• Here, the effects were assessed of local climate and canopy thinning on the carbon isotope composition (δ13C) signatures of different plant groups in the understorey of a beech (Fagus sylvatica) forest in southern Germany. The relationship between δ13C and environmental parameters, within different time integrals, was also studied.
• δ13C was analysed in different tissues of beech regeneration, and herbaceous and woody understorey vegetation, in thinned and untreated control stands differing in aspect and, hence, local climate, on three dates during the growing season.
• Generally, tissues were 13C-depleted on the north-east, compared with the south-west aspect. Thinning had variable effects on δ13C, depending on plant group and sampling date. δ13C in beech leaves and roots in control stands was mostly influenced by mean soil water potential in 4- and 8-wk integrals before sampling, respectively, and in leaves, additionally, by mean radiation in a 4-wk time integral. Shoot water potential and transpiration influenced foliar δ13C of beech whereas δ13C in the beech wood was modulated by soil temperature integrated over a 4-wk period before sampling. Above-ground tissues of woody and herbaceous plants were 13C-enriched in mid-summer; their δ13C was poorly related to environmental factors.
• δ13C of various tissues of beech regeneration appears to be indicative of recent environmental conditions within the forest understorey and, consequently, this easy-to-determine physiological parameter could be used widely to assess effects of silvicultural treatments.
The carbon isotope composition (δ13C) of plant tissues is an information-rich signal providing useful insights into different plant functions (Adams & Grierson, 2001). δ13C of plant tissues depends on δ13C of the ambient atmospheric CO2 and on the isotope fractionation within the plant (Farquhar et al., 1989). In C3 plants, the slower diffusion of the heavier 13C isotope, compared with 12C, from the atmosphere to the site of carboxylation, and the strong discrimination of Rubisco against 13C, are largely responsible for the depletion of plant material in 13C relative to the atmosphere. The δ13C composition of plant tissues is described by Farquhar et al. (1989) as follows:
δ13Cplant = δ13Catm − α − (b−α)Ci/Cα
where δ13C is expressed in ‰, α is the discrimination during diffusion (c. 4.4‰), b is the discrimination during carboxylation by Rubisco (c. 29‰), Ci is the CO2 concentration inside the stomatal cavities, and Cα is the atmospheric CO2 concentration. Due to their effects on Ci, intercepted radiation as well as atmospheric and soil water deficits modify δ13C in plant carbon (Leavitt & Long, 1986; Livingston & Spittlehouse, 1996; Korol et al., 1999). Stomatal closure as a consequence of water deficits reduces Ci, leading to an increase in δ13C (Lauteri et al., 1997). As a consequence, plant water potential can be related to δ13C in leaves and wood and in turn to the availability of water (Damesin et al., 1998; Warren et al., 2001). Alternatively, under light-limiting (but not water-limiting) conditions, photosynthesis and, consequently, Ci depends on radiation. Therefore, organic carbon in leaves from the shaded part of the crown of trees is more depleted in 13C compared with the sun-exposed crown (Leavitt & Long, 1986).
In addition, assessment of δ13C in different plant tissues provides information on the reaction of plants to environmental parameters within various scales. For example, δ13C of wood is considered to be a spatially and temporally integrated measure of leaf internal CO2 concentration of the growing season the wood was produced in and/or of one or more growing season(s) before wood formation (Schleser et al., 1999; Geßler et al., 2001). Furthermore, foliar δ13C is regarded as an indicator of water-use efficiency during longer periods of time (Farquhar et al., 1989) depending upon δ13C of first carbon storage pools remobilised during bud break, second carbon assimilated during leaf expansion and third assimilates produced later in the growing season (Damesin et al., 1998). Carbon isotope composition of phloem sap sampled during the growing season is assumed to be a strong guide to short-time changes in Ci : Cα during the actual growing season (Pate & Arthur, 1998; Adams & Grierson, 2001; Geßler et al., 2001; Keitel et al., 2003).
Owing to its dependence on water availability δ13C in plant organic matter can be a useful tool for assessing plant responses to climate change. In the near future summer droughts are expected to increase both in frequency and duration in Central Europe (Linder et al., 1996; Peñuelas, 1996; Lawlor, 1998; IPCC, 2001; Saxe et al., 2001). Natural regeneration of the drought sensitive European beech (Fagus sylvatica L.) – one of the most important deciduous tree species in central Europe – may be affected intensively by such climate alterations, especially since the area of distribution includes sites with shallow limestone-derived soils with low water storage capacity (e.g. Schwäbische Alb, Fränkische Alb, Schweizer Jura and French Jura).
