deposition of nitrogen-containing compounds to an

ELSEVIER 0 2 6 9 - 7 4 9 1 ( 9 5 ) 0 0 0 3 6 - 4

Environmental Pollution, Vol. 91, No. 1, pp. 21 34, 1996 Copyright ©1995 Elsevier Science Limited

Printed in Great Britain. All rights reserved 0269-7491/96 $09.50 + 0.00

DEPOSITION OF NITROGEN-CONTAINING COMPOUNDS TO AN EXTENSIVELY M A N A G E D GRASSLAND

IN CENTRAL SWITZERLAND

R. Hesterberg, a A. Blatter, a M. Fahrni , a M. Rossett i A. Neftel , a* W. Eugster b & H. W a n n e r b

~Swiss Federal Research Station for Agricultural Chemistry and Environmental Hygiene, CH-3097 Liebefeld-Bern, Switzerland blnstitute of Geography, University of Bern, CH-3012 Bern, Switzerland

(Received 17 August 1994; accepted 1 May 1995)

Abstract During four intensive observation periods in 1992 and 1993, dry deposition of nitrogen dioxide (NO2) and ammonia (NH3), and wet deposition of nitrogen (N) were determined. The measurements were carried out in a small, extensively managed litter meadow surrounded by intensively managed agricultural land. Dry deposition of NHs was estimated by the gradient method, whereas eddy correlation was used for NO2. Rates of dry deposition of total nitrate ( = nitric acid (HNOs) + nitrate (NO3-)), total nitrite (= nitrous acid (HONO) + nitrite (NO2-) ) and aerosol-bound ammonium ( NH4 + ) were estimated using deposition velocities from the literature and measured concentrations. Both wet N deposition and the vertical NH3 gradient were measured on a weekly basis during one year.

Dry deposition was between 15 and 25 kg N ha -1 y-l, and net wet deposition was about 9.0 kg N ha -1 y-1. Daily average NO2 deposition velocity varied from 0.11 to 0.24 cm s -1. Deposition velocity of NH3, was between 0.13 and 1.4 cm s -1, and a compensation point between 3 and 6 ppbV NH3 (ppb = 10 -9) was found. Between 60 and 70% of dry deposition originated from NH3 emitted by farms in the neighbourhood. It is concluded that total N deposition is exceeding the critical load for litter meadows, is highly correlated to local NH3 emissions, and that NH~ is of utmost importance with respect to possible strategies to reduce N deposition in rural regions.

Keywords: Dry nitrogen deposition, wet nitrogen deposition.

INTRODUCTION

The increase in the emission of the primary NO and NH3 due to anthropogenic causes strongly increased total N deposition, which may severely influence the N status of natural ecosystems. Accumulation of inorganic N leads to soil acidification (van Breemen et al., 1982; Stumm et al., 1982; Fabian, 1987) followed by an increase in NO3- leaching into the groundwater and

*To whom correspondence should be addressed.

21

acidification of surface waters as well as a washout of plant nutrients, such as potassium, calcium and mag- nesium (Ulrich & Pankrath, 1983; Fabian, 1987; Bobbink et al., 1992). Furthermore, N acts as fertilizer and may damage plants and decrease biodiversity (Ellenberg, 1985, 1989; Mohr, 1993a,b, 1994; Bobbink et al., 1988).

Before strategies to reduce N loading of ecosystems can be developed, it is necessary to know the amount and composition of N inputs. Within the Cooperative Program for the Monitoring and Evaluation of the Long Range Transmission of Air Pollutants in Europe (EMEP), the concept of critical loads has been introduced. Nilsson and Grennfelt (1988) defined the critical load as 'a quantitative estimate of an exposure to one or more pollutants below which significant harmful effects on specified sensitive elements of the environment do not occur according to our present knowledge'. The critical load has been set at 5-10 kg N ha-i y-1 for moors, at about 15 kg N ha -1 y l for unman- aged grasslands and at 20-30 kg N ha 1 y-i for forests (Grennfelt & Th6rnel6f, 1992). In recent years, exten- sive efforts have been made to produce regional maps of the critical load of N as a function of vegetation type and soil characteristics. To calculate the exceedance of critical loads, accurate information concerning the amount and speciation of N deposition is now required.

Numerous results from flux measurements of gaseous N species are reported in the literature (Erisman et al., 1989; Hanson et al., 1989; Stocker et al., 1993; Meixner et al., 1985). However, in regions with complex topo- graphy suitable field sites with sufficient fetch hardly exist. In Switzerland, semi-natural ecosystems are typi- cally located in hilly areas not suitable for intensive farming, or they are small protected areas surrounded by intensively managed agricultural land. In spite of these difficulties, it is necessary to obtain reliable esti- mates of the yearly N deposition under such conditions. In this paper, we report results from measurements of the amount and speciation of total N deposition to an extensively managed litter meadow located in an area surrounded by intensive agriculture.

