Study site and experimental setup
This study was conducted at the forest meteorological research site Hartheim of the Meteorological Institute of Freiburg, a 38-year-old pine plantation in the southern upper Rhine valley, Germany (47°56′N, 7°36′E, elevation 201 m). The forest was mostly planted with Scots pine (P. sylvestris L.), with only a few patches of Austrian pine (Pinus nigra L.). All measurements were made in P. sylvestris plots. Most of the tree foliage was between 11 m above ground level and the top of the canopy at ca. 15 m. Plant area index (PAI) was 1.9 m2m−2, owing to recent thinning (Schindler, Türk & Mayer 2006). A detailed description of the experimental site and its management was given by Mayer et al. (2000) and Schindler et al. (2006).
The measurement campaign took place from 6 June 2005, 1200 h (all times are expressed as Central European time, with a 0 to 24 h notation) to 10 June 2005, 0600 h. Samples and measurements were taken every 6 h throughout this period, at 1200, 1800, 2400 and 0600 h. Three adjacent dominant or co-dominant individuals of P. sylvestris were selected within reach of a truck-mounted 25 m hydraulic lift, and were used as replicates for tree-based measurements.
Meteorological data [air temperature, relative humidity, photosynthetically active radiation, precipitation (PAR) and wind speed] were determined continuously at a measurement tower. Air temperature, humidity and vapour pressure deficit (VPD) were determined using a psychrometer according to Frankenberger (Mayer & Gietl 1976) at 2, 12 and 19 m height. Wind speed was measured with cup anemometers at 2, 6, 12, 15 and 19 m height. PAR was determined with a Li-190SZ sensor (Li-Cor, Inc., Lincoln, NE, USA) at 16 m; above-canopy precipitation was measured at 29 m. Volumetric soil moisture (TDR probes) was determined at 30 cm depth, and soil temperature (thermocouples) was measured at 3 cm depth. Data were recorded every 30 s and averaged over 30 min periods.
Collection of plant material
At each measurement time, twigs were sampled from the sunlit upper third of the canopy and were used for collection of needles, xylem sap and twig phloem. Needles were separated into current growth season (N needles) and previous growth season (N-1) needles, and were used for the extraction of needle water and needle water-soluble organic matter. In addition, trunk bark samples were taken from three heights below the live crown: high-, mid- and low-stem, at 10.0, 6.0 and 1.5 m from the ground, respectively, and were used for phloem collection. Bark samples (ca. 150 mg) were taken from the twig bark with a scalpel, and from the trunk bark with a core borer (13 mm diameter).
Brandes et al. (2006) observed no intracanopy gradient of δ13C at the same stand, and found a strong and significant correlation between canopy-integrated stomatal conductance and leaf-level stomatal conductance that was determined in the upper part of the canopy. We conclude that the twigs sampled here adequately represent the environmental conditions of the canopy, as a prerequisite for comparing leaf level measurements, such as leaf δ18O, with parameters that integrate a canopy signal, such as phloem δ18O.
Leaf, xylem and atmospheric water
N and N-1 needles were transferred in glass tubes and immediately frozen in liquid N2. Bulk leaf water was extracted from the needles by cryogenic vacuum distillation: the frozen tubes containing the needles were placed in a 80 °C water bath, connected to a vacuum system (ca. 4.10−2 mbar) including water traps that were cooled with liquid N2. The water was then transferred into 2 mL vials and kept frozen until δ18O analysis (see further discussion). The dried needles were ground and analysed for bulk δ13C (see further discussion).
Xylem sap samples were taken according to the method presented in Keitel et al. (2006). One centimetre of bark was removed from the cut end of a twig. A polyethylene tube was then fitted onto the shoot at one end, and equipped with a hypodermic needle at the other. The needle was inserted in a 2 mL vial with an airtight seal, together with another needle that was connected to a hand vacuum pump. A gentle vacuum was applied, while the needles were subsequently cut off the end of the twig to facilitate xylem sap collection. The xylem sap sampled in the vials was immediately frozen in liquid nitrogen and stored at −20 °C until analysis (see further discussion).
