Plant-mediated nitrous oxide emissions from beech (Fagus sylvatica) leaves


Author for correspondence: Mari Pihlatie Tel: +358 9 191 50578 Fax: +358 9 191 50860 Email:


  • • Nitrous oxide (N2O) emission estimates from forest ecosystems are based currently on emission measurements using soil enclosures. Such enclosures exclude emissions via tall plants and trees and may therefore underestimate the whole-ecosystem N2O emissions.
  • • Here, we measured plant-mediated N2O emissions from the leaves of potted beech (Fagus sylvatica) seedlings after fertilizing the soil with 15N-labelled ammonium nitrate (15NH415NO3), and after exposing the roots to elevated concentrations of N2O.
  • • Ammonium nitrate fertilization induced N2O + 15N2O emissions from beech leaves. Likewise, the foliage emitted N2O after beech roots were exposed to elevated concentrations of N2O. The average N2O emissions from the fertilization and the root exposure experiments were 0.4 and 2.0 µg N m−2 leaf area h−1, respectively. Higher than ambient atmospheric concentrations of N2O in the leaves of the forest trees indicate a potential for canopy N2O emissions in the forest.
  • • Our experiments demonstrate the existence of a previously overlooked pathway of N2O to the atmosphere in forest ecosystems, and bring about a need to investigate the magnitude of this phenomenon at larger scales.


Nitrous oxide (N2O) is a strong greenhouse gas contributing approx. 6% to the anticipated global warming (IPCC, 2001). Microbial production by denitrification and nitrification processes in soils is considered to be the main source of N2O into the atmosphere (Kroeze et al., 1999). Currently, the emission estimates from natural ecosystems are based solely on soil emission measurements using soil enclosures. In the enclosure technique, the soil is covered with a shallow chamber, and the emission from the soil is calculated from the increase in N2O concentration inside the chamber (e.g. Butterbach-Bahl et al., 1997). Such enclosures exclude tall plants and trees and may therefore underestimate the whole soil–plant N2O emissions.

Agricultural and wetland plants have been found to contribute significantly to the total N2O emissions from soil–plant systems (Chang et al., 1998; Rusch & Rennenberg, 1998; Yan et al., 2000; Smart & Bloom, 2001; Chen et al., 2002; Müller, 2003). Rusch & Rennenberg (1998) found that black alder (Alnus glutinosa), a wetland tree species, emitted N2O through the bark of the tree when the gas concentration in the soil solution was above the ambient concentration. They suggested that the gases diffuse through the aerenchyma of the bark. Yan et al. (2000) found that 87% of the N2O emissions from a rice field occurred through rice plants when the soil was flooded and 18% when the soil was unsaturated, whereas Chen et al. (2002) and Zou et al. (2005) found that N2O emissions from maize, soybean and wheat plants accounted for up to 11, 16 and 62% of the total soil–plant N2O emissions. All of these field studies, however, failed to identify the mechanisms responsible for plant-mediated N2O emissions.

In a laboratory study, Chang et al. (1998) found that barley (Hordeum vulgare) and canola (Brassica napus) plants can serve as a conduit for dissolved N2O from the root zone to the atmosphere. They suggested that this transpiration-mediated N2O emission may be a common phenomenon with several upland plants. Smart & Bloom (2002) and Hakata et al. (2003) found, also in the laboratory, that N2O can be formed enzymatically inside wheat leaves during nitrite (NO2) photoassimilation. According to Smart & Bloom (2002), this enzymatic production of N2O in the leaves could account for 5–6% of the total N2O emissions from agricultural soil–plant systems.

In summary, the studies reporting plant-mediated N2O emissions have all been conducted with agricultural or wetland plants, whereas studies with upland forest trees are lacking. We hypothesize that tree leaves can be a significant source of N2O into the atmosphere and that N2O produced in the soil can be transported to the atmosphere via the transpiration stream.

To test our hypothesis, we conducted two laboratory experiments with potted beech (Fagus sylvatica) seedlings to examine whether beech leaves emit N2O under controlled conditions and whether N2O can be transported from the soil solution to the atmosphere via transpiration stream.