In a changing climate, the application of forest management practices may have effects on the physiology of beech regeneration different from the ones observed today (Fotelli et al., 2002). At present, silvicultural techniques such as selective thinning of the mature canopy are routinely applied in beech forests (Tarp et al., 2000) in order to support natural regeneration (Dertz, 1996, Ministerium für Ländlichen Raum, Ernährung, Landwirtschaft und Forsten in Baden-Württemberg (ed.), 1997). Thinning is known to improve abiotic conditions (light intensity, nutrient availability, temperature; e.g. Aussenac, 2000; Mizunaga, 2000; Thibodeau et al., 2000) for beech seedlings, but due to its effects on interspecific relations within the forest understorey (Madsen, 1995; Lof, 2000) it may favour drought-tolerant species under altered climatic conditions. Since δ13C is a useful indicator for comparing the ecophysiology of functional groups in complex forest ecosystems (Huc et al., 1994; Brooks et al., 1997; Guehl et al., 1998; Bonal et al., 2000), it may also be used for identifying differences in water balance between beech regeneration and other – competing – functional groups of the understorey vegetation, induced by varying climate and thinning.
This study aimed at assessing the relationship of δ13C in different tissues of different functional groups of a beech forest understorey to environmental parameters, and at comparing these relationships between the functional groups. Moreover, we intended to validate δ13C signature as a time-integrating indicator of changes in environmental factors, as modulated by climate and thinning. The study was performed with naturally regenerated beech seedlings and other understorey species growing in beech forests on two opposite-exposed aspects (SW compared with NE), differing in local climate, and within each site, in a thinned and an unthinned stand. Our initial hypothesis was that δ13C would largely reflect variations in water availability, which is higher at the NE-facing site (Geßler et al., 2001) and probably also in the thinned stands (Breda et al., 1995). Moreover, we expected to identify differences in δ13C patterns among beech seedlings and neighbouring understorey plants, attributable to differences in their water status (Fotelli et al., 2001).
Materials and Methods
The experimental sites are located in southern Germany (longitude: 8°40′ E; latitude: 48°00′ N), about 100 km south-south-west from Stuttgart in a low mountain range (Schwäbische Alb, 740–760 m a.s.l.). Mean annual regional air temperature measured at a climate station of the DWD (Deutscher Wetterdienst, Offenbach, Germany), about 4 km from the experimental sites, is about 6.6°C, and mean temperature during the growing season (May to October) about 11.5°C. Mean annual precipitation is 856 mm with monthly maxima in June and July The sum of precipitation during the growing season (May to October) amounts to 410 mm. During the 2000 growing season the sum of precipitation was of about the same level, but mean air temperature was higher than long-term averages (Table 1).
Table 1. Climatic characterization of the experimental sites during the growing season (May–Sept) 2000
Site exposure Treatment
One experimental site faces to the north-east (NE) and the other to the south-west (SW). On each site a silvicultural treatment (thinning; T) plus an unthinned control (C) were established in March 1999. Climatic parameters were recorded at four climate stations c. 1.5 m above the forest floor. *Sum of precipitation (mm) refers to the amount of precipitation throughfall passing through the canopy and reaching the forest floor. **All means shown are averages over the entire growing season 2000 (May–Sept.); the values of Tsoil, RH and soil water potential are averages of mean daily values. The values of midday Tair are averages of mean midday temperatures, measured from 12 : 00 to 17 : 30 hours.
The experimental sites are located on two opposite-exposed sides (not more than 1000 m apart) of a single, narrow valley. One experimental site faces to the north-east (NE) and the other to the south-west (SW). Rainfall does not vary significantly across the valley (Geßler et al., 2001) but differences in the amount of throughfall, that is the precipitation reaching the forest floor, could be observed between the different plots (Table 1). The slope at the SW-exposed site is very steep (36–58%) and at the NE-exposed site precipitous (58–100%). Soil profiles are characterized as Terra fusca – Rendzina derived from limestone (Weissjura beta and gamma series) and are shallow on both sites, averaging to less than 0.20 m depth of topsoil before becoming dominated by parent rock interspersed with pockets of organic matter and mineral soil. The soil profile at the SW site is especially rocky, containing more than 40% (volumetric basis) rocks and stones (> 63 mm diameter) in the top 0.20 m of the soil and rising to 80% below 0.50 m. The soil at the NE site contains 15% rocks and stones in the uppermost 0.20 m of the soil and c. 30% below 0.50 m. Soil pH (H2O) is 5.7 in the surface organic layer and 7.5 at 0.60 m depth.
On both sites beech (Fagus sylvatica L.) is the dominant species making up > 90% of the total basal area of adult trees. The average age of the adult beech trees is 70–80 yr. The difference in aspects (NE, SW) produces a difference in radiation interception at the canopy level. The maximum daily radiation above the canopy of adult trees on the NE site amounts to 79% of the radiation available on the SW site in July, and 47% in October (Geßler et al., 2001). According to retrospective analyses of meteorological data, as well as of growth and water status of adult beech trees (Geßler et al., 2001), the SW-exposed site has permanently lower water availability and higher air temperatures than the NE-exposed site.
