Abstract

The study of the fate of assimilated carbon in respiratory fluxes in the field is needed to resolve the residence and transfer times of carbon in the atmosphere–plant–soil system in forest ecosystems, but it requires high frequency measurements of the isotopic composition of evolved CO2. We developed a closed transparent chamber to label the whole crown of a tree and a labelling system capable of delivering a 3-h pulse of 99% 13CO2 in the field. The isotopic compositions of trunk and soil CO2 effluxes were recorded continuously on two labelled and one control trees by a tuneable diode laser absorption spectrometer during a 2-month chase period following the late summer labelling. The lag times for trunk CO2 effluxes are consistent with a phloem sap velocity of about 1 m h−1. The isotopic composition (δ13C) of CO2 efflux from the trunk was maximal 2–3 days after labelling and declined thereafter following two exponential decays with a half-life of 2–8 days for the first and a half-life of 15–16 days for the second. The isotopic composition of the soil CO2 efflux was maximal 3–4 days after labelling and the decline was also well fitted with a sum of two exponential functions with a half-life of 3–5 days for the first exponential and a half-life of 16–18 days for the second. The amount of label recovered in CO2 efflux was around 10–15% of the assimilated 13CO2 for soil and 5–13% for trunks. As labelling occurred late in the growing season, substantial allocation to storage is expected.

Introduction

The anthropogenic perturbation of the earth carbon cycle and the associated climate changes are partly mitigated by carbon sequestration in the terrestrial biosphere, that is itself affected by the atmospheric CO2 and the climate (Cox et al. 2000). This is a major issue for international policy on climate change and carbon budgets within the Intergovernmental panel on climate change (Houghton 2001). Among the terrestrial ecosystems, forests largely contribute to the biospheric carbon exchange and carbon stock (Saugier et al. 2001). Ecosystem respiration almost balances photosynthetic carbon assimilation in forest ecosystems, the net ecosystem exchange being one order of magnitude smaller than the two gross fluxes (Granier et al. 2000, Valentini et al. 2000). Small variations of one of these two fluxes due to climate changes could therefore strongly affect the carbon budget of the ecosystem and its ability to sequester carbon, possibly leading to positive or negative feedbacks (Valentini et al. 2000, Ciais et al. 2005).

Although ecosystem respiration includes contributions from the aerial part of trees (leaves, branches and stems), two-thirds of its sources are located on the ground and below the ground (Janssens et al. 2001). These sources include roots and associated symbiotic organisms such as mycorrhizal hyphae, rhizospheric microbes that feed on root exudates and saprotrophic bacteria and fungi that feed upon above-ground and below-ground litter such as coarse woody debris (Nadelhoffer and Raich 1992, Raich and Schlesinger 1992, Epron et al. 1999, 2001, Ngao et al. 2005). In particular, the potential importance of the carbon flux to the rhizosphere through roots and mycorrhiza has been recognized for forest ecosystems but is poorly documented by in situ measurements (Högberg and Read 2006). A precise partitioning of these fluxes would require a better knowledge of the carbon residence time in each of the ecosystem compartments, as well as a better understanding of carbon allocation among these compartments. As Trumbore (2006) recently pointed out, carbon allocation patterns between growth, storage and respiration of different tree compartments and the soil are not yet well understood.

Stable isotopes, especially 13C, are widely used in ecology as tracers in trophic webs (Dawson et al. 2002, Cerling et al. 2007). The fate of carbon in the soil plant system can be followed by pulse-labelling plants in the field with 13CO2 for a short period of time (< 1 day). The 13C assimilated by plants during the pulse labelling is then tracked in the respiratory fluxes during the following days and weeks (chase period). Up to now, this approach has been mostly used under controlled conditions, but a few in situ studies concerned herbaceous species (Ostle et al. 2000, Johnson et al. 2002, Rangel-Castro et al. 2004, Leake et al. 2006) or small tree saplings, either transplanted (Phillips and Fahey 2005) or grown in pots (Lacointe et al. 1993). Recently, Carbone et al. (2007) and Högberg et al. (2007) pulse labelled small boreal conifers growing in the field (with 14CO2 and 13CO2, respectively), to resolve the relative roles of new photosynthetic products as sources of below-ground and above-ground respiration.

The accurate determination of residence and transfer times of carbon in the atmosphere–plant–soil system requires frequent measurements of the isotopic composition of evolved CO2 during the chase period following the short-term labelling with 13CO2. Up to now, both cost and time required for analysing air samples by mass spectrometry in the laboratory limit frequency and duration of isotopic measurements in experiments studying either variations of natural isotopic abundance during seasons or changes in isotopic enrichment after a pulse labelling. The recent development of tuneable diode laser absorption spectrometers (TDLAS) allows in situ simultaneous measurements of effluxes of 13CO2 and 12CO2 at a high frequency. TDLAS have recently been used to examine ecosystem functioning (Bowling et al. 2003, Griffis et al. 2004, Barbour et al. 2007, Bahn et al. 2009, Marron et al. 2009), and this is a promising tool for tracking 13C in respiratory fluxes after pulse labelling (Bahn et al. 2009).

The aim of this study was to develop and test a closed transparent chamber that was designed to label the whole crown of a tree and a labelling system capable of delivering a 3-h pulse of 13CO2 in the field. The isotopic compositions of trunk and soil CO2 effluxes were recorded continuously by TDLAS using a trace gas analyser during a 2-month chase period. The high temporal resolution enables accurate assessments of the time lag between photosynthetic assimilation by the crown and CO2 release from the trunk and the soil, as well as the precise determination of the half-life of labelled carbon in a given respiratory efflux.

Materials and methods

Study site

The study was conducted in the state forest of Hesse (Hesse 2, 48°40′40″ N and 7°04′05″ E, 300 m elevation, Northeast of France) in September 2008. The study site is located in a mixed stand of 20-year-old beeches (Fagus sylvatica L., 63% of basal area, G), hornbeams (Carpinus betulus L., 26% of G), oaks [Quercus petreae (Matt.) Lielb. and Quercus robur L., averaging 8% of G], willows (Salix caprea L., 2% of G) and birches (Betula pendula Roth, 1% of G). This stand originated from natural regeneration (Ngao et al. 2007). The soil is classified as luvisol and is covered by mull humus. The average stand height is 10 m.

Incident photosynthetic photon flux density (PPFD) was measured above the stand at 16 m height (top of eddy covariance flux tower) with a photosynthetically active radiation (PAR) sensor (Delta T-BF2, Cambridge, UK). Air temperature and relative humidity were also monitored (Vaisala HMP45, Helsinki, Finland). Soil temperature at 10 cm depth was measured with four home-made copper-constantan thermocouples. Data acquisition was made with a datalogger (CR7, Campbell UK) at 10-s intervals, and 30 min averages were calculated and stored. The datalogger failed to store data between 27 October and 4 November. Volumetric soil water content at 15 cm depth was measured with a TDR probe (Trase, SoilMoisture Equipment Corp., Goleta, CA) at five locations in the vicinity of the eddy covariance flux tower.

In March 2008, three codominant trees were selected in a pure beech patch (8–9 m tall trees). Two of these trees were selected for labelling (LT1 and LT2) and one tree was selected as a control (CT). Trunk diameters at 1.3 m above soil level were 21, 28 and 20 cm, respectively. Suppressed trees that were in the vicinity of the selected trees were removed and 60 cm deep and 1.5 m long trenches were dug around each tree. Because of the presence of a compact clay horizon at 50 cm depth, almost no roots were found < 50 cm (Zapater and Granier, unpublished results). The trenches were lined with a polyethylene film and refilled with the displaced earth. The root system of each tree was thus confined in a soil volume below a square surface area of 3.56, 2.06 and 3.45 m2 for LT1, LT2 and CT, respectively. All roots and root exudates within this area therefore originated from the isolated tree, and all roots and root exudates of this tree were contained in this volume. This trenching process ensured that there was no dilution of root respiration of the labelled trees by unlabelled ones.

