Biogeosciences Carbon flux to woody tissues in a beech / spruce forest during summer and in response to chronic O 3 exposure

The present study compares the dynamics in carbon (C) allocation of adult deciduous beech ( Fagus sylvatica) and evergreen spruce ( Picea abies ) during summer and in response to seven-year-long exposure with twice-ambient ozone (O3) concentrations (2 × O3). Focus was on the respiratory turn-over and translocation of recent photosynthates at various positions along the stems, coarse roots and soils. The hypotheses tested were that (1) 2 × O3 decreases the allocation of recent photosynthates to CO 2 efflux of stems and coarse roots of adult trees, and that (2) according to their different O3 sensitivities this effect is stronger in beech than in spruce. Labeling of whole tree canopies was applied by releasing 13C depleted CO2 (δ13C of −46.9 ‰) using a free-air stable carbon isotope approach. Canopy air δ13C was reduced for about 2.5 weeks by ca. 8 ‰ in beech and 6 ‰ in spruce while the increase in CO 2 concentration was limited to about 110 μl l−1 and 80 μl l −1, respectively. At the end of the labeling period,δ13C of stem CO2 efflux and phloem sugars was reduced to a similar extend by ca. 3–4 ‰ (beech) and ca. 2– 3 ‰ (spruce). The fraction of labeled C ( fE,new) in stem CO2 efflux amounted to 0.3 to 0.4, indicating slow C turnover of the respiratory supply system in both species. Elevated O3 slightly stimulated the allocation of recently fixed photosynthates to stem and coarse root respiration in spruce (rejection of hypothesis I for spruce), but resulted in a significant reduction in C flux in beech (acceptance of hypotheses I and II). The distinct decrease in C allocation to Correspondence to: T. E. E. Grams (grams@tum.de) beech stems indicates the potential of chronic O 3 stress to substantially mitigate the C sink strength of trees on the longterm scale.


Introduction
Tropospheric ozone (O 3 ) is a major component of global climate change (IPCC, 2007), mitigating the carbon (C) sink strength of forest trees and ecosystem productivity (Sitch et al., 2007;Matyssek et al., 2010b).Along with increased emissions of anthropogenic precursors, in particular nitrogen oxides, tropospheric O 3 concentrations are predicted to rise over Central Europe and at the global scale (Fowler et al., 1999(Fowler et al., , 2008;;Prather et al., 2001).Elevated O 3 concentrations are known to negatively affect the metabolism and growth of a wide range of tree species, including deciduous European beech (Fagus sylvatica) and evergreen Norway spruce (Picea abies; Matyssek et al., 2010a, b;Wieser et al., 2002;Nunn et al., 2006).Photosynthetic decline, impaired phloem loading, and increased C demand for repair have all been observed in response to ozone exposure.Detoxification may curtail the tree-internal assimilate flux to stems, roots and soils in response to O 3 (Andersen, 2003;Matyssek and Sandermann, 2003;Wieser and Matyssek, 2007).
Since the flux of current photosynthates is considered an important driver of woody tissue and soil respiration in forests (Ryan et al., 1996;Högberg et al., 2001), limited C availability caused by O 3 stress may affect the respiratory activity and growth of stems and total belowground C allocation (Matyssek et al., 1992;Günthardt-Goerg et al., 1993;Coleman et al., 1996;Spence et al., 1990).As a result, root biomass and sugar concentrations may be reduced (Grulke et al., 1998(Grulke et al., , 2001)).Highlighting the phototoxic potential of O 3 to Central-European forests, Pretzsch et al. (2010) reported a 40 % decrease in stem growth of adult beech upon eight years of twice-ambient O 3 exposure, whereas spruce showed no significant growth response.Likewise, in phytotron experiments on juvenile beech, reduced allocation of recent photosynthates to stems was identified as the mechanistic basis for reduced stem growth and competitiveness in response to 2 × O 3 (Kozovits et al., 2005a, b;Ritter et al., 2011).
In the present paper, we investigate dynamics in C allocation of adult trees in response to chronically elevated O 3 concentrations.Clarification is particularly needed for respiratory C fluxes of woody tissues.We compare the allocation of recent photosynthates to the respiratory turn-over in stems, coarse roots and soils in adult beech and spruce in a naturally grown forest.
In accordance with their contrasting O 3 sensitivity, we hypothesized that (1) 2 × O 3 decreases allocation of recent photosynthates to stem and coarse root CO 2 efflux of adult trees and (2) that this effect is stronger in beech than in spruce.To this end, we took advantage of a unique free-air O 3 fumigation experiment employed in a mixed forest with adult beech and spruce trees (Matyssek et al., 2010b).Stable carbon isotope labeling was performed on these trees using the isoFACE exposure system (Grams et al., 2011).In view of hypothesis evaluation, focus was on translocation of recent photosynthates and CO 2 efflux at various positions along the stems and coarse roots.