On each site, measurements were conducted in a stand subjected to a silvicultural treatment (thinning) and in an unthinned (control) stand. The selective felling was performed in March 1999. Total basal area of the mature beech trees was reduced from c. 27 m2 ha−1 on the NE site and 20 m2 ha−1 on the SW site (controls), to 10 m2 ha−1 on both sites to obtain a high variability of light conditions on the forest floor. Thinning decreased the leaf area index (LAI) from 5.16 (control) to 1.68 (thinned stand) on the NE site, and from 5.12 to 2.12 (thinned stand) on the SW site. One year after thinning, the density of understorey vegetation, other than natural regeneration of beech, increased in the thinned stands, compared to the controls (c. 25% on the NE site, and c. 8% on the SW site; T. Paul & A. Reif, pers. comm.).
Environmental conditions in the forest understorey during the growing season
Soil water potential was measured with pressure transducer tensiometers (UPT-4; UMS GmbH, Munich, Germany). On each aspect (NE and SW), soil water potential was measured in the control stands under the closed canopy at two soil depths (0.20 and 0.40 m) with two tensiometers per depth. Field observations showed that the vast majority of the root system of young beech seedlings, as well as of the other woody and herbaceous species, was restricted to soil depths between 0 and 0.20 m. However, soil water potential at 0.40 m depth was also recorded in order to test its possible importance as a determinant of variations in plant δ13C. The tensions were automatically recorded and stored as 4 h mean values (Delta T Logger; Delta T-Devices Ltd, Cambridge, UK). The deviation between the two tensiometers installed per soil depths and aspect amounted to – as an average during the growing season – less than 20% (relative to the more negative value).
Micrometeorological measurements were conducted in the centre of the control and of the thinned stand on both (NE and SW) aspects. Photosynthetically active radiation (PAR) in a wavelength range of 400–700 nm was measured with LI-190SA Quantum Sensors (LI-COR Inc., Lincoln, NB, USA) 1.5 m above ground. At the same ground level air temperature (Tair) and relative humidity of the air (RHair) were determined with Vaisala HMP45D-probes (Vaisala, Helsinki, Finland) shielded from direct radiation. Soil temperatures (Tsoil) at 0.05 and 0.10 m depth were measured with PT-100 thermometers (W-EYK 6; Heraeus GmbH, Kleinostheim, Germany). Precipitation, defined as the amount of throughfall passing through the canopy and reaching the forest floor, was collected with rain gauges, constructed according to DVWK-rules (DVWK, 1986), and quantified using a precipitation sensor (QMR 102; Vaisala, Helsinki, Finland). Tair, RHair, Tsoil and precipitation were determined every 30 s and calculated as daily averages (CR23X Micrologger, Campbell Scientific Inc., Logan, UT, USA). Detailed information about the meteorological measurements is given by Holst et al. (2000).
Three field campaigns were performed in 2000: in May at the beginning of the growing season, in July (mid-summer), and in September at the end of the growing season. During each campaign, six c. 2-yr-old, nonneighbouring beech seedlings with similar phenological characteristics (height, diameter, above-ground biomass) per site and silvicultural treatment were randomly selected for water potential and gas exchange measurements as well as for tissue sampling. In addition, all plants of the neighbouring vegetation within a radius of 0.15 m around each beech seedling were collected. Due to random selection, individual beech seedlings of each site and treatment were surrounded by different species. For that reason, the species of the adjacent vegetation were divided into two functional groups, that is woody plants (other than beech) and herbs (Table 2). Functional grouping of species – an approach crucial for understanding and modelling the functions of complex ecosystems (Körner, 1994; Chapin et al., 1996; Brooks et al., 1997) – was based on a combination of morphological and physiological parameters. The parameters, which were determined at the beginning of the growing season were: (a) total biomass per plant, which was significantly higher in woody plants (3–4.6 g) than in herbs (0.2–0.8 g), and (b) transpiration rates at 09 : 00 a.m., which were significantly lower in woody plants (1.1–1.4 mmol/m2 s−1) compared with herbaceous species (1.5–2.1 mmol/m2 s−1).
Table 2. Species that made up the two functional groups of vegetation surrounding the beech (Fagus sylvatica) seedlings studied
All neighboring vegetation was harvested during each measurement campaign from a cycle of c. 0.15 m around the selected beech seedling. The functional group ‘woody vegetation’ consisted of tree seedlings (other than beech) and shrub species. Mainly perennial herbs, with the exception of I. parviflora (annual) made up the group of ‘herbaceous vegetation’. On average, 1–2 plants of woody vegetation and 1–4 plants of herbaceous vegetation were present within each 0.15 m-radius-plot around the beech seedlings studied.
Acer platanoides L.
Anemone nemorosa L.
Acer pseudoplatanus L.
Asarum europeum L.
Cornus sanguinea L.
Fragaria vesca L.
Fraxinus excelsior L.
Galium odoratum (L.) Scop.
Prunus avium L.
Impatiens parviflora Dc.
Rubus fruticosus Agg.
Mycelis muralis (L.) Dumort
Rubus idaeus L.
