Respiration is a substantial driver of carbon (C) flux in forest ecosystems and stable C isotopes provide an excellent tool for its investigation. We studied seasonal dynamics in δ13C of CO2 efflux (δ13CE) from non-leafy branches, upper and lower trunks and coarse roots of adult trees, comparing deciduous Fagus sylvatica (European beech) with evergreen Picea abies (Norway spruce).
In both species, we observed strong and similar seasonal dynamics in the δ13CE of above-ground plant components, whereas δ13CE of coarse roots was rather stable. During summer, δ13CE of trunks was about −28.2‰ (Beech) and −26.8‰ (Spruce). During winter dormancy, δ13CE increased by 5.6–9.1‰. The observed dynamics are likely related to a switch from growth to starch accumulation during fall and remobilization of starch, low TCA cycle activity and accumulation of malate by PEPc during winter. The seasonal δ13CE pattern of branches of Beech and upper trunks of Spruce was less variable, probably because these organs were additionally supplied by winter photosynthesis. In view of our results and pervious studies, we conclude that the pronounced increases in δ13CE of trunks during the winter results from interrupted access to recent photosynthates.
European forests (including Siberia) represent 25% of the world's forest area, covering 44% of the total European land mass (FAO 2006). Forest ecosystems play a key role within the global carbon (C) cycle, both via C uptake during plant photosynthesis and C loss during ecosystem respiration (Goodale et al. 2002; Nabuurs et al. 2003). The balance between these two processes determines whether European forests act as carbon sinks or sources (Valentini et al. 2000). Plant respiration is a major component of forest ecosystem respiration (Luyssaert et al. 2010) and comprehensive understanding of the complex dynamics behind plant respiratory processes is crucial to predict carbon cycling in forests under future environments (Trumbore 2006). However, these dynamics are not well understood.
Stable C isotopes are of increasing relevance in the field of carbon cycle research and provide new insights, for example, into the dynamics of plant metabolic processes (Dawson et al. 2002). Stable C isotopic composition (δ13C in ‰) is altered by a variety of processes along the C transfer pathway from the atmosphere through the plant to the soil, such as discrimination during gross photosynthesis, tree C allocation, or C loss to the atmosphere via respiration (Bowling, Pataki & Randerson 2008). Currently, fractionation by photosynthesis is well understood (Farquhar, Ehleringer & Hubick 1989), but large uncertainties remain concerning post-photosynthetic fractionation processes such as tree-internal C transport or plant respiration (Gessler et al. 2009a). For instance, distinct seasonal changes in the δ13C of respiratory CO2 efflux (δ13CE) from non-leafy branches and trunks of European forest trees have been observed (Damesin & Lelarge 2003; Maunoury et al. 2007).
Only few studies have investigated seasonal dynamics on δ13CE of adult forest trees (Damesin & Lelarge 2003; Maunoury et al. 2007), with the winter period hardly covered. Simultaneous measurements on seasonal δ13CE of different tree organs, including both above and below-ground components are still lacking (Bowling et al. 2008). The present study addresses these shortcomings by investigating the δ13CE from 60-year-old forest trees of one evergreen conifer (Picea abies Karst.) and one deciduous broad-leaved species (Fagus sylvatica L.). Rates of CO2 efflux and δ13CE of upper and lower trunks, coarse roots and branches were compared (the latter only in Beech). The most prominent dynamics in δ13CE were expected in the deciduous species due to the switch in respiratory C supply from recent photosynthates to storage carbon during winter, similar to observations during germination of Phaseolus vulgaris (Bathellier et al. 2008). Consequently, the seasonal dynamics were expected to be damped in the evergreen species, which may profit from photosynthates formed during mild winter periods.