22 R. Hesterberg et al.

8°°II300

0 5 10m

M3

M4

M2 • u m H | n l

6 0 1.

30.30 32.30 35.50 36.00 38.00 62.00

~----9.10

r- ' l

R

MI "

P

i D

Fig. 1. Arrangement of instruments during IOP-2 (distances in m). D = direction of nearest road, H = hedge, MW = instrument trailer, R = wet deposition collecter, M1 = measuring mast, L = lift, M2 = mast for NHH, HONO and HNO3 sampling, G = container for NH3, HONO and HNO3 instruments, M3 = sonic anemometer mast, M4 = mast for wind speed, temperature and

humidity profiles, and location of Bowen-ratio measuring system, B = drainage canal.

I N S T R U M E N T A T I O N

Experimental site The exper imenta l site was loca ted in the lower Reuss valley near the village o f Merenschwand (20 k m SW o f Zurich, 385 m a.s.1.). The li t ter m e a d o w (Junco-Mol i - n ie tum typicum; K16tzli, 1969) was a small p ro tec ted area o f a b o u t 60 m x 350 m with some scat tered Alde r Buck thorn (Frangula alnus, 25% o f area). The grass and bushes were harves ted once in the fall. The long axis o f the field was or iented N W - S E (Fig. 1). To the NE, the site was separa ted f rom ad jacen t fields by a hedge o f 5-6 m height and 5 m width. To the SW, the l i t ter m e a d o w was separa ted f rom ne ighbour ing fields by a d ra inage canal o f a b o u t 1 m dep th and 2 m width. Pre- vai l ing winds were f rom N W (up-val ley winds) and SE

(down-va l ley winds), i.e. para l le l to the Reuss val ley (Hut t e r & Bfirki, 1991), and in the ideal d i rec t ion relative to the measur ing site. Da i ly mean speed o f wind f rom the ma in di rect ions and f rom N E were genera l ly be low 2.5 m s -l, and f rom all o ther d i rect ions be low 1.0 m s -~. Best fetch o f over 200 m existed for the wind sector 2850-345 ° (NW). Due to he te rogeneous proper t ies

o f soil, vegeta t ion, and land managemen t , the measur - ing site represented a typical non- idea l terra in , according to the descr ip t ion by Hicks and M c M i l l a n (1988).

F r o m June 1992 to M a y 1993, ver t ical profiles o f NH3 concent ra t ion , wet depos i t ion , and meteoro log ica l pa rame te r s (solar rad ia t ion , air t empera tu re and humidi ty , wind speed and direct ion, air pressure) were recorded cont inuously . The profile o f NH3 concent ra - t ion and wet depos i t ion were de te rmined on a weekly

Table 1. Summary of the instruments used during the four intensive observation periods (lOP), measuring heights in m, and characteristics of the vegetation cover

Parameter Method Instrument IOP-1 June IOP-2 August IOP-3 February IOP-4 May 1992 1992 1993 1993

NO2 flux Eddy correlation LMA-3, sonic anemometer 2.05 2.59 2.11 2.00 NH3 flux Gradient method WEDD, FIA 0.89 / 2.61 1.60 / 3.20 no 0.90 / 2.60 HNO3, HONO flux Inferential method WEDD, IC no 1.60 / 3.20 no 0.90 / 2.60 NO3 flux Inferential method WEDD no no 0.86 / 2.50 0.90 / 2.60 NO2 flux Inferential method WEDD no no 0.86 / 2.50 0.90 / 2.60 NH4 ÷ flux Inferential method WEDD no no 0.86 / 2.50 0.90 / 2.60 NH3 flux Gradient method Passive samplers During one year NO3-, NH4 ÷ wet deposition Wet deposition AM 301 During one year Wind speed profiles, Aerodynamic method 6-cup-anemometers 0.57 - 2.61 1.14 - 4.30 0.30 - 4.20 0.50 - 4.22 Temperature profiles, Thermo conductors Mass transfer coefficient Mass transfer coefficient Eddy correlation Sonic anemometer sonic 2.05 2.59 2.11 2.00

temperature Mass transfer coefficient Bowen ratio method Bowen-ratio-system 0.57 - 2.61 1.60 / 3.20 no 0.67 - 3.97 Height of vegetation Grass (75 % cover) 0.30 m 0.60 m < 0.05 m 0.25 m Bushes (25% cover) 0.50 m 1.40 m < 0.05 m 0.25 m

no = not measured

Deposition o f N-containing compounds 23

basis. Rates of dry deposition of NO2 were determined during intensive observation periods (IOP) in June 1992 (IOP-1), August 1992 (IOP-2), February 1993 (IOP-3), and May 1993 (IOP-4), and rates of dry deposition of NH3 were measured in IOP-I, IOP-2 and IOP-4. The attempt to determine dry deposition rates of HNO3 and HONO was not successful. During IOP-3 and IOP-4, concentrations of aerosol-bound NO3-, NO2- and NH4 + have been measured together with the respective gases. Table 1 summarizes the techniques and instruments used and the condition of the vegetation during each IOP. Deposition of NH3, HONO and HNO3 was measured by the gradient method, whereas the NO2 flux was determined by the eddy correlation technique. Fig. 1 shows the arrangement of the instruments during IOP-2.