Experimental design
The study was carried out during August/early September 2006 in a 60 to 70-year-old mixed beech/spruce stand at "Kranzberger Forst" in southern Bavaria, near Freising, Germany (elevation 485 m a.s.l., 48 • 25 N, 11 • 39 E; Pretzsch et al., 1998).Trees of European beech (Fagus sylvatica [L.]) and Norway spruce (Picea abies [L.] Karst.), about 25 to 28 m high, were exposed to either unchanged ambient (1×) or experimentally increased twice-ambient (2×) O 3 concentrations.The 2 × O 3 regime had experimentally been enhanced since 2000, using a free-air O 3 exposure system (Werner and Fabian, 2002;Karnosky et al., 2005).To prevent risk of acute O 3 injury in the 2× O 3 regime, maximum O 3 concentrations were restricted to <150 nl l −1 (cf.Matyssek and Sandermann, 2003).The exclusion of untypically high O 3 peaks resulted in a chronically enhanced 2× O 3 regime with a higher frequency of O 3 levels that currently occur sporadically at the site, by this, simulating the widely observed trend of currently increasing O 3 background concentrations (Fowler et al., 2008;Sitch et al., 2007;Vingarzan, 2004).The forest grew on luvisol derived from loess over tertiary sediments with high nutrition and water supply.Long-term mean (1970Long-term mean ( -2000) ) annual air temperature and rainfall were 7.8 • C and 786 mm, respectively (monitored by Deutscher Wetterdienst at climate station "Weihenstephan", at 4 km distance from the research site; DWD Offenbach, Germany; Matyssek et al., 2007).Scaffoldings and a canopy crane provided access to the tree canopies.

Climate conditions and stable carbon isotope labeling
After a warm and dry period in July 2006 air temperature decreased during the labeling experiments in August and September (Table 1, Fig. 1).Correspondingly, highest O 3 concentrations occurred during July, and AOT40 (i.e.accumulated O 3 concentrations above a threshold of 40 nl l −1 ) exceeded the critical level of 5 µl O 3 l −1 h under the 1 × O 3 regime already in May (LRTAP Mapping Manual, 2004;Nunn et al., 2005a).O 3 concentrations in the 2 × O 3 treatment were enhanced by a factor of 1.6 because of the maximum level of 150 µl l −1 (see above).Continuous stable carbon isotope labeling was performed from 18 August through 5 September and 26 August through 12 September in beech and spruce, respectively, using a free-air stable carbon isotope exposure system ("isoFACE", for details see Grams et al., 2011).In brief, from 07:00 through 19:00 LT, 13 C-depleted CO 2 (δ 13 C of ca.−46.9 ‰) was homogenously released into the canopy of three study trees in each O 3 regime and species (total of 12 trees) by means of micro-porous tubes.During label exposure, O 3 concentrations (means ± SE) were 29.7 ± 6.9 (1 × O 3 ) and 49.3 ± 11.9 nl l −1 (2 × O 3 ; Fig. 1a).Photosynthetic photon flux density (PPFD) was moderate due to frequently overcast sky and occasional precipitation (48 and 32 mm during beech and spruce labeling period, respectively, Fig. 1b).

Isotope-ratio mass spectrometry (IRMS)
Gas samples were analyzed for δ 13 C within 48 h by IRMS (GVI-Isoprime, Elementar, Hanau, Germany) coupled to a gas autosampler (Gilson 221 XL, Gilson Inc. Middleton, USA).Dried plant material was analyzed in a combined elemental analyzer (EA3000, Euro Vector, Milan, Italy) and IRMS.Carbon isotope ratios are expressed in delta notation (δ 13 C) using the Vienna PeeDee Belemnite (VPDB) as a standard.For gaseous and solid samples, the iterated measurements of a laboratory working standard showed a precision of δ 13 C < 0.1 ‰ (SD, n = 10).to sun and shade leaves).Canopy air from all sampling positions was sucked through PVC tubes by means of membrane pumps, analyzed for CO 2 concentration (infra-red gas analyzer (IRGA), Binos 4b.1, Rosemount AG, Hanau) and sampled once a day (∼12:00 LT) using a 100 ml syringe.Gas samples were flushed through 12 ml Exetainer vials and analyzed as detailed above.

