of the nature and generality of acclimation responses to temperature and CO2 concentration among diverse taxa. Given a lack of information on the nature of respiratory acclimation, our objective was to compare the temperature response functions of foliar dark respiration of seedlings of these species grown in various temperature and CO2 environments. We explore the extent that temperature acclimation is associated with adjustments in intercepts or slopes of temperature–response functions. We test the hypothesis that acclimation in rates of dark respiration in foliage grown in varying CO2 concentrations and temperature environments is associated with changes in leaf structure and chemistry by comparing respiration rates at a common temperature with changes in SLA, leaf carbon, TNC, and N concentrations. We test whether a leaf N-based model of respiration is altered by growth in contrasting CO2 concentrations and temperature environments and examine a model of leaf respiration as a function leaf N and TNC.
METHODS AND MATERIALS
We obtained seeds of aspen (P. tremuloides), paper birch (B. papyrifera), tamarack (L. laricina), black spruce (P. mariana) and jack pine (P. banksiana) originating from native populations in northern Minnesota, USA. Species are hereafter referenced by their generic names. Seeds were sown into 2·7 dm3 pots filled with a 4 : 1 ratio (v/v) mixture of pure silica sand to a soil medium of equal proportions of loam, sand, and peat. Germination occurred in uniform conditions (370 μmol mol–1 CO2, 20 °C, 80% relative humidity, 16 h photoperiod at 577 μmol m–2 s–1, PPFD) in controlled-environment chambers (Conviron E15; Controlled Environments, Inc., Winnipeg, Manitoba, Canada). The pots were watered daily with deionized water and a modified half-strength Hoagland's solution. Plants of each species were grown separately in 10 to 15 pots in each growth chamber. The pots were periodically repositioned to prevent shading and to randomize any effect of position across all pots within a chamber.
We selected CO2 treatments of approximate current and future (≈ 1·5 × ambient) mean atmospheric concentrations and three light/dark period temperatures of 18/12, 24/18, and 30/24 °C. The CO2 and temperature treatments were applied as a complete factorial in a set of six identical growth chambers (Conviron E15). Treatments began about 18 d after germination and lasted 97 d. The CO2 concentrations averaged 370 and 580 μmol mol–1 in the ambient and elevated treatments.
Relative humidities were set at 60/65% (light/dark period) in the 18/12 °C treatment, 65/70% at 24/18 °C, and 70/75% at 30/24 °C to partially offset the increased vapour pressure gradients at higher temperatures. In each chamber, lighting consisted of metal halide and sodium high-intensity-discharge lamps, providing a maximum of about 1200 μmol m–2 s–1 (PPFD) at plant height. Given varying light habitat requirements among the species, the 16 h photoperiod was divided into two 5 h periods of one-half PPFD (532 ± 72 μmol m–2 s–1) at the beginning and end of the photoperiod with the middle 6 h period at full PPFD (1200 ± 98 μmol m–2 s–1). PPFD, temperature, CO2 concentration, and relative humidity were measured regularly to verify treatment conditions. In a test for chamber uniformity, a prior factorial experiment in the same set of chambers with seedlings of these species showed that neither plant dry mass growth nor measured environmental factors (i.e. CO2 concentration, temperature, PPFD, and relative humidity) differed between replicate chambers. Further details are provided in Tjoelker et al. (1998a).