NO2 measurements NO2 was measured with a LMA-3 detector (Scintrex/ Unisearch) with a detection limit of 50 pptv (Schiff et al., 1986; Drummond et al., 1989a,b, 1991; Fehsenfeld et al., 1990). Comparisons with other NO2 detectors are reported in literature (Gregory et al., 1990; Fehsenfeld et al., 1990; Fahey et al., 1986). The LMA-3 is fast enough (time constant 0.3 s) to be used for NO2 flux determinations in combination with a sonic anemometer (see below). The following modifications to the original instruments were necessary, the first two allowed eddy correlation measurements:

(1) Stronger pumps were used for both the solution (ISMATEC Reglo 100, 8 rolls, maximal pressure difference 700-800 mbar) and for air (Brey G12/ 45, maximal pressure difference about 620 mbar) regulated by a flow sensor. In order to avoid flow of the luminol solution to the air exhaust when the air inlet was blocked, the pressure difference between the reaction chamber and ambient conditions produced by the air pump had to be smaller than the pressure difference produced by the liquid pump.

(2) The air flow through the reaction chamber was set to 2.2 litre min ~. At flow rates higher than 2.5 litre min -l the window between the wick and the photomultiplier could have become contaminated by luminol.

(3) In order to prevent temperature effects during field measurements, the LMA-3 was enclosed in a thermally insulated box and kept at about 10°C above the maximum daily temperature.

(4) An absorber for NO2 and O3 was placed in the air inlet of the luminol feed bottle.

Calibration of the LMA-3 was carried out every second or third day with zero air and span gas between 2 and 20 ppbv NO2. Reproducibility was better than 10 % of the NO2 span value. A variability of the luminol sensitivity of up to 15 % must be expected (Kelly et al., 1990). In order to correct for the non-linearity of the calibration curve in the low ppbv range, multiple point calibration was necessary (Hesterberg, 1994).

The reaction of luminol and NO2 is known to be interfered with 03 and PAN. According to Drummond et al. (1989b) this interference can be described by the equation:

£i-Cl Ck ~-Cl -- So Co3 -- C~pCpAN (1)

ci

where Ck = corrected NO2 concentration, c~ = linearly corrected LMA-3 signal, ao = 0.01 ppbv NO2 (ppbv O3) -l, ci = 15 ppbv NO2 and ap = 0.25 ppbv NO2 (ppbv PAN) -l. However, Gehrig and Baumann (1992) reported a PAN interference of up to ap = 0.5 ppbv NO2 (ppbv PAN) -1. Our own determination of the parameters ci and So yielded an 03 interference of So = 0.008 ppbv NO2 (ppbv O3) -l and ci = 6.2 ppbv. Some luminol batches purchased in 1991 (batch numbers less than 920300) showed an interference of up to ao -- 0.06 ppbv NO2 (ppbv O3) -1.

Determination of NH3, HNO3 and HONO NH3, HONO and HNO3 were sampled with the newly developed wet effluent diffusion denuder (WEDD) technique (Vecera & Dasgupta, 1991). NO 3- and NO2- were analysed by ion chromatography (IC) (Genfa & Dasgupta, 1989). The response time for the system was 5 min. The detection limit of the wet denuder for HNO3 and HONO was 0.05 ppbv. The precision of the signal was 5 % at 1 ppbv in ambient air. NH4 + was analysed as the fluorescence signal produced by reaction with o-phthalaldehyde in the presence of sodium sulphite (Genfa & Dasgupta, 1989). With a 1-min averaging period the detection limit of the continuous-flow instrument (FIA) was 0.1 ppbv. A two-channel instrument was used for gradient measurements.

Determination of aerosol-bound NH4 +, NO3 and NO2 In order to measure the aerosol-bound quantity of NH4 + , NO3- and NO2 the mist chamber system (MCS) developed by Simon and Dasgupta (1995) was used (see also Blatter et al., 1993) with a collection efficiency for NaNO3 aerosols of 100%, and of 98% for NaESO4 aerosols using 600 mg min -~ of steam (aerosol concentration 2.7 x 10 5 g m -3, mean diameter 2.5 x 10 -7 m). The calculated efficiency based on Henry's law was more than 99 % for gaseous HNO3 and more than 95 % for NH 3. This system may be used to measure total nitrate (HNO3 + NO3- ) and total nitrite (HONO + NO2-). The response time was 5 min with a detection limit of 2 nmol m -3. A single measuring system was used in combination of both the WEDD and the MCS collection systems.

