Vertical structure and diurnal variability of ammonia exchange potential within an intensively managed grass canopy

Stomatal ammonia compensation points ( χs) of grass species on a mixed fertilized grassland were determined by measurements of apoplastic [NH +4 ] and [H ] in the field. Calculatedχs values were compared with in-canopy atmospheric NH3 concentration ( χa) measurements. Leaf apoplastic [NH+4 ] increased by a factor of two from the lowest level in the canopy to the top level. Bulk leaf [NH+4 ] and especially [NO − 3 ] slightly increased at the bottom of the canopy and these concentrations were very high in senescent plant litter. Calculated χs values were below atmosphericχa at all canopy levels measured, indicating that the grassland was characterized by NH 3 deposition before cutting. This was confirmed by the χa profile, showing the lowestχa close to the ground (15 cm above soil surface) and an increase inχa with canopy height. Neither χs nor χa could be measured close to the soil surface, however, the [NH +4 ] in the litter material indicated a high potential for NH 3 emission. A diurnal course in apoplastic [NH +4 ] was seen in the regrowing grass growing after cutting, with highest concentration around noon. Both apoplastic and tissue [NH +4 ] increased in young grass compared to tall grass. Following Correspondence to: A. Neftel (albrecht.neftel@art.admin.ch) cutting, in-canopy gradients of atmospheric χa showed NH3 emission but since calculated χs values of the cut grass were still lower than atmospheric NH 3 concentrations, the emissions could not entirely be explained by stomatal NH 3 loss. High tissue [NH+4 ] in the senescent plant material indicated that this fraction constituted an NH 3 source. After fertilization, [NH+4 ] increased both in apoplast and leaf tissue with the most pronounced increase in the former compared to the latter. The diurnal pattern in apoplastic [NH +4 ] was even more pronounced after fertilization and calculated χs values were generally higher, but remained below atmospheric [NH3].


Introduction
Several investigations have revealed the bidirectional character of NH 3 exchange between vegetation and the atmosphere with large fluctuations on annual, seasonal and daily time scales (Sutton et al., 1995(Sutton et al., , 2007Bussink et al., 1996;Herrmann et al., 2001;Horvath et al., 2005;Walker et al., 2006). In a non-fertilized managed grassland in The Netherlands, NH 3 emission fluxes were frequent (about 50% of the time) during a warm and dry summer period, while in a wet and cool autumn period deposition fluxes dominated (80% of the time; Kruit et al., 2007).
Published by Copernicus Publications on behalf of the European Geosciences Union.
The direction of the NH 3 flux between plant leaves and the atmosphere depends mainly on the stomatal NH 3 compensation point (χ s ) of leaves, which is the atmospheric NH 3 concentration where NH 3 emission and deposition are balanced and no net exchange occurs (Farquhar et al., 1980;. On the canopy level the apparent compensation point will also be influence by all other surfaces (soil, leave litter) and will depend on the pH of these surfaces Flechard et al., 1999) In chamber studies χ s was shown to be influenced by the N status of the plant (Sharpe and Harper, 1995;Mattsson et al., 1998;Mattsson and Schjoerring, 2002;Sommer et al., 2004) and by environmental factors such as temperature (Mattsson et al., 1997), photosynthetic photon flux density and air humidity (Mattsson and Schjoerring, 1996;Husted et al., 2002).
Measurements of vertical NH 3 concentration gradients within a grass/clover canopy (Denmead et al., 1976) and a quackgrass (Agropyron repens L.) canopy (Lemon and van Houtte, 1980) showed a sharp increase of the NH 3 concentration towards the soil surface, resulting in a upward NH 3 flux from the soil to the base of the grass canopy. Similarly, a more recent study based on the inverse Lagrangian source/sink analysis for an oilseed rape (Brassica napus) canopy also revealed highest NH 3 concentrations at the ground level, which was suggested to originate from decomposing litter leaves (Nemitz et al., 2000). This was supported by a very high ammonium (NH + 4 ) concentration measured in senescent plant material from oilseed rape compared to the concentration in intact leaves (Husted et al., 2000). It is not known whether corresponding NH + 4 gradients between leaves of different age may occur in perennial grass species.
A diurnal pattern of the NH 3 exchange has been observed in Brassica napus (Husted et al., 2000), barley (Schjoerring et al., 1993) and grassland (Trebs, et al. 2006), with highest NH 3 emission rates typically occurring during the daytime and low rates at night. Reported diurnal variations in apoplastic NH + 4 and H + concentrations are small (Husted et al., 2000;van Hove et al., 2002). Consequently changes in NH 3 emission were attributed to temperature effects on NH 3 solubility and NH + 4 dissociation in the apoplast due to varying canopy temperature during the diurnal course . In addition, fluctuations in leaf surface wetness will affect the NH 3 exchange (Walker et al., 2006;Kruit et al., 2007). Diurnal variations of NH 3 emission have also been observed over grassland, but correlation between the measured atmospheric χ a and χ s , calculated from flux density measurements, was low .
The experiment presented here was carried out in May and June 2000 in Braunschweig, Germany and was part of a joint investigation within the EU GRAMINAE project (for a detailed description of the experiment see Sutton et al., 2008). The aim was to estimate the NH 3 exchange potential of the vegetation on a vertical gradient within a fertilized grass canopy and its diurnal variations by means of χ s mea-surements. The vacuum infiltration technique for apoplast extraction was directly applied in the field and calculated χ s was related to in-canopy NH 3 concentrations. It is discussed whether leaf bulk tissue [NH + 4 ] could be a useful indicator of χ s , since measuring this parameter would be more convenient and less time-consuming than the determination of χ s .
A priori knowledge of χ s or a simple parameterisation of it is important for modelling NH 3 exchange in ecosystem models using the canopy compensation concept.