MATERIALS AND METHODS
The study was conducted at the forest research site Kranzberger Forst (48°25′ 08" N, 11°39′ 41" E, 485 m a.s.l., Freising, Germany). The mixed forest consisted of 60–70 year old individuals of the late-successional tree species European beech (F. sylvatica) and Norway spruce (P. abies), respectively. Beech trees were established in dense groups of 50–60 individuals (Pretzsch, Kahn & Grote 1998). The mean tree height was 28.8 ± 0.8 m in Spruce and 25.9 ± 0.3 m in Beech (Table 1). Stand density was 829 trees ha−1, and basal area amounted to 46.4 m2 ha−1 (Wipfler et al. 2005). The long-term averages (1970–2000) of annual mean air temperature and precipitation were 7.8 °C and 786 mm, respectively (Matyssek et al. 2007). The dominant soil type was Luvisol with high nutrition and water supply. Slope inclination was orientated North, not exceeding 1.8 °C. Tree canopies were accessible by scaffolding between 17 to 25 m aboveground and by a canopy crane (Häberle et al. 2004). The current study was conducted on seven adult individuals each of Beech and Spruce from May 2008 through June 2009.
Table 1. Tree size, height of CO2 efflux chambers, diameter of corresponding tree component and sap wood depth behind chambers (means ± SE)
Position within tree
Height above ground
DBH = diameter breast height. Sample size n = 7 except Beech branches (n = 5).
25.9 ± 0.3
22.6 ± 1.0
1.7 ± 0.1
0.8 ± 0.0
35.9 ± 7.9
12.6 ± 2.4
17.9 ± 3.9
6.6 ± 0.9
1063.6 ± 249.7
1.3 ± 0.0
24.6 ± 3.5
8.5 ± 0.7
0.7 ± 0.1
25.5 ± 3.5
8.8 ± 0.9
2935.6 ± 433.1
−0.08 ± 0.02
1.5 ± 0.1
0.7 ± 0.0
18.1 ± 2.7
28.8 ± 0.8
6.3 ± 0.1
27.5 ± 2.0
6.1 ± 0.7
1539.9 ± 319.8
1.3 ± 0.0
33.7 ± 2.4
7.7 ± 0.5
0.7 ± 0.1
34.9 ± 2.7
7.9 ± 0.9
2839.6 ± 113.1
−0.05 ± 0.01
1.6 ± 0.2
0.8 ± 0.1
27.9 ± 10.8
Air temperature (TAir), solar radiation, air pressure and absolute air humidity were measured above the canopy on top of the scaffolding. Precipitation was assessed at a forest climate station of the LWF (Bayerische Landesanstalt für Wald und Forstwirtschaft, Freising, Germany), approximately 1 km distant from the research site.
The mean annual TAir during the experiment was 8.2 ± 0.4 °C (±SE, n = 365), and the sum of annual rainfall amounted to 724.4 mm (both calculated from 20 May 2008 through 20 May 2009; see also Fig. 1). Compared to long-term observations (1970–2000), TAir was higher by 0.4 °C and precipitation lower by 62 mm. Winter snow cover during the study period lasted from December 2008 until mid-March 2009.
CO2 efflux rate
The CO2 efflux rate of both upper and lower trunk and of one coarse root was determined on each study tree. In addition, CO2 efflux was measured on one non-leafy branch per Beech tree (n = 5). Plexiglass® (Röhm GmbH, Darmstadt, Germany) chambers were permanently installed at each position (Table 1), resulting in a total of 47 chambers (Beech n = 26, Spruce n = 21, for details see Grams et al. 2010). Trunk chambers were laterally attached, enclosing a trunk surface area of 203 to 534 cm2. Roots and branches were completely enclosed by chambers. Each chamber was equipped with a NTC temperature sensor (negative temperature coefficient sensor, Pt 100, SEMI 833 ET, Hygrotec® Messtechnik GmbH, Titisee, Germany) installed 5 mm into the bark to measure cambial temperature (TBark in °C). All chambers were covered with aluminized polyester foil.