During IOP-2 and IOP-4 the two systems were fixed to two separate platforms mounted to opposite sides of a 6-m long ladder. A computer-controlled motor mounted to the top of the ladder was used to move the two platforms up and down in opposite direction. This allowed the platform to be switched or to be placed at identical height. NH3 concentrations measured at identical levels mostly did not differ by more than 5 %. Larger deviations were observed during rapid

24 R. Hes terberg et al.

fluctuations in concentration. The heights at which measurements were made are listed in Table 1.

NH3 measurements with passive samplers The passive NH3 samplers described by Blatter et al. (1992) were used for NH3 concentration determinations on a weekly basis. Laboratory tests yielded a standard deviation for the slope of the calibration curve of 2.5 %. In field conditions, the passive samplers showed deviations of less than 5 % in comparison with the WEDD. Three to five boxes, each containing four samplers, were mounted at levels between 0.5 and 5.5 m above ground. The precision (2a) was better than 7 %. Outliers (1 value out of 4) were rejected.

Meteorological measurements Self-constructed 6-cup-anemometers described by Ros- set (1990) were used for the measurement of horizontal wind speed. The detection limit was about 0.15 m s l , and stable rotation could be obtained above 0.2 m s -1. Four to six anemometers were installed at the heights given in Table 1. Dry bulb temperature was measured with active semi-conductors (Analog Devices AD590 JF). Deviations between sensors were smaller than + 0.05 ° C.

A sonic anemometer of type Solent (former Gill, Great Britain) was used for the measurement of three- dimensional wind speed fluctuations (Eugster, 1994). The coordinate system of the wind vector measurements was rotated to be in line with the mean streamlines of the averaging period of 30 min (Zeman & Jensen, 1987). The sampling rate was approximately 21 Hz.

Bowen-ratio measurements Description of the technical details of the Bowen ratio measuring system Campell was given by Rosset (1990). Due to the use of water for the determination of the wet bulb temperature, measurements below the freezing point were not possible. Global radiation was measured at 1 m above ground, and three soil heat flux plates were installed at 1.5 cm below the soil surface.

Measurements of wet deposition of nitrogen Rain was sampled on a weekly basis with a wet/dry precipitation collector (Model 301 Aerochem Metrics, Florida). Concentrations of NH4 + and NO3- were determined by ion chromatography. Measurements were repeated when the balance between anions and cations differed by more than 10 % of the sum of ion equivalents. When the second measurement still yielded an imbalance of more than 10%, the data from this sample were rejected. This occurred occasionally with very small rain water samples.

METHODS

Gradient method Under steady-state conditions the vertical flux Fc of a chemically inert substance is independent of height:

F~ = - K A-f Az (2)

where K = turbulent mass transfer coefficient, c = concentration and z = height along the vertical axis. The coefficient K must be calculated from micro- meteorological measurements and the gradient is determined by at least two concentration measurements at different heights. Using the aerodynamic method (Businger et al., 1971; Monteith & Unsworth, 1990; Zeller et al., 1990), K is calculated from the friction velocity u*, the Karman constant k = 0.4, the non- dimensional vertical gradient of mass ~ and the zero plane displacement d by

K = u . k ( z - d ) c k - l (3)

is a function of the Monin-Obukhov-Length L = ¢ (Monin & Obukhov, 1954) and can be parametrized (Panofsky & Dutton, 1984) according to

{ (1 + 5~-d) stable, L~>0

~b= (1 16~d) -1/2 (4) - unstable, L< 0

where 0. is the turbulent potential scaling temperature, g the acceleration due to gravity, and 0 the average potential temperature. With the aerodynamic gradient method, u. and 0. were determined from measured profiles of horizontal wind speed and temperature.

Often, d is assumed to be 0.67 x h (h = vegetation height). The relationship between d and h was tested by calculating d, u., and 0. from vertical profiles of wind speed and temperature by numerical iteration (Hester- berg, 1994; method proposed by Marquardt, 1963). The average of these calculations confirmed the relationship, but d tended to decrease with increasing friction velocity u.. By using d = (a u. + h), a ranged between 2.4 and 4.0 s.

Alternatively, K could be determined directly from 0. by measuring the sensible heat flux H with a sonic anemometer and a fast-response temperature sensor (Businger et al., 1971; Zeller et al., 1990), or with the Bowen-ratio measuring system. Applying the Bowen- ratio method, K is calculated from the energy balance above the surface from net radiation, soil heat flux and profiles of temperature and humidity (Stull, 1988; Monteith & Unsworth, 1990).