One month after labelling (mid-October), the crowns were enclosed in a net to collect all the leaf litter. The leaf litter in the net was collected in mid-November and weighed just after collection and after oven drying at 60 °C for 48 h. A subsample (4–6% of fresh weight) was taken before drying and leaf area (LA) was determined, allowing us to estimate the total LA of the crown. The total LA of the crown was converted into green leaf mass by dividing LA by SLA, the specific leaf area of leaves sampled just after labelling (see below).

Whole crown labelling chamber

Two 10-m height stainless steel scaffoldings were built in parallel to each other around the tree to install the chamber at the top of the tree (Figure 1). Ropeholes, isolated from the inside chamber by a seal, were placed every 15 cm at the top of the chamber to accommodate a nylon rope that secured the top of the labelling chamber on four stainless steel bars. The bars were firmly attached to the top of the two scaffoldings. The base of the chamber was also secured on three stainless steel bars that were attached to the scaffoldings.

Figure 1.

Diagram of the whole crown labelling chamber. Air for 12CO2 and 13CO2 monitoring was continuously sampled at the inlet of the cooling device. 13CO2 was injected at the outlet of the cooling device. This figure appears in colour in the online version of Tree Physiology.

Figure 1.

Diagram of the whole crown labelling chamber. Air for 12CO2 and 13CO2 monitoring was continuously sampled at the inlet of the cooling device. 13CO2 was injected at the outlet of the cooling device. This figure appears in colour in the online version of Tree Physiology.

The top of the tree crown labelling chamber was a 6.25 m2 square made of 200-μm-polyane film (celloflex 4TT; Prosyn polyane, Saint Chamond France), and the base of the chamber (1 m2) was made with two stainless steel half-plates (2 mm thick) with circular openings to accommodate the trunk (92 mm diameter) and the two cooling device tubes (251 mm diameter). The two half-plates were joined using aluminium tape (Pure aluminium adhesive tape, Plasto SAS, Dijon, France) and sealed to the trunk with putty (terostat VII, Henken, Düsseldorf, Germany). The four sides of the chamber were 4.8 m height for LT1 and 7.1 m height for LT2. The first 1.3 m for LT1 and 4.8 m for LT2 were vertical, whereas the following parts of the sides were oblique to reduce the width from 2.5 m at the top to 1 m at the base (see Figure 1). According to the side dimension of each tree crown, the maximum volume of the labelling chamber was 19.5 m3 for LT1 and 37.5 m3 for LT2. The chamber could not be fully opened because of the surrounding trees, which means that the true chamber volumes were lower than the maximum values.

The top and the four sides of the chamber were sealed and three of the four sides were also sealed together before installation on the tree. The last side was stuck to the other after installing the chamber around the tree using 100 mm width transparent tape (Nitto PVC 21, Nitto Europe NV, Genk, Belgium). The four sides were stuck to the stainless steel base similarly. Air-tightness of the chamber was evaluated by comparing the amount of 13CO2 delivered and the amount of 13C recovered in the leaves after the labelling (see Results section).

Air temperature and air humidity inside the tree crown labelling chamber were recorded with one probe (HMP50, Vaisala, Finland) attached to the tree crown. Two PAR sensors (JYP 1000, SDEC France, Reignac sur Indre, France) were installed at the top of the tree. Data acquisition was made with a datalogger (CR1000, Campbell UK) at 30-s intervals. We failed to record the data during the first hour of the first labelling experiment, but some data were recorded manually. The chamber air was controlled and maintained at the temperature of the outside air by circulating air at a rate of 1200 m3 h–1 through an air- and water-cooled condensing axial fan outdoor unit (AXCZ 221, HITECSA, Vilanova i la Geltrú, Spain) connected to an air conditioner (BSH-BZ 221, HITECSA, Vilanova i la Geltrú, Spain). The axial fan was modified (Frige Clim Zalesky, Nancy, France) to connect tubes directly to the outlet and inlet of the fan, avoiding exchange with the outside air. Air was circulating from the bottom of the chamber to the cooling device and from the cooling device to the top of the chamber through insulated aluminium tubes (250 mm diameter), which ensured good mixing of the air within the chamber. The volume of tubes and cooling device was 0.5 m3. Including tubes and cooling device, the total volume were estimated at 11.4 and 20.1 m3 for LT1 and LT2, respectively. We did not correct the chamber volume for the trunk and branches as they accounted for < 0.2% of the total chamber volume (see Damesin et al. 2002, for a set of allometric equations for beech trees in Hesse). Chamber air was recirculated through the cooling unit at least every minute, which was enough to ensure good temperature control and good mixing of the air.

Pulse labelling

LT1 and LT2 were labelled on 9 September (day of year 253) and on 16 September (day of year 260), respectively. The labelling chamber was closed at 10:00 UT for the first tree and at 9:30 UT for the second one. Evolution of 12CO2 and 13CO2 concentrations ([12CO2] and [13CO2], respectively) inside the chamber was monitored simultaneously with a 12CO2/13CO2 infrared gas analyser (S710, SICK/MAIHAC, Germany). The CO2 concentration declined after the chamber was closed because of photosynthesis (see Results section).

When [12CO2] reached values < 120 μmol mol−1 for the first tree (10:50 UT) and 190 μmol mol−1 for the second tree (10:30 UT), 25 l of pure 13CO2 (99.299 atom %, Eurisotop, Cambridge Isotope Laboratory Inc., Andover, MA) was injected at a flow rate setting between 0.11 and 0.18 l min−1 using a mass flow controller (Bronkhorst, Netherlands) in the air stream coming from the cooling device. We considered that the beginning of the labelling period was also the beginning of the chase period (time zero on graphs). Average [13CO2] inside the chamber, measured at the outlet of the chamber, was around 425 μmol mol−1 ([12CO2] around 90 μmol mol−1) for LT1 and around 520 μmol mol−1 ([12CO2] around 165 μmol mol−1) for LT2. Abundances were 83 atom%13C (δ = 420,000‰) and 76 atom%13C (δ = 280,000‰), respectively, as compared to ambient air that is 1.1 atom%13C (δ = −8‰).

After 3- (first labelling) and 2.5-h (second labelling), the gas cylinder was emptied. The [13CO2] of the chamber then declined progressively to 95 and 160 μmol mol−1 for LT1 and LT2 before we opened and removed the tree crown labelling chamber at 14:35 UT and 15:05 UT for LT1 and LT2, respectively. The labelling period lasted 03:45 h for LT1 and 04:35 h for LT2.

Assuming that the leak of CO2 from the chamber to the outside air was negligible, the rate of change of [13CO2] (d[13CO2]/dt) depends on the net rate of 13CO2 uptake by the crown (U) and on the amount of 13CO2 injected in the chamber (I): 

(1)
formula

The total amount of 13CO2 taken up by the crown was calculated by integrating U over the labelling period.

Soil and trunk CO2 effluxes and 13C composition

Soil and trunk CO2 effluxes (FS and FT) and their isotopic composition (δ13CFS and δ13CFT) were measured by TDLAS with a trace gas analyser (TGA 100A, Campbell Scientific, Inc., Logan, UT) coupled to flow-through chambers (see Marron et al. 2009 for details). The diode laser produces linear wavelength scans centred on the selected absorption lines (2291.680 cm−1 for 13CO2 and 2291.542 cm−1 for 12CO2).

