• In the present study NH 3 and NO 2 exchange between the atmosphere and needles of adult spruce ( Picea abies ) trees at a field site (‘Höglwald’) exposed to high loads of atmospheric nitrogen was assessed.
• Twigs were fumigated with different NH 3 (C NH3 ) or NO 2 (C NO2 ) concentrations using the dynamic chamber technique. Beside fluxes of NH 3 (J NH3 ) and NO 2 (J NO2 ), transpiration (J H2O ), leaf conductance for water vapour (g H2O ), photosynthetic activity (J co2 ), photosynthetic photon fluence rate (PPFR), air temperature (T) and relative air humidity (RH) were determined.
• Both fluxes, J NH3 and J NO2 , depended linearly on C NH3 and C NO2 in concentration ranges representative for the field site and g H2O as a measure of stomatal aperture. For both trace gases compensation points could be determined amounting to 2.5 nmol mol −1 for NH 3 and to 1.7 nmol mol −1 for NO 2 .
• The fluxes of NH 3 and NO 2 could not be explained exclusively by exchange through the stomata. In both NH 3 and NO 2 fumigation experiments additional deposition onto the needle surface was observed and increased with increasing C NH3 and C NO2 . 15 N[NH 3 ] fumigation experiments with adult spruce trees confirmed the results of gas exchange measurements and revealed that NH 3 -N deposited to spruce needles is subjected to long distance transport within the plant, supplying the plant with additional nitrogen from the atmosphere.
NH3 and NO2 exchange between the atmosphere and plants can occur as a bi-directional flux, the direction of which is determined by the NH3 and NO2 concentration gradient between leaf interior and ambient air (Husted & Schjoerring, 1996; Thoene et al., 1996; Hereid & Monson, 2001). Hence, depending on the plant internal and the ambient concentration, plants can act as both, sources and sinks for NH3 and NO2.
With the present study we intended to assess – isolated from other ecosystem fluxes – the mechanisms of NH3 and NO2 exchange between twigs of adult spruce subjected to chronically high atmospheric N input and the ambient atmosphere using the dynamic chamber technique. We wanted to test the hypothesis, that mainly the concentration of trace gas and stomatal conductance determines NH3 and NO2 exchange of adult trees in the field as observed in previous studies with spruce (Kesselmeier et al., 1993; Thoene et al., 1996) and other species (e.g. NH3: Populus euamericana (Van Hove et al., 1989); Hordeum vulgare; Husted et al. (1996)); NO2: Triticum aestivum (Weber & Rennenberg, 1996); sunflower (Latus et al., 1990)) under controlled conditions. However, due to the chronically high N-input into the ecosystem examined (Rennenberg et al., 1998) and the colonisation of spruce needles with chemolithoautotrophic nitrifiers at this site, which are able to consume NH3, NH4+ and NO2−(Papen et al., 2002), compensation point, internal (mesophyll) resistances and/or cuticular adsorption may differ significantly from those determined under laboratory conditions and with other species in the field. Hence, it is important to understand the magnitude of canopy effects and the primary environmental and physiological controls over NH3 and NO2 exchange in the field, in order to accurately quantify and parameterise regional NO2 and NH3 exchange inventories (Sparks et al., 2001).
Therefore, we determined the effects of (a) CNH3 and CNO2, respectively, and (b) different climatic (T, PPFR, RH) and physiological (JH2O, JCO2 and gH2O) parameters on JNH3 and JNO2 between twigs of mature spruce trees (Picea abies) – the most common coniferous tree species in Central Europe – and the atmosphere. Twigs of spruce trees were exposed to different CNH3 or CNO2 representative for the field site. JNH3 (in the NH3 exposure experiments) and JNO2 (in the NO2 exposure experiments) were determined in the field under varying natural climatic conditions and, as a consequence, varying physiological parameters. Additional 15N-NH3 fumigation experiments were performed in order to trace the fate of NH3 taken up and validate the NH3 exchange measurements.
Materials and Methods
Gas exchange experiments in the field were carried out with twigs from the sun exposed part of the crown of dominant c. 90-yr-old-spruce trees grown at the field site ‘Höglwald’. The site is located 50 km west-north-west of Munich (Germany, longitude: 11°10′ E; latitude 48°30′ N) in the prealpine region at 540 m a.s.l. Annual average air temperature amounts to 8°C and annual precipitation to 800 mm. The soil is an acidic podsolic para brown earth (hapludalf, US Soil Taxonomy). In 1770 the natural vegetation, a Luzulo Fagetum typicum, was replaced by spruce trees (Picea abies[L.] Karst). When the fumigation experiments were performed, the second generation of spruce had reached an age of c. 90 yr. Stand density amounts to 612 trees per ha, basal area is 79 m2 ha−1. A detailed site description is given by Rennenberg et al. (1998).
