Diurnal variation of the δ13C of pine needle respired CO2 evolved in darkness


James L. Prater. Fax: +1 850 644 2581; e-mail: jprater@ocean.fsu.edu


The δ13C of pine needle CO2 evolved in darkness (δ13Cr) for slash pine trees (Pinus elliottii) was determined by placing recently collected pine needles in darkness and collecting respired CO2 over a short time period (< 15 min). δ13Cr measurements were made over several 24 h periods to test the hypothesis that significant variation in δ13Cr would be observed during a diurnal cycle. The δ13Cr measurements from the 24 h time series trials showed a consistent midday 13C-enrichment (5–10‰) relative to bulk biomass. The δ13Cr values became more 13C-depleted at night and following shading, and approached bulk-biomass δ13C values by dawn. The effect of night-time respired 13C-enriched CO2 on the δ13C value of the remaining assimilate is shown to be minimal (13C depleted by 0.22‰) under field conditions for P. elliottii needles.


As C3 plants fix carbon through photosynthesis, the δ13C of CO2 is modified by isotopic fractionations caused by diffusion and by enzymatic reactions converting CO2 to biomass. Fractionation occurs during the diffusion of CO2 through stomata because of the faster transfer of the lighter isotope (the fractionation for diffusion through the stomata, a, is about 4.4‰), whereas the isotopic effects of diffusion and dissolution into the liquid phase and transfer to the chloroplast, where photosynthesis takes place, has been considered to be minimal (O’Leary 1981; Farquhar, O’Leary & Berry 1982). As enzymes associated with the C3 photosynthetic pathway act on CO2, fractionation related to these enzymes (b), primarily from Rubisco, further depletes the heavy isotope by approximately 28‰. Fractionation effects associated with respiration have been considered to be less important. Isotopic discrimination during photosynthesis (δ13C) can be expressed as a function of the ratio of internal to atmospheric CO2 partial pressure (ci/ca) with Eqn 1 (Farquhar et al. 1982; Evans et al. 1986; von Caemmerer & Evans 1991; Gillon & Griffiths 1997)

Δ13C = a + (b − a)(ci/ca) − (b − ai)A/(gwca) − (eRd/k + fΓ *)/ca(1)

a simplified version of the complete equation, where ai represents the fractionation during diffusion and dissolution of CO2 into the liquid phase (1.8‰), A represents the assimilation rate, gw is the leaf internal CO2 conductance, e is the fractionation during day respiration, Rd is the day respiration rate, k is the carboxylation efficiency, f is the fractionation during photorespiration and Γ * is the CO2 compensation point in the absence of day respiration. The terms representing the effect of CO2 transfer to the chloroplast and fractionation during respiration, (bai)A/(gwca) and (eRd/k +  *)/ca, respectively, are often considered to be negligible. An abridged version of Eqn 1, used to calculate isotopic fractionation during photosynthesis while neglecting the effect of CO2 transfer to the chloroplast and fractionation during respiration (Δ13Ci, Eqn 2), has proved useful in linking variation in Δ13C with stomatal regulation and ci/ca (Bowling et al. 2002; Fessenden & Ehleringer 2003; McDowell et al. 2004). The δ13C value of assimilated carbon (δ13Ci) can be calculated by subtracting Δ13Ci from the δ13C value of CO2 in ambient air (δ13Cair, Eqn 3).

Δ13Ci = a + (b − a)(ci/ca)(2)
δ13Ci = δ13Cair − Δ13Ci(3)

During foliage respiration, plants convert previously assimilated material back into CO2 to provide energy for cell functions. Foliage respiration occurs both in darkness by dark respiration and while exposed to light through two processes, photorespiration and day respiration (Sharkey 1988; Gillon & Griffiths 1997). Although fractionation during photorespiration is thought to be significant (Sharkey 1988; Gillon & Griffiths 1997), fractionation during dark respiration has been considered to be negligible (Lin & Ehleringer 1997).

Recent investigations have, however, indicated that fractionation-like processes appear to occur during dark respiration. Duranceau et al. (1999) found that the δ13C of dark-respired CO2 (δ13Cr) was 13C-enriched by approximately 6‰ relative to bulk sucrose. Since then, other investigations have shown similar results. δ13Cr has been found to be 13C-enriched relative to a wide variety of metabolites, across multiple species, with some variability across species and environmental conditions (Duranceau, Ghashghaie & Brugnoli 2001; Ghashghaie et al. 2001; Xu et al. 2004). Tcherkez et al. (2003) showed δ13Cr was not constant in darkness over a 5 day period and suggested that variation could be caused by changes in the respiratory substrate coupled with associated changes in the relative contribution from the different metabolic pathways, nonstatistical distribution of 13C in glucose molecules (Rossman, Butzenlechner & Schmidt 1991) or by kinetic effects of respiratory enzymes (reviewed in Ghashghaie et al. 2003).