Atmospheric water vapour was collected by cryogenic condensation. Air was pumped at 35 L h−1 for 2 h (centred on target measurement time) from four locations at 13 m height (mid-canopy) through a trap filled with ethanol and liquid N2 (ca. −70 °C). The collected water was immediately transferred into 2 mL vials and kept frozen until δ18O analysis (see further discussion).
Organic matter in leaf and phloem
Bark samples were washed with demineralized water immediately after collection and placed in 6 mL vials containing 2 mL of demineralized water. The samples were left to exude for 5 h as described by Rennenberg, Schneider & Weber (1996). This method was identified as the most suitable technique for assessing δ18O in phloem sap (Gessler, Rennenberg & Keitel 2004). Sucrose is expected to account for more than 90% of phloem sap in trees (e.g. Pate et al. 1998). No direct exchange between phloem-transported sucrose and phloem water or water from the exudation solution will occur, because the sucrose molecule contains no carbonyl bounds. We therefore assume no change in the δ18O of phloem sap as a consequence of the exudation procedure. Contamination of phloem exudates with cellular constituents was shown to be negligible under the experimental conditions applied (Schneider et al. 1996). A volume of 75 to 100 µL of phloem exudation solution was transferred in silver capsules (IVA Analysentechnik; Meerbusch, Germany), and water was evaporated at 60 °C in an oven before isotope analysis (see later discussion). We tested the issue of sealing the dried phloem samples under argon immediately after removing them from the drying oven on a separate set of samples, as addressed by Cernusak et al. (2003a). We indeed found that the δ18O in phloem sap tended to be lower if the samples were not sealed under argon immediately, both sealed and unsealed data sets being well linearly correlated (r = 0.980, P < 0.0001, n = 24). We applied the following correction to our unsealed samples: δ18Ocorrected = 1.1 × δ18Ounsealed − 2.9.
The N and N-1 needles, frozen in liquid N2 immediately after harvest, were microwaved to stop physiological activity and then freeze-dried. The samples were homogenized with mortar and pestle in liquid N2. Water-soluble organic matter was extracted as follows: 50 mg of homogenized sample were incubated for 60 min at 5 °C in 1 mL demineralized water, heated at 100 °C for 1 min to precipitate proteins. The samples were cooled on ice and then centrifuged (12 000 g at 5 °C for 5 min). The 75 to 100 µL of supernatant was transferred in silver capsules, and the water was evaporated at 60 °C in an oven before isotope analysis (see later discussion).
Mass spectrometry measurements
The determination of δ18O in xylem, bulk needle and atmospheric water vapour samples was established according to Gehre et al. (2004), using a TC/EA (high temperature conversion/elemental analyser, ThermoFinnigan, Bremen, Germany) coupled with a DeltaPlus XP mass spectrometer via a ConFlo III interface (Werner, Bruch & Brand 1999). The precision was < 0.15‰. In bulk needle organic matter and needle water-soluble organic matter, δ18O was determined as follows: 75 to 100 µL phloem exudation solution or of water-soluble organic matter extracted from needles was transferred in silver capsules (IVA Analysentechnik) and water was evaporated at 60 °C in an oven. For bulk leaf material, 0.5 mg of homogenized dried sample was transferred in silver capsules. The samples, which contained ca. 300 µg organic O on average, were combusted in a TC/EA coupled to an isotope ratio mass spectrometer (DeltaPlus XL; Finnigan MAT GmbH, Bremen, Germany). The 18O/16O oxygen stable isotope ratio (R = 18O/16O) is expressed using small delta notation in parts per thousand, relative to the international Vienna Standard Mean Ocean Water(VSMOW) standard, as , where Rsampleand Rstandard refer to the isotope ratio of the substance of interest and of the standard, respectively.
Leaf gas exchange
Leaf gas exchange was measured in the upper canopy by inserting small twigs (with their needles attached) into a conifer leaf chamber connected to a portable gas exchange system (GFS3000; Heinz Walz GmbH, Effeltrich, Germany). The measurements were conducted under ambient light and temperature conditions. Twigs with both N and N-1 needles attached were placed in the chamber, with N-1 needles representing on average ca. 65% of the total needle area in the chamber. Net CO2 and H2O exchange rates were measured, and stomatal conductance (gs) was subsequently calculated according to von Caemmerer & Farquhar (1981). Leaf water content was measured for both N and N-1 needles.