Twigs from the adult spruce trees with a length of c. 0.2 m with 2–3 needle age-classes were placed through a slot – that was sealed gas-tight with an inert PTFE (Teflon®) paste – into the gas exchange chambers without excising the twigs from the tree. The gas exchange equipment was set up in the crown using scaffoldings. Elemental analyses of the needles did not show nutrient deficiencies or imbalances (Rennenberg et al., 1998); current and previous year's needles and phloem exudate, however, contained high amounts of arginine (Geßler et al., 1998a), assumed to be an indicator of excess N supply (Näsholm & Ericson, 1989). Since the experimental site is situated between the industrialised regions of Munich and Augsburg (Bavaria, Germany) and is bordered by intensively used agricultural land with applications of liquid manure from pig and cattle farming (Rennenberg et al., 1998), the forest examined is exposed to high atmospheric inputs of ammonium and nitrate from wash- and rain-out processes (Kreutzer, 1995). During the present investigation throughfall within the spruce forest contained 10 kg NO3–N ha−1 yr−1 and 20 kg NH4–N ha−1 yr−1. The annual average NH3 and NO2 concentrations in the air of the canopy region amounted to 4.1 (Huber, 1997) and 6.92 nmol mol−1, respectively, in 1995 (R. Gasche & H. Papen, unpublished data). The range of NH3 concentrations was c. 0.4–240 (Geßler & Rennenberg, 1998) and NO2 concentrations amounted to 0.02–44 nmol mol−1 (Gasche & Papen, unpublished data).
Gas exchange measurements and calculation of gas fluxes
Fluxes of H2O, CO2, NH3 and NO2 between spruce and the atmosphere were studied with the dynamic chamber technique (Wesley et al., 1989; Thoene et al., 1991; Rennenberg et al., 1996) using the gas handling system and the analytical devices described in detail by Geßler et al. (2000). Two identical dynamic chambers were continuously flushed with processed ambient air that was filtered to remove reactive trace gases (Rennenberg et al., 1996), humidified (cp. Ball, 1987), and enriched with defined amounts of NH3 or NO2 from pressurised standard mixtures (100 µmol mol−1 NH3 in N2, or 20 µmol mol−1 NO2 in N2, Messer-Griesheim, Germany). The chambers made of borosilicate glass (> 90% transmission in the wavelength range of PAR; Schott Glaswerke, Mainz, Germany) were equipped with PTFE coated ventilators in order to minimise boundary layer resistance and to force rapid mixing of the incoming air. The boundary layer conductance, as calculated according to Von Willert et al. (1995), amounted to c. 3200–3400 mmol m−2 s−1, hence, being 500–1000 fold higher compared with leaf conductance (see, e.g. Figure 3). To avoid chemical reactions between NH3 or NO2 and the inner surfaces of the gas exchange device the tubing consisted of stainless steel (NH3 exchange experiments) or PTFE (NO2 exchange experiments). The gas stream regulated by mass flow controllers (Tylan General, Eching, Germany) was maintained at a rate of 3.0 l min−1 that was equal to an exchange of the air inside the chamber of once per minute. Inside the chambers RH, T (Humicap 113y, Vaissala, Finland) and PPFR (LI-COR, USA) were measured continuously. The air temperature within the chambers was adapted to ambient air temperature by means of Peltier elements (Thoene et al., 1991; Rennenberg et al., 1996) in order to avoid time-dependent chamber effects due to increased temperatures inside the chambers. RH was kept at values below 40% to avoid condensation resulting from rapid temperature change.
JNH3, JNO2, JH2O and JCO2 between the enclosed branch and the atmosphere were derived from the difference in concentrations (CCO2 and CH2O (BINOS, Rosemount, Hanau, Germany); CNH3 (continuous wet flow denuder AMANDA, ECN, The Netherlands); CNO2 (TECAN CLD 770 AL/PLC 760, Eco-Physics Dürnten, Switzerland and AC 30 Environment SA, France)) between the outlet ports of the chamber containing the twig and the empty control chamber (Thoene et al., 1991; Rennenberg et al., 1996; Geßler et al., 1998b; 2000). The additional gas volume produced by transpiration in the chamber containing the twig was kept low by applying high gas flow rates and restricting the needle surface area enclosed, and was considered for flux calculations according to von Caemmerer & Farquhar (1981).
The data obtained were related to the needle surface area inside the chamber. The projection area of the needles was measured at the end of the experiments using an area meter (Delta-T Devices, Cambridge, UK); needle surface area was calculated according to Oren et al. (1986). Fluxes into or onto the leaves are given by negative values, fluxes from the leaves into the atmosphere by positive values.
where RD is the ratio of the diffusivity of NH3 and NO2, respectively, in air to the diffusivity of water vapour in air. According to Massman (1998) RD amounted to 0.91 and 0.62 for NH3 and NO2, respectively.