The effect of the nonstatistical distribution of 13C in glucose molecules (Rossman et al. 1991; Tcherkez et al. 2003) and the associated 13C-enriched δ13Cr (Duranceau et al. 1999, 2001; Ghashghaie et al. 2001, 2003; Tcherkez et al. 2003; Xu et al. 2004) on assimilated carbon should be investigated. Also, recent findings have shown that the isotopic composition of ecosystem respiration varies on short time scales (Knohl et al. 2004) and that ecosystem and leaf respiration draws from at least two distinct carbon sources with different characteristics and residence times (Schnyder et al. 2003; Nogues et al. 2004), leading to the conclusion that temporal variability of leaf respiration or variation in the substrate used for respiration could be responsible for differences in reported fractionation values in the literature. Temporal variability in δ13Cr is not well understood and more work is needed to answer questions regarding the implications of 13C-enriched respiration.

Assimilated carbon δ13C varies with environmental conditions over the course of a day (i.e. changes in ci/ca in response to changes in photosynthetically active radiation or vapour pressure deficit, Eqn 1) leading one to expect variation in the δ13C of day-respired CO2, assuming that recently assimilated C is utilized for autotrophic respiration (Bowling et al. 2002; Mortazavi & Chanton 2002b; Mortazavi et al. 2005). This is regardless of any respiration-related fractionation effects (Ghashghaie et al. 2001, 2003) or shifts in metabolic pathways (Tcherkez et al. 2003). Furthermore, δ13Cr in darkness has been shown to vary over the course of multiple days under constant environmental conditions, presumably because of changes in carbohydrate concentrations and shifts in the relative contributions of the two major decarboxylation processes (i.e. pyruvate dehydrogenase activity and the Krebs cycle) that occur in the dark (Tcherkez et al. 2003). We tested the hypothesis that δ13Cr would vary over the course of a day and continue to vary overnight under field conditions, where significant changes in environmental conditions occur, by measuring δ13Cr for slash pine trees (Pinus elliottii) under field conditions to determine the extent of diurnal variation in δ13Cr. Ghashghaie et al. (2003) suggested that discrimination during night-time respiration and the resulting 13C-enriched δ13Cr would cause 13C depletion of the remaining plant material. We use our measured night-time respiration rates and δ13Cr values, and daytime photosynthetic rates, to estimate the impact of night-time δ13Cr on the δ13C of the remaining plant material.


Study sites and conditions

The August and November (2002) 24 h trials of δ13Cr measurements were conducted at the same height on the same slash pine tree at Tallahassee, Florida. Temperatures varied from 23 to 38 °C for the August trial and from 5 to 19 °C for the November trial. Sunrise and sunset were at 0703 and 2022 h (EDT) for the August trial and at 0657 and 1745 h (EST) for the November trial.

The April (2003) trial, which included δ13Cr, photosynthetic rate and respiration rate measurements, was conducted on a slash pine tree at the same height in the canopy and within 10 m of the tree used for the August and November trials. Temperatures varied from 19 to 35 °C. Sunrise and sunset were at 0706 and 2007 h (EDT) for the April trial.

The shading experiment in November (2003) included δ13Cr measurements and was conducted on a slash pine tree within 2 km of the trees used for the August, November and April trials. Temperatures varied from 19 to 33 °C. Sunrise and sunset were at 0700 and 1742 h (EST).

Pine needle sampling

Leaf-chamber measurements (see below) utilized pine needles still attached to the limb, while δ13Cr measurements used detached needles. For each δ13Cr measurement new pine needles were detached from the tree. All of the pine needles came from approximately the same elevation. In the November (2003) shading experiment, two subgroups were determined by the shading regime, and the pine needles for each subgroup measurement came from the same limb.

Each δ13Cr measurement was made with the sequenced-air-sampling (SAS) system Fig. 1, which is designed specifically for the collection of foliage respired CO2 (see below), within 15 min of needle detachment. The short time interval from detachment to measurement was necessary to minimise variations caused by dehydration or other physiological changes associated with detachment. After the δ13Cr measurement was made, the needles were refrigerated and taken back to the laboratory where they were dried at 60 °C and ground to a fine powder. An isotope-ratio mass spectrometer (IRMS) coupled to a Carbon–Hydrogen–Nitrogen (CHN) analyser was used to determine δ13C values of bulk-organic matter.

Figure 1.

Schematic view of the sequenced air sampler (SAS) used for collecting CO2 samples of foliage respiration. The 0.5 L system consists of a 120 mL serum vial containing the respiring material and eight 30 mL gas-collection vials. Circulation is provided by an air pump with a flow rate of approximately 2 L min−1. The three-way valves allow the flow to be circulated through or to bypass each vial. The system neglects temperature or humidity control and is intended for measurements on short time scales.