Description and management of the measurement site
The measurement site was located near Braunschweig (52 • 18 N , 10 • 26 E , 79 m a.s.l.) in Lower Saxony, Germany. The field was 600×300 m in size and consisted of a mixed sward dominated by Lolium perenne L. It has been an intensively managed grassland for 4 years, typically receiving 250 kg N ha −1 a −1 . Prevailing wind directions were SW to W and E. A farm with 300 cattle and 3000 pigs was located in the W of the field. The field was cut on 29 May and N fertilizer (100 kg N ha −1 ) was applied as calcium ammonium nitrate on 5 June.

NH 3 concentration measurements
Instruments for the measurement of χ a were placed in the centre of the field. χ a was measured continuously on-line by Mini Wet Effluent Denuders (mini-WEDD), as described by Neftel et al. (1998), connected to a four-channel fluorescent analyzer. Before cutting three of the Mini-WEDDs were placed within the plant canopy and one directly above the canopy. Air flow rates of 200 ml min −1 and 800 ml min −1 were used for the lowest two mini-WEDDs and for the two above, respectively. A liquid flow of 0.12 ml min −1 was used and the detection limit was 0.1 µg NH 3 m −3 .

Sampling of plant material
During the first period of the experiment, a few days before the field was cut on 24 and 25 May, plant material was collected from different layers within the plant canopy and separated into flowers, stems and leaf sheaths and green and brown leaf laminae.
The samples for the analysis of the diurnal variation have been taken on the 26 May, while samples for the in canopy profiles have been taken on the 29 May in an uncut small plots. The cut of the whole field took place in the morning of the 29 May.
The fully developed green leaf laminae were used for apoplast extraction as described below. After the cut it was no longer possible to properly divide plant material into different species. Therefore a mixture of cut leaves from all Biogeosciences, 6, 15-23, 2009 www.biogeosciences.net/6/15/2009/ the species was collected. The plant material was randomly collected in the field and immediately brought to an adjacent field lab. Some of the leaves were used for extraction directly after sampling and the plant material used for the determination of tissue NH + 4 and NO − 3 was immediately frozen in liquid nitrogen and stored at −20 • C.

Apoplast extraction
Apoplast liquid was extracted by means of vacuum infiltration (Husted and Schjoerring, 1995) modified as follows: Whole leaf laminas were infiltrated with 280 mM sorbitol solution at a pressure of 16 bar and under vacuum for 5 s. This procedure was repeated 5 times. After infiltration, solution on leaf surfaces was removed by use of paper towels, where upon the leaves were packed into plastic bags and left to equilibrate for 20 min in daylight in order to reach complete homeostasis of the apoplastic NH 4 + concentration. Thereafter the leaves were centrifuged for 10 min at 4 • C and 800 g. During the night the samples were extracted in the same way as during the day, but green artificial light was used instead of white light.
Concentrations of NH + 4 in the extracted solution were determined by flow injection analysis (FIA) or HPLC analysis (Waters Corp., Milford, USA) using o-phthalaldehyde (OPA) as reagent as described by Genfa and Dasgupta (1989). pH of the diluted apoplastic solution was measured with a Micro-Combination pH electrode (type 9810, Orion, Beverly, USA). It is assumed that the dilution with sorbitol is not changing the pH. In order to assess cytoplasmic contamination of the apoplasts, malate dehydrogenase (E.C. 1.1.1.38) activity was determined and compared with the activity measured in bulk leaf extracts (Husted and Schjoerring, 1995). Cytoplasmic contamination was below 1.5% for all considered plant species.