During three field campaigns (May 1995, September 1995 and July 1996) branches from the crown of three adult spruce trees – one branch of one individual tree per campaign – were exposed to at least two different NH3 concentrations. Altogether seven different NH3 concentrations between 2.4 and 135 nmol mol NH3 were applied in order to cover the whole NH3 concentration range measured in ambient air close to the forest stand in 1995 and 1996 (Huber, 1997). Concentrations of 2.4, 13 and 135 nmol mol−1 where applied in May 1995, 21 and 135 nmol mol−1 in September 1995 and 6.2, 45 and 72 mmol mol−1 in July 1996. The concentration of 135 nmol mol−1 was applied during two campaigns (May and September 1995) in order to test if patterns of Jnh3 were comparable between the different points of time. During two additional field campaigns (July and September 1997) branches from two adult spruce trees – two branches of one individual tree (in succession) per campaign as replicates – were exposed to five different CNO2 between 0.2 and 25 nmol mol−1 in order to cover at least 95% of the NO2 concentration range measured in the canopy of the forest stand in 1995 and 1996 (H. Papen & R. Gasche, pers. comm.). Concentrations of 0.2 and 20 nmol mol−1 were applied in July 1997, 5, 12 and 25 nmol mol−1 in September 1997. For regression analyses (cp. Fig. 4) the results of the replicate measurements (each NO2 concentration regime was applied with two branches of one individual tree in succession) were combined for each concentration.
The fumigation of a branch with one particular CNH3or CNO2 was started with an acclimation period of 6 (NH3 fumigation) or 2 h (NO2 fumigation), respectively. After acclimation, gas exchange experiments at the given concentration were performed for at least 7 h during daylight plus for at least 2 h during night in order to avoid transient effects and to achieve a high variation of climatic parameters and, as a consequence, of JH2O, gH2O and JCO2. The gas exchange experiments were continued by increasing the NO2 or NH3 concentrations from the lower to the higher level in order to avoid memory effects with an acclimation period between two concentrations as explained above.
[15N]NH3 fumigation experiments
The reliability of the gas exchange measurements described above for the determination of NH3 fluxes was assessed by a comparison of this method with the 15N enrichment approach. 15N enrichment was also applied to examine the distribution of 15N deposited in different organs or tissues fumigated and in subsequent branch sections outside the fumigation chambers. For this purpose three twigs of three different adult spruce trees were exposed to c. 140 nmol mol−1[15N]NH3 (calibration gas: 100 ppm [15N]NH3 (99% enriched) in N2, Messer-Griesheim, Germany) for 24–72 h during additional field campaigns in September 1994, May 1995 and September 1995. During these experiments JNH3 was determined as described above. At the end of the experiment the twig inside the chamber as well as the three following sections outside the chamber with a total length between 0.24 and 0.32 m, were collected separately, subdivided into needles (current year's and older needles), bark and wood, and stored in liquid N2 until analysis. Total N and 15N abundance were determined in ground and dried material using a C/N analyser (Roboprep® CN, Europa Scientific, Northwich, UK) and a mass spectrometer (TracerMass®, Europa Scientific) connected in series. N deposition was determined by calculating the 15N enrichment of the twigs exposed to 15N compared with control twigs of the same trees harvested immediately before the fumigation experiment was started.
Data recording and statistical analyses
During gas exchange experiments data were recorded once per minute and calculated as 60 min (NH3 exchange) and 15 min (NO2 exchange) average values. For the examination of the stochastic relationship between different parameters measured and/or calculated from the gas exchange experiments, bivariate correlation analyses were applied. Since the different parameters measured in the field did not vary independently additional partial correlation analyses were carried out. This procedure computes partial correlation coefficients that describe the linear relationship between two variables while controlling for the effects of one or more additional variable(s). For the correlation analyses between JNH3 and JH2O, JCO2, gH2O the influence of T, PPFR and RH was excluded. The control variable for the other analyses was JH2O. In order to compare the regression lines between gH2O and JNO2 or JNH3 at different concentrations or at same concentrations but different points of time, covariance analyses were applied. All statistical analyses were carried out with SPSS® for Windows® (release 5.0–9.0; SPSS Inc., Chicago, IL, USA).
Diurnal variations of NH3 and NO2 exchange
Fig. 1 shows as a typical example for the diurnal courses of J NH3 , J H2O , J CO2 , T, PPFR and RH during NH 3 fumigation at a given C NH3 (140 nmol mol −1 ) determined in May 1995. J NH3 showed a diurnal maximum of up to −1.4 nmol m −2 s −1 between 10 : 00 and 13.00 (CET) and a minimum early in the morning before sunrise. Comparable diurnal patterns were also observed for J H2O and J CO2 showing maximum values of 0.08 mmol m −2 s −1 and –0.2 µmol m −2 s −1 , respectively. The positive values of J CO2 at night indicate CO 2 emission due to respiration.