Leaf-chamber measurements

Photosynthetic rates, respiration rates and the ci/ca ratio were measured on needles still attached to the tree using a portable photosynthesis system and a 0.25 L leaf chamber (LI-6200, Li-Cor Inc., Lincoln, NE, USA). Surface areas of needles were approximated using a two-dimensional model and were determined by doubling the length multiplied by the width of the needles. Ambient CO2 concentrations were measured with the LI-6200, and net photosynthetic rates (P) and ci were calculated by using the LI-6200 software. Respiration rates (R) were also calculated using the LI-6200, by darkening the leaf chamber during data collection. Each rate and ci/ca measurement reported is the average value of eight consecutive measurements and the errors reported are one standard deviation. No isotopic measurements were made with the Li-Cor leaf-chamber system.

Δ13Cr measurements

Determination of δ13Cr at each time point was made through the analysis of eight consecutive air samples collected over 11 min using detached needles. Approximately 20 needles were contained in a closed system with a background CO2 concentration of ca and a δ13C value of δa. The δ13C of CO2 respired into the system was assumed to be constant (δr). Isotopic mass balance equations are as follows:

ct = cr + ca(4)
δtct = δrcr + δaca(5)

where δt and ct are the δ13C and CO2 concentrations in the closed system at any relative mixture of respired and background CO2. If Eqn 5 is solved for δt and Eqn 4 is solved for cr and substituted into Eqn 5, the result,

δt  = (ca(δa − δr))/ct + δr(6)

represents the line formed when the inverse of concentration is plotted against the corresponding δt values, and the y-intercept will be δr, in this case δ13Cr. This technique is commonly referred to as a Keeling plot (Keeling 1958; Miller & Tans 2003; Pataki et al. 2003), and similar methods have been used previously to determine δ13Cr (Fessenden & Ehleringer 2003; McDowell et al. 2004; Xu et al. 2004).

The gas samples needed for each δ13Cr calculation were collected within the closed system of the SAS, which can be subsampled eight consecutive times without pressurization or depressurization of the system. Flushing the SAS with a consistent background air source eliminates variability associated with ambient background air conditions. Once the system is closed, the valves to each sample vial allow that vial to be bypassed and isolated from the system. Bypassing a sample vial permits the removal of the vial, and causes a decrease in the remaining system volume but no change in pressure. Leak tests have shown that the system can remain closed while circulating air for over an hour with no measurable alteration to the enclosed δ13C-CO2 value (data not shown).

We collected the samples necessary for the determination of δ13Cr using the bottle method, in which pine needles were collected and placed in a 120 mL glass serum vial. The bottle was then capped with a butyl rubber stopper, wrapped in aluminium foil for the elimination of light, and placed on-line with the SAS. The needles were isolated in the foil-wrapped bottle for 2 min while the system was flushed with a constant background air source. After the system was closed, the eight samples were taken at regular time intervals while the CO2 concentration rose as a result of respiration by the pine needles. Because relative humidity and air temperature were not regulated by the SAS during measurements, limiting the time interval between samples to 1.5 min minimised unwanted variations in the closed system's air mass, such as drastic changes in relative humidity, temperature or CO2 concentration.

Within one week of collection, samples were analysed with a Hewlett Packard 5890 Series II gas chromatograph (Agilent Technologies, Inc., Palo Alto, CA, USA) coupled to a Finnigan Delta-S IRMS (GC-IRMS) for the determination of δ13C (‰, relative to Vienna PDB, VPDB) and concentration of the contained CO2 (p.p.m.). Injection of identical amounts of each sample along with known standards during GC-IRMS analysis permitted simultaneous determination of the CO2 concentration and the δ13C value by comparison of the CO2 peak amplitude to concentrations on a standard curve (Mortazavi & Chanton 2002a). Each set of eight samples was plotted on a Keeling plot (Fig. 2), and a Model I linear regression was performed; the y-intercept represented the δ13C of the CO2 that was added to the closed system (δ13Cr). Errors reported for δ13Cr values are standard errors from the y-intercept of the eight-sample regression. The average r2 value for the Keeling plot data reported in this study is 0.982 ± 0.015 (n = 49), which is sufficient to minimise differences between results generated via a Model I or Model II regression (Miller & Tans 2003; Pataki et al. 2003).

Figure 2.

Keeling plot from the SAS standardization and an example of a δ13Cr calculation. (a) Our sampling system, the SAS, was tested by injecting a known standard into background air (δ13C ≈ −8‰) to simulate respiration. The Keeling plot yielded a value of −1.5 ± 0.1‰ for the standard, which had a known value of −1.67‰. (b) A typical Keeling plot δ13Cr calculation, −20.8‰ in this case.