Temperature response of dark respiration
Specific rates of respiration in plants are negatively correlated with dry mass of plants or plant parts (Poorter & Pothmann 1992), including these boreal tree species grown from seed (Reich et al. 1998b). To account for this phenomenon, we measured shoot respiration rates at comparable plant sizes and stage of leaf development as well as five to seven common harvest dates between 19 d and 91 d of treatment (Tjoelker, Oleksyn & Reich, 1999). To minimize differences in plant size and provide comparable amounts of mature sample tissue for determination of temperature response of shoot dark respiration, species were measured separately. Faster-growing Populus and Betula were measured prior to the slower-growing conifers, which developed mature foliage later. Measured rates of mature leaves predominately reflect maintenance respiration. Populus were measured at 30 d, Betula 68 d, Larix 82 d, Picea 84 d and Pinus at 97 d of treatment. For each species, leaves or needles had fully developed in the treatment conditions and shoot samples were more comparable in terms of dry mass (mean of 0·3 to 0·7 g among species versus five- to 10-fold variation at common harvests) and stage of development (cotyledons versus mature foliage) than if measures were conducted at a common time. The degree of temperature acclimation was consistent among treatments and species whether compared at a common size or age (Tjoelker et al., 1999).
Individual shoot samples consisted of entire shoots or intact stem segments (7–10 cm length) with attached leaves or needles from each CO2 and temperature treatment combination. For the temperature–response samples, stem dry mass as a proportion of total sample mass constituted, on average, 18% for Populus, 22% for Betula, 17% for Larix, 7% for Pinus, and 9% for Picea. Before measurement, shoot samples were cut from the plants. Needles and leaves were left attached to stems to avoid injuring the tissue and minimize desiccation upon repeated measures at the different temperatures. Prior checks revealed no difference in CO2 exchange rates of attached or cut shoots. We measured the temperature response of three shoots of each species and treatment combination, except Populus, for which two shoots were sampled in the 18/12 and 24/18 °C treatment combinations.
At the end of the 8 h dark period on each measurement date, pots from each treatment combination were moved to a darkened growth chamber (Conviron E15). Rates of CO2 exchange were measured using infrared gas analysers and cuvettes (LCA-3 and PLC-C; Analytical Development Co. Ltd, Hoddesdon, UK), operated in an open configuration. Water vapour was removed from the analyser air stream using columns of magnesium perchlorate. Temperature response of respiration was determined at growth chamber temperatures of 12, 18, 24, and 30 °C. The sequence of measurement temperatures was reordered for each of the three replicate groups, consisting of one shoot from each of the six treatment combinations. At each measurement temperature, net CO2 exchange of each shoot was measured at both treatment CO2 concentrations by changing the concentration of CO2 in the cuvette, averaging 365 and 592 μmol mol–1. Between measurements the cut shoots were kept in the dark and hydrated. Measures for each replicate group were completed within 3 h.
Leaf structure and chemistry
Leaf and needle areas (one-sided projected) were measured using a video image analysis system (AgVision; Decagon Devices, Inc., Pullman, WA, USA). Leaves, needles and stems were separated and oven dried (65 °C) and masses determined. Nitrogen and carbon (C) concentrations of dried and ground leaf and needle tissues were measured by a CHN element analyser (Perkin Elmer Corp., Norwalk, CT, USA). Total non-structural carbohydrate (TNC) concentrations of leaves were determined using the methods of Haissig & Dickson (1979) and Hansen & Møller (1975). Sugars were extracted from oven-dried and ground tissue in methanol–chloroform–water, and tissue residuals were used for determination of starch content. Following addition of anthrone, soluble sugars were determined spectrophotometrically at 625 nm. Starch was gelled and converted to glucose with amyloglucosidase (Sigma, St. Louis, MO, USA). Glucose concentrations were determined by assaying with glucose oxidase. The sample was mixed with peroxidase-glucose oxidase-o-dianisidine dihydrochloride and absorbance measured at 450 nm after a 30 min incubation at 25 °C. TNC concentrations are the sum of the soluble sugar and starch concentrations and are presented as glucose equivalents (mg glucose g dry mass–1). Respiration rates were calculated on the basis of shoot (leaf + stem) mass (shoot Rsm, nmol g–1 s–1), leaf mass (leaf Rsm, nmol g–1 s–1), leaf area (leaf Rsa,μmol m–2 s–1), and leaf nitrogen (leaf RsN, μmol mol N–1 s–1). Leaf-based rates are approximate, since they express shoot respiration rate with respect to leaf mass, area or N.