Separate values of gs were calculated for N and N-1 needles, based on (1) the overall gs of N and N-1 needles that were inserted in the gas exchange chamber together; (2) a summertime value of 0.53 for the ratio of gs of N-1 needles to gs of N needles, based on the measurements of Beadle et al. (1985) on upper-canopy needles of P. sylvestris; and (3) the relative leaf area of each type of needle in the chamber. Separate transpiration rates were calculated for both needle cohorts, based on these gs values, and assuming comparable relative humidity (RH) of ambient air and comparable leaf temperature for N and N-1 needles (gs should not influence leaf temperature or internal water vapour pressure, because pine needles are assumed to be strongly coupled with the environment; Barbour, Walcroft & Farquhar 2002). According to Beadle et al. (1985), we used a summertime value of 0.61 for the ratio of assimilation rate of N-1 needles to that of N needles.
In order to test our assumptions for the ratios of transpiration rate, assimilation rate and stomatal conductance between N-1 and N needles, we performed additional experiments with approximately 4-year-old saplings collected at the field site. Saplings were collected from the site during winter and grown in a greenhouse in the natural soil from the stand. Approximately 3 months after bud break, we performed gas exchange measurements (n = 3 trees; 3 twigs per tree) on N and N-1 needles separately, at midday (light intensity of approximately 500 µmol m−2 s−1, air temperature of 24 °C). We observed the following mean ratios for gas exchange parameters of N-1 to N needles: stomatal conductance, 0.51; transpiration, 0.55; and photosynthesis, 0.64. These values are comparable with the ratios given by Beadle et al. (1985), which we used for calculating stomatal conductance, transpiration and photosynthesis of the two different needle cohorts.
The projected area of the N and N-1 needles that were inserted into the chamber was determined with a leaf area meter (ΔT Devices, Cambridge, UK). Mean needle length was determined from a subsample of the needles. Three-dimensional leaf area was calculated assuming the needle has the shape of a half cylinder, and was used as a basis for the gas exchange values (Luoma 1997; Haberer 2002).
Leaf water Δ18O
The oxygen isotope ratio of oxygen in the substance of interest can be expressed as enrichement above source water, using an upper case delta notation (Δ18O) in parts per thousand:
In the present study, xylem water was considered to be source water.
The measured enrichment of bulk needle leaf water above source water is denoted Δ18OB. According to Farquhar & Gan (2003), the 18O enrichment of bulk leaf water is given by
where ϕx, ϕv andϕL are the proportions of total water associated with the longitudinal xylem, the veinlets and the lamina mesophyll, respectively, and Δ18Ox, Δ18Ov and Δ18OL denote the evaporative enrichment of xylem, veinlet and lamina leaf water above source water, respectively. Because the models applied here (which include the Péclet effect) calculate lamina leaf water enrichment, and because lamina leaf water is also the reaction water in which assimilates are produced, it is necessary to estimate Δ18OL from measured Δ18OB. According to Farquhar & Gan (2003), we assumed the water volume in the veinlets to be negligible and proposed the following procedure to estimate ϕx. In their careful and extensive study on Scots pine needle anatomy, Lin, Jach & Ceulemans (2001) showed that the contribution of xylem to the cross-sectional area of current year needles (sampled in October) was 2.2%. We prepared cross-sectional cuttings of N and N-1 needles and compared the contribution of xylem area to the cross-sectional area by visual inspection under the microscope (50-fold magnification). There was no obvious difference in the relative xylem area between the two needles classes. We therefore assumed the value of 2.2% contribution of xylem to the cross-sectional area to be valid for N and N-1 needles. For 10 needles of each cohort, we then calculated the cross-sectional area, based on needle thickness and assuming the needle to be a half cylinder. The estimated cross-sectional xylem area values were multiplied by needle length to obtain the xylem volume of a needle, which was assumed to equal the volume of vascular water. In addition, the total water content of the needles was estimated from the fresh to dry weight ratio. Xylem volume contributed 4.5 and 3.1% to total water volume of N-1 and N needles, respectively. Based on these ϕx values, we calculated Δ18OL from Δ18OB, assuming needle xylem water not to be 18O enriched compared with the twig xylem water that we sampled. This assumption may introduce a slight error as xylem water gets a little bit enriched as it moves along the needle (Farquhar & Gan 2003; Gan et al. 2003).