To determine whether the detachment of needles or sampling with the SAS caused deviations in δ13Cr, we compared the bottle method of collecting respired CO2 with two other methods; using a Mylar balloon and using a low-volume 220 mL leaf chamber. The balloon method used by McDowell et al. (2004) was the least intrusive sampling procedure, causing very little physical or environmental disturbance to the needles and permitting withdrawal of larger gas samples without pressurization or depressurization of the closed system (McDowell et al. 2004; Xu et al. 2004). During the balloon procedure a portion of the limb with a group of pine needles was sealed inside a Mylar balloon with a fan inside to provide air circulation. The CO2 concentration was monitored with an LI-6200, and eight samples were collected at 5 min intervals using a 60 cc syringe and stored in 30 mL glass serum vials capped with butyl-rubber stoppers as pressurized samples. The samples were analysed with the same procedures as those for the bottle method, and a Keeling plot was used to determine δ13Cr.

The low-volume 220 mL leaf chamber tests were designed to determine whether needle detachment or subsampling with the SAS caused variation in δ13Cr. Procedures used for the low-volume leaf chamber test were identical to those for the bottle method, with the exception that needles were not detached from the limb and were contained in a low-volume leaf chamber specifically designed for pine needles instead of being detached and placed in a 120 mL bottle.

Twenty four hour time series

The time series measurements were initiated at sunrise; δ13Cr measurements were taken at approximately 1 h intervals near sunrise and sunset, and 2 h intervals otherwise, for a 24 h period. All CO2 samples required for calculation of δ13Cr for the time series were collected by means of the bottle method described above, and each δ13Cr value represents the Keeling plot y-intercept for the series of eight respired-CO2 samples collected at the specified time point. Errors reported for δ13Cr measurements are the standard errors from the y-intercept of the Keeling plot regression. Each δ13Cr measurement was taken with a new set of needles from the same tree. The needles from each time point were saved, and their biomass δ13C was determined when necessary.

The April (2004) time series included leaf-chamber measurements that were made at 1 h intervals during daylight and at 2 h intervals during the night. To determine the total amount of carbon fixed by needles during the April (2004) time series, the 24 h period from 0700 h 19 April to 0700 h 20 April 2003, was divided into consecutive 0.5 h intervals, and each interval was assigned a value for P (daylight hours, 0700–2000 h) or R (night-time, 2000–0700 h). During intervals where measured values of a term needed were not available, values were extrapolated from the measured values taken before and after the time interval by linear interpolation. In the case of P, sunrise and sunset end-members were each assumed to have a rate of zero. The integrated values of P and R (Ptotal and Rnight) were determined by the sum of P or R multiplied by the interval time (0.5 h) for all 0.5 h time intervals within the time period of interest (daylight hours for Ptotal, night-time hours for Rnight).

The impact of night-time needle respiration on assimilated carbon for the April (2003) trial was determined by a needle budget mass balance approach (Eqn 7).

(1 − f)δ13C24 h net = δ13CP total − (f)δ13CR night(7)

δ13C24 h net is defined as the δ13C value of remaining assimilated carbon after the effects of needle night-time respiration have been accounted for. Assimilated carbon δ13C (δ13CP total) for the daylight hours was determined by substituting the rate-weighted average of ci/ca into Eqn 2 and the resulting Δ13Ci and δ13Cair into Eqn 3. Total night-time needle-respired carbon δ13C (δ13CR night) is the rate-weighted average of δ13Cr for all night-time data points. Fraction of assimilation used in needle night-time respiration (f) was determined by dividing Rnight by Ptotal.

Shading experiment

The shading experiment was initiated at sunrise on 12 November 2003. Measurements were taken on two adjacent limbs at the same elevation on the same tree. Both limbs received full sun until noon, when one of the limbs was subjected to partial shading. The partial-shade limb was covered with landscaping shade cloth, which reduced photosynthetically active radiation (PAR) to 40% of the full sun value with minimal effect on temperature (2–3 °C cooler than full sun). δ13Cr measurements were made on 2 h intervals and the results from the two groups were compared.

PAR and vapour pressure deficit (VPD)

Air temperature (T) was measured at the August and November trials, and relative humidity (RH) and PAR were also measured at the November trial. RH and T were measured with an LI-6200, and PAR was measured and integrated over 30 min intervals with an LI-1000 equipped with an LI-190SA quantum sensor (Li-Cor Inc.). VPD was calculated from the T and RH measurements.