CO2 efflux rates were assessed by a computer-controlled flow-through gas-exchange system (Fig. 2a). A stable reference gas (406.6 ± 0.9 µmol CO2 mol−1; δ13C = −3.06 ± 0.02‰, mean ± SE, n = 325) was continuously pumped via PVC tubes through the chambers at a rate of approx. 0.5 L min−1. Outgoing chamber air was analysed by an IRGA (Infra-red gas analyser, Binos 4b, Emerson Process Management, Weißling, Germany). Two parallel systems allowed for the simultaneous measurements of a maximum of 28 chambers, automatically interchanging between sampling chambers every 8 min. Measurements on Beech and Spruce were interchanged approximately every two weeks.
where R15 represents the efflux rate at 15 °C and Q10 the temperature sensitivity of the CO2 efflux. It was assumed that R equals RM during periods without growth (Amthor 1984) and accordingly, we determined R15 and Q10 during winter (November through February) individually for each chamber and week. Individual Q10 and R15 were used to calculate RM for the whole year on the basis of TBark. We are aware that CO2 efflux may not necessarily represent the actual respiration due to possible contributions of CO2 transported within the xylem sap (Teskey et al. 2008). However, regarding potential CO2 contamination of the respiration rate by the sap flow, the coefficient of determination between CO2 efflux and sap flow were low (general linear model: R2Beech = 0.18, R2Spruce = 0.07, both P ≤ 0.001, n > 4900, each, see also discussion). The calculated difference between RM and the actual CO2 efflux rate was used to identify periods of active growth respiration (RG).
CO2 sampling for stable carbon isotope analysis
The stable C isotope composition (δ13C) of the CO2 efflux was measured once a month in Beech and Spruce throughout the whole experiment (i.e. 13 months). Beech trees were monitored from May 2008 through May 2009 and Spruce from June 2008 through June 2009. All chambers were leak-tested one week before each sampling by applying a slight over-pressure (ca. 2000 Pa) while tolerating pressure drops of < 100 Pa min−1. In the case of trunk chambers, the stable carbon isotope composition of the CO2 efflux was calculated using a two-end-member mixing model (see Dawson et al. 2002). To this end, both the δ13C and the [CO2] of chamber and reference air were sampled (δ13CS, [CO2]S, δ13CR and [CO2]R, respectively). The δ13C of CO2 efflux (δ13CE) was then calculated as:
An automated gas sampler (Gilson 221 XL, Gilson Inc. Middleton, WI, USA) sampled both chamber and reference air into 12 mL glass vials (Exetainer, Labco Limited, High Wycombe, UK) after the IRGA of the flow-through gas-exchange system (see Fig. 2a). Samples were then analysed for δ13C by IRMS (isotope-ratio mass spectrometry, GVI-Isoprime, Elementar, Hanau, Germany). Previous testing at individual points of the system revealed that the whole flow-through gas exchange system (including the IRGA) had no contaminating effect on δ13CR (see Grams et al. 2010). Each trunk chamber was sampled every 3 h.
The amount of CO2 released from coarse roots and non-leafy branches was too small to be analysed by the flow-through systems. Instead δ13CE was assessed by means of a custom-built closed-respiration system that included a total of six sampling vials (Fig. 2b, cf. Prater, Mortazavi & Chanton 2006). For sampling, the system was operated in a closed mode for about one hour, and CO2 release from the coarse root or branch resulted in an increase of (CO2). About every 10 min one sampling vial was isolated by simultaneously closing valves directly in front of and behind the vial. To obtain reliable estimates of δ13CE, we adjusted the sampling interval to attain a (CO2) range of ≥ 100 µmol mol−1 between the first and last sample. The δ13C of CO2 efflux was calculated from the six gas samples taken at each measurement according to the ‘Keeling Plot Approach’ (Keeling 1958; Zobitz et al. 2008). Keeling plots with coefficients of determination (R2) < 0.97 were rejected. Root and branch δ13CE were sampled twice a day during early morning and onset of night.
Sap flow measurements
Sap flow was assessed for each tree at a height of 1.3 m. Firstly, sap flow density was assessed using two Granier-type sensors per tree (Granier 1985) installed into trunks at 2 cm depth. Decline functions of sap flow density towards central sapwood parts (>2 cm) were available for the study site (U. Metzger, Ecophysiology of Plants, Technische Universität München, pers. com.) and used to scale measured sap flow density to whole-tree sap flow. Sapwood depth was assessed on wood cores next to sensors.