Oxalis acetosela L.
Sambucus nigra L.
Paris quadrifolia L.
Sambucus racemosa L.
Viola hirta L.
Ulmus glabra Huds.
Viola mirabilis L.
Viola reichenbachiana Jordan Ex Boreau
Urtica dioica L.
After each campaign, the plant material of the beech seedlings was divided into leaves, stem and roots, whereas woody plants (other than beech) and herbs were separated into shoots (stems plus leaves) and roots.
Leaf transpiration rates
Leaf transpiration rates and stomatal conductance of the selected beech seedlings and the neighbouring woody and herbaceous species, grown within a 0.15-m-radius around each beech seedling, were determined during the three field campaigns (May, July, September) in the 2000 growing season. Transpiration of three fully expanded leaves per plant positioned at different heights within the plant canopy was measured with a steady state porometer (LI-1600; Li-Cor, Inc.) at 9 : 00. Per aspect treatment and measurement campaign transpiration of six beech seedlings, 6–12 woody plants and 6–24 herbs were determined.
Plant water potential of beech seedlings
Whole shoot water potential of the selected beech seedlings (six plants per aspect and silvicultural treatment) was measured at the bottom of the shoot directly after harvest applying the ‘Scholander Pressure Chamber’ technique (Scholander et al., 1965). Measurements were performed between 9 : 00 and 10 : 00 a.m.
Carbon isotope composition
Plant material from all species and plant parts was oven-dried for 3 d at 65°C. Subsequently, plant material was ground and homogenized with a ball mill into fine powder. Samples of 1–2 mg were transferred into tin capsules (Type A; Thermo Quest, Milan, Italy) and injected into an elemental analyser (NA 2500; CE Instruments, Milan, Italy) coupled to an isotope ratio mass spectrometer (Delta Plus; Finnigan MAT GmbH, Bremen, Germany) by a Conflo II interface (Finnigan MAT GmbH, Bremen Germany). The δ13C values were defined as: δ13C(‰) = [(Rsample : RPDB) − 1] × 1000, where Rsample and RPDB are the 13C : 12C ratios of the measured sample and of Pee Dee belemnite (PDB; Craig, 1957), respectively.
All statistical analyses were carried out using SPSS 10.05 (SPSS, Inc., Chicago, IL, USA). The effects of site (NE, SW) and treatment at each date were analysed with a one-way anova procedure, including Duncan's posthoc test. The same statistical procedure was used for assessing significant differences between functional groups.
Regression analyses were conducted with the linear regression procedure of SPSS. Not significant estimators were eliminated by backward selection of the independent factors. We tested the relationship between δ13C in different tissues of all plant groups studied (collected at the three dates during the 2000 growing season from both aspects and canopy densities), and various climatic/edaphic parameters (Soil temperature at 0.05 m and 0.10 m depth, air temperature, relative humidity, and PAR, measured at 1.5 m above ground, precipitation). Regression analyses between δ13C of different plant tissues and soil water potential in 0.20 and 0.40 m were restricted to the control plots of both aspects since no soil water potential data were available for the thinning treatments. For the regression analyses, daily means of all climatic and edaphic parameters were averaged over a 1, 2, 4 and 8 wk-time interval before harvest in order to identify the time integral within which δ13C of organic compounds is an indicative measure of these environmental influences (cp. Keitel et al., 2003). In addition, the relationship of δ13C in beech seedlings to their water potential and transpiration at the time of harvest was tested.
Climatic conditions in the forest understorey during the studied growing season
The assessment of soil water potential during the 2000 growing season revealed substantially lower values on the SW, compared with the NE aspect (Fig. 1). However, even on the SW site, no substantial drought was observed, with the soils close to field capacity throughout the monitoring period.
PAR was about three times higher in the SW-control stand compared with the NE-control stand (Fig. 1). This difference in radiation resulted in higher soil temperatures at 0.05 and 0.10 m depth in the control stand of the SW compared with the NE site (Table 1). Enhancement in radiation due to thinning on both aspects (Fig. 1) caused mean daily soil temperatures to be consistently higher in the thinned than in the control stands (Table 1).
Effect of site and thinning on δ13C composition
Beech seedlingsFigure 2 shows the patterns of δ13C in leaves, wood and roots of beech seedlings from all treatments during the growing season. During the entire growing season, leaves from the NE aspect were significantly depleted in 13C compared with those from the SW aspect. With one exception, thinning did not influence foliar δ13C. However in July δ13C increased significantly from −31.7 to −30.7‰ on the NE aspect as a consequence of thinning. Similar to leaves, wood was significantly 13C-enriched in the control stand of the SW aspect compared with the NE control. However, wood of beech seedlings in the NE-thinned stand was substantially enriched in 13C, particularly in July and September, as compared to the control stand of the same aspect. The effects of aspect and silvicultural treatment on δ13C in roots were similar to the ones observed for leaves: Roots on the SW aspect were less depleted in 13C and thinning had no significant effect on δ13C, either on the SW or on the NE aspect.