Standardization of CO2 collection

We tested the bottle method δ13Cr measurement technique by simulating foliage-respired CO2 with seven 1 mL injections of Ultra High Purity working CO2 standard, calibrated against NIST NBS-19 calcite (Coplen 1996), into the system at 1.5 min intervals. Analysis of the samples was consistent with the procedures for field samples and yielded a δ13Cr of −1.5 ± 0.1‰ compared to −1.67‰, the known value of the standard (Fig. 2a). This result indicates that the SAS sampling procedure does not cause fractionation and yields δ13C values similar to the δ13C value of introduced CO2.

Comparisons with other methods further validated the technique. In two replicate trials on adjacent limbs, the Mylar-balloon, leaf-chamber and bottle methods were all used to determine δ13Cr. The samples were taken in the early morning, just before sunrise. The results from four out of the six measurements were consistent, and leaks are suspected to have caused the variations observed in the limb-1 leaf-chamber and balloon measurements (Table 1). The balloon and leaf-chamber methods were problematic in the field because of difficulties in sealing the system. The longer time between successive samples collected from the balloon method (5 min) aggravated the effect of any leaks. The remaining four measurements were all within ± 0.2‰. The comparison indicates that detaching the pine needles did not cause an unrealistic physiological response, altering the respired-CO2δ13C value within the short duration of the measurement (< 15 min).

Table 1. δ13C values for respired CO2 (‰, VPD). Three methods for collecting respired CO2 were applied to needles from the same tree limb. Values shown are the eight-sample Keeling plot y-intercepts, representing the δ13C of respired CO2. The error shown is the standard error from the y-intercept of the Keeling plot regression
 Leaf chamberBalloonBottle
Limb 1−26.5 ± 0.2−26.5 ± 0.1−27.7 ± 0.2
Limb 2−27.5 ± 0.2−27.5 ± 0.2−27.6 ± 0.1

Twenty four hour time series

Respired CO2δ13C values from needles placed in darkness during δ13Cr measurements varied from −30‰ to −20‰ in the August, November and April trials (Fig. 3). The δ13Cr values for the August and November trials had similar magnitudes of variation and bimodal shapes, with midday maxima at 1500 and secondary maxima at around 1700 h, even though the temperatures were on average 20 °C higher and sunset was 2.5 h later for the August trial. After the second maximum, δ13Cr values for the August trial returned to the August pre-dawn level (−27‰) rapidly (by 2300 h) compared to those of the November trial, where the night-time air temperature was on average 18 °C cooler and δ13Cr values did not return to the November pre-dawn level (−28‰) until 0300 h. The April trial exhibited a similar pattern in diurnal 13C enrichment but the magnitude of variation exceeds that of the previous trials. The night-time return of δ13Cr to the pre-dawn level (−29‰) was rapid, similar to the return to baseline in the August trial. Night temperatures were similar in April and August, and cooler in November.

Figure 3.

Twenty four hour time series of respired CO2 from needles placed in darkness (δ13Cr) and biomass δ13C. Closed symbols represent δ13Cr and open symbols represent bulk-biomass δ13C: August (2002) trial by triangles, November (2002) trial by squares, April (2003) trial by circles. Each δ13Cr value shown represents the Keeling intercept from the regression of a set of eight samples, and errors shown for δ13Cr values are standard errors from the Keeling plot regression. Biomass δ13C represents pine needles collected for the respiration experiments. Biomass δ13C errors are half of the range between two replicates. Sunrise was at approximately 0700 h for all trials, and sunset was at 2022 h in August, 1745 h in November, and 2007 h in April. Values of δ13CP total (open circle), δ13CR total (closed circle) and δ13C 24 h net (grey circle) for the April (2003) are shown (I). The error bars associated with δ13CP total represent the maximum range of variability for 0.05 < f < 0.20 assuming δ13CP total of −29‰ (see Table 3).

The pine needles used for δ13Cr measurements in the August and November trials were analysed for bulk biomass δ13C (Fig. 3). No significant variation was seen over the diurnal cycle, and the average values for each trial were within 0.5‰.

The November trial VPD ranged from 0.2 to 1.4 kPa over the diurnal cycle, and its midday maximum corresponds well with the first peak of δ13Cr(Fig. 4). After sunset, VPD quickly declined to its early morning minimum of 0.2 kPa. PAR had a bimodal shape over the diurnal cycle, caused by shading from other nearby trees. The onset and cessation of sunlight marked the initial enrichment trend and the decline from the first peak in δ13Cr (Fig. 4).

Figure 4.

Respired CO2δ13C (shaded squares), vapour pressure deficit (VPD, dashed line), and photosynthetically active radiation (PAR, dotted line) plotted against time for the November (2002) 24 h series. PAR and VPD lines are smoothed data from measurements taken at 30-min intervals during daylight hours and 2-h intervals otherwise.