Materials and Methods

Beech (Fagus sylvatica L.) seedlings were collected in 2001 from the forest Lille Bøgeskov near Sorø on the island of Zealand, Denmark (55°29 ′N, 11°39′ E) (Pilegaard et al., 2003). The seedlings were planted in pots and kept outdoors until the experiment in May 2004. Two separate laboratory experiments were conducted to examine whether beech trees possess the potential for leaf-based emissions of N2O to the atmosphere.

Soil 15N-fertilization experiment

Beech seedling pots (two replicates, S1 and S2) were watered with 200 ml of a solution containing 15N-labelled NH4NO3 (15NH415NO3) (5 atom%) and glucose to provide approx. 100 mg of N and 500 mg of C per kg of soil, respectively. The purpose of 15NH415NO3 and glucose addition was to activate soil microbes to produce N2O (N2O + 15N2O) in the soil, because the 15NH415NO3 serves nitrogen and glucose as a substrate for denitrifying bacteria. The seedlings were stored at 21°C over night and the emission of N2O + 15N2O from the foliage was measured during two subsequent days.

Root chamber experiment

One day before the start of the experiment, the beech seedlings were removed from the pots and gently washed to remove all soil around the roots. The root compartment of each seedling was enclosed in a gas-proof root chamber made of polyvinyl chloride (PVC) with a total volume of 1400 ml (Formánek & Ambus, 2004). Each chamber contained 300 ml of deionized water, and immediately after the enclosure 2 ml of 100% N2O was injected to result in a concentration of 1800 ppmv in the gas phase. To dissolve N2O into the water, the chamber was shaken carefully and left to equilibrate overnight at 21°C. The concentration of N2O in the solution at equilibrium was calculated to be 2300 µg N2O l−1 using the Bunsen constant of 0.631 for gas solubility at 20°C (Tiedje, 1982; Heincke & Kaupenjohann, 1999). The experiment was conducted with four beech seedlings (S3–S6), each of them measured once per day on two subsequent days.

A dose exposure test was conducted to assess the response of N2O emissions to the concentration in the root solution. Leaf emissions of N2O were measured from the two replicate seedlings after exposing the beech roots first to 2300 and then to 10 400 µg N2O l−1 in the root solution.

Emission measurements

Leaf N2O emissions were measured using a metacrylat chamber enclosing only the foliage of a beech seedling (Fig. 1). During a measurement, the chamber air was kept at 25°C and at relative humidity below water saturation by circulating the air via a fridge and a magnesium perchlorate [Mg(ClO4)2] drier. Air and beech-leaf temperatures and relative humidity inside the chamber were measured with a Testo 615 thermohygrometer (Testo AG, Lenzkirch, Germany) and a Raytek® RAYMX2 infrared thermometer (Raytek, Santa Cruz, CA, USA). Photosynthetically active radiation at the top of the chamber was measured with a Li-250 light meter (Li-Cor Inc., Lincoln, NE, USA). Carbon dioxide (CO2) concentration was monitored with a Li-820 gas analyzer (Li-Cor, Inc., Lincoln, NE, USA), and pure CO2 was injected at time intervals to keep the concentration between 300 and 400 ppmv of CO2.

Figure 1.

Measurement set-up for measuring N2O emissions from the leaves of beech (Fagus sylvatica) seedlings. Leaf N2O emissions were measured with a metacrylat shoot chamber under constant irradiance (300 µmol s−1 m−2) and temperature (24°C). The chamber was equipped with a fan and two closed loops in order to cool the chamber air and keep relative humidity below saturation, and to monitor CO2 concentration inside the chamber.