Assessment
During labeling, δ 13 C of canopy air was effectively decreased.Compared to the unlabeled beech control, mean reductions in sun and shade crowns under 1 × O 3 were 8.1 ± 0.2 and 8.9 ± 0.3 ‰, respectively, and under 2 × O 3 9.2 ± 0.4 and 8.4 ± 0.5 ‰, respectively, (Table 2b).In spruce, mean reductions under 1 × O 3 were 6.0 ± 0.6 ‰ and 6.3 ± 0.8 ‰, respectively, and under 2 × O 3 , 7.5 ± 0.9 ‰ and 6.5 ± 0.7 ‰, respectively (Table 2a).CO 2 concentration in the canopy air of beech under both O 3 regimes was increased by about 110 µl l −1 and in spruce by about 80 µl l −1 (Table 2a).In both species, [CO 2 ] and δ 13 C of canopy air were each similar before and after labeling.Release of CO 2 and thus label application in beech exceeded that of the spruce experiment.The increase in CO 2 concentration of the canopy air did not affect the sap flow of labeled trees, suggesting unchanged stomatal conductance at the leaf level (Grams et al., 2011).Hence, the rate of CO 2 uptake was assumed to rise to some extent, while the increase in leaf internal to external CO 2 concentration was estimated to be small (<0.02).Therefore, changes in photosynthetic discrimination against 13 C were calculated to stay below 0.4 ‰ (Grams et al., 2011).

Assessment of stem and coarse root CO 2 efflux
Stem and coarse root CO 2 efflux (E) of labeled and unlabeled control trees was assessed by means of a computercontrolled open gas exchange system (for details see Grams et al., 2011).Plexiglas chambers (Plexiglas ® , Röhm GmbH, Darmstadt, Germany) were attached at a lower and upper stem position and at one coarse root per tree (except for the unlabeled control spruce tree).Chambers were darkened with aluminized polyester foil to avoid refixation of efflux CO 2 by corticular photosynthesis.For assessment of CO 2 efflux, chambers were connected through PVC tubing to an IRGA (Binos 4b, Emerson Process Management, Weißling, Germany).Stem CO 2 efflux was based on the volume (V in m 3 ) of the stem sector behind the chamber (i.e.living tissue of bark and sapwood) and coarse root CO 2 efflux on the totally enclosed coarse root volume, respectively (Desrochers et al., 2002;Saveyn et al., 2008).

δ 13 C of stem and coarse root CO 2 efflux
Data on δ 13 C of CO 2 efflux (δ 13 C E ) sampled from stems and coarse roots are shown as 24 h-means (±SE).Coarse root δ 13 C E was assessed once per day (between 10:00 and 13:00 LT) by means of a closed respiration system (for details see Grams et al., 2011).A total of six 12 ml Exetainer vials were subsequently flushed with chamber air of increasing CO 2 concentration and δ 13 C E of coarse roots was calculated according to the "Keeling Plot approach" (Keeling, 1958(Keeling, , 1961)).Air from stem respiration chambers was www.biogeosciences.net/8/3127/2011/Biogeosciences, 8, 3127-3138, 2011 automatically sampled in 12 ml Exetainer vials, which were flushed with sample gas for six minutes each, at a flow rate of 0.15 l min −1 .A total of eight samples per day and chamber were assessed.Isotopic signature of CO 2 efflux of the stem was calculated after Eq. ( 1) using a two end-member mixing model.
where, [CO 2 ] sample = CO 2 concentration of sample gas from a stem respiration chamber (µl l −1 ), [CO 2 ] reference = CO 2 concentration of reference gas from an empty chamber (µl l −1 ), δ 13 C sample = δ 13 C of sample gas from a stem respiration chamber (‰) and δ 13 C reference = δ 13 C of reference gas from an empty chamber (‰).We considered that stem CO 2 efflux may not only consist of local tissue-respired CO 2 , but may be biased by xylemtransported CO 2 deriving from lower stem parts and/or root respiration (Teskey et al., 2008).However, the absent correlation between xylem sap flow and stem respiration rate or δ 13 C E (data not shown) suggests xylem-transported CO 2 to only marginally interfere with sampled CO 2 or to originate from similar respiratory processes as the locally respired CO 2 behind the stem chamber.