[12CO2] and [13CO2] of available working standards (Air Product, SAS, Paris, France, 0.5% certified for CO2 concentrations) were 397.22/4.28, 496.94/5.36, 695.80/7.50 and 1189.17/12.83 μmol mol−1, respectively. They were computed from their isotopic composition measured by an isotope ratio mass spectrometer (IRMS, Delta S, ThermoFinnigan, Bremen, Germany).

A first manifold was used to switch between each of three working standards and the chamber inlet and outlet lines. A mean concentration was measured over 20 s for each working standard and for both the reference and sample air streams. A 30-s purge was used in each case. Five minutes were necessary for the complete measurement of the sequence (three working standards followed by reference – sample – reference) on one chamber. A second manifold was used to switch between the seven chamber inlets and outlets, allowing each chamber to be measured every 35 min.

The soil chambers were made of stainless steel and allowed the enclosure of 314 cm2 of soil (Marron et al. 2009). The chambers were composed of a 12.5-cm high collar, 2.5 cm of which was pushed into the soil, covered with mobile lids allowing the sequential measurements of several collars without removing them from soil. Collars were located 60 cm from the trunks. The trunk chambers were composed of two polymethyl methacrylate half-boxes with semicircular openings to accommodate the trunk (Damesin et al. 2002). The chambers were placed around trunk portions of 20 cm in length and were covered with an insulated aluminium sheet to avoid light and increase in temperature. The two half-boxes were screwed into place and sealed with putty. The inlet of the chamber was made of a 15 cm long PVC tube (internal diameter 41 mm). The surface area of the enclosed trunk segments was computed from the diameter of each extremity, assuming that the trunks were cone shaped. Two chambers were set up at the base of the crown (2.5 m) and at the base of the trunk (0.25 m) of LT1, one at 1.6 m height on the trunk of LT2 and one at 1.3 m height on the trunk of CT. The chamber at the base of the LT1 trunk was removed after 1 week, whereas the others remained for 2 months.

Air was drawn continuously through both the inlet and the outlet of all chambers using a pump (Membrane pump, GAST DAA-V505-GD). The air flow rate was measured and controlled using mass flowmeters (MC-2SLPM-D, Alicat Scientific) and needle valves, and ranged between 0.25 and 1.25 dm3 min−1 depending on the rate of respiration, to obtain a concentration difference between inlet and outlet > 50 μmol mol−1 for 12CO2 and 0.5 μmol mol−1 for 13CO2, and to remain within the range of [12CO2] and [13CO2] of our calibration tanks.

Total CO2 concentration ([CO2], μmol mol−1) was calculated from the concentrations of individual isotopologues by 

(2)
formula
where fother is the fraction of CO2 containing all isotopologues other than 12C16O16O and 13C16O16O, and is assumed to be 0.00474 (Griffis et al. 2004).

Soil or trunk CO2 effluxes (FS and FT, μmol m−2 s−1) were calculated using either each isotopologue concentration or total CO2 concentration 

(3)
formula
where P is the atmospheric pressure (Pa), F is the flow (m3 s−1), SC is the surface of soil or of trunk inside the chamber (m2), T is the temperature (°K) and 8.314 J mol−1 K−1 is the ideal gas constant.

The isotopic composition of soil or trunk CO2 effluxes (δ13CFS and δ13CFT, ‰) was calculated as 

(4)
formula
where RPDB is the isotopic ratio of Pee Dee Belemnite (0.0111798825).

Label recovered in CO2 effluxes

The cumulative label recovered in a given efflux (CLRFT and CLRFS), i.e., cumulative label respired by trunk or by root and rhizosphere (root-associated microbes and microbes feeding on root exudates), was calculated by summing the daily average 13CO2 effluxes, corrected for the background isotopic composition of effluxes measured on the control tree, and multiplied by the surface (S) of the trunk or the surface of soil beneath which the roots of the trees are 

(5)
formula
where CLR(d) is the cumulative label recovered at day d after the labelling, FD is the daily average CO2 efflux (in gC m−2 d−1) and AL and AC are the relative abundance of 13C in the efflux in LT and CT trees. The surface of the soil was delimited by the trenches and the surface of the trunk was estimated from the tree diameter according to allometric equations for beech trees in Hesse (Damesin et al. 2002).

The relative abundance of 13C in any efflux or compartment (A) can be calculated as  

(6)
formula

The CLR in an efflux can be fitted with a sum of two exponential functions. We chose a sum of two exponential functions rather than a single one because of the shape of the CLR kinetics. This implicitly suggests that the two components were contributing to the efflux. In addition, we added a lag between the pulse labelling and the beginning of its recovery in a given component. We therefore fitted a set of three equations 

(7)
formula
where t is time after labelling (in days), Ki are decay constants, Ci are asymptotes, i.e., the total amount of labelled carbon that could be recovered in a given component i, and Li are lags between labelling and the beginning of the recovery of the label in this component. Equations were fitted using non-linear least-squares regression (PROC NLIN of SAS software, Marquardt–Levenberg method, SAS Institute Inc., Cary, NC). Half-life and mean residence time were ln(2)/K and 1/K, respectively and were expressed in days. We compared the performance of models using a single exponential function or a sum of two exponential functions using an F test (Brown and Rothery 1994, Epron et al. 2004).

Isotopic composition of bulk leaf organic matter

Two twigs, each carrying about 10 leaves, were collected before and just after the pulse labelling. The surface area of the leaves was determined before they were frozen in liquid nitrogen and transported in a Dewar container to the nearby laboratory and then freeze-dried for at least 48 h. The leaves were then weighed and ground in fine powder using a laboratory mill (CB 2200, Cep Industrie - Département SODEMI, St Ouen l’Aumône, France). Total C content and δ13C of the bulk leaf organic matter were determined using mass spectrometry as above and expressed in ‰. The relative abundance of 13C of leaves was calculated as above (Eq. (6)). SLA was calculated as the ratio of leaf area over leaf dry mass.

Results

Climate and trunk and soil CO2 effluxes

Air temperature declined during the labelling and the chase period from summer values at the beginning of September (up to 25 °C during daytime) to winter values at the end of the chase period (< 0 °C during nighttime; Figure 2). During the same time, soil temperature decreased from 16 to 5 °C. A pronounced drop in air temperature was observed between the two labelling dates. The daily sum of PPFD also declined from 1.2 at the beginning of September to values < 0.1 mol m−2 d−1 in November. Despite some rainfall (84 mm), average soil water content was almost constant in September (19%, data not shown), and was not recorded later.

Figure 2.

Climatic conditions during the period of 1 September to 30 November 2008. Daily sum of rainfall (A), daily sum of PPFD (B) and daily minimum air temperature (C, closed circles), daily maximum air temperature (C, closed triangles) and daily average soil temperature at 10 cm depth (C, open diamonds). The dotted and the dashed lines indicate the labelling dates of LT1 and LT2, respectively.

Figure 2.

Climatic conditions during the period of 1 September to 30 November 2008. Daily sum of rainfall (A), daily sum of PPFD (B) and daily minimum air temperature (C, closed circles), daily maximum air temperature (C, closed triangles) and daily average soil temperature at 10 cm depth (C, open diamonds). The dotted and the dashed lines indicate the labelling dates of LT1 and LT2, respectively.

Both FT and FS (daily averages) exhibited a decline from September to November that followed the decline in temperature (Figure 3). The decline in FT was more pronounced compared to the declines in FS as expected from the different decline in air and soil temperatures.

Figure 3.