Syringe gas samples from the chamber air were taken with a needle through a septum inserted at the outlet of the chamber. Samples for N2O analysis were taken at 20–40 min intervals and injected into pre-evacuated Venojects (Terumo, Leuven, Belgium), which were stored in a cool room until analysis with a gas chromatograph equipped with an electron capture detector (Schimadzu GC-14B, Kyoto, Japan). Gas samples for isotopic 15N2O analysis were taken at 0, 2 and 4 h during the enclosure and immediately injected into pre-evacuated crimp-sealed serum bottles. The 15N2O contents were analyzed within 1 wk following chemical removal of H2O + CO2 and cryogenic focusing of the N2O by a Finnigan MAT PreCon trace gas concentration unit (ThermoElectron, Bremen, Germany). The PreCon was interfaced to a gas chromatographer coupled in continuous-flow mode to a Finnigan MAT Delta PLUS isotope ratio mass spectrometer (ThermoElectron).

Leaf N2O analysis

We measured leaf N2O concentrations from the seedlings in the 15N fertilization experiment and from beech trees in the forest Lille Bøgeskov. Leaves from the forest Lille Bøgeskov were sampled on 15 June from tree branches at 2 and 16 m heights in the canopy. Approximately 8 leaves per seedling and 13 leaves per canopy height in the forest were analyzed. Fresh leaves were cut into pieces and inserted into 3.5-ml glass vials (Venojects) – one to two leaves per vial. These and blank vials were then flushed with nitrogen gas (N2) for 2 min to remove N2O from the headspace. The vials were stored in the dark at +4°C for several days to inhibit possible N2O production and to allow N2O to diffuse from the leaf water to the vial gas space. Concentration of N2O inside the vials was analyzed with a gas chromatograph and expressed as µg N2O l−1.

We assumed that all N2O in the leaves was dissolved in the leaf water. Hence, the N2O concentration inside beech leaves (cLe) at equilibrium between the liquid and gas phases was calculated by multiplying the headspace gas concentration (cG) in the vial with the Bunsen gas solubility constant of 1.05 for N2O at +5°C (Tiedje, 1982; Heincke & Kaupenjohann, 1999):

image(Eqn 1 )

The initial leaf N2O concentration (cLi) before equilibrium with headspace gas was calculated as follows:

image(Eqn 2 )

(VL and VG, leaf and gas volumes inside the vial, respectively; W, leaf water content.) Leaf volumes were calculated by multiplying the leaf area (LA) in each vial (cm2) with an estimate of the beech leaf thickness of 0.015 cm, according to Bussotti et al. (2005). Leaf water content was assumed to be 42 and 53% for sun and shade leaves, respectively (Pilegaard et al., 2003). Mean leaf water content between the sun and shade leaves was used for the laboratory seedlings.

Data processing and emission calculations

Emissions of N2O from the beech leaves were calculated for each seedling separately from the slope of N2O concentration change during the chamber enclosure. The leaf emissions are expressed as µg of N m−2 of leaves h−1. The LA was determined individually for each seedling. A sample of 10 leaves per seedling was collected, weighed (to give weight ws) and the LA (m2) of each leaf in the sample was determined (to give area LAs). Then the rest of the leaves in the seedling were collected and weighed (to give weight wr). The leaf area of each seedling was thereafter calculated as follows:

image(Eqn 3 )

The 15N-enrichment of the emitted N2O was calculated from Keeling regressions, for example the y-axis intersection of the linear correlation line of 15N enrichment vs the reciprocal of N2O concentration observed in the enclosures during the experimental time span (Formánek & Ambus, 2004).


Soil-fertilization-induced N2O emissions

Nitrous oxide concentration and the percentage of 15N-labelled N2O (15N2O) inside the shoot chamber increased in three out of four chamber measurements shown in Fig. 2. The average fertilization induced N2O emission from the foliage of S1 and S2 from four chamber measurements was 0.4 ± 0.4 µg N h−1 m−2 of leaves (mean ± SE). The average N2O emission 1 d after fertilization was 0.6 µg N h−1 m−2 of leaves (S1: 0.5 µg N h−1 m−2, S2: 0.7 µg N h−1 m−2). Two days after the fertilization, no N2O emission was detected from S1, but S2 continued to emit N2O at a rate of 0.7 µg N h−1 m−2 of leaves.

Figure 2.