The temperature response of dark respiration generally follows Van’t Hoff's reaction rate–temperature rule that a reaction rate increases exponentially with temperature. The temperature coefficient Q10 describes the increase in rate with a 10 °C increase in temperature. Values of Q10 (between 12 and 30 °C) were estimated separately for each species and treatment combination using non-linear regression of shoot respiration rate (Rsm) against the four measurement temperatures (T) of 12, 18, 24 and 30 °C:
Rsm(t) = R12× Q10(T– 12)/10,
where R12 is the estimated specific respiration rate at the reference temperature of 12 °C. Inspection of the log–linear model fits revealed a small but consistent departure from linearity, suggesting that the responses were not strictly exponential over this temperature range. Consequently, Q10 values declined with increasing measurement temperature. Here, we present Q10 values determined over the entire measurement temperature range of 12 to 30 °C as the basis for comparing the temperature response functions among species and treatments. A separate-slopes analysis of covariance was used to determine if slopes (i.e. Q10) of the log–linear temperature response functions were homogeneous among treatment groups. Subsequently, a same-slopes analysis of covariance was used to test the equality of the intercepts among the treatment groups. Since the determination of the temperature responses required repeated measures on individual plants, analyses included both between-plant and within-plant variation.
The specific respiration rate of plants grown at the high temperature (30/24 °C) and at the lowest temperature (18/12 °C) measured at their respective dark-period growth temperatures was used to calculate the acclimation ratio in the form of a Q10 (see equation), where Rsm(t) denotes the rate at 24 °C and R12 the rate at 12 °C. The acclimation ratio Q10 values were compared to the Q10 of the instantaneous response of dark respiration to measurement temperature to assess the degree of acclimation of respiration to thermal environment using equivalent indices. A value of 1·0 indicates complete acclimation, reflecting no difference in respiration rate of plants grown and measured in their different thermal environments. Values greater than 1·0 and less than the Q10 of the instantaneous temperature response function indicate partial acclimation. No acclimation has occurred if the acclimation ratio and Q10 values are equal, since the Q10 of the short-term temperature response curve would result in similar differences in respiration across growth temperatures. An equivalent approach is described in detail elsewhere (Larigauderie & Körner 1995).
Regression and analysis of variance were used to examine responses to long-term growth environment. We analysed each species separately. For each of the six CO2 and temperature treatment combinations, the three measured plants were considered subsamples and were used to calculate means for each treatment combination. Most leaf traits showed linear responses to growth temperature. An inspection of the means of the treatment combinations and a test for differences between CO2 concentration treatments in the homogeneity of slopes of the growth temperature responses both indicated that the CO2–temperature interactions were not significant (P > 0·05 in separate-slopes analysis of covariance). Thus, we concluded that CO2 concentration and temperature effects were independent in this study. Therefore, in analysis of variance we restricted our statistical tests to the replicated main effects of CO2 concentration and temperature. In constructing F-tests of the effects of CO2[1 degree of freedom (d.f.)] and temperature treatment (2 d.f.), the CO2× temperature treatment interaction (2 d.f.) was used as the mean square error term (2 d.f.), on the assumption that the added effect of the interaction term was zero (Sokal & Rohlf 1981). The main effect of temperature treatment was partitioned into single-degree-of-freedom contrasts for linear and lack of fit terms to examine responses to long-term growth temperature. All analyses were conducted with statistical analysis software (JMP 3·2; SAS Institute, Cary, NC, USA).
Short-term response of dark respiration to temperature and CO2
Specific rates of shoot dark respiration (shoot Rsm) increased approximately exponentially with increasing measurement temperature in each species and treatment combination (Fig. 1). The slopes of the log–linear temperature response curves, and hence Q10 (12–30 °C), did not differ statistically among the growth temperature and CO2 treatment combinations (P = 0·14 for Betula, otherwise P≥ 0·49, Table 1). However, species exhibited different Q10 values; Q10 in Populus shoots was the lowest, averaging 1·8 compared with about 2·1 for the other four species.