Eddy correlation and aerodynamic method yielded similar K values within a range of 20 %. The average K from the aerodynamic method and from eddy correlation measurements was routinely used. In comparison to those two methods a larger scatter was observed in the Bowen-ratio data (Buwal, 1994a), even after testing the quality of the data according to Ohmura (1982). We therefore used the K values from the Bowen-ratio measuring system only in cases where no data was available from the other two methods.

NO2 flux measurements by eddy correlation The turbulent flux of a gas, such as NO2, in any direction can be expressed by the covariance coy < uj,c > of

Deposition of N-containing compounds 25

the time series of the wind speed along the direction of uj and the concentration c. Thus, the vertical turbulent flux of the gas is

F~ : c o v < w,c > =w' c' (5)

where w is the vertical wind speed and w' and c' are the momentaneous deviations from the averages • and ?. Due to the alignment with the mean streamlines (Zeman & Jensen, 1987) the mean vertical wind speed w is always zero, such that eqn (5) is valid. The time-series of concentration was detrended by linear regression, then shifted by 1-2 s relative to the wind speed time series to correct for the lagtime caused by the gas inlet system of the LMA-3. This lagtime was determined from the clear peak of the cross-correlation of w and c.

The LMA-3 is not an ideal instrument for measuring NO2 fluctuations at high frequencies. To estimate the error of the eddy correlation flux measurement obtained from the LMA-3, Eugster and Senn (1995) developed a cospectral correction model which was used to correct the measured fluxes. The underestimation of the NO2 flux was in the range of 11-22 % on average. The correction model used the measured cospectrum by Kaimal et al. (1972). A similar model using a different approach can also be found in Moore (1986). The reported NO2 fluxes in this work were corrected for these damping losses. For the determination of the mean NO2 deposition, the data from the unusable wind sectors (64°-138 ° ) were not used. Furthermore, data were rejected in case of an observed positive momentum flux, u'w'>~ O, where u' is the momentaneous deviation from the average (Eugster et al., 1993). The distance between the points of wind speed measurement and the inlet system of the LMA-3 was 8 cm, which was small enough for eddy correlation measurements at 2 m above ground as suggested by Dyer et al. (1982).

The reproducibility of the LMA-3 was tested during IOP-4 with two instruments working in parallel. Linear regression analysis of the two concentration measurements yielded a slope of 0.979 + 0.006 and an intercept of (0.29 + 0.04) ppbv NO2 (r 2 = 0.96). Regression analysis of NO2 fluxes yielded a slope of (1.058 i 0.007) and an intercept of (0.89 ± 0.10) x 10 -3 ppbv m s -1 (r 2 = 0.96). Thus the eddy correlation technique with use of a LMA-3 detector is capable of determining NO2 fluxes with a reproducibility of 6 %.

Resistance analogy A commonly used approach in describing fluxes is the resistance analogy. This approach assumes that the flux is proportional to c. The proportionality coefficient VO is generally known as the deposition velocity:

Fc = --Vd-C (6)

Both Vd and c are referenced to a fixed height above ground. Vd can be expressed as the reciprocal value of a series of resistances:

1 (7)

Vd -- Ra + Rb + Rc

where Ra = aerodynamic resistance, Rb -- boundary layer resistance, and Rc = canopy resistance. Ra and Rb can be calculated from friction velocity and wind speed (Monteith & Unsworth, 1990). Rb and Rc depend on the nature of the gas, and on surface characteristics.

For some gases, such as NH3 or NO, emission may occur depending on the condition of the vegetation/soil system which has a finite partial pressure. Formally this can be accounted for in eqn (6) by a compensation point Co and introducing Vd* as a deposition velocity:

Fc = - v d * . (c - Co) (8)

Net emission occurs when the concentration drops below Co, otherwise net deposition takes place.

RESULTS

Nitrogen dioxide flux Concentrations and fluxes of NO2 were corrected for the O3 and PAN interference of the LMA-3 according to eqn (1) (Hesterberg, 1994). 03 concentrations were measured on site, and PAN concentrations observed by Wunderli and Gehrig (1991) near the city of Zurich were used. 03 and PAN fluxes were estimated from concentration and deposition velocity (eqn (6)). 03 deposition velocity was assumed to be equal to the observed NO2 deposition velocity, PAN deposition velocity was chosen at half the observed NO2 deposition velocity (Erisman et al., 1994). On average the calculated 03 and PAN interferences to the NO2 deposition were quite small with -2 % in IOP-1, 1% in IOP-2, -1 % in IOP-3, and 1 % in IOP-4 with extreme values between - 12 and + 50 %.