Leaf net CO2 exchange
The net rate of leaf CO2 exchange (JCO2 in mmol CO2 m−2 leaf area d−1) was measured on one sun branch per species with two permanently installed gas exchange chambers (Götz 1996). The CO2 concentration at the chamber inlet and outlet was assessed each by an IRGA (BINOS 4b.2, Emerson Process Management, Weißling, Germany). Net CO2 exchange rates were related to the projected leaf area and calculated according to von Caemmerer & Farquhar (1981).
All data were analysed using SPSS 16.0 (SPSS GmbH Software, Munich, Germany). Paired and independent Student's t-tests were used to identify significant differences in mean δ13CE between two consecutive measurements at the same position (branch, upper and lower trunk, coarse root), or between species/sampling positions during leaf fall and leaf flush. For longer time periods, that is, the main vegetative period and the winter dormancy, this was analysed by repeated measures analysis of variance. General linear models (GLM) were used to explain the amount of variance for the whole seasonal patterns in δ13CE. Covariates used were TBark, sap flow and CO2 efflux rate (the latter only in the case of lower trunk δ13CE). The day of the year (DOY) and the tree individual (TI) were treated as constant factors.
CO2 efflux, sap flow and net photosynthetic rate
In general, CO2 efflux of all tree components, sap flow and JCO2, followed the annual course of TAir (Figs 1 and 3). In Beech, branch CO2 efflux peaked at the end of May, followed consecutively by upper trunks (mid-July), lower trunks and coarse roots (both end of July). Similarly, the CO2 efflux in Spruce displayed maxima at upper and lower trunks during mid-June followed by roots at the end of July. CO2 efflux rates of both species were similar, except for lower trunks during summer and for coarse roots in early spring. In both cases, Spruce CO2 efflux was nearly twice as high as in Beech. RM followed a similar annual course as the CO2 efflux. Growth respiration (RG = differences between the CO2 efflux rate and RM) was observed from mid-May through mid-October, indicating active growth processes in the tree components during this time (Fig. 3).
Both species had comparable rates of tree sap flow. Only during April and May 2009, Spruce sap flow rate was higher compared to Beech. During summer, JCO2 in Spruce was lower than in Beech. However, during winter, significant periods of leaf net CO2 uptake were observed in Spruce.
Intra- and interspecific differences in δ13CE
No significant within-species differences were found between δ13CE of non-leafy branches and upper trunks or between upper and lower trunks (Fig. 4, Table 2a). During leaf flush and leaf fall, the δ13CE of Beech coarse roots was enriched in 13C compared to lower trunks. During the main vegetative period, Spruce coarse roots were significantly enriched in 13C compared to lower trunks (Table 2a). The opposite was the case during winter dormancy (for more details see Fig. 6).
Table 2. Differences in δ13C of CO2 efflux between two sample positions (δ13CB= branch, δ13CUT = upper trunk, δ13CLT = lower trunk and δ13CR = coarse root) within the same species (A) and between species (B)
B Difference in δ13C between Spruce and Beech δ13CSpruce–δ13CBeech
Significant differences were calculated by apaired Student's t-test, bindependent Student's t-test and crepeated measures anova for the whole phenological period. Data are shown as mean differences ± SE. Significance levels of P ≤ 0.05, P ≤ 0.01 and P ≤ 0.001 are represented by *, ** and ***. Sample size n for period (I) and (III) = 10–14, for period (II) = 40–56 and for period (IV) = 60–84.
The two species displayed significant differences in δ13CE of the same tree components (Fig. 5, Table 2b). The δ13CE of Spruce trunks was enriched in 13C compared to Beech, except for the upper trunk position during winter. The same trend was observed in coarse roots.
Seasonal dynamics in δ13CE
In both species, distinct seasonal changes in δ13CE were observed (Fig. 6). Changes in δ13CE of non-leafy branches and trunks were most pronounced outside of the main growing season, whereas δ13CE of coarse roots was generally less dynamic (Fig. 6).