Development of N2O (N2O  + 15N2O) concentration inside the shoot chamber during three chamber enclosures. Increase in N2O (open circles) and 15N2O (closed circles) abundance (percentage of the total N2O) inside the chamber during chamber enclosures of beech (Fagus sylvatica) (a) seedling S1, day 1 (b) S2, day 1, and (c) S2, day 2 after the fertilization. On day 1 of S2, the N2O sampling was stopped after 90 min of the measurements owing to a technical problem in the sample collection system.

The abundance of 15N2O inside the shoot chamber increased nearly linearly, and similar to the N2O emission rates, the signal of 15N2O was greater with S2 than with S1 (Fig. 2). The N2O emitted from S2 and S1 were 6% and 0.41% enriched with 15N2O, respectively.

Root-solution-induced N2O emissions

Nitrous oxide concentration inside the shoot chamber increased in all the chamber measurements: six of them conducted after exposing the roots of the seedlings to 2300 µg N2O l−1 in the root solution and two after exposing the roots to 10 400 µg N2O l−1 in the root solution (Fig. 3). Development of N2O concentration inside the shoot chamber one day after the exposure of the roots to 2300 µg N2O l−1 was started is shown in Fig. 3.

Figure 3.

Nitrous oxide concentrations inside the shoot chamber after exposing the beech roots to 2300 µg N2O l−1 in the root solution. The panels represent emission measurements with different beech (Fagus sylvatica) seedlings (S3, S4, S5 and S6) 1 d after the start of the root exposure. The N2O emission is calculated from slope of the linear regression line fitted to the N2O concentrations.

Beech seedlings exposed to 2300 µg N2O l−1 in the root solution emitted N2O at an average rate of 2.0 µg N h−1 m−2 of leaves (range 1.0–3.4 µg N m−2 h−1). Emissions of N2O increased with increasing N2O concentration in the root solution (Fig. 4). When the root solution concentration was 10 400 µg N2O l−1, the average leaf N2O emission was 6.4 µg N h−1 m−2 of leaves (range 4.3–8.5 µg N m−2 h−1). The response in N2O emission rate to increased N2O concentration in the root solution was similar in both of the seedlings S4 and S5; however, they emitted N2O at different levels.

Figure 4.

Leaf N2O emissions after exposing beech (Fagus sylvatica) roots to different N2O concentrations in the root solution. Nitrous oxide emissions (µg N2O-N m−2 h−1) from beech leaves of two seedlings (S4, black bars; S5, grey bars) after exposing the beech roots to 2300 and 10 400 µg N2O l−1 in the root compartment solution.

Nitrous oxide concentration in the leaves

Leaf N2O concentrations decreased in the following order: laboratory seedlings > forest at height 2 m > forest at height 16 m. The average leaf N2O concentration of the laboratory seedlings and of the forest trees at heights 2 and 16 m were 108, 20 and 3.5 µg N2O l−1, respectively. The average leaf N2O concentrations were above the ambient atmospheric concentration of 0.57 µg N2O l−1 (equivalent to 0.314 ppmv) (IPCC, 2001).


The increased N2O and 15N2O concentrations inside the shoot chamber during the enclosure period demonstrated that soil fertilization with 15NH415NO3 and glucose-induced N2O emissions from the beech foliage. The addition of mineral nitrogen and organic carbon substrates activates soil denitrifying bacteria to produce N2O in the soil (Azam et al., 2002). Hence, we assumed that denitrifying bacteria in the soil produced N2O and 15N2O using the glucose and 15NH415NO3 as substrates. Part of this N2O and 15N2O was then taken up by beech roots from the soil water and transported to the leaves in the transpiration stream. The 15NH415NO3 added to the soil was approximately 5% enriched with 15N, whereas the emitted N2O into the shoot chamber was approx. 6% enriched with 15N for S2 and 0.4% enriched for S1. The results from S2 indicate that the emitted N2O was derived solely from the added inorganic nitrogen, whereas the results from S1 indicate that an additional unlabelled inorganic N pool already present in the pot–root environment was available for N2O production.