Fraction of labeled C in stem respiration
The fraction of labeled carbon (f E,new ) in CO 2 efflux (E) was calculated following Lehmeier et al. (2008) and Gamnitzer et al. (2009): where, δ 13 C old represents the δ 13 C of E before labeling and δ 13 C new the δ 13 C of E of a tree grown (theoretically) continuously with labeled CO 2 .The labeling period of 18 to 19 days was too short to fully achieve new isotopic equilibrium in Eand therefore δ 13 C new was derived from C isotope discrimination ( 13 C) before labeling, following Eqs.( 3) and ( 4): 13 C=([δ 13 C unlabeled air −δ 13 C old ]/[1000+δ 13 C old ])•1000(‰) (3) where, δ 13 C unlabeled air and δ 13 C labeled air represent the δ 13 C of canopy air before and during the labeling, respectively.Day-to-day variation in δ 13 C E may occur from variations in label incorporation and in 13 C depending on weather conditions (Pate and Arthur, 1998;Bowling et al., 2008).Thus, δ 13 C E of the labeled trees were corrected for the dayto-day variations in 13 C (being rather small, i.e. <0.5 ‰) of the unlabeled control trees, which showed rather stable δ 13 C E throughout the experiment, i.e. 22.4 ± 0.1 and 21.4 ± 0.1 ‰ for the upper and lower stem positions of beech, respectively, and 19.4 ± 0.1 ‰ for the lower stem position of spruce.

Assessment of δ 13 C of phloem sugars
Phloem sap was sampled on day 0 and during the last labeling day from the lower stem position following the method of Gessler et al. (2004).Small pieces of bark with adherent phloem tissue (∅ 5 mm) were cored in the vicinity of the lower stem chamber and incubated (5 h at 4 • C) in 15 mM sodium polyphosphate buffer (Sigma-Aldrich, Munich, Germany).After centrifugation (12 500 rpm, 5 min), phloem sap was analyzed for water soluble sugars (sum of sucrose, fructose, glucose, raffinose and pinitol; i.e.C PS in mg) by means of HPLC (CARBOsep CHO-820 calcium column, Transgenomic, 219 Glasgow, UK).Freeze-dried phloem sap was analyzed for stable carbon isotope (δ 13 C sample in ‰) and element composition (C sample in mg), and δ 13 C of phloem sugars (δ 13 C PS in ‰) was calculated according to Eq. ( 5): with δ 13 C NPS representing δ 13 C of non-sugar C (assuming δ 13 C NPS to correspond to δ 13 C sample before labeling, cf.Grams et al., 2011) and C NPS (in mg) denoting the nonsugar C content after labeling (calculated as difference between C sample and C PS ) in the phloem sap.

Sampling of leaves and fine roots
Leaves and fine roots were sampled before and during the last labeling day.Leaves were collected with different exposure to compass directions in sun and shade crowns.Recently grown fine roots (≤2 mm diameter) of labeled and control trees were sampled from organic soil horizons (<10 cm soil depth) and cleaned from soil with distilled water.Dried plant material (72 h at 65 • C) was fine-ground and weighed into tin capsules for δ 13 C analysis.

Assessment of δ 13 C of soil respired CO 2
Soil gas samples were collected as detailed by Andersen et al. (2010).In brief, specific soil-gas sampling wells were placed belowground prior to tree labeling (distance from bole base of about 0.2 to 0.5 m) at 8 cm and 15 cm depth.Teflon tubing was used to draw 5-8 ml of soil gas from each sampler using a gas-tight syringe.Each beech and spruce tree served as its own control by following the change in δ 13 C of soil-respired CO 2 throughout 2.5 weeks of labeling.In the case of beech, a total of four soil-gas sampling wells were additionally installed at an unlabeled control plot.Gas samples were subsequently filled into 12 ml Exetainer vials and analyzed for δ 13 C. Calculation of δ 13 C of soil-respired CO 2 follows Eq. (1), while CO 2 of ambient air above the soil served as reference.Note that soil CO 2 efflux was not adjusted by −4.4 ‰ to account for the more rapid diffusion of 12 C compared to 13 C (Andersen et al., 2010).δ 13 C analysis of additional gas samples taken directly above the forest floor indicated that CO 2 label was restricted to the crown and did not reach the forest soil (Grams et al., 2011).

Statistical analyses
Statistical analysis was performed using the SPSS 16.0 software package (SPSS Inc., Chicago, USA).Individual study trees were regarded as experimental units, and beech and spruce were analyzed separately.Data were statistically analyzed using General Linear Model (GLM) approach and ttests where appropriate.Statistical evaluation of the course in δ 13 C E of stems and coarse roots and the fraction of labeled C in stem CO 2 efflux and coarse root CO 2 efflux of labeled trees was performed using repeated measures analysis of variance.Differences at p ≤ 0.05 were regarded as statistically significant, and at p ≤ 0.1 as marginally, and denoted by * and (*), respectively.