Daily average CO2 efflux from the trunk (A, FT) and from the soil (B, FS). Values are mean of three chambers for trunks (one on each labelled tree and one on the control tree) and of four of six collars for soil (two on each labelled tree and two on the control tree but only four collars were recorded simultaneously. The four soil respiration chambers were moved every week with one chamber on each tree and with one chamber moving between the three trees). Vertical bars are standard deviations. The dotted and the dashed lines indicate the labelling dates of LT1 and LT2, respectively.

Figure 3.

Daily average CO2 efflux from the trunk (A, FT) and from the soil (B, FS). Values are mean of three chambers for trunks (one on each labelled tree and one on the control tree) and of four of six collars for soil (two on each labelled tree and two on the control tree but only four collars were recorded simultaneously. The four soil respiration chambers were moved every week with one chamber on each tree and with one chamber moving between the three trees). Vertical bars are standard deviations. The dotted and the dashed lines indicate the labelling dates of LT1 and LT2, respectively.

Labelling

Mean air temperature, air humidity and PPFD within the chamber during the labelling period were 21.3 °C (range 12.0–24.4), 72% (50–92) and 974 (360–1960) μmol m−2 s−1, respectively, for LT1. The weather was cooler and less sunny for LT2 with mean values of 14.5 °C (6.2–16.9), 90% (60–99) and 502 (135–1470) μmol m−2 s−1, respectively. The PPFD measured inside the chamber above the tree was 75% of that measured at the top of the eddy covariance flux tower.

The 12CO2 concentration decreased rapidly between the closure of the chamber and the beginning of the injection, at a rate of 0.050 μmol mol−1 s−1 or 1.48 gC h−1 for LT2 (Figure 4B). The 13CO2 concentration decreased at an average rate of 0.113 μmol mol−1 s−1 (2.34 gC h−1; Figure 4A) for LT1 and of 0.063 μmol mol−1 s−1 (1.86 gC h−1; Figure 4B) for LT2 after the gas cylinder was emptied.

Figure 4.

Evolution of 12CO2 (closed circles) and 13CO2 (open circles) concentrations within the crown labelling chamber (A) during the labelling of the first tree on 9 September and (B) during the labelling of the second tree on 16 September. Vertical arrows indicate from left to right (1) the closing of the chamber, (2) the start of the labelling, (3) the exhaustion of the 13CO2 cylinder and (4) the opening of the chamber. We failed to record the data during the first hour of the first labelling experiment.

Figure 4.

Evolution of 12CO2 (closed circles) and 13CO2 (open circles) concentrations within the crown labelling chamber (A) during the labelling of the first tree on 9 September and (B) during the labelling of the second tree on 16 September. Vertical arrows indicate from left to right (1) the closing of the chamber, (2) the start of the labelling, (3) the exhaustion of the 13CO2 cylinder and (4) the opening of the chamber. We failed to record the data during the first hour of the first labelling experiment.

The total amount of 13CO2 taken up by the crown (U, Eq. (1)), assuming negligible leaks, was calculated for LT2 only because initial data were lacking for LT1. Cumulative U was 21.7 L (i.e., 11.96 g of 13C). U could not be estimated precisely for LT1 because of the loss of data.

The residual amount of 13CO2 in the chamber at the end of the labelling can be calculated from chamber volume and [13CO2] recorded just before opening the chamber. The amount of injected 13CO2 is the difference between the total cylinder volume and this residual amount. It gives 22.6 L (12.45 g of 13C) for LT1 and 21.9 L (12.05 g of 13C) for LT2, respectively.

The isotopic composition of leaves collected just after the end of labelling (30 min) was 1872 ± 233‰ for LT1 and 1075 ± 246‰ for LT2. The foliar biomass of each crown was estimated at 0.90 and 1.98 kg of dry matter for LT1 and LT2, and the carbon content of leaf dry matter was 0.48 ± 0.006 gC gDM−1. This represents a total amount of 13C in the leaves of labelled crown of 13.5 and 21.6 gC for LT1 and LT2. The background of 13C in leaves was obtained knowing the isotopic composition of leaves before labelling and of the control tree (−29.0 ± 0.16‰). The differences between total amount and background were 8.8 and 11.4 gC for LT1 and LT2. This is equivalent to 16.4 and 20.7 L of 13CO2.

δ13C of trunk CO2 efflux

Before labelling, the isotopic composition of the trunk CO2 efflux (δ13CFT) of LT1 and CT was almost identical (−30.5‰ below the crown and −30.4‰ at the base of the trunk for LT1, and −30.1‰ at 1.3 m height for CT). A 10‰-enrichment of δ13CFT compared to that of the control appeared 11 h after the beginning of the pulse labelling at the base of the crown of LT1 and 13 h at the base of the trunk (Figure 5A). The maximum δ13CFT was observed at the base of the crown after 39 h (> 4600‰) and at the base of the trunk after 41 h (> 3500‰). This delay between the base of the crown and the base of the trunk corresponds to a phloem sap velocity of about 1 m h−1. δ13CFT of LT1 declined exponentially during the first two weeks to values around 700‰ and then linearly and slowly during the next weeks to values around 250‰ in mid-November (Figure 5C).

Figure 5.

Time courses of the isotope composition of trunk CO2 efflux (δ13CFT, ‰) after labelling. Comparison of δ13CFT at the base of the crown (closed symbol) and at the base of the trunk (open symbol) for LT1 during the first week (A), first week of the chase period for LT2 (B) and the 2-month chase period for LT1 (C) and LT2 (D). The top horizontal bars on the upper panels indicate diurnal and nocturnal periods. The vertical lines on the upper panels indicate the beginning of labelling (solid), the beginning of 13C recovery (dotted) and the maximum values of δ13CFT (dashed), respectively. The base line on the lower panels is the δ13CFT of CT. Note that the Y scale is not the same for LT1 (left) and LT2 (right).

Figure 5.

Time courses of the isotope composition of trunk CO2 efflux (δ13CFT, ‰) after labelling. Comparison of δ13CFT at the base of the crown (closed symbol) and at the base of the trunk (open symbol) for LT1 during the first week (A), first week of the chase period for LT2 (B) and the 2-month chase period for LT1 (C) and LT2 (D). The top horizontal bars on the upper panels indicate diurnal and nocturnal periods. The vertical lines on the upper panels indicate the beginning of labelling (solid), the beginning of 13C recovery (dotted) and the maximum values of δ13CFT (dashed), respectively. The base line on the lower panels is the δ13CFT of CT. Note that the Y scale is not the same for LT1 (left) and LT2 (right).

δ13CFT enrichment of 10‰ for the LT2 (1.6 m height) compared to that of the control tree occurred after 22 h (Figure 5B). The maximum δ13CFT for LT2 was also delayed compared to LT1 and was observed after 70 h. The maximal δ13CFT for LT2 (1000‰) was lower than that for LT1, and decreased to < 100‰ after 2 months (Figure 5D). This might reflect the differences in tree size, the pulse label being more diluted in LT2 than in LT1. δ13CFT of both labelled trees exhibited large diurnal variations, with maximum values at night and minimal values during daytime (Figure 5A and B). These diurnal fluctuations were dampened after several days but were still visible at least 1 month after labelling.

δ13C of soil CO2 efflux

Before labelling, the isotopic composition of soil CO2 efflux (δ13CFS) on LT1 was −26.3‰ and −27.3‰, respectively, for the two collars, in the same range as the values measured on CT (−28.4 and −25.2‰). A 10‰-enrichment in δ13CFS of LT1 compared to that of control was observed 22 h after the beginning of the labelling (Figure 6A), i.e., about 9 h after the base of the trunk. The maximum enrichment (> 1000‰) was observed 64 h after the beginning of the labelling period, 1 day later than at the base of the trunk.

Figure 6.