In the fertilization experiment, the shoot emissions are a sum of all processes that could lead to N2O emissions from the leaves and shoot, such as transpiration, N2O formation in the leaves and N2O diffusion through the bark (Rusch & Rennenberg, 1998; Smart & Bloom, 2001; Hakata et al., 2003). We considered gaseous N2O diffusion through the bark of the beech seedlings as a minor factor contributing to the total N2O emissions, because beech trees lack aerenchyma structure in the stem to facilitate gas diffusion from the stem to the atmosphere. However, part of the N2O emitted from the leaves may have been produced inside the leaves during NO2 assimilation, as reported by Smart & Bloom (2001) and Hakata et al. (2003).

To test whether transpiration could account for the N2O emissions from beech leaves, we conducted another experiment. In this experiment, the beech roots were exposed to elevated concentration of N2O in the root solution. The solution in the root compartment was made with deionized water, which is depleted with NO3. This way, we were able to eliminate the possible NO3 assimilation and consequent production of N2O in the leaves. As a water-soluble gas, N2O can theoretically be taken up by roots of the trees and transported to the leaves via transpiration stream. We discovered that N2O was taken up by beech roots and transported in the transpiration stream to the leaves and further to the atmosphere.

High N2O concentrations in the leaves of the laboratory seedlings confirmed that the beech leaves were sources of N2O to the atmosphere. Our finding that the leaf N2O concentrations were much higher in the laboratory seedlings than in the forest trees indicates that the natural N2O emissions are much smaller than those measured in the laboratory. However, the leaf N2O concentrations in the beech forest were higher than the ambient atmospheric N2O concentration, indicating that beech leaves may also be sources of N2O in the forest.

Using the leaf area index (LAI) value of 5 measured at the Lille Bøgeskov beech forest (P.T. Sørensen, Risø National Laboratory, pers. comm.), the observed leaf-based N2O emissions from the fertilization and the root chamber experiments correspond to area-based emissions of 5.2 and 10.0 µg N2O-N m−2 h−1, respectively. These rough estimates of the canopy emissions in a forest ecosystem are of the same order of magnitude as soil-derived N2O emissions from north and central European forest ecosystems (Ambus et al., 2001; Beier et al., 2001; Butterbach-Bahl et al., 2002). In the root chamber experiment, the beech seedlings were exposed to a root solution containing 2300–10 400 µg of N2O l−1 of solution. These N2O concentrations were similar to what Chan et al. (1998) used in their studies with canola and barley and also similar to the concentrations measured from agricultural soils (Heincke & Kaupenjohann, 1999). However, the concentrations were 1–3 orders of magnitude higher than those measured from natural forest ecosystems (Heincke & Kaupenjohann, 1999; Papen & Butterbach-Bahl, 1999). Hence, our results from the root chamber experiment more likely represent a potential for transpiration-mediated N2O emissions in forest ecosystems. The emission estimate from the 15N fertilization experiment represents a situation of N-affected forest ecosystem and, hence, may apply to forests exposed to high atmospheric N-deposition or fertilization management, such as poplar plantations.

Current emission estimates from forest ecosystems are based on soil enclosure measurements and the effect of forest canopies on N2O emissions has been ignored. Our results indicate that the reported N2O emissions from forest ecosystems are an underestimate of the total N2O emissions from forest ecosystems. Field measurements are needed to quantify the potential N2O emissions from forest canopies. Measurements of N2O concentration in the soil solution and inside tree leaves may help to estimate the magnitude of canopy exchange of N2O. Nevertheless, our experiments demonstrate the existence of a previously unknown pathway of N2O to the atmosphere in forest ecosystems, and bring about a need to investigate the magnitude of this phenomenon in larger scales.


European Science Foundation (ESF) Stable Isotopes in Biospheric-Atmospheric Exchange (SIBAE) program, Nordic Centre of Excellence, Research Unit Biosphere–Aerosol–Cloud–Climate Interactions (BACCI), European Commission through the project NOFRETETE (EVK2-CT2001-00106) and the Academy of Finland are acknowledged for financial support.