Stomatal NH 3 compensation points
The stomatal NH 3 compensation point (χ s , mol NH 3 mol −1 air or ppbV) χ s , was calculated by use of Eq. (1) derived from  taking into account that K d [H + ] apoplast within the range of apoplastic pH values: is the dimensionless ratio between the apoplastic NH 4 + and H + concentrations, and K H and K d are thermodynamic constants of 10 −1.76 l mol −1 and 10 −9.25 mol l −1 at 25 • C, respectively. values represent a measure of the NH 3 exchange potential independent of temperature The calculated χ s at 25 • C (T ref ) was adjusted to the actual canopy temperature T a by the following equation derived from : χ s T a is the NH 3 compensation point at the actual canopy temperature T a ( • K), H 0 dis the enthalpy of NH + 4 dissociation (52.21 kJ mol −1 ), H 0 vap the enthalpy of vaporization (34.18 kJ mol −1 ), and R the gas constant (0.00831 kJ K −1 mol −1 ).
Stomatal compensation points are normally expressed as dimensionless mol fraction, whereas atmospheric ammonia concentrations are expressed in this special issue as µg m −3 . Conversion of the mol fraction into concentrations is given by 0.2 g of the frozen plant material was homogenized to powder and was extracted in 2 ml 10 mM formic acid in a cooled mortar containing a little quartz sand. The extract was centrifuged at 25 000 g and 4 • C for 10 min. The supernatant was transferred to 500−µl 0.45µm polysulphone centrifugation filters (Micro VectraSpin; Whatman Ltd., Maidstone, UK)

Vertical structure of NH 3 exchange potential
In order to characterise the vertical structure of [NH + 4 ] and [NO − 3 ] of the plants, plant material was collected from four different layers on the same day when the field was cut (29 May). The fully developed canopy was 76 cm high at that stage. Green leaf laminae, which were used for apoplast extraction, were found in all the layers except in the top level (60-70 cm) (Fig. 1). Brown senescent leaves constituted an additional fraction in the lowest canopy layer (0-20 cm), but uncontaminated apoplast liquid could not be obtained from this fraction. Apoplastic [NH + 4 ] was more than double in the leaves occurring at the upper layer of the plant compared to the lowest canopy level (Fig. 2a). Due to the relatively large variability between the replicates, the increase cannot be well quantified. Leaf apoplastic pH ranged between 6.3  (n=8±SE) whereas for the other levels a mixture of all species was considered (n=4±SE). χ a represent mean concentrations over three days before cutting (10:00 a.m.-16:00 p.m.). and 6.6 in all the layers (Fig. 2b). Tissue [NH + 4 ] was much higher in brown senescing leaves close to the soil surface compared to green leaves at the same canopy height (Fig. 2c).
[NO − 3 ] of stems and green leaves decreased with canopy height (Fig. 2d) and was highest in the stems except in the layer closest to the ground where [NO − 3 ] was higher in the leaves. Similar to apoplastic [NH + 4 ], χ s increased by a factor of two from the bottom to the top layer. Values were below the measured in-canopy χ a (Fig. 3).

Diurnal course of NH 3 exchange potential
Before the cut, the most abundant plant species Lolium perenne and Phleum pratense were selected for determination of the NH 3 exchange potential during a diurnal course.  Fig. 4a and c did not show any particular pattern whereas apoplastic pH was higher during the night than during the day (Fig. 4b). After the field was cut, apoplastic [NH + 4 ] of grass leaves was generally higher and a distinct diurnal course could be seen on the first day, with highest apoplastic [NH + 4 ] before noon and a decrease during the night (Fig. 4a). However, apoplastic [NH + 4 ] remained low on the following day. parallel to the lower canopy temperature on the second day compared to the day before. However, the increase in [NH +