Time courses of the isotopic composition of the soil CO2 efflux (δ13CFS, ‰) after labelling. First week of the chase period for two collars (open and closed symbols) on LT1 (A) and LT2 (B) and the 2-month chase period for two collars (open and closed symbols) on LT1 (C) and LT2 (D). The top horizontal bars on the upper panels indicate diurnal and nocturnal periods. The vertical lines on the upper panels indicate the beginning of labelling (solid), the beginning of 13C recovery (dotted) and the maximum values of δ13CFT (dashed), respectively. The base line on the lower panels is the δ13CFT of CT.

Figure 6.

Time courses of the isotopic composition of the soil CO2 efflux (δ13CFS, ‰) after labelling. First week of the chase period for two collars (open and closed symbols) on LT1 (A) and LT2 (B) and the 2-month chase period for two collars (open and closed symbols) on LT1 (C) and LT2 (D). The top horizontal bars on the upper panels indicate diurnal and nocturnal periods. The vertical lines on the upper panels indicate the beginning of labelling (solid), the beginning of 13C recovery (dotted) and the maximum values of δ13CFT (dashed), respectively. The base line on the lower panels is the δ13CFT of CT.

Recovery of the label in δ13CFS of LT2 was delayed when compared with that of LT1. δ13CFS of LT2 was 10‰ above the control 30–35 h after labelling depending on the collars (Figure 6B). The maximum enrichment also differed between collars. It occurred after 72 h (almost at the same time as in the trunk) for the collar exhibiting the highest enrichment (> 1400‰) and after 84 h for the collar exhibiting the lowest enrichment (< 700‰). This might reflect the differences in root contribution to soil CO2 efflux between the two collars. δ13CFS initially declined sharply, then slowly during the two chase period months, reaching 20–60‰ (Figure 6C–D).

Irrelevant (i.e., positive) nocturnal values of δ13C were recorded for both soil and trunk CO2 effluxes on CT the night following the labelling (Figures 5 and 6). These values were surely due to nocturnal accumulation of 13CO2 below the canopy because of very stable atmospheric conditions (no wind). Similar, but less pronounced, positive values were also observed the following night.

Label recovery in CO2 efflux

The rate of increase of the cumulative amount of label recovered in trunk and soil CO2 effluxes was well predicted using Eq. (7) (Figure 7). A sum of two exponential functions was used, which might suggest that two different pools of substrates with different half-lives are respired by the trunk (Table 1). The sum of two exponential functions predicted CLR in trunk respiration better than that of a single exponential one (Fobs = 1721 for LT1 and 92 for LT2, P < 0.05). The half-lives of the fast components were of 2.1 and 8.1 days for LT1 and LT2, while the half-lives of the slow components were longer, 15.2 and 16.2 days for LT1 and LT2. The first component appeared with a lag shorter than a day, while the second component was observed after 6 or 18 days. The total amounts of labelled carbon recovered in trunk CO2 efflux were 1.17 and 0.51 gC for LT1 and LT2 (sum of C1 and C2 in Table 1).

Figure 7.

The CLR in the trunk CO2 efflux (diamonds) and in the soil CO2 efflux (triangles) for LT1 (A) and LT2 (B). The lines show the fit of Eq. (7) to the experimental data. Adjusted model parameters and the coefficient of determination between observed and simulated values (R2) are presented in Table 1.

Figure 7.

The CLR in the trunk CO2 efflux (diamonds) and in the soil CO2 efflux (triangles) for LT1 (A) and LT2 (B). The lines show the fit of Eq. (7) to the experimental data. Adjusted model parameters and the coefficient of determination between observed and simulated values (R2) are presented in Table 1.

Table 1.

Parameters and results of the fit of Eq. (7) (CLR as a function of time after labelling) in trunk and soil CO2 effluxes. Adjusted parameters ± 95% confidence interval and the coefficient of determination between observed and simulated values (R2) are given. K are decay constants, C are asymptotes, i.e., the total amount of labelled carbon that was recovered in a given component and L are lags between labelling and the beginning of the recovery of the label for each component.

 First component
 
Second component
 
 
 C (gCL (d) K (d−1C (gCL (d) K (d−1R2 
LT1 Trunk 0.66 ± 0.01 0.23 ± 0.04 0.330 ± 0.015 0.51 ± 0.01 6.46 ± 0.28 0.046 ± 0.001 0.999 
Soil 0.82 ± 0.01 0.76 ± 0.03 0.213 ± 0.003 0.48 ± 0.01 21.49 ± 0.23 0.038 ± 0.002 0.999 
LT2 Trunk 0.41 ± 0.00 0.95 ± 0.03 0.085 ± 0.002 0.10 ± 0.01 18.00 ± 0.30 0.043 ± 0.002 1.000 
Soil 0.81 ± 0.03 1.38 ± 0.09 0.134 ± 0.011 0.36 ± 0.03 12.66 ± 0.51 0.041 ± 0.004 0.999 
 First component
 
Second component
 
 
 C (gCL (d) K (d−1C (gCL (d) K (d−1R2 
LT1 Trunk 0.66 ± 0.01 0.23 ± 0.04 0.330 ± 0.015 0.51 ± 0.01 6.46 ± 0.28 0.046 ± 0.001 0.999 
Soil 0.82 ± 0.01 0.76 ± 0.03 0.213 ± 0.003 0.48 ± 0.01 21.49 ± 0.23 0.038 ± 0.002 0.999 
LT2 Trunk 0.41 ± 0.00 0.95 ± 0.03 0.085 ± 0.002 0.10 ± 0.01 18.00 ± 0.30 0.043 ± 0.002 1.000 
Soil 0.81 ± 0.03 1.38 ± 0.09 0.134 ± 0.011 0.36 ± 0.03 12.66 ± 0.51 0.041 ± 0.004 0.999 

The cumulative amount of label recovered in soil CO2 effluxes clearly showed a break point for both trees, and the sum of two exponential functions predicted CLR in soil respiration better than that of single exponential models (Fobs = 2134 for LT1 and 37 for LT2, P < 0.05). This effect could once more indicate that two different components were contributing to the label. A lag of more or less 1 day was observed for the first component, while the second component was delayed by 13 (LT2) to 21 (LT1) days. A third component was evident 60 days after labelling on LT1, but the number of data collected after that date was too small for satisfactory fitting. The half-lives of the initial and delayed components of soil CO2 efflux were 3.3 and 18.2 days for LT1 and 5.2 and 16.7 days for LT2, respectively. The total amounts of labelled carbon recovered in soil CO2 efflux were 1.30 for LT1 and 1.17 gC for LT2. The total amounts of carbon recovered in both trunk and soil CO2 effluxes accounted for 28% and 15% of the 13C retrieved in the foliage just after labelling.

Discussion

Labelling

The crown labelling chamber appeared to be a useful alternative to soil-covering chambers that were used in previous pulse-labelling experiments in the field (Carbone et al. 2007, Högberg et al. 2007, Bahn et al. 2009). Firstly, there was no contamination of the soil atmosphere by diffusion of 13CO2 into the soil pores that would later have back-diffused into the atmosphere leading to an artefact in soil CO2 efflux. Secondly, only the target tree was labelled, understorey vegetation remained unlabelled. This is important for studying patterns of below-ground allocation because tree roots, especially root respiration, cannot be easily separated from those of the understorey vegetation. Thirdly, the fact that the chamber only includes the crown enables the labelling of tall trees while maintaining a reasonable chamber volume, and therefore requiring a reasonable amount of 13CO2. In addition, as long as the width of the crown increases less than the height of the tree, this type of chamber can be used on tall trees, the limitations being the height of the scaffolds and the access to electricity for the cooling device in remote areas. As far as we know, this is the first in situ pulse-labelling experiment of trees exceeding 8 m in height. Previous in situ labelling experiments were conducted on much smaller trees (2–4 m, Horwath et al. 1994, Carbone et al. 2007, Högberg et al. 2007).