Upon completion of leaf flush in April (period I), daily means of δ13CE for Beech branches, upper trunks and lower trunks were rather stable throughout the vegetative period (II) (−28.2 ± 0.2‰, −28.1 ± 0.4‰ and −28.4 ± 0.2‰, respectively, see Table 3), but they significantly declined during leaf senescence (period III, −30.5 ± 0.8‰, −31.1 ± 0.4‰ and −31.7 ± 0.3‰, respectively). From this minimum onwards, a steady increase occurred throughout the winter dormancy (period IV, by 5.6‰, 8.0‰ and 9.1‰, respectively). Except for branches, this increase was significant over the whole period (see repeated measures analysis of variance in Table 3) and was followed by a distinct decline at the beginning of leaf flush (period I, −6.5‰, −9.0‰ and −8.7‰, respectively). Values observed during leaf flush were similar to those during leaf fall of the previous year. The observed decline in spring occurred firstly in branches (March 2009). In contrast to above-ground organs, coarse root δ13CE of Beech was less dynamic but nevertheless significantly increased during winter dormancy (period IV, from −28.3 ± 1.0‰ to −25.4 ± 0.3‰, Fig. 6, Table 3).
Table 3. Influence of the time point of measurement on the δ13C of branch, trunk and coarse root CO2 efflux of beech and spruce trees, detected with repeated measures analysis of variance
Main vegetative period (II) (P value)
Winter dormancy (IV) (P value)
Significance levels of P ≤ 0.05, P ≤ 0.01 and P ≤ 0.001 are represented by *, ** and ***. Significance levels are shown as P-values. Sample size n = 20–28 for period (II) and 30–42 for period (IV).
In Spruce, the decline in δ13CE of above-ground organs during leaf fall was smaller than in Beech (period III). The δ13CE of Spruce lower trunks increased stronger during winter dormancy (period IV) than in upper trunks (8.4‰ and 5.9‰, respectively).
In both species, the variance in δ13CE of upper and lower trunks was found to be significantly related to the day of the year (DOY, strongest influence), the tree individual (TI), TBark and the CO2 efflux rate (Table 4). Sap flow did not cause a significant effect on δ13CE. Including all parameters within one model explained a high amount of variance (R2 upper trunk: Beech: 0.74, Spruce: 0.70, R2 lower trunk: Beech: 0.83, Spruce: 0.59, all with P ≤ 0.001, n > 350). Again, DOY had the strongest influence.
Table 4. Influence of different parameters on annual variability in δ13C of trunk CO2 efflux, given by their coefficients of determinants. Parameters were tested individually by GLM
DOY = Day of the year, TI = tree individual, TBark = bark temperature. Significance levels of P ≤ 0.05, P ≤ 0.01 and P ≤ 0.001 are represented by *, ** and ***. Sample size n > 350 per trunk position.
The main objective of the present study was to identify seasonal dynamics in δ13CE for a deciduous (F. sylvatica) and an evergreen (P. abies) European tree species. Distinct changes in δ13CE were recorded between summer and winter periods. The most pronounced seasonal dynamics were found in deciduous Beech during phenological changes and winter dormancy.
Contribution of sap flow to δ13CE
In addition to local tissue respiration, CO2 transported with the xylem sap was recently recognized to contribute to CO2 efflux (Höltta & Kolari 2009, Saveyn et al. 2008; Teskey et al. 2008). Plant-respired and soil CO2 may enter the xylem stream and be transported up the tree (up to 60% and 10%, respectively, see Teskey & McGuire 2007; Moore et al. 2008; Aubrey & Teskey 2009). In the present study, only a weak positive correlation between rate of sap flow and CO2 efflux was observed in trunks (Beech R2: 0.18, Spruce: R2: 0.07 both P ≤ 0.001). Moreover, δ13CE was not significantly correlated with sap flow (Table 3). Thus, we conclude that in our study the δ13CE in European beech and Norway spruce was not affected by CO2 transported with the xylem sap, which is in accordance with recent findings (Ubierna et al. 2009; Grams et al. 2010).