Woody vegetation The patterns of δ13C in shoots and roots of woody plants (other than beech) are presented in Fig. 3. In May and September, shoots of woody plants from the control stand on the SW aspect were significantly enriched in 13C compared with NE controls. This trend was also observed for the thinning treatments in July and September. Thinning had no effect on δ13C of the shoots on the SW aspect but resulted in 13C enrichment in May and September, and in depletion in July on the NE aspect. Root δ13C of woody vegetation showed aspect-dependent patterns comparable to the shoots, with 13C enrichment on the SW site in May (control and treatment) and September (control) compared with the NE aspect. On both aspects, thinning did not alter δ13C in the roots of this functional group.
Herbaceous vegetation In the shoots of herbaceous species there was a tendency towards 13C enrichment on the SW compared with the NE aspect (Fig. 4). A significant difference in δ13C between SW and NW aspects was observed in May for both, the control and the thinning treatments, in July only for the thinned stand and in September for the control. Thinning did not influence shoot δ13C on the SW aspect. At the NE exposed site, thinning resulted in 13C depleted shoots in July and in δ13C enriched shoots in September.
During the whole growing season, roots of herbaceous plants from the SW site were enriched in 13C compared with the NE site (Fig. 4). Thinning caused δ13C of roots to increase significantly on both aspects in September, whereas no thinning effect was observed in July In May an effect of the silvicultural treatment was only observed on the NE aspect, where thinning resulted in a significant decrease in δ13C.
Differences among functional groups Whereas root δ13C did not differ significantly between plant groups, there was a tendency towards increased 13C depletion of beech shoots in May (NET) and July (NEC, SWC and SWT) (Table 3). In September, however, the herbaceous species had most 13C-depleted above-ground tissues in the thinned stand of the NE and in the control stand of the SW aspect.
Table 3. Comparison of carbon isotope composition (δ13C) in the shoots of the different functional groups ‘beech (Fagus sylvatica)’, ‘woody species’ and ‘herbaceous species’
The beech seedlings were located in the understorey of two differently treated forest stands (control-C, thinned-T) at two opposing sides of a valley (NE vs. SW-aspect). Woody and herbaceous species were collected in a 0.15-m radius around each beech seedling. Plants were sampled at three different dates during the 2000 growing season (May, July and Sept.). One-way ANOVA was applied to compare mean values of different plant groups for each aspect, treatment and sampling date. Statistically significant differences (P < 0.05) are indicated by different letters. Values shown are means ± SE of whole shoot (leaves + stems) δ13C of 6 (beech) 6–12 (woody species) and 6–24 plants (herbs).
−30.72 ± 0.45 (a) n= 6
–30.42 ± 0.33 (a) n= 6
−28.71 ± 0.16 (a) n= 6
−28.25 ± 0.34 (a) n= 6
−31.49 ± 0.16 (a) n= 6
−29.31 ± 0.25 (a) n= 6
−29.01 ± 0.16 (a) n= 6
−29.64 ± 0.23 (a) n= 6
−31.68 ± 0.45 (a) n= 6
−29.62 ± 0.33 (b) n= 6
−28.81 ± 0.25 (b) n= 6
−28.58 ± 0.51 (a) n= 6
−30.63 ± 0.66 (a) n= 6
−28.70 ± 0.41 (b) n= 6
−27.90 ± 0.34 (a) n= 8
−27.72 ± 0.29 (a) n= 9
−28.86 ± 0.13 (b) n= 6
−29.99 ± 0.37 (a) n= 7
−28.02 ± 0.25 (b) n= 6
−28.51 ± 0.31 (b) n= 6
−31.89 ± 0.34 (b) n= 7
−30.26 ± 0.30 (b) n= 11
− 28.75 ± 0.85 (a) n= 8
−27.59 ± 0.58 (a) n= 12
−29.78 ± 0.19 (a) n= 9
−29.36 ± 0.18 (b) n= 24
−27.99 ± 0.25 (a) n= 15
−28.13 ± 0.32 (a) n= 14
−29.04 ± 0.38 (b) n= 9
−30.86 ± 0.27 (a) n= 18
−28.27 ± 0.39 (b) n= 7
−28.24 ± 0.15 (b) n= 10
−32.83 ± 0.37 (a) n= 6
−31.04 ± 0.25 (a) n= 12
−30.17 ± 0.25 (a) n= 8
−29.23 ± 0.55 (a) n= 6
Variation of transpiration rates
The transpiration rates of all plant groups studied (beech seedlings, woody plants and herbs) are shown in Fig. 5. During the whole growing season transpiration rates of beech seedlings were significantly lower in the NE control stand compared with the other stands. Between May and July transpiration of beech increased significantly independent on aspect or treatment. Until the end of the growing season transpiration rates of beech increased slightly on the NE control, whereas they remained constant on the other plots. Transpiration of the individuals of the functional group ‘woody species’ did not differ between aspect and treatment in May. In July, a strong increase in transpiration of woody plants from the SW thinning treatment resulted in significant differences between this plot and the control and thinned stand on the NE aspect. In September the woody species on the NE control plot transpired significantly less than the plants from the other stands.