The amount of 13CO2 taken up by the crown was calculated by three independent methods: (i) from the change in [13CO2] with time, (ii) the difference between the cylinder volume and the residual amount of 13CO2 in the chamber at the end of labelling and (iii) the isotopic composition of leaves just after labelling. The first two estimations were very similar but both could have been affected by leaks, whereas method (iii) gave slightly lower values. This was expected because sugar export, leaf respiration and 12CO2 uptake were likely to dilute the 13C content of leaves. According to these estimates, the labelling device was very efficient since more than 70–90% of the injected 13CO2 (corrected for the amount of 13CO2 remaining in the chamber) was recovered in the bulk leaves 30 min after the end of labelling.

High-frequency tracking of 13C in respiratory effluxes during the chase period

The TDLAS has proved its reliability for the measurement of the natural abundance of 13C in the atmosphere in the field. Its high temporal resolution allows in situ determination of isotopic CO2 mixing ratios, the isotopic signature of nocturnal ecosystem respiration and the net ecosystem exchange of 12CO2 and 13CO2 (Bowling et al. 2003, Griffis et al. 2004, 2005). It has also given more insight into the temporal variations of the isotopic composition of soil respiration, suggesting changes in the relative contribution of respiratory sources (Marron et al. 2009). TDLAS measures 12CO2 and 13CO2 concentrations separately, rather than isotopic ratios. Values of δ13C of respired CO2 as high as +5000‰ can be measured in flow-through chambers as long as the 13CO2 concentration in the outlet of the chamber remains within the [13CO2] range of the calibration gas. As TDLAS measures specifically 12C16O16O and 13C16O16O, it does not take into account all the CO2 isotopologues, such as 12C18O16O and 12C17O16O, for example, (Griffis et al. 2004). The calculation of total CO2 concentration required a correction (see Eqn. (2)), by assuming the contribution of all other isotopologues to be 0.00474 (fother). This value comes from isotopic abundances used in the high-resolution transmission molecular absorption database (HITRAN, http://www.cfa.harvard.edu/hitran//) and it is related to natural abundance. One might expect that fractionation occurring during photosynthesis, carbohydrate metabolism, transport and respiration would affect fother in the respiratory efflux. More importantly, fother of the labelling gas (0.0116) and of the atmosphere are different. As the respiratory substrates are a mixture of molecules produced before, during and after the pulse labelling, and as CO2 at the outlet of the chambers is a mixture of atmospheric CO2 and respired CO2, fother at the chamber outlets is almost unpredictable. However, using 0.0116 for respired CO2 and 0.00474 at the inlet for calculating [CO2] at the outlet of the chamber only slightly changed the absolute values of CO2 efflux (< 1%).

The TDLAS approach allows the capture of temporal variability at an infra-daytime scale (Bowling et al. 2003, Bahn et al. 2009, Marron et al. 2009). The large diurnal variation we observed in δ13C of both the soil and trunk CO2 effluxes was also observed in the soil respiration of alpine grassland (Bahn et al. 2009). It was consistent with diurnal fluctuations of δ13C (0.5−1.5‰, natural abundance) of the soil CO2 efflux in a previous study on this site, which also showed maximal values in the middle of the night and minimal values in the early afternoon (Marron et al. 2009). The diurnal variations we observed might reflect a remobilization of stored carbohydrates to sustain respiration during the night and the use of new unlabelled assimilate derived from ongoing assimilation during daytime (Gessler et al. 2008, Bahn et al. 2009). It is also possible that the diurnal variations in δ13C reflect diurnal changes of the contribution of different sources of CO2. Day-time FT might be a mixture of high-labelled CO2 respired locally and less-labelled CO2 transported by the xylem sap from the root or the soil, while night-time local respiration would contribute to night-time FT (Teskey and McGuire 2005). Similarly, diurnal shifts in the relative contributions of both autotrophic (more labelled) and heterotrophic (less labelled) components of FS might account for the diurnal variations in δ13CFS (Kodama et al. 2008), but higher relative contributions of root respiration are expected during daytime than during nighttime (Marron et al. 2009).

Kinetics of label recovery and carbon allocation

The high measurement frequency of 12CO2 and 13CO2 (one measurement every 35 min) enables us to describe the kinetics of label recovery in trunk and soil respiration. The time lag for detecting 13CO2 in FT and FS is thought to reflect the time required to transfer carbohydrates from the crown to the trunk tissue or to the roots. This time lag was higher for LT2 than for LT1, supporting the hypothesis of a reduced velocity of photosynthate transport due to much lower air temperature at the time of LT2 labelling than at the time of LT1 labelling. In LT1, the 2-h delay observed between the chamber at the base of the crown and the chamber at the base of the tree (2.25 m distance) is consistent with a phloem sap velocity of about 1 m h−1 (Peuke et al. 2001, Keitel et al. 2003). Temperature affected the velocity of photosynthate transport between the crown and the trunk (0.2 and 0.95 days for LT1 and LT2, respectively) much more than in the root system, as the difference in time lag between FT and FS was almost the same for the two labelled trees (about half a day). This was expected because the drop in maximal air temperature between the two labelling experiments (10 °C) was much more important than the change in soil temperature (3 °C). The decrease in phloem sap velocity could be explained by a direct effect of temperature on phloem sap mobility and by an indirect effect through a decrease in photosynthesis and phloem loading. Time lags of 2–4 days were reported for FS in 3–4 m tall black spruce (Carbone et al. 2007), poplar (Horwath et al. 1994) and Scots pine (Högberg et al. 2007) trees. The discrepancies between previous estimates and our results could be attributed to difference between gymnosperm and angiosperm trees or to a better time resolution in the TDLAS-based measurements. A time lag of 0.8 day and a phloem sap velocity of about 1 m h−1 for LT1 (see above) suggest a path length of about 20 m that is twice the size of the trees. These findings corroborate previous results showing a close coupling between photosynthesis and soil CO2 efflux (Ekblad and Högberg 2001, Bowling et al. 2002, Knohl et al. 2005).

The fact that the CLR in an efflux was better described with a sum of two exponential functions than with a single one, implicitly suggests that two components were contributing to both trunk and soil effluxes. The first components had a shorter mean residence time for the first labelled tree than for the second one (MRT 2–3 days and 5–8 days, respectively), and might reflect a direct use of labelled photosynthate in the respiratory metabolism that is rapidly diluted by recent photosynthate. The difference in MRT between the two labelled trees might reflect a slowdown of photosynthesis (i.e., a lower dilution) when temperature dropped just before the second labelling. Transitory storage and rapid remobilization cannot be excluded (see above discussion on the infra-day variations). The second component that contributed to the effluxes after a lag time of 6–21 days and had a MRT of 22–26 days might correspond to the use of stored carbohydrate later in the season when the supply of current photosynthesis no longer supported trunk or root respiration. Similar patterns were observed for FT and FS suggesting that the second component of FS could also be related to the root compartment. However, we cannot fully exclude that the second component of FS could be associated to the heterotrophic component of soil respiration and to the onset of decomposition of labelled organic substrate (turnover of mycorrhiza hyphae or fine roots, turnover of microbes fed on labelled root exudates).