Seasonality in δ13C of Beech CO2 efflux
After the completion of leaf flush (period I), δ13CE of above-ground organs was rather stable during times of growth respiration (main vegetative period (II): −28.3 ± 0.2‰) and significantly declined during leaf senescence (period III: −31.2 ± 0.4‰). During period (II), large fractions of recent photosynthates are used for wood formation, that is, lignin and cellulose synthesis, whereas in September, photosynthates are assumed to be mainly used for the build-up of reserves, such as starch (Barbaroux & Breda 2002; Helle & Schleser 2004; Vaganov et al. 2009). Starch is enriched in 13C compared to phloem sugars (up to 4‰, see Bowling et al. 2008) due to fractionation processes during its synthesis (Tcherkez et al. 2004; Gessler et al. 2008). Consequently, during starch accumulation, the remaining respiratory substrate and the respired CO2 are assumed to be relatively 13C depleted (Maunoury et al. 2007). The observed drop at the end of the growing season from about −28.3 to −31.2‰ fits to this consideration and could be interpreted as a switch from woody tissue formation to the accumulation of storage C.
During winter dormancy (period IV), we observed a steady increase in δ13CE of branches and trunks. This may partly be attributed to a switch of respiratory C supply from recent photosynthates to storage C, that is, from sugars to more 13C enriched starch (Damesin & Lelarge 2003; Maunoury et al. 2007). Because starch is typically 1–2‰ enriched in 13C compared to sugars (Bowling et al. 2008), this switch alone is unlikely to account for the observed increase in δ13CE of 5.6–9.1‰. We note that starch may become continuously 13C enriched during winter, but we are not aware of a corresponding biochemical mechanism for this. On the other hand, the TCA cycle is often assumed to be sensitive to low temperatures, in particular in comparison to PDH activity (Atkin et al. 2000; Kodama et al. 2008). Thus, prevailing PDH activity during winter may release 13C enriched CO2 (Tcherkez et al. 2003; Maunoury et al. 2007; Priault et al. 2009). However, δ13CE was not correlated to TBark during winter dormancy (data not shown) and reached its highest levels in March when temperatures were already rising. Consequently, the slow but constant increase of δ13CE is unlikely caused only by the temperature dependent metabolic branching between TCA and PDH activity alone. Recently, high phosphoenolpyruvate carboxylase (PEPc) activity has been demonstrated in trunks (Berveiller & Damesin 2008). In this reaction, PEPc re-fixes respired CO2 to supply the TCA cycle with 13C enriched malate (Berveiller et al. 2007b), favouring 13CO2 by 5.7‰ (Raven & Farquhar 1990). While TCA cycle activity is low, PEPc derived malate may not be metabolized completely and may accumulate in trunks (Gessler et al. 2009b). This may gradually enrich the respiratory substrate pool in 13C as this malate may be metabolized in the TCA cycle. To address this hypothesis, future research should focus on the simultaneous measurements of δ13CE and respiratory substrate pools in trunks during winter. During leaf flush (period I) we observed a rapid decrease in δ13CE to a minimum of about −31.6‰, similar to levels during leaf senescence (period III). This suggests that the pool of malate was already completely metabolized at rising temperatures (Barbour et al. 2007). Additionally, soluble C reserves (i.e. sugars and transitory starch) may be mobilized and allocated from lower tree parts to branches (Lacointe et al. 2004), while in the trunk, 13C depleted lipids may serve as respiratory substrates (Tcherkez et al. 2003).
Conversely, coarse root δ13CE of Beech was relatively stable throughout the year and increased during winter dormancy only by about 2.8‰. Our data from adult forest trees are consistent with results from Bathellier et al. (2009) who observed δ13CE of root respiration in P. vulgaris to be almost constant (about −27.5‰), even under C starvation (i.e. 4 d of continuous darkness). According to model calculations, this was due to a high contribution of the pentose phosphate pathway (PPP) to root δ13CE (up to 25%, Dieuaide-Noubhani et al. 1995; Bathellier et al. 2009). Like the TCA cycle, PPP produces 13C depleted CO2. Under starvation, the relative contribution of the TCA cycle changes, accompanied by the use of different respiratory substrate, that is, 13C depleted lipids/proteins. However, while this may affect the δ13CE of leaf respired CO2 (Tcherkez et al. 2003), respective changes in roots are damped by PPP derived CO2 (Bathellier et al. 2009). As opposed to other findings (Klumpp et al. 2005; Schnyder & Lattanzi 2005) root CO2 efflux was not depleted in 13C compared to trunks and branches (Table 2a). However in the present study, root δ13CE originates from coarse roots (diameter of about 1.5 cm) and does not reflect respiration of fine roots or whole root systems as it is the case in the above mentioned studies.