Transpiration of herbaceous species did not differ between aspect and treatment in May, similar to woody species. In July, transpiration increased significantly on the SW aspect whereas it remained constant on the NE aspect resulting in maximum values on the SW control plot. With the exception of the control stand on the NE aspect transpiration rates decreased significantly until September leading to minimum values in the thinning treatments on both aspects.
Relationship between δ13C and climatic or physiological parameters
Beech seedlings The highest regression coefficient for a model describing the variation of foliar δ13C in beech was obtained with soil water potential at 0.20 m depth and PAR as independent variables (Table 4). Averaged over a 4-wk time integral before plant sampling these parameters together explained 95% of the variation in foliar δ13C on the control plots of both aspects (P = 0.01). When the two variables were averaged over 1, 2 or 8 wk before sampling R2 was lower (0.41, 0.61 and 0.62, respectively; Table 4). Soil water potential alone calculated as the mean value of a 4-wk time-integral before sampling explained 87% of the variation in foliar δ13C of beech seedlings from the controls. Soil water potential at 0.40 m depth, did not account significantly for the variation in foliar δ13C. PAR was a significant determinant of foliar δ13C only for beech seedlings grown in the control stands (R2 = 0.61, P = 0.045; Fig. 6b), which received 5–15 times less radiation compared with the thinned stands. In addition, shoot water potential measured immediately before sampling accounted partially for the variation in foliar δ13 (R2 = 0.56, P < 0.015; Fig. 7a). When mean values from both aspects and treatments were included as independent variables into a regression model, leaf transpiration did not influence δ13C in foliage of beech significantly (R2 = 0.16; P = 0.18; Fig. 7b). However, when the aspects were considered separately, leaf transpiration explained 78 and 36% of the variation in δ13C on the SW and the NE exposed site. On both aspects, a decrease in transpiration resulted in an increase in 13C enrichment. When stomatal conductance was used as independent variable, patterns comparable to transpiration were observed (Fig. 7c).
Table 4. Regression models describing the relationship between carbon isotope composition (δ13C) of different tissues from different functional groups and mean values of environmental parameters in 1-, 2-, 4- and 8-integrals before the sampling date
Determinants of the model
Time integral before sampling
The levels of significance (P) are given only for the statistically significant models (P < 0.05). Where it is not explicitly mentioned results from both aspects and treatments were included in the calculations. For beech leaves all determinants tested are shown, for all other plant groups and tissues only variables with R2 > 0.15 are listed.
Soil water potential (0.20 m depth) and PAR (only controls)
Soil water potential (0.20 m depth) (only controls)
Soil water potential (0.40 m depth) (only controls)
PAR (only controls)
Tsoil (0.05 m depth)
Soil water potential (0.20 m depth) (only controls)
Tsoil (0.05 m depth)
Tsoil (0.05 m depth)
Soil water potential (0.20 m depth) (only controls)
Soil water potential (0.20 m depth) (only controls)
Soil water pot. (0.20 m depth) (only controls)
PAR (only controls)
Soil water potential (0.20 m depth) (only controls)
PAR (only controls)
Soil water potential averaged over an 8-wk time interval before sampling, accounted for 97% of the variation in δ13C of roots in the control stands (P < 0.001; Fig. 6c). Mean soil water potential (at 0.2 m depth) in a time integral of 1, 2 or 4 wk before sampling produced lower regression coefficients (Table 4).
In contrast to leaves and roots, soil water potential did not significantly affect δ13C of wood. Furthermore, no significant influence of PAR was observed. δ13C in wood was mainly controlled by average soil temperature in a 4-wk integral before sampling (R2 = 0.51, P = 0.006; Table 4; Fig. 6d).
Woody vegetation Soil water potential was the only environmental parameter tested that produced a significant regression model (Table 4) describing the variation in δ13C of shoots. No significant regression coefficient was obtained for the regression analysis between root or shoot δ13C and transpiration rate (for shoots: R2 = 0.06 P = 0.44; for roots: R2 = 0.09 P = 0.35).
Herbaceous vegetation Significant regression coefficients (shoots: R2 = 0.46, P = 0.048; roots: R2 = 0.76, P = 0.02) were produced for the regression models, which included PAR (averaged over the 4-wk-period before sampling) as determinant for the variation of δ13C in the control stands (Fig. 8; Table 4). The other environmental (soil water potential, soil or air temperature, precipitation) and physiological (transpiration) parameters tested did not significantly account for δ13C variations in roots or shoots.