Stem and coarse root CO 2 efflux
In general, both species displayed up to 4 times higher (beech) and up to 2 times higher (spruce) CO 2 efflux rates at the upper compared to the lower stem position (Table 3), whereas rates of coarse roots were 10 to 60 times higher than in stems.In beech, 2 × O 3 significantly diminished the CO 2 efflux rate of the upper stem (by ca.−60 %), but caused a pronounced, but only marginally significant (p = 0.065), increase in coarse roots (by ca.+65 %).In spruce, CO 2 efflux rate of the upper stem position was significantly increased under 2 × O 3 (by ca.90 %), whereas this effect was much smaller (ca.20 %) and statistically not significant at the lower stem position.However, long-term exposure to 2 × O 3 reduced the CO 2 efflux rate of spruce coarse roots by ca. 25 % (not statistically significant, p = 0.157).
Beech Spruce Upper Stem 14.1 ± 2.7 a 5.5 ± 1.1 a * 12.8 ± 0.6 a 24.6 ± 1.6 a * Lower Stem 3.8 ± 1.8 b,c 4.9 ± 1.9 a,c 11.9 ± 0.9 a,c 14.While unlabeled control trees displayed minor day-today variations in δ 13 C E of the various organs during labeling (SD < 0.3 ‰), labeled trees displayed decreasing val-ues upon label application (Fig. 2).In beech, δ 13 C E of the stems decreased from day 2 onwards under both O 3 regimes (Fig. 2a), with a significantly more pronounced decline under 1 × O 3 .Likewise, coarse root δ 13 C E decreased from day 2 onwards (Fig. 2c), although this effect was less prominent than in stems.Similar to beech, δ 13 C E of stems in spruce decreased from day 3 onwards under both O 3 regimes (Fig. 2b).Contrasting with beech, the decline was significantly stronger under 2 × O 3 and more pronounced in the upper compared to the lower stem position (p < 0.05, except for day 3).In coarse roots, the decline in δ 13 C E was somewhat delayed, in particular under 1 × O 3 and somewhat stronger under 2 × O 3 (p = 0.085 at day 5, Fig. 2d).

Fraction of labeled C in stem and coarse root CO 2 efflux
In beech, the fraction of labeled carbon (f E,new ) in stem CO 2 efflux started to increase during labeling day 2 and was significantly lower in 2 × O 3 compared to 1 × O 3 from day 3 onwards (Fig. 3a).At the end of the labeling period (day 19), f E,new had approached maximum levels of 0.40 ± 0.01 under 1 × O 3 , whereas under 2 × O 3 only 0.33 ± 0.06 and 0.26 ± 0.06 at the upper and lower stem position, respectively, were reached.Lowest f E,new was observed for coarse roots (maximum of 0.2), being significantly reduced under 2 × O 3 from day 5 onwards (Fig. 3c).In spruce, f E,new of stem CO 2 efflux started to increase on labeling day 2, reaching maximum levels of 0.37 ± 0.03 (upper stem) and 0.25 ± 0.05 (lower stem) under 1 × O 3 , and 0.39 ± 0.06 and 0.30 ± 0.02, respectively, under 2 × O 3 at the end of the labeling period (day 18, Fig. 3b).Increase of f E,new in spruce coarse roots started somewhat delayed (day 3) but reached levels similar to those of the lower stem position (Fig. 3d).
Contrasting with beech, 2 × O 3 did not result in a consistently reduced f E,new in stems and coarse roots.3.4 δ 13 C in leaves, phloem sugars, fine roots and soil respired CO 2 before labeling Before labeling, no apparent differences in δ 13 C caused by the long-term 2 × O 3 exposure were found in the foliage, phloem sap of the stem, fine roots and soil respired CO 2 in either species (Table 4).In general, δ 13 C in the sun leaves was significantly increased by ca. 3 ‰ (beech) and 2 ‰ (spruce) compared with shade leaves each.The δ 13 C of soil-respired CO 2 underneath beech of about −24 ‰ was not affected by the O 3 treatment.In comparison with beech, all samples from spruce were enriched in 13 C by 1 to 2 ‰ (p ≤ 0.05).
In spruce, δ 13 C of soil respired CO 2 was reduced by about 1.2 ‰ under 2 × O 3 and increased by about 1.0 ‰ at a soil depth of 15 cm compared to 8 cm.