In the presence of 03, NO emitted from soils is rapidly oxidized to NO2, producing an upward NO2 flux. Consequently, NO2 fluxes are decreasing with height. At ambient 03 levels, the time constant for the oxidation of NO to NOz is in the same order of magnitude as the time necessary for transporting emitted NO to the measuring height by turbulent eddies. Hence, measured fluxes of NO2 reflect approximations of net NO~ (= NO + NO2) fluxes. Assuming a constant rate of NO emission throughout one IOP and a constant NO oxidation efficiency, an indirect approach was used to estimate the order of magnitude of the NO emission rate. The slopes of the linear regression analysis according to eqn (8) reflect mean deposition velocities, the intercepts correspond to NO2 fluxes. Denitrification and nitrification processes in soils lead to negligible NO2 emission rates in comparison to NO emission rates (Galbally, 1989). Hence, the estimated NO2 emission corresponds to the release of NO, when an oxidation efficiency of 100 % is assumed. This approach yielded mean NO emission rates of 1.6, 0.7 and 0.7 kg NO-N ha-I y-1 during IOP-1, 2 and 4, respectively (Table 2 and

26 R. H e s t e r b e r g et al.

Table 2. Number of observations (A), percentages of accepted data (pc), concentration of NO2, the finear regression analysis using eqn (8), dry deposition of NO2, (positive sign corresponds to emission), and deposition velocity of NO2 (eqn (6)) for different lOPs

( ± standard deviation on concentrations, depositions and deposition velocities, ± standard error on slopes and intercepts)

lOP A pc Conc a (ppbv) Linear regression analysis Deposition a Vd a (cm s -I) Intercept Vd*'Co Slope va r 2 (kg N ha -I y-l) (kg N ha -I y-i) (cm s "l}

IOP-1 514 74.3 7.4 + 1.2 1.55 + 0.23 4).245 + 0.016 0.39 3.1 + 1.3 0.234 + 0.075 IOP-2 831 64.3 5.2 ± 1.6 0.68 ± 0.16 4).206 q- 0.013 0.31 1.8 ± 1.3 0.188 ± 0.102 IOP-3 1137 63.2 19.9 ± 2.4 -0.11 ± 0.17 4).101 ± 0.004 0.46 3.7 ± 0.6 0.105 -4- 0.016 IOP-4 1081 69.8 4.9 ± 1.6 0.67 ± 0.15 --0.238 ± 0.015 0.26 1.9 ± 1.0 0.242 ± 0.132

aAverage of mean daily variation.

5 t '7 1992 oP1,

• ~ i . : ¢

-l° t E ~

-15 . . . ." . . . 0 10 20 30 40

NO2 [ppbv]

Fig. 2.

5

0

-1

-1

August 1992 (IOP-2)

.!i

10 20 30 40 NO2 [ppbv]

54 February 1993 (IOP-3)

0 ~- ' . .~-- ' - .- ' : : .:, ........ : .......... • ~..~,- : .. •

-I0 . . . . . :....:.:.-.

- 1 5 ~ 0 10 20 30 40

NO2 [ppbv]

5i M~vY..1993 (IOP-4) '

~:" i .

- 1 5 1 " " 0 10 20 30 40

NO2 [ppbv]

Linear regression analysis of the relationship between measured NO2 flux (neg. = deposition) and NO2 concentration using eqn (8) ( - - regrssion 1ine;--95% confidence intervals on the slopes)•

Fig. 3.

25 ] NO2 [ppbv]

20

:i 0 6 12 18 0

time

~ June 1992 0 o a - 0

Average diurnal variations in concentration, flux

t . , I . . , . , , , , I NO2 flux [kg N ha ~ y-~]

0 - ¥ . . . . . . . .

-2 ', /

-3 ' ,~

0 6 12 18 0 time

*-- August 1992 ~ February 1993 (IOP-2) (IOP-3)

0.6 ] Vd NO2 [cm s"]

0.4

0.3

0.2 0.I 0.0 4 -~ . . . . . . . . . . . . . . . . . . . . . . .

-0.11 6 12 18

time

---, May 1993 (IOP-4)

(neg. = deposition) and deposition velocity of NO2.

Fig. 2). N o significant N O flux could be observed in winter (IOP-3). An annual total o f 0.7 kg N O - N ha -l y~ lies in the range of N O emission rates o f 0.1-4 kg N ha-~ y-l given in, the l i terature for non-cul t ivated, unfertilized soils (Betlach & Tiedje, 1982; Johansson & Galbal ly , 1984; Ander son & Levine, 1987; Shepherd e t al . , 1991)• NO2 fluxes repor ted in this s tudy were corrected for this effect. The largest correct ion had to be applied dur ing summer ( lOP- 1) when abou t 50 % o f the NO2 deposi t ion was compensa ted for by oxidized N O leaving the soil. N o correct ions were applied during winter ( IOP-3) when the intercept was not significantly different f rom zero.