The present study aimed to assess the effects of different meso-climatic conditions and of thinning on the C-isotope composition in different tissues of young beech seedlings and two additional functional plant groups in the understorey of a beech forest in Southern Germany. Our initial hypothesis was that δ13C of the studied plants reflects the general aspect- (NE compared with SW) and treatment-specific differences in water and light availability (Geßler et al., 2001). Furthemore, we expected to identify differences in the factors that mainly affected δ13C, among the plant groups studied.
Above-ground tissues and roots of beech seedlings and other understorey species were constantly 13C-enriched on the SW, compared with the NE aspect (Figs 2, 3 and 4). Regression analyses revealed that soil water potential – as a measure of soil water availability – at 0.20 m depth was the main single environmental factor responsible for the significant 13C-enrichment in leaves (R2 = 0.87, P = 0.005; Table 4) and roots (R2 = 0.97; P < 0.001) of beech and in shoots of woody plants (R2 = 0.47; P = 0.048) in the control stand of the SW, compared with the NE aspect. It is a well-established finding that water shortage causes stomatal closure, thus, resulting in lower discrimination against 13C (Farquhar et al., 1989; Lauteri et al., 1997; Arndt et al., 2000; Fotelli et al., 2001; Warren et al., 2001). Since there were also significant lower transpiration rates in beech seedlings grown in the NE compared with the SW control stand (Fig. 5), it is concluded that differences in the water status of understorey species between aspects are mainly responsible for the observed patterns in δ13C of different tissues. This hypothesis is supported by the observation that decreasing plant water potential from the NE to the SW aspect accounted for 56% of the concurrent increase in δ13C in beech foliage (Fig. 7).
Soil water potential and, hence, water availability in deeper soil layers (40 cm) had no significant effect on δ13C, since the rooting zone of beech seedlings and other understorey species was not below 0.20 m.
Within a comparable range of transpiration rates leaves of beech seedlings were significantly 13C enriched on the SW aspect (Fig. 7). The observed differences in δ13C patterns among aspects could not be attributed to differences in stomatal conductance at comparable transpiration rates: regression analysis between stomatal conductance and foliar δ13C showed a relationship comparable with transpiration. As a consequence, δ13C could not be used as a general indicator for instantaneous water use or stomatal reactions for all aspects and treatments. Bearing in mind that a mixture of different carbon pools with different turn-over times contributes to δ13C of bulk leaf material (Adams & Grierson, 2001) it is obvious that its dependency on parameters measured immediately before sampling may be confounded by a whole set of other physiological and environmental variables and their influence over different integrals of time.
Warren et al. (2001) found isotopically heavier tissues in conifers grown in thinned stands, compared with plants in closed stands, and attributed this response to increased radiation interception after thinning which mediates higher photosynthesis and lower discrimination against 13C (Farquhar et al., 1989). Also in the present study, radiation modulated the effect of water availability on the δ13C. Foliar δ13C signatures of beech seedlings on the control plots reflected, to a great extent, the combined effects of soil water availability and light intensity integrated over a 1-month period (R2 = 0.95, P = 0.01; Table 4). PAR alone accounted for 61% of the variation in foliar δ13C in beech and was the main determinant of δ13C in shoots and roots of herbs grown in the control stands (Table 4) where apparently light intensity was a limiting factor (Fig. 1). This supports the hypothesis that light intensity modifies the effect of water availability on δ13C at the site studied when the former is a limiting factor for CO2 assimilation (Geßler et al., 2001). The low foliar δ13C signatures in the NE-control stand (Fig. 2) may, thus, be a result of the restricted light intensities (Fig. 1, NEC), according to the well-documented effect of PAR on Ci and, thus, δ13C (Leavitt & Long, 1986; Farquhar et al., 1989; Broadmeadow & Griffiths, 1993; Livingston & Spittlehouse, 1996). On the other hand the possibility can not be excluded that light intensity increases leaf temperature and, as a consequence, leaf-to-air water vapour pressure difference causing stomatal closure. A decrease in δ13C with increasing radiation could, thus, also be a result of stomatal reaction. The additional assessment of δ18O in plant organic matter and a correlation between δ13C and δ18O as proposed by Scheidegger et al. (2000) and Keitel et al. (2003) allows a differentiation between effects of stomatal reaction and/or photosynthesis on δ13C, and should be applied in future studies.
Different from the observations of Warren et al. (2001), no consistent trend towards an increase in δ13C in the thinned stands could be observed in the present study, although radiation increased intensively as a consequence of thinning. On the SW aspect there was almost no difference in δ13C between treatments. The effect of thinning on δ13C of plant tissues from the NE aspect was not constant among sampling dates and species with both, increase and decrease of δ13C as a consequence of treatment. As described by Buchmann et al. (2002), differences in plant δ13C in the forest understorey may be due to differences in δ13C of CO2 assimilated. As a result of microbial and root respiration δ13C of CO2 near the forest floor is typically more negative than values of above-canopy CO2 (Buchmann et al., 1998). The variable effects of thinning on δ13C of understorey species growing on the NE aspect may, thus, be a result of variable influence of the silvicultural treatment on the intensity of soil respiration during the growing season and, in addition, of generally high spatial and seasonal variation of CO2 efflux from the soil (Buchmann, 2000). Still, we lack data on δ13C of the CO2 assimilated for verifying this hypothesis.