In July and September 1997 twigs of adult spruce trees were exposed to different NO2 concentrations. Fig. 2 shows a diurnal course of JNO2, JH2O, JCO2, PPFR, RH and T determined in September 1997 applying 12 nmol mol−1 NO2. As observed for JNH3, JNO2 showed a distinct diurnal course with maximum NO2 flux rates of –0.05 nmol m−2 s−1 between 11 : 00 and 15 : 00 when JH2O (0.15 mmol m−2 s−1); at that time, JCO2 (−2.0 µmol m−2 s−1), PPFR (890 µmol m−2 s−1) and T (35°C) were also highest. Comparable diurnal patterns were also observed during measurements in July.
Correlation between JNH3 or JNO2 and different climatic and physiological parameters
In order to assess the climatic and physiological parameters that determine the patterns of NH3 and NO2 fluxes at the different concentration levels of NH3 (2.4, 6.2, 13, 21, 45, 72, 140 nmol mol−1) and NO2 (0.2, 5, 12, 20, 25 nmol mol−1) bivariate correlation analyses between JNH3 and JNO2 and the climatic and physiological parameters were carried out. Table 1 shows the Pearson correlation coefficients for the linear relation between JNH3 and JNO2 and the climatic and physiological parameters summarising the data obtained at the different NH3 and NO2 concentrations applied (total range) and for each concentration separately. When the dataset was not subdivided into the different concentrations a high correlation was obtained only between JNH3 and CNH3 (r = −0.985; P < 0.01) and between JNO2 and CNO2 (r = −0.929; P < 0.01). Temperature and gH2O showed significant but low correlation with JNH3 and JNO2, whereas RH only correlated with JNH3. For six of the seven NH3 concentrations applied high correlation coefficients were obtained between JNH3 and JH2O (−0.678 to –0.897) and between JNH3 and gH2O (–0.645 to –0.943). The low correlation observed for 2.4 nmol NH3 mol−1 is supposed to be due to the very low net flux of NH3 at this concentration (cp. Fig. 3). For the other climatic and physiological parameters significant (P < 0.05) correlation was observed between JNH3 and JCO2 (6.2, 21, 45 nmol mol−1), PPFR (13, 21, 45, 72, 135 nmol mol−1), RH (72, 135 nmol mol−1) and T (6.2, 13 nmol mol−1). For all 5 NO2 concentrations applied significant correlation was calculated between JNO2 and the parameters gH2O and JH2O. From the other parameters determined a correlation with JNO2 was observed for RH at three NO2 concentrations and for JCO2, PPFR and T at two NO2 concentrations.
Table 1. Bivariate correlation analyses between J NH3 (a) and J NO2 (b) and physiological and climatic parameters at different NH 3 and NO 2 concentrations. The table shows Pearson's correlation coefficients as a measure of the association of two variables that are linearly related. In the first row (‘total range’) data were not subdivided into the different concentrations [nmol mol −1 ] applied, in the following rows only data obtained at the NH 3 or NO 2 concentration indicated were used for the analyses
Correlation coefficients significant at the 0.05 level are identified with a single asterisk, those significant at the 0.01 level are identified with two asterisks.
(a) JNH3 at the following CNH3
2.4–135 (total range)
(b) JNO2 at the following CNO2
0.2–25 (total range)
From the bivariate correlation analysis it may be assumed that in addition to trace gas concentration (CNH3, CNO2), JH2O and gH2O, JNO2 and JNH3 are influenced by PPFR, T and RH at least under particular conditions. To test this assumption, partial correlation coefficients were computed for the dependency of JNO2 on one particular climatic or physiological parameter, whilst the influence of other variables was excluded (see ‘Data recording and Statistical Analyses’). Partial correlation analyses indicated significant correlation between JNH3 and the parameters gH2O and JH2O at six of seven NH3 concentrations (coefficient of partial correlation between –0.661 and −0.811). JNO2 showed significant correlation to the parameters JH2O and gH2O at every concentration level applied (coefficient of partial correlation between 0.387 and –0.941). All other parameters (JCO2, PPFR, RH, and T) showed distinct lower partial correlation with JNH3 or JNO2 and were in almost all cases not significant (Table 2).
Table 2. Partial correlation analyses between J NH3 (a) and J NO2 (b) and physiological and climatic parameters at different NH 3 and NO 2 concentrations. The table shows partial correlation coefficients as a measure of the association of two variables that are linearly related while the influence of other parameters (variables) is controlled. For the correlation analyses between J NH3 and J H2O , J CO2 , g H2O the influence of T, PPFR and RH was excluded . The control variables for the other analyses was J H2O .
Correlation coefficients significant at the 0.05 level are identified with a single asterisk, those significant at the 0.01 level are identified with two asterisks.