Photosynthetic and respiration rate data from the April (2003) trial are shown in Table 2. The integrated value for night-time needle respiration (Rnight = −0.038 ± 0.002 mol C m−2, Table 2) is approximately one-tenth (f = 0.101) of the integrated value for daytime needle assimilation (Ptotal = 0.375 ± 0.019 mol C m−2, Table 2). Other studies have shown similar ratios of carbon uptake and respiration in pine needles (Cropper & Gholz 1991; Will 2000). The assimilation rate-weighted average of ci/ca (0.73, Table 2) corresponds to a δ13CP total of −29.7‰ (δ13Cair =−8.2‰, Eqns 2 and 3). This estimate of assimilation is 13C-depleted relative to the rate-weighted average of night-time needle δ13Cr (δ13CR total, −27.8‰, Table 2).

Table 2.  Rate, ci/ca, and δ13Cr data from the April (2003) trial. P, R and ci/ca were measured with a portable photosynthesis system equipped with a 0.25 L leaf chamber (LI-6200) on needles still attached to the limb. δ13Cr values were measured on detached needles using the sequenced air sampler (SAS). The integrated total of respiration (Rnight = −0.038 ± 0.002 mol C m−2) is approximately one-tenth of assimilation from the daylight hours (Ptotal = 0.375 ± 0.019 mol C m−2). The rate-weighted average of ci/ca is 0.73. The rate-weighted average of δ13Cr is −27.8‰. Each rate and ci/ca measurement reported is the average value of eight consecutive measurements and the errors reported are one standard deviation. Errors reported for δ13Cr values are standard errors from the Keeling plot regressions
(P; µmol CO2 m−2 s−1)
(R; µmol CO2 m−2 s−1)
δ13Cr (‰)
1006 5.5 ± 0.20.76 ± 0.012123−0.8 ± 0.3−22.0 ± 0.4
120211.9 ± 0.80.63 ± 0.042304−0.9 ± 0.05−26.4 ± 1.0
141515.3 ± 3.40.76 ± 0.090104−1.1 ± 0.1−28.4 ± 0.3
1512 8.5 ± 0.70.73 ± 0.030301−1.2 ± 0.2−29.0 ± 0.6
160311.0 ± 0.80.73 ± 0.010504−0.9 ± 0.04−31.0 ± 0.6
171011.4 ± 0.70.72 ± 0.040700−0.7 ± 0.1−29.3 ± 0.6
1903 5.3 ± 2.40.87 ± 0.03   

The δ13C of remaining needle-assimilated carbon (δ13C24 h net) estimated with Eqn 7 (−29.9‰) is only 13C-depleted by 0.22‰ compared to the value of daytime-assimilated carbon (δ13CP total) of −29.7‰. This estimate assumes that 10.1% of assimilated carbon is used for night-time needle respiration and uses the estimate for δ13CP total that is dependent on ci/ca measurements. Sensitivity analysis of this calculation is shown in Table 3. Over the 24 h period when δ13Cr was determined, the measurement obtained just prior to dawn (−29.3 ± 0.6‰) is most representative of the 13C of net assimilated carbon (δ13C24 h net) (Fig. 3).

Table 3.  Change in δ13CP total caused by night-time respiration (δ13C24 h netδ13CP total), where f is the fraction of assimilation used for respiration at night. δ13CP total is the integrated value of assimilation during daylight (dawn to dusk). Appropriate values for the April (2003) trial based on leaf-chamber measurements are f = 0.101 and δ13CP total of −29.73‰ (ci/ca = 0.73, Table 2, Eqns 2 and 3)
δ13CP totalChange in δ13CP total from night-time respiration (‰)

Shading experiment

During the November 2003 shading experiment, δ13Cr varied from −34‰ to −15‰ and both subgroups exhibited similar 13C-enrichment until the initiation of shading at noon (Fig. 5). The limb subjected to a partial shading regime exhibited a decline in δ13Cr following shading, while the full-sun control limb exhibited a slower decline in δ13Cr, similar to the pattern for the previous 24 h δ13Cr data (Fig. 3). However, while initial 13C-enrichment and night-time 13C-depletion were similar to previous 24 h trials, sampling resolution was not sufficient to show whether the bimodal pattern observed in the daytime data from the previous 24 h trials was present.

Figure 5.

Results from the shading experiment, November (2003). Needles from two limbs of the same tree were subjected to full sun until noon, then one limb was left in full sun (open circles) and one was subjected to partial shading (shaded circles).