The finding that foliar δ13C of beech seedlings is an indicator of environmental conditions within a month integral before sampling contradicts the results of other studies with adult trees. In deciduous trees such as European beech, foliage that develops in spring is formed mainly from stored carbon and nutrients (Kozlowski & Pallardy, 1997) – newly assimilated carbon is mixed with the previously stored carbon to form the new leaves. It has been reported that foliar δ13C largely reflects conditions of the previous growing season(s) (Damesin et al., 1998; Geßler et al., 2001; Le Roux-Swarthout et al., 2001). Still, the close relationship between foliar δ13C and water potential of beech seedlings in a greenhouse-experiment (Fotelli et al., 2001), as well the results of the present study, support that foliar δ13C signature of young beech seedlings reflects – to a great extent – the isotopic composition of recently produced assimilates.
Not only δ13C of beech foliage but also of roots proved to be measure for soil water availability: averaged over a 2-month period soil water potential accounted for c. 97% of δ13C variation in the control stands (Table 4). This finding supports the assumption that δ13C in the roots is – with a considerable time lag – mainly defined by the carbon isotope composition of carbon exported from the leaves. This observation is further strengthened by the results of Högberg et al. (2002) who demonstrated that there is a significant flux of recently fixed carbohydrates from the leaves to the roots, which greatly affects root respiration, and thus carbon balance of roots. Furthermore, Ekblad & Högberg (2001) showed that newly produced photo-assimilates significantly accounted for the variation in root respiration and in its δ13C composition.
In contrast to leaves and roots, soil water availability had no direct effect on the δ13C signature of beech wood. The carbon isotope composition of wood is an integrator of environmental conditions over longer periods of time (Livingston & Spittlehouse, 1996; Pate & Arthur, 1998; Geßler et al., 2001; Porté & Loustau, 2001). Therefore, the positive relationship between δ13C signature of wood and soil temperature (Fig. 6d Table 4) most probably reflects the differences in the light environment of aspects and treatments, which induce long-term differences in soil temperature (Table 1). In agreement with our results, Dupouey et al. (1993) found an increase in δ13C of wood cellulose of European beech with increasing temperature.
The observed relationships between δ13C of herbs or woody plants, and environmental factors were weaker, compared with beech seedlings. This may be due to the diversity of species forming these functional groups, and/or the complexity of the conditions occurring close to the forest ground. Since species within both groups differ in life form and phenology, the time when the major part of carbon-containing material is laid down may vary significantly among members of each functional group. Thus, effects of environmental parameters on δ13C may be concealed.
Tissues of above-ground parts from herbs and woody plants had a tendency to 13C-enrichment – compared with beech seedlings – in May and, especially, in July (Table 3). This difference may be attributed to the higher root biomass and, hence, probably better water access of beech seedlings compared with plants of the other groups. However, the conditions in the field seem to be more complex, as herbs were the plants with highest 13C depletion in autumn. Since most of the herbaceous species assessed are vernal (cp. Oberdorfer, 1983) they show signs of senescence in mid-September before visible yellowing of leaves occurs in beech (Kirchgäßner, 2001). Decreasing rates of photosynthesis (Proietti, 1998), and, thus, increasing Ci in senescing leaves might be the reason for the observed increase in 13C depletion in shoots of herbaceous species at the end of the growing season (Fig. 4).
In beech seedlings δ13C of leaf carbon was a measure for the instantaneous water balance, indicated by the proxy variables shoot water potential and transpiration. However, since the carbon pool of plant tissues does not only consist of directly assimilated carbon but, in addition, of structural carbon with longer turn-over times (Adams & Grierson, 2001) generally higher correlation was obtained when soil water-potential – as a measure for water supply – integrated over 4- or 8-wk-periods was used for regression analyses. In addition, the light environment as modified by stand density significantly influenced δ13C in beech and herbs. Above-ground tissues of woody and herbaceous understorey plants had a tendency to 13C-enrichment in mid-summer, compared with beech seedlings; a response that can probably be attributed to the larger root biomass and, thus, to better water access of the latter.
Owing to its sensitivity towards variations in recent environmental conditions δ13C in different tissues of beech is a promising tool for assessing responses of beech regeneration to silvicultural treatments under conditions of Global Climate Change.
This study was part of the project SFB 433, funded by DFG (Deutsche Forschungsgemeinschaft). We thank P. Escher for technical assistance during the measurements of isotopes composition. Dr Sabine Augustin and Prof. Ernst E. Hildebrand, Institute of Soil Science and Forest Nutrition, are acknowledged for providing the data on soil water potential. MN Fotelli thanks the DAAD (Deutscher Akademischer Austauschdienst) and LGFG (Landesgraduiertenförderungsgesetz) for financial support during the present study.