(a) JNH3 at the following CNH3
(b) JNO2 at the following CNO2
From the results of the bivariate (Table 1) and partial correlation analyses (Table 2), it is concluded that JNH3 and JNO2 are determined by (1) CNH3 and CNO2 (Table 1; total range), and (2) JH2O and gH2O, the latter calculated from JH2O and being a measure of stomatal aperture (Von Willert et al., 1995). The results of the partial correlations procedure shows that the bivariate correlation found between JNH3 or JNO2, and PPFR, JCO2, RH and T (Table 1) is mainly due to the linear interconnection between PPFR, JCO2, RH and T with gH2O and/or JH2O.
Regression analyses between JNH3 or JNO2 and gH2O
JNH3 (Fig. 3) and JNO2 (Fig. 4) were plotted against gH2O at different CNH3 and CNO2. At the lowest NH3 (2.4 nmol mol−1) and NO2 (0.2 nmol mol−1) concentration applied slight emission of NH3 and NO2, was determined as indicated by positive values of JNH3 (Fig. 3) and JNO2 (Fig. 4). At the next higher CNH3 (6.2 nmol mol−1) and CNO2 (5 nmol mol−1) deposition of NH3 and NO2 occurred indicating the existence of compensation points for both trace gases. At a given CNH3 ≥ 6.2 nmol mol−1 and CNO2 ≥ 5 nmol mol−1, JNH3 and JNO2 increased linearly with increasing gH2O. This relation remained constant at different times of measurement in the growing season, since there was no significant difference between May and September 1995 (Fig. 3) at a CNH3 of 135 nmol mol−1. Changes in NH3 and NO2 concentrations resulted in significant changes of the slopes and intercepts of the regression lines that characterise the relation between JNH3 and gH2O and between JHO2 and gH2O, respectively. Since the intercepts can be regarded as a measure for JNH3 and JNO2 at total closure of the stomata (Geßler et al., 2000), it is concluded that cuticular deposition of both nitrogen trace gases occurs in spruce and increases with increasing NH3 and NO2 concentration.
Regression analyses between JNH3 or JNO2 and trace gas concentration
When JNH3 and JNO2 were calculated for a constant gH2O (6.25 mmol m−2 s−1) and plotted against CNH3 and CNO2, a significant linear regression between the fluxes and the concentrations could be observed (Fig. 5, R2 = 0.977 for NH3; R2 = 0.998 for NO2). JNO2 was about one order of magnitude smaller than JNH3 at comparable concentrations. For both trace gases compensation points could be calculated that amounted to 2.5 nmol mol−1 for NH3 and to 1.7 nmol mol−1 for NO2.
From the results shown in Figs 3–5 it is concluded that JNH3 and JNO2 depend linearly on gH2O, provided CNH3 and CNO2 were constant. For constant gH2O a linear correlation between JNH3 and JNO2 and CNH3 and CNO2 could be determined.
Calculation of the quotient between measured and predicted conductance for NH3 and NO2
Provided fluxes of NH3 and NO2 proceed exclusively through the stomata measured leaf conductance (gNH3; gNO2) should be similar to predicted conductance (pNH3; pNO2) derived from gH2O. Consequently, the quotient (q) between measured and predicted conductance is then equal to 1. If additional sinks for NH3 or NO2 on the needle surface are present, measured conductance is higher than predicted (q > 1). In contrast, the presence of internal resistances for NH3 or NO2 leads to q smaller than 1. Fig. 6 shows q calculated for the NH3 and NO2 exchange experiments in the field, where different NH3 and NO2 concentrations were applied. In both, NH3 and NO2 exchange experiments, RH and T had no influence on q in the T and RH range applied in the experiments. Increasing NH3 and NO2 concentration, however, caused a more or less linear increase of q confirming the increasing importance of cuticular deposition. In the NH3 fumigation experiments the mean value of q for a CNH3 of 6.2 nmol mol−1 amounted to c. 1 and increased to 2.3 and 3.7 at a CNH3 of 45 and 135 nmol mol−1, respectively. In the experiments applying NO2 to spruce q amounted to 1.0 at a CNO2 of 0.2 nmol mol−1 and increased to 3 at a CNO2 of 25 nmol mol−1. In both, NH3 and NO2 exchange experiments, there was a slight tendency for q to increase with decreasing PPFR.
[15N]NH3 fumigation experiments
NH3 deposition onto/into spruce needles was assessed with a second approach using 15N[NH3] for fumigation. This method does not allow the calculation of deposition rates with the high resolution in time like the gas exchange measurement described above, but it was used first to examine the allocation of deposited 15N within the twig and second to double-check the results of NH3 gas exchange measurements.
Fig. 7 shows the specific 15 N enrichment [µg 15 N g −1 d. wt] in needles, bark and wood of different age within the gas exchange chambers and, in addition, of subsequent sections outside the chambers. In each experiment the last year's and older bark showed the highest 15 N contents as compared to the other tissues. The enrichment in current year's needles was comparable to older needles: The specific 15 N enrichment of needles was between 15% and 33% lower compared with the previous year's and older bark. However, total 15 N-enrichment on a tissue base was highest in the needles due to the highest biomass of this tissue. Not only in the sections inside the gas exchange chambers, but also in segments outside, 15 N enrichment was detected ( Fig. 7 ). Highest specific enrichment was found in the bark and only traces in the woody tissue.