In the presence o f active vegetat ion (IOP-1, 2, and 4), the diurnal pa t te rn o f NO2 flux reflected the influence o f

s tomata l up take (Fig. 3). Largest deposi t ion rates between 3 and 6 kg N ha -1 y-1 were observed dur ing the late morn ing hours. S tomata l closure could have been responsible for the decline to abou t 2 kg N ha -1 y-1 around noon. The average deposition was between 1.8 and 3.1 kg N ha -1 y-i dur ing these IOP ' s (Table 2). Dur ing wintert ime, only a small diurnal var ia t ion p ropor t iona l to the NO2 concentra t ion was observed and the mean flux of 3.7 kg N ha "1 y-i was much higher than dur ing the campaigns in spring and summer , due to higher NO2 concentra t ions o f 20 p p b v with daily means o f up to 25 p p b v dur ing stable, foggy weather situations.

The mean diurnal var ia t ion o f the measured deposi- t ion velocity dur ing the growing season was control led by s tomata l activity• M i n i m u m values o f a round 0.1 cm


s -1 were reached during the night and maximum values between 0.35 and 0.55 cm s -1 before noon. In August, Vd decreased due to stomatal closure in the afternoon to 0.2 cm s -l (IOP-2) and to around 0.33 cm s -1 in June and May (IOP-1 and IOP-4). It can be assumed that the decrease was more pronounced in August due to a larger leaf area index leading to enhanced evapotranspiration, rather than due to lower amounts of rainfall. During winter (IOP-3) Vd was nearly constant throughout the day and the daily mean of0.11 cm s -1 was about the half of the averages of 0.194).24 cm s -1 observed during the summer IOP's.

NI-I a deposition The successful use of the gradient technique requires stationary conditions, i.e. the turbulent surface layer above the ground must be in equilibrium with the ground. Hence, the data must be evaluated according to the following criteria. (a) Wind direction must be in the appropriate sector where fetch requirements are fulfilled. (b) Concentration differences must be larger than the detection limit (0.2 ppbv for NH3). (c) Wind speed must be larger than 0.3 m s -1 in order to get a reliable measurement of eddy diffusivities K. (d) No large variations in NH3 concentration due to nearby emissions should occur, i.e. 30-min averages should not change by more than 20 %. (e) The NH3 compensation point Co of our experimental field is likely to be lower than 10 ppbv so therefore, emission should not have occurred during periods with a concentration above 10 ppbv NH3.

In summary, NH3 flux data were rejected when one of the following criteria was met:

(1) wind direction in bad sector (0°-285 ° or 345 °- 360 ° )

(2) Ac < 0.2 ppbv between the two heights (3) mean wind speed at 2 m smaller than 0.3 m s -~ (4) Ac/c > 20 % for 30-min averages (5) NH3 emission occurs above 10 ppbv NH3 Table 3 lists the parameters of the analysis from

which fluxes were calculated from mean concentrations as function of the criterion passed. Only 5-10 % of the data passed all criteria. Up to 95 % of all data were rejected applying criteria (a) and (d), but large changes in NH3 fluxes only occurred during IOP-1. Based on linear regression analysis and by accounting for errors of fluxes and concentrations (Mandel, 1964) the arithmetic average NH3 flux varied only between 8.2 and 8.9 kg N ha-~ y-l when a fraction of 0.02 of the variance of concentration determination to the variance of the flux determination was used. Because of the relatively large uncertainty of Fc compared to the uncertainty of the concentration, simple linear regression analysis was applicable.

Figure 4 shows a plot of the data with the regression lines from the final calculation. The compensation point was determined to be 5.6 + 6.9 in lOP-l , 2.9 + 4.2 in IOP-2. and 4.8 + 2.7 ppbv NH3 in IOP-4 (+ standard error). Unfortunately, the uncertainty of these values was of similar magnitude as the mean value, and in some cases the sign of Co changed depending on the criterion passed. Only for IOP-4, was the standard error of Co smaller than Co itself, indicating clearly the existence of a compensation point. Over unfertilized grassland Co is reported to be not significantly different from zero (Andersen et al., 1993; Sutton & Fowler, 1993; Sutton

Table 3. Number of observations (A), percentages of accepted data (pc), concentration of NH3, the linear regression analysis using eqn (8) (pos. = emission), and concentration-weighted NH3 flux (pos. = emission) depending on quality criteria used during each lOPs

(4- standard deviation on concentrations, 4- standard error on slopes and intercepts)

IOP Criterion A pc Conc (ppbv) Linear regression analysis Flux Intercept v0"-Co Slope v d (kgNha-1 y-l) (kg N ha "l y-l) (cm s -1) r 2