Shift in δ 13 C of CO 2 efflux and organic material by the end of labeling
During the 2.5 week labeling period, the δ 13 C of stem and root CO 2 efflux, soil-respired CO 2 and organic samples (phloem sugars, leaves and fine roots) in the unlabeled control trees of both species was only marginally affected (<0.5 ‰, Fig. 4).In labeled beech, the drop in δ 13 C E at the end of label application in the upper stem position was unaffected by O 3 (3.5 ± 0.2 ‰ in both O 3 treatments), but less pronounced at the lower stem position under 2 × O 3 (3.3 ± 0.1 ‰ and 2.3 ± 0.5 ‰ under 1× and 2 × O 3 , respectively) (Fig. 4b, c).Phloem sugars sampled from the lower stem position displayed similar shifts in δ 13 C of 4.0 ± 1.4 ‰ and 3.5 ± 0.6 ‰ under 1× and 2 × O 3 , respectively.In consistency with the reduced label strength in spruce canopy air (about 6.0 ‰ compared to 8.2 ‰ in beech), the drop in stem δ 13 C E of spruce was lower than in beech (Fig. 4e,  f).Conversely to beech, the drop was somewhat increased by 2 × O 3 : upper and lower stem position of 2.4 ± 0.2 ‰ and 1.8 ± 0.3 ‰ under 1 × O 3 , respectively, and 2.8 ± 0.2 ‰ and 2.1 ± 0.2 ‰ under 2 × O 3 , respectively.Again, a similar shift was observed in phloem sugars (3.2 ± 0.3 ‰ and 2.5 ± 0.2 ‰ under 1× and 2 × O 3 , respectively).Corresponding changes of δ 13 C in leaf bulk material were much smaller (about 1.5 ‰).Upon labeling, belowground allocation of recent photosynthates was not affected by the O 3 treatment and, in general, was reduced compared to stem CO 2 efflux and phloem sugars.The decline upon labeling in δ 13 C E of coarse roots was 1.8 ± 0.1 ‰ and 1.4 ± 0.1 ‰ in beech and 1.7 ± 0.9 ‰ and 2.1 ± 0.8 ‰ in spruce under 1× and 2 × O 3 , respectively.Under beech, changes in δ 13 C of soil-respired CO 2 were similar to coarse roots δ 13 C E (about 1.5 to 2.5 ‰), whereas soil CO 2 under spruce remained unchanged (Fig. 4e, f).Similar to leaf bulk material, δ 13 C of fine roots displayed smaller changes than sampled CO 2 efflux and was in the range of 0.5 ‰, irrespective of the O 3 treatment.