The amounts of label recovered in soil CO2 efflux and in trunk CO2 efflux were different between the two trees (15% of the assimilated 13CO2 in soil respiration for LT1 and 10% for LT2, and 13% of the assimilated 13CO2 in trunk respiration for LT1 and only 5% for LT2). Therefore, the fraction of the label that was allocated to both trunk and soil respiration represented 28% of the assimilated 13CO2 for LT1 and 15% for LT2. Allocation to crown (branches and leaves) respiration was not estimated in this study. On an annual basis, branch and leaf respiration of beech accounted for 43% of autotrophic respiration at this site, and root and stem respiration accounted for 57% (Granier et al. 2000, Damesin et al. 2002). Using these ratios, the fraction of label allocated to whole tree respiration could be estimated at 26–49% of the assimilated 13CO2. As labelling occurred late in the growing season (radial growth surveyed by means of micro-dendrometers ended mid-September at this site), substantial allocation to storage was expected (Mordacq et al. 1986, Barbaroux et al. 2003), especially for the second tree that was labelled later in the season after a marked drop in air temperature.

Conclusions

The large crown labelling chamber we described was suitable to deliver a 3-h pulse of 13CO2 for the labelling of tall trees in the field, without causing contamination of the soil atmosphere that was observed with soil-covering chambers. The high temporal resolution offered by tuneable diode laser spectrometers enables accurate assessments of the time lag between photosynthetic assimilation by the crown and CO2 release in the trunk and in the soil, as well as precise determinations of the half-life of labelled carbon in respiratory effluxes. A stimulating question is whether the time constant can be predicted in mechanistic terms.

The TDLAS approach allows the capture of temporal variability at a scale that could not be reached with classical IRMS or AMS approaches. But there is a compromise between the frequency of measurements and the number of chambers that can be measured sequentially. Our chase experiments were not designed to account for the spatial variability of soil CO2 efflux in the vicinity of the labelled trees. This could lead to high uncertainties in the estimation of the fraction of the label that was allocated to root respiration.

The present pulse-labelling experiment was carried out in September, late in the growing season when growth in both height and diameter has ceased. Allocation patterns to below-ground respiration and trunk respiration may be different during the periods of active growth. The experiments were conducted on only two trees, and are therefore ‘case studies’ from which general statements on carbon allocation should be derived with caution. This study has opened the door for future experiments on several species. Pulse-labelling experiments over the whole growing season are required to capture seasonal changes in carbon allocation patterns.

Funding

The site belongs to the European flux-monitoring network (CarboEurope IP) and to the French network of forest ecosystems (ORE ‘fonctionnement des écosystèmes forestiers’). Financial support was provided by the French National Research Agency (ANR) through the CATS Project (ANR-07-BLAN-0109).

The skilful technical assistance of Thomas Aiguier for help in building the labelling chamber, of Patrick Gross for calibration of the SICK, of Pascal Courtois, Jean Marie Gioria and Bernard Clerc in the field and of Christian Hossann and Claude Brechet running the IRMS, is gratefully acknowledged. The authors thank Christophe Robin for lending them the SICK analyser.