where ε+ is the equilibrium fractionation between liquid water and vapour, εk accounts for the kinetic fractionation during the diffusion of water vapour from the leaf to the atmosphere, Δ18Ov is the isotopic difference of atmospheric water vapour compared with source water, ea and ei represent the water vapour pressure in the atmosphere and the leaf intercellular air space, respectively.
where rs and rb represent the resistance to water vapour of leaf stomata (based on measured gs values) and boundary layer, respectively, and their respective associated fractionation factor (32 and 21‰, Cappa et al. 2003). Boundary layer resistance was calculated from the wind speed in the canopy measured at 12 m height and from mean needle diameter, according to Jones (1992). Calculated boundary layer resistance was generally comprised between 0.2 and 0.8 m2 s mol−1 and was higher only at one time point (10 June, 0600 h: 4 m2 s mol−1).
Average lamina mesophyll water is, however, expected to be less enriched than the water at the evaporative sites, resulting in an isotopic gradient between the leaf vein and the evaporative sites. The Péclet effect is the net ratio of (1) the unenriching convection of water to the leaf evaporative sites via the transpiration stream to (2) the effect of the 18O-enriching diffusion of water away from the sites of evaporation. Taking into account this effect (Farquhar & Lloyd 1993), the steady state enrichment of mean lamina mesophyll water above source water (Δ18OLs) can be expressed as:
where the Péclet number is ℘ = LE/CD, calculated from the scaled effective path length L (m), evaporation rate E (mol m−2 s−1), molar concentration of water C (55.5 103 mol m−3) and diffusivity of H218O in water D (2.66 10−9 m2 s−1). The scaled effective path length was estimated by fitting the non-steady state model to the measured Δ18OL under expected steady state conditions that typically occur in the end of the afternoon.
Under non-steady state conditions, the enrichment of mean lamina mesophyll water above source water (Δ18OLn) can be calculated following Farquhar & Cernusak (2005):
where α+ = 1 + ε+ and αk = 1+εk; W is the lamina leaf water concentration (mol m−2), t is time (s), g is the total conductance to water vapour of stomata and boundary layer , and wi is the mole fraction ofwater vapour in the leaf intercellular air spaces (mol mol−1). W was estimated based on bulk leaf water content per unit area corrected for the proportion of vascular water (as described previously). The Péclet number used in the non-steady state model was estimated with the steady state model. Because the Δ18OLn term occurs on both sides of Eqn 7, it is simpler to solve the equation iteratively.
The non-steady state model requires initial values for Δ18OLn and W for a time point (t0 − 1) preceding the first observation. We estimated this initial value for Δ18OLn based on the observed diel patterns of measured Δ18OL during the entire measurement period. Our first measurement started at 1200 h (t0) on 6 June 2005. From 3 d of measurements (7, 8 and 9 June), we calculated the fraction between Δ18OL at 1200 h and the preceding time point (0600 h). On average, the 0600 h value amounted to 67 and 53% of the 1200 h values for N and N-1 needles, respectively. Based on these fractions, we estimated the initial Δ18OLn for t0 − 1 from Δ18OL at t. The initial value of W was assumed to equal the one measured at 0600 h on the second day of measurement.
To calculate the mean daytime oxygen isotope enrichment above xylem water (Δ18O), the Δ18O values of each daytime measurement time was weighted by the corresponding CO2 assimilation rate measured (A, mol m−2 s−1), according to Cernusak et al. (2005):
where the numerator is the daily integral of the product of A and Δ18O (‰ mol m−2), and the denominator is the daily integral of photosynthesis (mol m−2).