Discussion
Our study compares the flux of recent photosynthates to the CO 2 efflux of stems and coarse roots in adult deciduous beech and evergreen spruce during summer and in response to seven-year long 2 × O 3 treatment.The hypothesis I that long-term exposure to elevated O 3 reduces the flux of recently fixed C to CO 2 efflux of stems and coarse roots was accepted for beech but rejected in the case of spruce, which is in accordance with their contrasting O 3 sensitivities (support for hypothesis II).
Long-term exposure to 2 × O 3 for seven years did not significantly affect the δ 13 C of beech and spruce leaves or sugars transported in the phloem sap during late summer (Table 4, cf. Grams et al., 2007;Gessler et al., 2009).Nevertheless, δ 13 C of beech sun leaves displayed a tendency similar to that reported by Kitao et al. (2009) in that 2 × O 3 increased δ 13 C of leaf dry matter caused by O 3 -induced stomatal closure.Likewise, spruce displayed some photosynthetic and stomatal limitation under 2 × O 3 although varying from year to year (Nunn et al., 2005b(Nunn et al., , 2006)).In general, δ 13 C of leaf and fine root biomass was about 2 ‰ higher in spruce compared to beech, likely resulting from higher leaf-level water-use efficiency in the evergreen conifer compared to deciduous trees www.biogeosciences.net/8/3127/2011/Biogeosciences, 8, 3127-3138, 2011 Statistical evaluation was performed using the t-test for paired comparisons.
Beech Spruce  In addition, data from one unlabeled control beech and spruce tree are included to confirm no effect of weather conditions on δ 13 C during experimentation.Overall, the t-test for paired comparisons indicated no significant differences in δ 13 C shift between O 3 regimes within CO 2 and solid samples of labeled beech and spruce.(Matyssek, 1986;Garten and Taylor, 1992;Diefendorf et al., 2010).
In both beech and spruce, labeled photosynthates were detected in the upper and lower stem CO 2 efflux from day 3 onwards (Figs. 2 and 3).Thus, slower C transfer in the phloem of gymnosperms compared to angiosperms could not be confirmed (Kuptz et al., 2011a;Dannoura et al., 2011).The fraction of labeled C (f E,new ) in the CO 2 efflux of beech stems was significantly reduced under 2 × O 3 (support of hy-pothesis I), indicating a higher dependency on C stores of the respiratory supply under 2 × O 3 (cf.Ritter et al., 2011).Such a response may be caused by (1) a direct adverse effect of O 3 on beech photosynthesis and thus reduced label uptake, although reductions were typically small (Nunn et al., 2005b;2006), or (2) a changed C allocation pattern by e.g. an O 3 -inhibited assimilate transport from the leaves.As a consequence the respiratory activity of stem tissues may be restricted (Matyssek et al., 2002) and C stores in stems and roots may decrease towards the end of the growing season (McLaughlin et al., 1982).Consequently, re-growth and bud development in spring may become limited (Matyssek and Sandermann, 2003).The significantly decreased flux of recent photosynthates to beech stems represents the mechanistic basis for the observed loss in stem productivity of 40 % under long-term exposure of 2 × O 3 (Pretzsch et al., 2010).In consistency with model predictions (Sitch et al., 2007), this indicates the potential of chronic O 3 stress to substantially mitigate the C sink strength of trees (Matyssek et al., 2010b).Contrasting with beech, exposure to 2 × O 3 in tendency increased the fraction of labeled C (f E,new ) stem CO 2 efflux of spruce, rejecting hypothesis I for spruce.At the same time, the rate of stem CO 2 efflux was significantly increased under 2 × O 3 .Such a stimulation following O 3 exposure has been reported in several studies on herbaceous plants (Grantz and Shrestha, 2006;Reiling and Davison, 1992) and is known to sustain repair-and detoxification processes (Matyssek et al., 1995;Rennenberg et al., 1996).The slightly increased C allocation to such processes in spruce may relate to its overall lower O 3 sensitivity compared to beech (Kozovits et al., 2005a, b;Matyssek et al., 2010b;Pretzsch et al., 2010).Whereas under 2 × O 3 allocation of C to reserves in beech stems may be restricted (Ritter et al., 2011) putatively reducing C supply for stem growth in the following year.
We do not expect the observed O 3 effects to be counteracted by the short-term increase in CO 2 concentration during labeling as CO 2 × O 3 interactions in beech are typically related to reductions in stomatal aperture (Grams et al., 1999;Grams and Matyssek, 1999) that were absent during labelling (Grams et al., 2011).Moreover, adjustments of beech in response to the long-term exposure (i.e. 7 yr) to the 2 × O 3 regime are unlikely to be ameliorated by short-term (i.e.2.5 weeks) increases in CO 2 concentration by about 100 µl l −1 .
Reduction of δ 13 C in canopy air for 2.5 weeks by about 8 and 6 ‰ resulted in a drop of stem δ 13 C E in beech of 3-4 ‰ and in spruce by 2-3 ‰, respectively (Fig. 4b-f).Correspondingly, f E,new of stem CO 2 efflux amounted to about 0.3 to 0.4 in both species.In parallel, δ 13 C of labeled phloem sugars was reduced to a similar extent by about 4 and 3 ‰ in beech and spruce, respectively, suggesting respiration of phloem sugars to be the main C source for stem CO 2 efflux (Kuptz et al., 2011a).Unlabeled C in phloem sugars after 2.5 weeks of continuous labeling may derive from "old C" atoms in C skeletons of currently synthesized sucrose as a consequence of slow turnover of precursor molecules or from remobilized C stores (Gessler et al., 2008;Tcherkez et al., 2003).We note that CO 2 efflux sampled from stems (and roots) may be affected by xylem-transported CO 2 deriving from lower stem region and/or root respiration (Teskey et al., 2008).We did not find a correlation between sap flow and both rates of stem CO 2 efflux and stem δ 13 C E in our study (cf.Grams et al., 2011;Kuptz et al., 2011a, b).Hence, contribution of xylem transported CO 2 to sampled CO 2 efflux may be small or originate from similar respiratory processes as at the sampled stem position.In fact, the contribution from soil CO 2 to stem CO 2 efflux was recently concluded to be rather small (Gebhardt, 2008;Aubrey and Teskey, 2009;Ubierna et al., 2009).However, contribution of respiratory CO 2 from lower parts of the stem or roots to sampled CO 2 efflux can not be ruled out completely and the extent of this putative influence remains obscure.
In consistency with the findings on δ 13 C E in stems, 2 × O 3 distinctly reduced f E,new of coarse root efflux of beech, supporting hypothesis I.The decrease in coarse root δ 13 C E during the labeling in summer was about 1-2 ‰ smaller than in stems, indicating a lower dependence of root CO 2 efflux on recent photosynthates (Wingate et al., 2008;Bathellier et al., 2009;Kuptz et al., 2011a).However, soil-respired CO 2 , which includes large contributions by root-respired CO 2 of unlabeled neighboring trees and heterotrophic soil respiration (Högberg et al., 2001;Andersen et al., 2005Andersen et al., , 2010)), was reduced in δ 13 C by 1.5 to 3 ‰.Hence, beech fine roots and associated micro organisms appear to be a relatively strong sink for recently fixed C during summer (Högberg et al., 2001;Plain et al., 2009;Epron et al., 2011).Slightly pronounced shifts in soil-respired CO 2 under 2 × O 3 fit well with previously reported increases in fine-root turn-over of beech under long-term O 3 exposure (Nikolova et al., 2010).Similar to C flux in spruce stems, elevated O 3 did not reduce the allocation of recent photosynthates to coarse root CO 2 efflux during summer (cf.Andersen et al., 2010).However, the C label was hardly detectable in the soil-respired CO 2 around the trees (Andersen et al., 2010), which may indicate favored allocation of labeled C to storage and/or structural pools in the fine roots during summer (cf.Kuptz et al., 2011a), resulting in the observed drop of δ 13 C in the fine root tissue during labeling (Fig. 4e, f).
In conclusion, the transfer of recently fixed C from beech and spruce crowns to stem and coarse root CO 2 efflux within 2 to 3 days displays tight coupling with canopy photosynthesis during summer.Our labeling approach for tracking of individual, isotopically labeled sugar molecules through tall beech and spruce trees should not be confused with the faster propagation of phloem pressure-concentration waves (Kuzyakov and Garvrichkova, 2010;Mencuccini and Hölttä, 2010).Chronic exposure to 2 × O 3 reduced allocation of photosynthates to the stem and coarse roots of beech and spruce in contrasting ways.The conifer spruce increased the flux of photosynthates to stems (rejection of hypothesis I for spruce), whereas this flux was restricted in stems and coarse roots of deciduous beech (acceptance of hypotheses I and II).The observed patterns in translocation of recent photosynthates are interpreted as a mechanistic basis for observed reductions in beech stem growth, highlighting the potential of chronic O 3 stress to substantially mitigate the C sink strength of trees.