contrast, [NO −
3 ] seemed to decrease during the day and an increase was observed during the night (Fig. 5b). Like before the cut, highest apoplastic pH was measured in the night (Fig. 4b). Due to generally lower apoplastic pH of the cut grass mix compared to the grass before cutting was similar before and after the cut (Fig. 4c). After fertilization [NH + 4 ] increased in both the apoplast and the tissue (Figs. 4a and 5a). The diurnal pattern in apoplastic [NH + 4 ] and was more pronounced after N application than before. Before fertilization a relatively good correlation was seen between leaf tissue and apoplastic [NH + 4 ], which was significant (p<0.01) after cutting but not before cutting (Fig. 6). Because apoplastic [NH + 4 ] increased while tissue [NH + 4 ] was rather unaffected after fertilization, the correlation between tissue and apoplastic [NH + 4 ] was very low. Before the field was cut the vertical profile of χ a was predominantly characterised by decreasing χ a towards the ground as shown for a diurnal course in Fig. 7. This χ a profile would therefore indicate NH 3 deposition from the atmosphere to the plant canopy. Calculated χ s of both Lolium perenne and Phleum pratense, which corresponded to the upper two χ a measuring heights, were below the in-canopy χ a . The increase in χ a during the night was not reflected in χ s . An inverseχ a profile was observed after the canopy had been cut. At the lowest measuring height χ a reached  10 µg m −3 in the morning and χ a decreased with measuring height (Fig. 8). χ a was lower during the night than during the day. Accordingly, highest NH 3 emission was measured during the day . Generally, χ s of the cut grass were much lower than χ a above the plant canopy. The same direction of the slope of the vertical χ a gradient but higher concentrations during the day was seen after N application (Fig. 9). A typical diurnal pattern with highest  concentration around noon was most pronounced after fertilization and was reflected in both calculated χ s and atmospheric χ a . Although χ s of the fertilized grass were about five times higher than before fertilization the values were still below atmospheric χ a of the lowest measuring height during the whole diurnal course.