Table 1. . Temperature response parameters1 of shoot dark respiration (Rsm, nmol g–1 s–1) of five boreal tree species grown in various CO2 concentrations and temperature treatment combinations
Changes in specific respiration rate at reference temperature (R12, Table 1) resulted in shifts in the overall elevation of the temperature response curves among the three growth temperature treatments, demonstrating that acclimation of respiration occurred in response to growth temperature environment (Fig. 1, Table 1). In general, growth in lower temperatures increased rates of dark respiration in each species when compared over a common range of measurement temperatures between 12 and 30 °C (Fig. 1). In both CO2 concentrations for each species, rates measured at common temperatures were higher in plants grown at the lowest temperature treatment (18/12 °C) than the highest temperatures (30/24 °C) with the 24/18 °C treatment usually intermediate. Estimates of R12, the specific rate of respiration at the reference temperature of 12 °C, declined with increased growth temperature in all species. Same-slopes analysis of covariance showed that the log–linear intercepts differed significantly among temperature treatments in each species (P≤ 0·02), except Populus (P = 0·24) Only Betula exhibited an effect of growth CO2 concentration on R12 which was about 25% lower in plants grown at 580 than 370 μmol mol–1 (P = 0·008). By comparison, short-term changes in measured CO2 concentration had no direct effect on rates of dark respiration in any species measured at growth temperatures (P≥ 0·15, Table 2) or other measurement temperatures (data not shown).
Table 2. . Specific respiration rates (nmol g–1 s–1) of shoots of five boreal tree species grown and measured in various CO2 concentrations and temperature treatment combinations and a test of a short-term inhibition of respiration by CO2 concentration
Acclimation of respiration to growth temperature
The extent that the instantaneous temperature response (e.g. 12–24 °C) of dark respiration (Q10) differs from the response of respiration of plants grown and measured at their respective growth temperature treatments (e.g. 12–24 °C) reflects the degree of acclimation to thermal environment. Shoot Rsm exhibited lower proportional increases with increased growth temperatures in comparison to short-term temperature responses. The acclimation ratio (analogous to Q10) of the mean shoot Rsm of plants grown and measured in dark period temperatures of the highest (24 °C) and lowest temperature treatments (12 °C), varied widely among the species (P = 0·04, Fig. 2), ranging from 1·1 (Picea) to 1·9 (Populus). The two broad-leaved species exhibited higher mean acclimation ratios than the three conifers, indicating less acclimation to growth temperature in broad-leaved than needle-leaved species. Populus did not acclimate, since the acclimation ratio and instantaneous Q10 were comparable, whereas Picea and Larix exhibited nearly complete acclimation to growth temperature environment, as the acclimation ratio was near 1·0, indicating a comparatively small proportional change in mean rate at the highest compared to lowest temperature treatment. For Betula and Pinus the acclimation ratio was less than the value of instantaneous Q10 (but greater than 1·0) suggesting a partial acclimation to thermal environment. As a consequence of partial temperature acclimation in Populus, Betula, and Pinus, specific rates of shoot respiration measured at growth temperatures increased linearly with increasing growth temperatures (P≤ 0·06, Table 2).
Leaf structure and chemistry
Inspection of treatment means suggested that CO2 concentration and growth temperature treatment effects on leaf N, C, and TNC concentrations and SLA were largely independent among the five boreal species (Table 3). All species had a lower mean leaf N (mg N g–1) in plants grown at 580 compared to 370 μmol mol–1 CO2 (averaged across temperature treatments); however, the effect was statistically significant only for Betula and Larix (P≤ 0·009, Table 3). Values of leaf N in plants grown at the elevated CO2 concentration, on average, ranged from 65% (Betula) to approximately 85% (Picea) of ambient-grown plants. In response to increased growth temperature treatment, each of the conifers had linear declines in needle N concentration (P≤ 0·04) which were about one-third lower in needles grown at the highest compared to lowest temperatures.