The δ13C of foliage-respired CO2 has been shown to be 13C-enriched relative to the substrates used for respiration and to bulk biomass (Duranceau et al. 1999, 2001; Ghashghaie et al. 2001, 2003; Tcherkez et al. 2003; Xu et al. 2004). Tcherkez et al. (2003) demonstrated that the 13C-enrichment of δ13Cr decreased during continuous darkness over several days, presumably as a result of the decrease in the carbohydrate pool size and changes in the relative contributions of the major decarboxylation processes of dark respiration. Over a 24 h period in the dark, glucose concentrations can decline by 50–100% at temperatures ranging from 20 to 30 °C (Tcherkez et al. 2003). Ghashghaie et al. (2003) suggested that discrimination during night-time respiration and the resulting 13C-enriched δ13Cr would cause 13C-depletion of the remaining plant material. We examined the hypothesis that significant variability of δ13Cr would be observed over a diurnal cycle and used our δ13Cr, P and R data to estimate the effect 13C-enriched night-time respiration could have on remaining assimilated carbon (δ13C24 h net).

Diurnal variation in δ13Cr

Similar diurnal patterns in δ13C for needles placed in darkness are evident in the August (2002), November (2002) and April (2003) 24 h series (Fig. 3). Despite temperature differences of approximately 20 °C between the August and November (2002) trials, the δ13Cr values are remarkably similar, with the exception that 13C-enriched CO2 was released for a longer period under the cooler conditions of the November (2002) time series (Fig. 3). The rate of change of night-time δ13Cr values is apparently related to respiration rates, demonstrated by the comparison of the August (2002) and November (2002) trials, where the night-time decline in δ13Cr values occurred at a slower pace for the November (2002) trial when air temperature was on average 18 °C cooler and associated respiration rates also would have been lower (Villar, Held & Merino 1995; Tcherkez et al. 2003). The rapid night-time decline in δ13Cr is similar to the work of Troughton, Card & Hendy 1974) who showed respiration that was 13C-depleted relative to biomass for Pinus radiata, possibly indicating that the respiratory metabolism of conifers produces CO2 that is 13C-depleted relative to night-time respiration of other C3 plants. However, unlike the results of Troughton et al. 1974), our results show night-time respiration approaching biomass δ13C values and include only one data point where respiration is 13C-depleted relative to biomass (Fig. 3).

The midday δ13Cr enrichment relative to plant organic matter 13C during the 24 h series (Fig. 3) is similar in magnitude to the maximum enrichment of ∼ 9‰ reported by Ghashghaie et al. (2001) for Nicotina sylvestris plants. Ghashghaie et al. (2001) found that under well-watered conditions, δ13Cr was enriched relative to even the most 13C-enriched substrates used for respiration. This enrichment could result from the non-statistical 13C distribution with the glucose molecules (Rossmann et al. 1991). Rossmann et al. (1991) demonstrated that C3 and C4 of glucose molecules extracted from sugar beets were enriched in 13C relative to the other carbon positions. During respiration, if a high percentage of catabolized carbon were used for lipid biosynthesis, for example, the 13C of respired CO2 would be highly enriched (Ghashghaie et al. 2003). Alternatively, if all of the catabolized carbon is used for respiration, all light and heavy carbon atoms will be carboxlyated. Ghashghaie et al. (2003) suggested that the degree of 13C-enrichment relative to the substrate used during respiration would depend on environmental conditions and the relative activities of different metabolic pathways.

The shading experiment results are consistent with the 24 h series data in that δ13Cr values become less 13C-enriched with time when assimilation rates are depressed (Fig. 5). The shaded subgroup's decline in δ13Cr after the midday maximum could result from respiration of 13C-depleted material produced in the low light (low assimilation rate, high ci/ca) conditions (Knohl et al. 2004; Scartazza et al. 2004), or it could be that respiration had consumed the available assimilate and had switched to a carbon pool with a residence time of a few days, as modelled by Schnyder et al. (2003); see also Nogues et al. (2004) with a more depleted 13C values.

Alternatively, the decline in δ13Cr could result from changes in the metabolic pathway of dark respiration (Tcherkez et al. 2003) over the diurnal cycle. The decline in δ13Cr values following sunset (Fig. 3) or shade (Fig. 5) to more 13C-depleted values prior to the following sunrise is consistent with, but more rapid than, the decline in the δ13C of foliage-respired CO2 in continuous darkness over a five day period (Tcherkez et al. 2003). Tcherkez et al. (2003) observed that δ13Cr values were 13C-depleted by as much as 9‰ compared to initial values following the decline in carbohydrate concentrations during prolonged exposure to darkness. They concluded that the decline in 13C could be accounted for by a shift in metabolism away from pyruvate dehydrogenation and towards fatty acid β-oxidation coupled to the Krebs cycle. Following depletion of the carbohydrate pool during prolonged exposure to darkness, the 13C of respired CO2 may become depleted even relative to the 13C of bulk organic matter (Tcherkez et al. 2003). However, during our diurnal experiments (Figs 3 & 5), the 13C of respired CO2 declined only to approach values similar to that of needle organic matter 13C.