15 N enrichment [µg 15 N] in the twig segments outside the gas exchange chamber amounted to 1.7, 1.4 and 1.7% of total 15 N enrichment of the twigs in September 1994, May 1995 and September 1995, respectively. Daily NH 3 deposition rates calculated from total 15 N enrichment amounted to 87.6, 62.1 and 76.1 µg N d −1 per twig in September 1994, May 1995 and September 1995, respectively, and were similar to those derived from gas exchange measurements (92.1, 65.4 and 72.1 µg N d −1 per twig).
The present study was aimed at assessing the effects of different climatic (T, RH, PPFR) and physiological parameters (JH2O, gH2O, JCO2) as well as the effects of different NO2 and NH3 concentrations on the JNO2 and JNH3 between the needles of spruce trees and the atmosphere under field conditions – isolated from other ecosystem fluxes of NH3 and NO2. A whole set of studies on NH3 and NO2 exchange have been performed with gas exchange chambers in the laboratory (e.g. Van Hove et al., 1990; Husted et al., 1996; Thoene et al., 1996; Hanstein et al., 1999; Hereid & Monson, 2001) with the aim of parameterising the effect of climatic patterns on trace gas exchange and understanding the physiological mechanisms that drive NH3 and NO2 emission and deposition. In addition, intensive micrometeorological research is performed with the aim of characterising ecosystem NO2 and NH3 fluxes between forests and the atmosphere and underlying mechanisms (e.g. Walton et al., 1997; Andersen et al., 1999; Sutton et al., 2000, 2001; Pryor et al., 2001). However, the assessment of NH3 and NO2 fluxes to and from an isolated compartment – the twigs/needles of adult spruce – is of particular interest in the forest ecosystem examined, since spruce needles have been found to be colonised by chemolithoautotrophic nitrifying bacteria (Papen et al., 2002). Since these organisms are able to consume NH3 as well as NH4+ and NO2− (Bock et al., 1989), that is produced when NH3 and NO2, respectively, dissolve in water, compensation points, internal resistances and surface deposition may be significantly different from plants grown and fumigated under controlled conditions or field grown plants without colonialisation by nitrifiers. From previous studies (Papen et al., 2002) it is known that NH3 deposition fluxes to spruce needles are increased significantly as a consequence of metabolic activity of nitrifying bacteria in and/or on the needles. However, until now no information has been available on how different climatic and physiological parameters influence NH3 and NO2 fluxes to nitrifier-colonised needles.
In the fumigation experiments twigs of spruce trees were exposed to defined concentrations starting from the lower to the higher concentration consecutively. In order to avoid time-dependent chamber effects, and, as a consequence changes in physiological performance of the incubated twigs, near-natural conditions were maintained in the chambers by adapting air temperatures within the chambers to ambient air temperature and using chamber material with > 90% PAR transmission. Experiments with beech (Geßler et al., 2000), where a comparable experimental design was applied, showed that over a period of up to 3 d transpiration, photosynthesis and JNH3 and JNO2 remained constant at a given trace gas concentration.
Fig. 3 shows that although absolute values of g H2O – which are known to vary with water availability, water pressure deficit of the air, nutrition, and needle age ( Wieser et al., 2000 ; Phillips et al., 2001 ) – and J NH3 may have been different between the two times of measurement (May 95; September 95), the regression between J NH3 and g H2O at a given concentration (135 nmol mol −1 ) remained constant as previously also observed for beech ( Geßler et al., 2000 ). Independent of the time of measurement, partial correlation procedures ( Table 2 ) that were used to differentiate between the influence of stomatal function and climatic parameters, produced comparably high (NO 2 : R between –0.657 and –0.941; NH 3 : R between –0.622 and –0.901) and highly significant correlations between J NH3 or J NO2 and g H2O at concentrations above c . 5 nmol mol −1 where significant fluxes of NH 3 or NO 2 could be observed ( Figs 3 and 4 ). Hence, it is concluded that – for the analysis of the factors determining J NH2 and J NO2 – measurements performed at different times can be compared.