Table 3. . Leaf structure and carbon, nitrogen, and carbohydrate concentrations of five boreal tree species grown in various combined CO2 concentrations (μmol mol–1) and temperature (°C) treatments. Means (± SE of plants sampled within treatments) of leaf nitrogen (N, mg g–1), carbon (C, mg g–1), total non-structural carbohydrates (TNC, mg glucose equivalents g–1), and specific leaf area (SLA, cm2 g–1) are shown
Compared with the effects on leaf N, the treatment combinations had minimal effects on leaf carbon concentration (mg C g–1) in these species. In Betula the mean C concentrations increased linearly with increasing growth temperatures (P = 0·05). However, in Larix mean C concentrations declined with increasing growth temperature (P = 0·03). The remaining species had no statistically significant differences in C content among growth environments. As a result of changes in N concentration among the CO2 and temperature treatments, the C : N molar ratio increased by 20–50% in plants grown at 580 compared with 370 μmol mol–1 CO2 in each species, except Pinus (data not shown). For the conifers, the C : N ratio increased by 40–60% with increased growth temperature.
The concentration of TNC in leaves of plants grown in elevated compared with ambient CO2 concentrations increased on average by 10% in Populus, 55% in Betula, 31% in Larix, and 42% in Picea, although differences were statistically significant only in Populus (P = 0·04). For Populus, TNC also decreased at higher growth temperatures (P = 0·004). TNC was predominately in the form of soluble sugars; the starch fraction was 39% of TNC in Betula and ranged from 10 to 18% of TNC for the other species.
Mean SLA was 10 to 40% lower in each of the five species grown in elevated compared to ambient concentrations of CO2. However, the effect of CO2 concentration was statistically significant only in Betula (P = 0·01). SLA increased with higher growth temperatures in Betula (P = 0·09) and Larix (P = 0·10) and the other three species exhibited increasing trends, especially for plants grown at the ambient but not elevated CO2 concentration. With the exception of Betula, reductions in SLA for plants grown at 580 compared to 370 μmol mol–1 CO2 were at least two times larger at the high (30/24 °C) than low (18/12 °C) temperature environment.
Overall, increases in leaf TNC concentration only accounted for a small fraction of the reduction in SLA in response to CO2 enrichment. Across all leaves, TNC content (g m–2) and specific leaf mass (1/SLA, g m–2) were weakly related (R2 = 0·15, P = 0·0002, data not shown). Leaf TNC concentration constituted, on average, 8% of leaf mass in the ambient CO2 concentration and 10% of leaf mass in the elevated CO2 treatment. Removing the diluting effect of TNC on leaf N concentration did not substantially alter the response of leaf N concentration to CO2 concentration and temperature treatments among the species (data not shown). Thus, changes in leaf TNC concentrations among the treatments had relatively minor effects on SLA and leaf N concentration.
Mass, area and N-based respiration
To examine the nature of acclimation of dark respiration to temperature and CO2 concentration in relation to leaf structure and chemistry, we compared shoot respiration rates at a common measurement temperature (18 °C) and measurement concentration of CO2 (370 μmol mol–1). Since species differed in SLA and leaf N concentrations, and growth environment affected these traits, we calculated net CO2 efflux of shoots on the basis of leaf mass, area, and nitrogen content (Table 4).