The 13C of needle organic matter represents the long-term integrated 13C values of net carbon fixed over the needle's life. Biomass carbon δ13C does not appear to represent respired carbon δ13C on short time scales (Fig. 3). Any changes in reservoirs that are influenced on short time scales are buffered by structural material and reservoirs of carbon with longer residence times. This conclusion supports the results of Brendel (2001), who showed that the response of bulk-foliage δ13C values to changing Δ13Ci was buffered by structural material, and the results of Knohl et al. (2004) and Scartazza et al. (2004) who showed that bulk organic carbon was unresponsive to short-term variations in ecosystem respiration δ13C.

Effect of 13C-enriched respiration

Our results suggest that δ13Cr is enriched relative to assimilated carbon δ13C estimated from Eqns 2 and 3 and measured ci/ca values (Table 2). Schnyder et al. (2003) observed that in mesocosm-scale experiments, integrated night-time respiration was 0.53‰ enriched relative to recent assimilate. The effect of 13C-enriched respiration on total assimilated carbon estimated with Eqn 7 suggests that δ13C24 h net would only be 13C-depleted by 0.22‰ compared to δ13CP total. The impact of night-time respiration on assimilated carbon appears to be minimal. The limitations of our estimate are that it assumes a value for assimilated carbon at dusk derived from ci/ca measurements (δ13CP total = −29.7‰, Table 2) and that 10.1% of assimilated carbon is used for night-time respiration (Table 2). We further assume that night-time respiration is fuelled by recent assimilate. Sensitivity analysis of this calculation (Table 3) shows that if f is lower than the projected 10% (i.e. that some respiration is fuelled by a longer-lived component) very little change in assimilated carbon δ13C will occur. Alternatively, if f was double our estimated value and assimilated carbon was significantly more 13C-depleted than our ci/ca measurements indicated, respiration could alter assimilated carbon δ13C by more than 1‰ (f = 0.2, δ13CP total = −32‰, Table 3).

Although night-time δ13Cr may be 13C enriched relative to assimilated carbon (δ13CP total = −29.73‰) by as much as 7.7‰ (Table 2), it is unclear how fractionation during day respiration affects Δ13C. Day respiration rates may be inhibited during photosynthesis (Villar et al. 1995), and our measurements only determine δ13Cr for dark respiration in the absence of light. Therefore further efforts are required to determine the effect of daytime respiration and associated fractionation effects on Δ13C.


Our primary objective was to examine the magnitude of variability for δ13C values of needle-respired CO2 in darkness over a diurnal cycle. Our results demonstrate that δ13Cr exhibits diurnal variability and is 13C-enriched compared to biomass during the day and early part of the night but approaches values similar to δ13CP total before sunrise. The 13C-enrichment apparent in δ13Cr shows a strong time dependency, with maximum enrichment immediately following assimilation and decreasing with time in darkness. Although night-time respiration has been shown to be 13C-enriched relative to assimilated carbon and bulk biomass, the relatively small quantity of 13C-enriched night-time respiration appears to have little effect on the 13C value of the remaining assimilated carbon. The dependency of δ13Cr on time and environmental conditions, such as temperature, may be responsible for some of the interspecies variation reported in the literature (reviewed in Ghashghaie et al. 2003). If species are to be compared, standard procedures need to be developed to address the dynamic temporal variability that exists in δ13Cr.

The enrichment in needle-respired CO2 and its gradual decay with time in darkness could potentially impact measurements of the 13C of ecosystem-respired CO2. Night-time samples of mid-canopy-respired CO2 samples consist of inputs from both autotrophic and heterotrophic respiration. While the signature of the heterotrophic component is expected to remain invariant on short time scales ( Trumobre 2000), the isotopic composition of autotrophic respiration contributed by the foliage apparently changes on a time scale of hours (Figs 3 & 5). If time series of night-time mid-canopy air samples are used to construct Keeling plots during periods when a large fraction of total ecosystem respired CO2 is contributed by foliage respiration, the intercepts of mid-canopy time series Keeling plots could potentially be biased towards more enriched values.


This research was supported by the National Science Foundation (NSF ♯ 0343604) and the US Department of Energy's Biological and Environmental Research (BER) program through the South-east Regional Center (SERC) of the National Institute for Global Environmental Change (NIGEC, Cooperative agreement No. DE-FC03–90ER61010). Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of NSF or DOE. We would like to express our appreciation to Dr David R. Bowling for advice on the balloon method, to Dr Anne B. Thistle for editorial review and to Dave Oliff for guidance on construction of the SAS. We would also like to thank Kelly Peeler, Travis Comer and Michael Harrison for technical assistance during fieldwork, and to editor Dr Dan Yakir and the three anonymous referees, whose constructive criticisms led to significant improvements in this manuscript.