Fig. 2 .
Fig. 2. Course in δ 13 C E of stems (triangles: upper stem, circles: lower stem) and coarse roots (diamonds) of labeled beech (a, c) and spruce (b, d) under 1× (white) and 2 × O 3 (black) (daily means ±SE, n = 3 trees) during labeling.Consideration was given to the initial difference in δ 13 C E by using data of day 0 as covariate.Dashed line indicates the initiation of the label application.Significant differences between O 3 regimes and stem positions at p ≤ 0.05 are indicated by * and • , respectively.Marginal significance at p ≤ 0.10 is denoted by (*).Statistical evaluation was performed using repeated measures analysis of variance.

Fig. 3 .
Fig. 3. Fraction of labeled C in stem CO 2 efflux (triangles: upper stem, circles: lower stem) and coarse root CO 2 efflux (diamonds) of labeled beech (a, c) and spruce (b, d) under 1× (white) and 2 × O 3 (black) (daily means ±SE, n = 3 trees).Dashed line indicates the initiation of the label application.Significant difference between O 3 regimes at p ≤ 0.05 is denoted by *.Marginal significance at p ≤ 0.10 is denoted by (*).Statistical evaluation was performed using repeated measures analysis of variance.

Fig. 4 .
Fig. 4. Shift in δ 13 C of canopy air, upper and lower stem CO 2 efflux, soil respired CO 2 at 8 and 15 cm soil depth, phloem sugars, sun and shade leaves as well as fine roots of beech(a-c) and spruce (d-f) after 2.5 weeks of labeling.Data are shown as means (±SE) for three labeled trees under 1× and 2 × O 3 , respectively.In addition, data from one unlabeled control beech and spruce tree are included to confirm no effect of weather conditions on δ 13 C during experimentation.Overall, the t-test for paired comparisons indicated no significant differences in δ 13 C shift between O 3 regimes within CO 2 and solid samples of labeled beech and spruce.

Table 1 .
Weather conditions and O 3 levels at the study site "Kranzberger Forst" during the growing season of 2006.Monthly sum of precipitation and average of daytime photosynthetic photon flux density (PPFD), relative air humidity (RH), air temperature (T air ), vapor pressure deficit (VPD) and soil moisture (±SE, n = 30 to 31).Ozone levels as monthly means ±SE (n = 30 to 31), AOT40 (i.e.accumulated O 3 concentrations above a threshold of 40 nl l −1 ) and SUM0 (i.e.daily sum of hourly O 3 concentrations).

Table 2 .
(A) CO 2 concentration (µl l −1 ) and (B) δ 13 C (‰) in canopy air of labeled beech and spruce trees under 1× and 2 × O 3 and one unlabeled control tree for each species.Data are presented for sun and shade crowns as means ±SE before (n = 12 h), during (n = 18 to 19 days) and after (n = 12 h) label exposure.

Table 4 .
δ 13 C (‰) of sun and shade leaves, phloem sugars, fine roots and soil respired CO 2 of beech and spruce before labeling.Data are shown as means ±SE (n = 3 trees).Lowercase letters denote significant differences between crown levels and soil depths (p ≤ 0.05).