Table 4. . Respiration rates of shoots of five boreal tree species on a leaf mass (Rsm, nmol g–1 s–1), area (Rsa,μmol m–2 s–1), and nitrogen (RsN, μmol mol N–1 s–1) basis measured at a common temperature (18 °C) and 370 μmol mol–1 CO2 concentration. Means (± SE of plants sampled within treatments) of the growth temperature and CO2 treatment combinations are shown
Inspection of the treatment means suggested that the effects of CO2 concentration and growth temperature on respiration rates were independent (Table 4). In each species leaf mass or area-based rates measured at 18 °C varied among the growth temperature environments. In general, mean respiration rates declined linearly with increasing growth temperature whether expressed as leaf Rsm (P≤ 0·12 in each species except Betula) or leaf Rsa (P≤ 0·06 in each species except Populus), reflecting the acclimation response evident in the temperature response curves (Fig. 1, Table 1). The magnitude and sometimes direction of CO2 effects on respiration among the five boreal species varied depending on the basis of expression of rates. For example, mean leaf Rsm was lower (although not statistically significant) in plants grown in the elevated CO2 concentration compared to ambient CO2 in Populus (P = 0·13) and Betula (P = 0·15). In contrast leaf Rsa showed an opposite trend, increasing on average in leaves grown at 580 compared to 370 μmol mol–1 in Populus (P = 0·09), Betula (P = 0·09), in part owing to decreases in SLA with CO2 enrichment. Unlike leaf Rsm or Rsa, RsN did not vary between CO2 concentration treatments in any species, although RsN exhibited declining trends with increased growth temperatures in Betula and Picea.
Respiration in relation to leaf N and TNC
To test the hypothesis that acclimation of dark respiration to temperature and CO2 concentration is related to changes in leaf N, we examined the relationship of leaf N and respiration both among and within species. Rates of shoot respiration expressed on a leaf mass basis (leaf Rsm at 18 °C, 370 μmol mol–1 CO2) were positively correlated with leaf N among species, treatments, and individual leaves (R2 = 0·55, P < 0·0001, n = 86). Likewise within each species except for Populus, correlations of leaf Rsm and leaf N across treatments were positive and statistically significant (R2≥ 0·26, P≤ 0·03).
Separate slopes analysis revealed that slopes of the leaf Rsm–leaf N relationship did not statistically differ between CO2 treatments (P = 0·17, Fig. 3), but that the intercept was higher in the elevated CO2 concentration (P = 0·0003). Therefore, for any given leaf N concentration plants grown in the elevated CO2 concentration had a higher rate of respiration than plants grown in the ambient CO2 concentration. In contrast, neither the slope (P = 0·53) nor intercept (P = 0·42) of the leaf Rsm–N relationship differed among the three temperature treatments (across CO2 concentrations and species). Thus, CO2 concentration effects on leaf Rsm occurred through a shift in the overall elevation (i.e. intercept or respiration rate at standard temperature) of the linear relationship, whereas growth temperature effects on leaf Rsm were primarily manifested through changes in leaf N, i.e. rates varied in proportion to leaf N concentration.
We explored whether acclimation in the leaf traits, SLA and TNC could account for the higher intercept of the Rsm–N relationship for plants grown in elevated compared to the ambient CO2 concentration. The residuals of an overall linear regression fit of leaf Rsm against leaf N were correlated with SLA (R2 = 0·10, P = 0·003), but more so with with leaf TNC (R2 = 0·35, P < 0·0001). Multiple regression analysis of leaf Rsm (across all sources of variation) using leaf N and TNC as independent variables jointly accounted for 71% of the total variation (P < 0·0001, Fig. 4). For a given leaf N, respiration increased linearly with increasing TNC; and for a given leaf TNC, respiration rates increased linearly with increasing leaf N. Thus, higher rates of respiration at a given leaf N in plants grown in elevated than ambient CO2 concentrations (Fig. 3) were, in large part, associated with increased TNC (Fig. 4).