Discussion
Application of the vacuum infiltration technique directly in the field enabled an immediate extraction of apoplast liquid and therefore frequent determination of the NH 3 exchange potential of the plants during a diurnal course. The measured apoplastic NH + 4 levels before fertilization were about 0.1 mM (Fig. 4a) matching values reported in pastures Fig. 9. Diurnal χ a gradient above the canopy and calculated χ s for grass stubbles 7 days after fertilization (12/13 June). χ s are means of 4 replicates ±SE. The dark period is indicated by the shaded area. under similar N conditions by Herrmann et al., 2001 and(Loubet et al., 2002. Considerably higher apoplastic NH + 4 concentrations, 0.2 to 0.9 mM, were observed in an intensively managed grassland in The Netherlands throughout the growing season (van Hove et al., 2002). The nitrogen availability in the soil, particularly that of ammonium, has a profound influence on apoplastic NH + 4 concentrations as also demonstrated by the increase following fertilization (Fig. 4a) (Mattson et al., 2008).
Determination of apoplastic [NH + 4 ] and pH is a labour intensive analysis. Consequently the analysis of the diurnal structure and the analysis of the vertical profiles were performed on different days. In the following discussion we assume that the determined values are representative for the grass canopy for the days before the cut. The vertical profile was measured in a remaining uncut plot on the same day as the rest of the field was cut. For the comparison with the atmospheric NH 3 in canopy concentration the mean values of them of the previous three days during daytime (10:00-16:00) have been taken and are shown in Fig. 3.
Apoplastic [NH + 4 ] and χ s increased by a factor of two from the bottom to the top of the intact plant canopy (Figs. 2a  and 3). Thus, young leaves had a relatively high NH 3 emission potential. At all in-canopy levels considered, χ s was below the measured atmospheric χ a , indicating that plants acted as NH 3 sinks. This was confirmed by the measured NH 3 flux which was characterized by NH 3 deposition (see Milford et al., 2008) and is in agreement with measurements carried out over a grass/clover canopy (Herrmann et al., 2001).
After the cut the apoplastic [NH + 4 ] decreased on the second day (Fig. 4) in parallel with the canopy temperature. This points to a temperature dependent physiological control of the apoplastic [NH + 4 ], because lower temperature would be in favour of higher apoplastic [NH + The NH 3 emission measured from the field after the cut (see Milford et al., 2008) could not be totally explained by a rise in χ s of the cut grass. χ s of the senescent plant material either attached to the stubbles or lying on the ground, however, could not be calculated since apoplastic infiltration of senescent plant material could not be achieved. Yet, very high tissue [NH + 4 ] measured in plant litter, which accounted for about 20% of the total above ground biomass after the cut, indicate that this fraction may represent an important NH 3 source. This might explain the NH 3 emission measured after cutting, when the litter fraction was not covered by a canopy and no re-capture by the intact leaves could occur anymore. Husted et al. (2000) showed that in an oilseed rape field, the plant litter fraction represented an NH 3 source, while intact leaves acted as NH 3 sinks. Similarly, in a grass/clover crop the highest in-canopy χ a was found towards the soil surface (Denmead et al., 1976). In the present investigation atmospheric NH 3 could not be measured below 15 cm and therefore NH 3 concentration directly above the soil surface is not known. However, using a tissue [NH + 4 ] value for brown leaves as presented in Fig. 2c and a measured pH of 7 (data not shown) would result in values for the litter of about 5000. Although this value cannot be considered as a direct measure of the effective NH 3 emission of plant litter it still indicates a high potential for NH 3 emission. Furthermore, NH 3 flux measurements carried out in a climate chamber study revealed a NH 3 emission of about 170 ng m −2 leaf area s −1 from cut senescent leaf material of Lolium perenne (Mattsson and Schjoerring, 2003). This would result in a NH 3 emission of about 80 ng m −2 s −1 using the amount of litter biomass per surface area of 20% of total as measured in the present investigation. While plant litter emission could explain the measured NH 3 emission after the cut it cannot entirely account for the high emission observed after fertilization. Directly after N application most of the NH 3 emission most probably originated from fertilizer particles lying on the ground (Herrmann et al., 2001). Yet, the NH 3 emission measured over the following days and its distinct diurnal pattern indicate that another NH 3 source than fertilizer must be involved. Although χ s of the grass considerably increased after fertilization (Fig. 4c) it still remained below measured atmospheric χ a and thus plants should represent an NH 3 sink.
A discrepancy between micrometeorological or cuvette studies and the bioassay approach in estimating χ s has been observed in several investigations. In most of these studies the bioassay approach yielded smaller estimates of χ s compared to the micrometerological or cuvette measurements (Mattsson et al., 1997;Hill et al., 2001;Mattsson and Schjoerring, 2002). Non stomatal exchange might be a reason for the observed discrepancies. Bioassay studies are a measure for the equilibrium NH 3 concentration in the stomatal cavity, whereas micrometeorological and cuvette measurements are indicating the NH 3 concentration in the surrounding atmosphere of the plants Considering a possible underestimation of χ s in the present study, NH 3 emission from the plants would become likely, especially after cutting and fertilization around midday, when the ratio between χ a and estimated χ s was smaller than during the rest of the day. However, the discrepancy between χ a and estimated χ s was still considerable for most of the collected data, indicating that also after fertilization other NH 3 sources might be involved in the NH 3 exchange of the canopy.
The diurnal measurements clearly showed that apoplastic [NH + 4 ] may change during the course of the day, with highest values around midday and decreasing concentrations during the night. This pattern was also reflected in which is an indicator for the NH 3 exchange potential of a plant but in contrast to χ s , it is independent of any change in canopy temperature. This is different from observations made in an oilseed rape field, where no diurnal variation in existed and where canopy temperature was the only factor influencing χ s on a diurnal scale (Husted et al., 2000).
Before fertilization a relatively clear linear relationship existed between leaf tissue [NH + 4 ] and apoplastic NH + 4 (Fig. 6), but this was not the case after fertilization. In addition, the ratio between tissue [NH + 4 ] and apoplastic [NH + 4 ] was much lower after fertilization compared to before fertilization. These findings differ from studies in a Scottish grassland, where the magnitude of increase in [NH + 4 ] after cutting was similar for the apoplastic and bulk tissue fraction (Loubet et al., 2002). Also in two grass species grown with different N supply the correlation between apoplast and leaf tissue [NH + 4 ] was fairly good  while in a wild perennial the same correlation was poor . The data presented here indicate that [NH + 4 ] in the tissue and in the apoplast may be regulated independently and thus the tissue [NH + 4 ] can not always be used as an indicator of χ s .

Conclusions
From the present investigation we conclude that the plants of a fully developed grassland acted as NH 3 sinks and that NH 3 was predominantly deposited to the tall canopy. NH 3 emission measured after the cut and after fertilization could not entirely be accounted for by stomatal loss. Yet, elevated tissue [NH + 4 ] and high values in especially senescent plant material indicated that NH 3 might be emitted from plant litter, which could explain the NH 3 emission measured after cutting. Although  showed a high inter-species correlation between and bulk leaf [NH +