References

Bahn
M.
Schmitt
M.
Siegwolf
R.
Richter
A.
Brüggemann
N.
Does photosynthesis affect grassland soil-respired CO2 and its carbon isotope composition on a diurnal timescale?
New Phytol.
 , 
2009
, vol. 
182
 (pg. 
451
-
460
)
Barbaroux
C.
Bréda
N.
Dufrêne
E.
Distribution of above-ground and below-ground carbohydrate reserves in adult trees of two contrasting broad-leaved species (Quercus petraea and Fagus sylvatica)
New Phytol.
 , 
2003
, vol. 
157
 (pg. 
605
-
615
)
Barbour
M.M.
McDowell
N.G.
Tcherkez
G.
Bickford
C.P.
Hanson
D.T.
A new measurement technique reveals rapid post-illumination changes in the carbon isotope composition of leaf-respired CO2
Plant Cell Environ.
 , 
2007
, vol. 
30
 (pg. 
469
-
482
)
Bowling
D.R.
McDowell
N.G.
Bond
B.J.
Law
B.E.
Ehleringer
J.R.
13C content of ecosystem respiration is linked to precipitation and vapor pressure deficit
Oecologia
 , 
2002
, vol. 
131
 (pg. 
113
-
124
)
Bowling
D.R.
Sargent
S.D.
Tanner
B.D.
Ehleringer
J.R.
Tunable diode laser absorption spectroscopy for stable isotope studies of ecosystem-atmosphere CO2 exchange
Agric. For. Meteorol.
 , 
2003
, vol. 
118
 (pg. 
1
-
19
)
Brown
D.
Rothery
P.
Models in biology: mathematics, statistics and computing
 , 
1994
Chichester
Wiley
688
Carbone
M.S.
Czimczik
C.I.
McDuffee
K.E.
Trumbore
S.E.
Allocation and residence time of photosynthetic products in a boreal forest using a low-level 14C pulse-chase labeling technique
Global Change Biol.
 , 
2007
, vol. 
13
 (pg. 
466
-
477
)
Cerling
T.E.
Ayliffe
L.K.
Dearing
M.D.
Ehleringer
J.R.
Passey
B.H.
Podlesak
D.W.
Torregrossa
A.-M.
West
A.G.
Determining biological tissue turnover using stable isotopes: the reaction progress variable
Oecologia
 , 
2007
, vol. 
151
 (pg. 
175
-
189
)
Ciais
P.
Reichstein
M.
Viovy
N.
, et al.  . 
Europe-wide reduction in primary productivity caused by the heat and drought in 2003
Nature
 , 
2005
, vol. 
437
 (pg. 
529
-
533
)
Cox
P.M.
Betts
R.A.
Jones
C.D.
Spall
S.A.
Totterdell
I.J.
Acceleration of global warming due to carbon-cycle feedbacks in a coupled climate model
Nature
 , 
2000
, vol. 
408
 (pg. 
184
-
187
)
Damesin
C.
Ceschia
E.
Le Goff
N.
Ottorini
J.-M.
Dufrêne
E.
Stem and branch respiration of beech: from tree measurements to estimations at the stand level
New Phytol.
 , 
2002
, vol. 
153
 (pg. 
159
-
172
)
Dawson
T.E.
Mambelli
S.
Plamboeck
A.H.
Templer
P.H.
Tu
K.P.
Stable isotopes in plant ecology
Annu. Rev. Ecol. Syst.
 , 
2002
, vol. 
33
 (pg. 
507
-
559
)
Ekblad
A.
Högberg
P.
Natural abundance of 13C in CO2 respired from forest soils reveals speed of link between tree photosynthesis and root respiration
Oecologia
 , 
2001
, vol. 
127
 (pg. 
305
-
308
)
Epron
D.
Farque
L.
Lucot
E.
Badot
P.M.
Soil CO2 efflux in a beech forest: the contribution of root respiration
Ann. For. Sci.
 , 
1999
, vol. 
56
 (pg. 
289
-
295
)
Epron
D.
Le Dantec
V.
Dufrêne
E.
Granier
A.
Seasonal dynamics of soil carbon dioxide efflux and simulated rhizosphere respiration in a beech forest
Tree Physiol.
 , 
2001
, vol. 
21
 (pg. 
145
-
152
)
Epron
D.
Nouvellon
Y.
Roupsard
O.
, et al.  . 
Spatial and temporal variations of soil respiration in a Eucalyptus plantation in Congo
Forest Ecol. Manag.
 , 
2004
, vol. 
202
 (pg. 
149
-
160
)
Gessler
A.
Tcherkez
G.
Peuke
A.D.
Ghashghaie
J.
Farquhar
G.D.
Experimental evidence for diel variations of the carbon isotope composition in leaf, stem and phloem sap organic matter in Ricinus communis
Plant Cell Environ.
 , 
2008
, vol. 
31
 (pg. 
941
-
953
)
Granier
A.
Ceschia
E.
Damesin
C.
, et al.  . 
The carbon balance of a young beech forest
Funct. Ecol.
 , 
2000
, vol. 
14
 (pg. 
312
-
325
)
Griffis
T.J.
Baker
J.M.
Sargent
S.D.
Tanner
B.D.
Zhang
J.
Measuring field-scale isotopic CO2 fluxes with tunable diode laser absorption spectroscopy and micrometeorological techniques
Agric. For. Meteorol.
 , 
2004
, vol. 
124
 (pg. 
15
-
29
)
Griffis
T.J.
Baker
J.M.
Zhang
J.
Seasonal dynamics and partitioning of isotopic CO2 exchange in a C3/C4 managed ecosystem
Agric. For. Meteorol.
 , 
2005
, vol. 
132
 (pg. 
1
-
19
)
Högberg
P.
Read
D.J.
Towards a more plant physiological perspective on soil ecology
Trends Ecol. Evol.
 , 
2006
, vol. 
21
 (pg. 
549
-
554
)
Högberg
P.
Högberg
M.N.
Gottlicher
S.G.
, et al.  . 
High temporal resolution tracing of photosynthate carbon from the tree canopy to forest soil microorganisms
New Phytol.
 , 
2007
, vol. 
177
 (pg. 
220
-
228
)
Horwath
W.R.
Pregitzer
K.S.
Paul
E.A.
14C allocation in tree-soil systems
Tree Physiol.
 , 
1994
, vol. 
14
 (pg. 
1163
-
1176
)
Houghton
H.J.
The carbon cycle and atmospheric carbon dioxide
Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change Ed. IPCC
 , 
2001
Cambridge University Press
(pg. 
183
-
238
)
Janssens
I.A.
Kowalski
A.S.
Ceulemans
R.
Forest floor CO2 fluxes estimated by eddy covariance and chamber-based model
Agric. For. Meteorol.
 , 
2001
, vol. 
106
 (pg. 
61
-
69
)
Johnson
D.
Leake
J.R.
Ostle
N.
Ineson
P.
Read
D.J.
In situ 13CO2 pulse-labelling of upland grassland demonstrates a rapid pathway of carbon flux from arbuscular mycorrhizal mycelia to the soil
New Phytol.
 , 
2002
, vol. 
153
 (pg. 
327
-
334
)
Keitel
C.
Adams
M.A.
Holst
T.
Matzarakis
A.
Mayer
H.
Rennenberg
H.
Gessler
A.
Carbon and oxygen isotope composition of organic compounds in the phloem sap provides a short-term measure for stomatal conductance of European beech (Fagus sylvatica L.)
Plant Cell Environ.
 , 
2003
, vol. 
26
 (pg. 
1157
-
1168
)
Knohl
A.
Werner
R.A.
Brand
W.A.
Buchmann
N.
Short-term variations in d13C of ecosystem respiration reveals link between assimilation and respiration in a deciduous forest
Oecologia
 , 
2005
, vol. 
142
 (pg. 
70
-
82
)
Kodama
N.
Barnard
R.
Salmon
Y.
, et al.  . 
Temporal dynamics of the carbon isotope composition in a Pinus sylvestris stand: from newly assimilated organic carbon to respired carbon dioxide
Oecologia
 , 
2008
, vol. 
156
 (pg. 
737
-
750
)
Lacointe
A.
Kajji
A.
Daudet
F.A.
Archer
P.
Frossard
J.S.
Mobilization of carbon reserves in young walnut trees
Acta Bot. Gall.
 , 
1993
, vol. 
140
 (pg. 
435
-
441
)
Leake
J.R.
Ostle
N.J.
Rangel-Castro
J.I.
Johnson
D.
Carbon fluxes from plants through soil organisms determined by field 13CO2 pulse-labelling in an upland grassland
Appl. Soil Ecol.
 , 
2006
, vol. 
33
 (pg. 
152
-
175
)
Marron
N.
Plain
C.
Longdoz
B.
Epron
D.
Seasonal and daily time course of the 13C composition in soil CO2 efflux recorded with a tunable diode laser spectrophotometer (TDLS)
Plant Soil
 , 
2009
, vol. 
318
 (pg. 
137
-
151
)
Mordacq
L.
Mousseau
M.
Deleens
E.
A 13C method of estimation of carbon allocation to roots in a young chestnut coppice
Plant Cell Environ.
 , 
1986
, vol. 
9
 (pg. 
735
-
739
)
Nadelhoffer
K.J.
Raich
J.W.
Fine root production estimates and belowground carbon allocation in forest ecosystems
Ecology
 , 
1992
, vol. 
73
 (pg. 
1139
-
1147
)
Ngao
J.
Epron
D.
Brechet
C.
Granier
A.
Estimating the contribution of leaf litter decomposition to soil CO2 efflux in a beech forest using 13C-depleted litter
Global Change Biol.
 , 
2005
, vol. 
11
 (pg. 
1768
-
1776
)
Ngao
J.
Longdoz
B.
Granier
A.
Epron
D.
Estimation of autotrophic and heterotrophic components of soil respiration by trenching is sensitive to corrections for root decomposition and changes in soil water content
Plant Soil
 , 
2007
, vol. 
301
 (pg. 
99
-
110
)
Ostle
N.
Ineson
P.
Benham
D.
Sleep
D.
Carbon assimilation and turnover in grassland vegetation using an in situ 13CO2 pulse labelling system
Rapid Commun. Mass Spectrom.
 , 
2000
, vol. 
14
 (pg. 
1345
-
1350
)
Peuke
A.D.
Rokitta
M.
Zimmermann
U.
Schreiber
L.
Haase
A.
Simultaneous measurement of water flow velocity and solute transport in xylem and phloem of adult plants of Ricinus communis over a daily time course by nuclear magnetic resonance spectrometry
Plant Cell Environ.
 , 
2001
, vol. 
24
 (pg. 
491
-
503
)
Phillips
R.P.
Fahey
T.J.
Patterns of rhizosphere carbon flux in sugar maple (Acer saccharum) and yellow birch (Betula allegheniensis) saplings
Global Change Biol.
 , 
2005
, vol. 
11
 (pg. 
983
-
995
)
Raich
J.W.
Schlesinger
W.H.
The global carbon dioxide flux in soil respiration and its relationship to vegetation and climate
Tellus
 , 
1992
, vol. 
44B
 (pg. 
81
-
99
)
Rangel-Castro
J.I.
Prosser
J.I.
Scrimgeour
C.M.
Smith
P.
Ostle
N.
Ineson
P.
Meharg
A.
Killham
K.
Carbon flow in an upland grassland: effect of liming on the flux of recently photosynthesized carbon to rhizosphere soil
Global Change Biol.
 , 
2004
, vol. 
10
 (pg. 
2100
-
2108
)
Saugier
B.
Roy
J.
Mooney
H.
Roy
J.
Mooney
H.
Saugier
B.
Estimations of global terrestrial productivity: converting toward a single number?
Terrestrial Global Productivity
 , 
2001
Academic Press
(pg. 
543
-
557
)
Teskey
R.O.
McGuire
M.A.
CO2 transported in xylem sap affects CO2 efflux from Liquidambar styreciflua and Platanus occidentalis stems, and contributes to observed wound respiration phenomena
Trees
 , 
2005
, vol. 
19
 (pg. 
357
-
362
)
Trumbore
S.
Carbon respired by terrestrial ecosystems – recent progress and challenges
Global Change Biol.
 , 
2006
, vol. 
12
 (pg. 
141
-
153
)
Valentini
R.
Matteucci
G.
Dolman
A.J.
, et al.  . 
Respiration as the main determinant of carbon balance in European forests
Nature
 , 
2000
, vol. 
404
 (pg. 
861
-
865
)