Species variation in temperature and CO2 acclimation of dark respiration
The five boreal species exhibited large variation in degree of acclimation of dark respiration to thermal environment. Our results suggest that rates of foliar dark respiration in seedlings of boreal conifers acclimate to changing temperatures to a greater degree than broad-leaved species. In a comparative study of alpine and lowland species grown in controlled environments, dark respiration of leaves among species differed in degree of acclimation to growth temperature (10 versus 20 °C), ranging from complete acclimation to no acclimation and averaging about 50% of the instantaneous temperature response (Larigauderie & Körner 1995). Respiration rates of seedlings of Pinus radiata D. Don. grown at 15/10 or 33/28 °C had partially acclimated 2 d after transfer to the reciprocal temperature treatment, as evident in the approximate doubling in respiration rates of plants grown at 33/28 °C or halving of rates of plants grown at 15/10 °C compared with pre-transfer rates at common measurement temperatures (Rook 1969). In Larix decidua Mill. seedlings, needle respiration rates were approximately 50% lower in plants acclimated for two weeks to 24 °C compared to 8 °C (Tranquillini et al. 1986). Species differences in respiration acclimation preclude simple predictions of the response of respiratory carbon losses at the ecosystem scale.
In four of five species, respiration rates measured at growth temperatures exhibited smaller increases in rates across thermal environments than expected from short-term temperature responses (Q10). Predicted changes in dark respiration to climate warming, if described by instantaneous temperature responses (i.e. Q10), are likely to overestimate carbon losses in dark respiration from zero to 100% depending upon species, since it is acclimation and not the instantaneous response to temperature that defines responses in the growth environment (Körner 1995).
Species differences in respiration acclimation responses to temperature and CO2 concentration were associated with changes in leaf structure and chemistry. Unlike the broad-leaved species, the needle-leaved species exhibited declines in leaf N concentration on a mass and area-basis in response to higher growth temperature. In broad-leaved species, concomitant changes in SLA and mass-based N concentration among treatment combinations resulted in no change in area-based N content. Lower N concentrations in conifer needles grown at higher temperatures were not associated with changes in SLA, which increased or stayed the same, while N concentrations declined, perhaps reflecting a metabolic adjustment to thermal environment. The increased needle N concentration in cold-grown conifers is consistent with findings of increased protein levels in leaves of cold-acclimated plants (Graham & Patterson 1982). In a similar pattern, plant species or genotypes originating in cold climates of high altitudes and latitudes, often exhibit higher leaf N concentrations or higher rates of respiration at a common temperature than counterparts from warmer environments (Lechowicz, Hellens & Simon 1980; Criddle et al. 1994; Reich et al. 1996; Oleksyn et al. 1998), perhaps reflecting metabolic adaptation to colder environments. Thus, adequate consideration of both temperature acclimation and adaptation of respiratory processes will be critical in modelling plant response to climate warming.
We found no evidence of an interaction effect between CO2 concentration and temperature growth environment on respiration. In contrast to our findings, growth temperature altered the long-term effect of CO2 concentration on leaves and plants of Dactylus glomerata L. cv. Potomac and Medicago sativa L. cv. Arc (Ziska & Bunce 1993, 1994). In those studies, reductions in CO2 efflux rates in response to CO2 enrichment were larger in plants grown at lower (15, 20 °C) than higher (25, 30 °C) temperatures.
Growth of plants in elevated concentrations of CO2 may result in decreased rates of leaf dark respiration, especially if rates are expressed on a leaf mass basis rather than area basis (Poorter et al. 1992; Curtis 1996). In the present study, increases in TNC in response to CO2 enrichment were relatively small compared with the reduction in leaf N. Thus, a reduced N concentration in response to CO2 enrichment was only partially accounted for by a dilution effect of increased TNC that has been observed in some studies (Thomas et al. 1993).