Differences between Spruce and Beech
Compared with Beech, the δ13CE of Spruce was consistently enriched in 13C. While we are not aware of studies comparing δ13CE from a conifer with that of a deciduous tree species, Garten & Taylor (1992) observed a similar offset in foliar δ13C. Likewise, soil CO2 efflux near conifers was reported to be enriched in 13C compared with soils under broad-leafed trees (Steinmann et al. 2004; Andersen et al. 2010). This difference most likely results from a lower ci/ca. ratio in conifer needles compared to leaves of deciduous species, resulting in a lower photosynthetic discrimination against δ13C (Farquhar et al. 1989).
Compared with Beech, the drop of δ13CE in Spruce during fall (period III) was less pronounced and may relate to less starch synthesis in Spruce compared to Beech. The increase of δ13CE in the upper trunk during winter (period IV) was less pronounced than in the lower part of the trunk. We hypothesize that upper trunk respiration was partially fed by winter photosynthates, as Spruce needles displayed prolonged periods with positive net photosynthesis during winter (Fig. 2). Likewise in Beech, 13C enrichment of the CO2 efflux in branches was smaller than in trunks (i.e. about 5.9 versus 8.0–9.1‰). In Beech branches, autotrophic C supply to respiration may result from corticular photosynthesis, producing photosynthates during mild winter days (Pfanz et al. 2002; Berveiller, Kierzkowski & Damesin 2007a). These photosynthates are mainly build by carboxylation of respired, 13C depleted CO2 (Pfanz et al. 2002). Since all chambers were darkened with silver foil, branch δ13CE was not biased by corticular photosynthesis in chambers directly, but photosynthetic refixation could have occurred in the surrounding tissue. These 13C depleted photosynthates could have been transported via the phloem along branches and respired into the chambers. The early decrease of branch δ13CE (compared to trunks) at the end of the winter dormancy under rising temperatures support the idea that corticular photosynthesis contributes to respiratory C pools in Beech branches.
In conclusion, strong and similar seasonal dynamics in δ13CE of above-ground plant components were observed in Beech and Spruce. The most pronounced changes in trunks were observed when growth respiration ceased and δ13CE was dominated by maintenance respiration (phase III, IV and I, Fig. 5). The observed seasonal patterns indicated a switch from woody tissue formation to starch accumulation at the end of trunk diameter growth. The large increase in δ13CE of 5.6–9.1‰ during winter dormancy could not completely be explained by remobilization of starch or by temperature sensitivity of the TCA cycle. We hypothesized an accumulation of 13C enriched malate by PEPc, which needs to be tested in future research. Although seasonal dynamics were similar in both species, seasonal changes in δ13CE of Beech were more pronounced, most likely due to a more abrupt switch in respiratory C supply from recent photosynthates to storage C in the deciduous species. Additional C supply by winter photosynthesis may explain the damped seasonal dynamics in δ13CE of non-leafy branches in Beech and upper trunks in Spruce. Overall, we conclude that the highest seasonal variation of δ13CE occurs in trunks that have no access to additional C from winter or corticular photosynthesis.
We thank Thomas Feuerbach, Peter Kuba, Sepp Heckmair, Hans Lohner, Ilse Süß, Johanna Lebherz and Christof Seidler for their technical assistance, both in the laboratory and in the field. In addition, we gratefully acknowledge Dr Karl-Heinz Häberle, Timo Gebhardt, Ursula Metzger, Dr Michael Leuchner, Nick Hoffmann and the ‘Bayerische Landesanstalt für Wald und Forstwirtschaft’ for providing data or information on the research site. Special thank goes to Dr Howie Neufeld for comments on an earlier version of this manuscript and help with the English text. The authors gratefully acknowledge the support by the Faculty Graduate Center Weihenstephan of TUM Graduate School at Technische Universität München, Germany. This study was funded by the DFG (German Research Foundation, Bonn, Germany) through SFB 607 ‘Growth and parasite defence – competition for resources in economic plants from agronomy and forestry’.