JNH3 and JNO2 are related linearly to trace gas concentration
In both, NH3 and NO2 fumigation experiments, trace gas fluxes (JNH3 and JNO2) depended linearly on CNH3 and CNO2 in the concentration range found at the field site studied. Apparently, the difference in NH3 and NO2 concentrations between the gaseous phase of the leaf interior and the atmosphere is the driving force for NH3 and NO2 exchange. In addition, it must be concluded that the capacity of spruce for NH3 and NO2 acquisition was not exceeded even at high NH3 and NO2 concentrations (Hanstein et al., 1999) different from observation of Sparks et al. (2001) for tropical tree species. The consecutive increase in NO2 or NH3 concentration applied in the present study may result in transient changes of apoplastic conditions and/or expression/activity of nitrate and ammonium transporters, hence, either increasing or decreasing fluxes of both trace gases. A decrease in NO2 or NH3 flux with increasing incubation time and, thus, increasing concentration can be ruled out, since no saturation was observed in the relation between JNH3 or JNO2 and trace gas concentration. However, the possibility that the observed linearity of this relation in a higher concentration range was due to a consecutive adaptation of the exposed twigs to higher NO2 or NH3 concentrations can not be ruled out. Nevertheless, in accordance with the present results linear dependencies between JNH3 and CNH3 have already been observed for Phaseolus vulgaris (Farquhar et al., 1980; Van Hove et al., 1987), Populus euamericana (Van Hove et al., 1989), Picea abies (Kesselmeier et al., 1993) and Hordeum vulgare (Husted et al., 1996) under controlled environmental conditions and for Fagus sylvatica (Geßler et al., 2000) and Picea abies (Peters & Bruckner-Schatt, 1995; Andersen et al., 1999) in the field. Due to the activity of chemolithoautotrophic nitrifiers, colonizing the needles of spruce at the field site ‘Höglwald’ the sink strength of the adult spruce trees for NH3 is increased by 24–37% compared with needles without bacterial activity (Papen et al., 2002). The bacterial activity can explain the c. 1.4-fold higher NH3 deposition fluxes at a given gH2O and CNH3 determined for spruce in the present study compared with beech (Geßler et al., 2000) from the same forest stand, but without colonisation of leaves with nitrifying bacteria (H. Papen, pers. comm.). However, JNH3 at a given CNH3 was about one order magnitude lower as observed for herbaceous species (Van Hove et al., 1990; Husted & Schjoerring, 1995; Husted et al., 1996, Husted & Schjoerring, 1996).
Needles of spruce show compensation points for NH3 and NO2
Spruce trees exhibited compensation points for both trace gases, NH3 and NO2, indicating the release/production of NH3 and NO2 inside the needles. Hence, the adult spruce trees are not only sinks, but also sources for NH3 and NO2 depending on atmospheric concentrations. The NH3 compensation point determined amounted to 2.5 nmol mol−1 and, thus, was somewhat lower than for beech trees at the same field site (up to 3.5 nmol mol−1; Geßler et al., 2000). In literature NH3 compensation points for trees and canopies of forest ecosystem are described to range between 0.8 nmol mol−1 (montane forest with low N supply; Langford & Fehsenfeld, 1992) and 48 nmol mol−1 (N-polluted forests; Wyers & Erisman, 1998) indicating the relevance of N supply and N status for this parameter (Sutton et al., 1995; Geßler & Rennenberg, 1998; Herrman et al., 2001). The dependence of NH3 compensation point on N supply is supposed to be the reason for a higher NH3 compensation point in spruce from the N oversaturated field site ‘Höglwald’ (Rennenberg et al., 1998) compared with individuals of the same species growing at another site in SE-Germany (Meixner et al., 1997). The NO2 compensation point observed for adult spruce in the present study amounted to 1.7 nmol mol−1 and therefore was in a range comparable to that found for young trees of the same species under controlled conditions (1.6 nmol mol−1; Thoene et al., 1996), for adult beech trees at the same field site (1.8–1.9; Geßler et al., 2000) and for different tropical trees (Sparks et al., 2001), but somewhat higher than reported for corn (c. 0.9 nmol mol−1Hereid & Monson, 2001) and for Scots pine and Norway spruce at less N exposed sites in Sweden (c. 0.1–0.7 nmol mol−1; Rondón et al., 1993; Rondón & Granat, 1994). This difference between individuals of the same species growing at different sites can be due to the excess N supply of the spruce at the field site studied, as indicated by high leaf N contents (Geßler et al., 1998a), which are known to increase NO2 compensation point (Sparks et al., 2001). At the field studied, NO2 emission from spruce is supposed to play only a minor role, since atmospheric NO2 concentrations in the canopy at the field site ‘Höglwald’ are usually higher than the NO2 compensation point (Papen & Gasche, pers. comm.).
NH3 and NO2 fluxes correlate with stomatal conductance at a given NH3 and NO2 concentration
Several authors have observed internal (mesophyll) resistances limiting NO2 deposition in different tree species (Johannson, 1987; Thoene et al., 1991; Sparks et al., 2001). Internal resistances (q < 1 cp pg. 14) were also observed under all environmental conditions in NO2 fumigation experiments with young spruce trees (Thoene et al., 1996). In the present study, however, indications of measured NO2 conductance to be smaller than NO2 conductance predicted from water vapour conductance were not found (Fig. 6E–H). Since the disproportionation reaction of NO2 in water resulting in equal amounts of nitrate and nitrite is slow at typical atmospheric NO2 concentrations, it is supposed that the reaction of ascorbate with NO2 to nitrite and dehydroascorbate is responsible for maintaining high fluxes of NO2 into the leaf and may prevent internal resistances (Ramge et al., 1993). Since the apoplastic ascorbate concentration does not only vary with species but also with environmental conditions (Polle et al., 1995, Schwanz et al., 1996) and stage of development (Luwe, 1996), the observed differences may be due to different apoplastic ascorbate concentrations. In addition, the removal of nitrite, formed in the reaction of NO2 with ascorbate, by chemolithoautotrophic nitrite oxidisers present and physiologically active in the needle apoplast of the adult trees can also contribute to the reduction of mesophyll resistance.