Q10 and temperature response
The evidence of common slopes of the exponential response of dark respiration to measurement temperature among the CO2 concentrations and growing temperature environments suggests that the temperature sensitivity of dark respiration (Q10, 12–30 °C) was not affected by growth environment for these five species, although intercepts and hence elevation of the relationships did vary among growth environments. The Q10 values at these temperatures were within the range of those commonly reported for herbaceous and woody plant species, averaging about 2·3 (Sprugel et al. 1995; Larigauderie & Körner 1995). The similarity of Q10 values among plants grown in thermal environments that differed by as much 12 °C suggests that Q10 may be relatively stable across growth environments; and therefore, useful in modelling the response of respiration in these species to short-term variation in temperature (e.g. diurnal) in a variety of environments, after adjustments are made for growth temperature effects on specific respiration rates at reference temperature and the dependence of Q10 on measurement temperature.
In contrast to our findings, some studies have reported a decrease in Q10 of dark respiration with increased temperature of growth environment (e.g. Fukai & Silsbury 1977; Lawrence & Oechel 1983; Ziska & Bunce 1994) and changes in Q10 with CO2 concentration of the growth environment (Ziska & Bunce 1994; Carey, DeLucia & Ball 1996). However, comparisons of temperature responses of respiration among studies are difficult, since Q10 values are affected by the exact range of measurement temperatures used in its determination (e.g. Sprugel et al. 1995) and temperature responses are sometimes linear rather than exponential (Brooks et al. 1991; Ziska & Bunce 1994). Our findings suggest that Q10 of dark respiration of tree seedlings is not affected by growth CO2 concentration or growth temperature.
Respiration versus leaf N and TNC
Much support exists for the hypothesis that mass-based respiration increases with higher tissue N (Merino, Field & Mooney 1982; Ryan 1995; Reich et al. 1996). Linear relationships of respiration rates and leaf N have been described for boreal trees and shrubs (Chapin & Tryon 1983; Ryan 1995; Reich et al. 1998b) and appear to be consistent across diverse plant functional types and biomes (Reich et al. 1998a). The slopes of respiration-N relationships reported for foliage of boreal trees and shrubs (Ryan 1995) and woody plants in the field (Reich et al. 1998a) were generally comparable with those of the present study. Thus, despite the vastly different growing conditions (field versus controlled environment) and age of trees (mature trees versus seedlings), our seedling data support the hypothesis of general linear relationships between foliar N content and rates of dark respiration.
In the present study rates of respiration at common leaf N concentrations were higher for foliage of plants grown at elevated compared to ambient CO2 concentrations among seedlings of five boreal tree species, in large part, a result of increased leaf TNC. Previous studies have demonstrated a positive correlation of leaf respiration rates and carbohydrate concentrations (Azcón-Bieto & Osmond 1983). This relationship is observed in response to CO2 enrichment (Thomas et al. 1993; Thomas & Griffin 1994) and is not associated with adjustments in total activities of respiratory enzymes of isolated mitochondria, but rather with a higher carbohydrate status (Hrubec et al. 1985). Thus, higher carbohydrate contents for a given leaf N concentration may result in higher rates of respiration observed in the present study and elsewhere (Mitchell et al. 1995).
Although increased leaf respiration rates in plants grown in elevated CO2 concentrations were associated with generally higher TNC concentrations, increases in specific respiration rates in elevated CO2-grown plants were partially offset by reduced leaf N. In contrast, growth in colder thermal environments resulted primarily in increased leaf N. Our findings suggest that changes in leaf N and TNC underlie acclimation of dark respiration to temperature and CO2 concentration, and may be useful in modelling foliar maintenance respiration in response to increased CO2 concentrations in contrasting thermal environments.
We thank Roma Zytkowiak and Piotr Karolewski of the Institute of Dendrology in Kórnik, Poland for conducting the carbohydrate analyses. We thank Mike Tobin and Dave Peterson for reviews. The study was supported in part through a doctoral dissertation fellowship award to M. G. Tjoelker from the Graduate School of the University of Minnesota, through the F. B. Hubachek, Sr. Endowment at the University of Minnesota, and through National Science Foundation (USA) grants IBN-9296005 and IBN-9630241.