Cuticular deposition increases with increasing concentrations of NH3 and NO2
Cuticular NO2 and NH3 deposition onto needles of spruce trees was observed as indicated by the intercepts of the regression lines in Figs 3 and 4, respectively, and by q > 1 given in Fig. 6 and increased with increasing NH3 concentration even at the low RH < 40% applied in the fumigation experiments. At low ambient CNH3 and CNO2 flux of NH3 and NO2 could be explained almost by stomatal uptake but at high concentrations measured conductance of both trace gases exceeded the one predicted from gH2O by a factor of up to 3. This finding is in contradiction to the calculations of Wellburn (1990) for Picea abies, who concluded that even at high NO2 concentrations (140 nmol/mol) stomatal flux exceeded deposition on the cuticles by one or two orders of magnitude. NO2 deposition rates onto cuticles were also minute for sun flower and tobacco (Neubert et al., 1993), for adult beech grown at the same field site examined in the present study (Geßler et al., 2000) and for spruce trees in a forest in Sweden (Rondon et al., 1993). It is known that NO2 is able to penetrate the cuticles within a certain range (Wellburn, 1990) and can be adsorbed to and/or react with components of the cuticle (Lendzian & Kerstiens, 1988). Differences in chemical composition between different species and, as a key factor, the metabolic activity of autotrophic nitrite oxidising bacteria colonising the needles at the site examined (Papen et al., 2002) may be responsible for the different concentration-dependent patterns of cuticular NO2 deposition. Especially the tendency of cuticular deposition to decrease with increasing PPFR indicates that nitrifying bacteria – known to be inhibited by light (Vanzella et al., 1989) – are involved in cuticular deposition.
For oilseed rape, no significant NH3 deposition on the cuticles even at a relative humidity of 80% could be observed (Husted & Schjoerring, 1995). However, significant cuticular deposition to a spruce forest was assumed by Andersen et al. (1993; 1999) and onto beech leaves by Geßler et al. (2000). NH3 deposition on the cuticles may be explained by the wetness of leaf cuticles observed even in the absence of free water at low air humidity (Van Hove & Adema, 1996) that may act as additional sink for highly water soluble NH3.
If it is supposed that at least a part of the NO2 and NH3 deposited on the cuticles can be re-emitted when ambient NO2 and NH3 concentration decreases, future research will have to achieve discrimination between net emission via stomata and cuticular re-emission.
NH3 taken up by needles of adult spruce contributes to the N demand of the whole trees
The 15N[NH3] fumigation experiments demonstrated that the deposition rates calculated from the gas exchange measurement were reliable, since daily deposition rates determined from both approaches were similar. In addition, it could be directly traced for the first time in adult trees grown in the field, that NH3 taken up by adult spruce trees is subjected to long distance transport, since the 15N tracer was found in significant amounts in bark of sections inside and outside the fumigation chambers. This is a prerequisite for atmospheric NH3 to serve as additional N supply for adult forest trees. The present finding of the field experiment is consistent with the results gained from NH3 exposition of Fagus sylvatica seedlings (Geßler et al., 1998b; Geßler & Rennenberg, 1998) and Pinus sylvestris (Pérez-Soba et al., 1994; Van der Eerden et al., 1990) under controlled conditions. A 3-d fumigation of beech with c. 55 nmol mol−1 NH3 resulted in a significant enrichment of soluble amino compounds in the leaves (mostly Arg) and in the phloem (Glu, Gln, Asp, Asn) (Geßler et al., 1998b). Exposure of Scots pine to NH3 stimulated glutamine synthetase (GS) and lead to a subsequent increase in amino compounds in the tissues exposed (Pérez-Soba et al., 1994). Hence, it must be concluded that NH3 similar to NO2 (Nussbaum et al., 1993, Muller et al., 1996) contributes to the pool of soluble nitrogen compounds circulating between shoots and roots that first provides different tissues and organs of the plant with nitrogen and second is involved in the adaptation of pedospheric N uptake to the plant's demand (Imsande & Touraine, 1994; Muller et al., 1996) in adult spruce.
This investigation is part of the ‘Höglwald’ ecosystem study and was financially supported by the Bundesminister für Bildung, Wissenschaft, Forschung und Technologie (BMBF) under contract No. BEO 51 0339614 and BEO 51 0339615.