IOP-1 None 242 100.0 4- 9.5 a 41 16.9 b 37 15.3 c 35 14.5 d 15 6.2 e 13 5.4

IOP-2

IOP-4

None 402 100.0 a 63 15.7 b 51 12.7 c 45 11.2 d 32 8.0 e 31 7.7

None 305 100.0 a 42 13.8 b 36 11.8 c 26 8.5 d 16 5.2 e 16 5.2

10.0 4- 9.5 2.7 ± 1.2 4).22 + 0.05 0.07 4).86 9.6 + 6.6 -4.4 + 1.6 -0.20 + 0.08 0.12 1.26

10.4 5:6.5 -5.4 5:1.9 4).23 + 0.09 0.16 1.58 9.6 5:5.4 -6.9 + 2.1 4).33 + 0.11 0.21 1.44

13.5 4- 5.6 -1.5 + 1.7 4).12 + 0.07 0.20 4).50 13.3 4- 6.0 -1.3 + 1.5 4).13 + 0.06 0.30 4).86 9.3 5:4.5 a -1.3 b -0.13 b -0.85 b 13 5:12 2.0 5:2.8 4).17 5:0.10 0.008 -5.74

9.0 5:6.9 -2.1 + 5.4 4).79 ± .27 0.12 -15.40 10.3 5:7.0 -1.0 5:7.2 4).77 5:0.33 0.10 -16.09 10.5 ± 7.4 4).8 5:8.0 4).79 + 0.35 0.10 -16.55 9.3 5:5.9 1.3 5:7.2 4).50 + 0.38 0.05 -12.31 8.8 5:5.3 -5.2 5:7.2 -1.01 5:0.39 0.17 -17.02

12.4 5:2.2 a -10.8 b -1.3 b -19.3 b 12 ± 12 -5.5 ± 2.1 4).58 ± 0.08 0.23 -6.58

9.3 ± 6.3 -2.3 ± 1.9 -0.40 ± 0.10 0,30 -6.03 9.9 5:6.4 -1.8 -4- 2.3 4).38 5:0.11 0.26 ~5.26 8.7 5:6.6 -1.5 5:2.6 4).39 ± 0.14 0.25 -6.59 8.9 ± 6.2 -4.5 5:2.3 4).53 5:0.12 0.56 -6.54 8.9 ± 6.2 -4.5 5:2.3 4).53 5:0.12 0.56 ~5.54

11.8 ± 4.4 a --4.7 b 4) .53 b -6.51 b

aAverage of mean daily variation. bTaking into account errors associated with concentrations and fluxes.


z ~ 0

z

50

25

0

-25

-50

-75

-100 0

June 1992 (lOP-l) 50 I August. .1992 (IOP-2)

25 . . . . "

-25 -

_5oi -75 -

• . "-, . ~ ~ O 0 ~

50 Ma..y 1993 (IOP-4)

25 -.C... .

• - • • . . . 0 :.'• • :""~?~b'- ............ ~.

-25

-50 ]

-75

-100 10 20 30 40 50 0 10 20 30 40 50 0 10 20 3'0 4 ' 0 50

NH 3 [ppbv] NH~ [ppbv] NH~ [ppbv]

Fig. 4. Linear regression analysis of the relationship between NH3 flux (neg. --- deposition) and NH3 concentration using eqn (8) (-- regression line; - - 95% confidence intervals on the slopes;ll data points used for regression analysis; • rejected data)•

25

20

15

10

5

0

NH3 [ppbv]

x x '~

/

I NH3 flux [kg N ha" y'] 5 !

-151\ / ;: :! :, ',~

-25 ~ . . . . . . . " . . . . 6 12 18 0 0 6 12

time time

• -'4 June 1992 ~-. August 1992 (lOP-l) (IOP-2)

18 0

5.0 4 v~' NH3 [cm s-']

4.0

3.0

2.0

1.0 ~ , . . i ,

0.0

-1.0

-2.0 . . , . . , . , . .

0 6 12 18 time

May 1993 0OP-4)

Fig. 5. Average diurnal variations in concentration, flux calculated by linear regression parameters (eqn (8)) (neg. = deposition), and deposition velocity Vd* of NH3. Because of the lack of data, the diurnal pattern of Va* is based on 3-h means•

et al., 1993) which was in disagreement with the observation made in the IOP-4. A compensation point possi- bly appears when a strong surface loading with NH3 and NH4 + in the presence of high concentrations in combination with low temperature and wet surfaces occurs. Parts of the already deposited NH 3 and NH4 are re-emitted with increasing temperature, decreasing air concentration of NH3 and drying surfaces. Probably, re- evaporation conditions in the early afternoon and large emissions from the nearby intensively managed fields increasing ambient NH3 concentrations up to 100 ppbv (30-min average) led to the high value of Co in comparison to values reported in the literature• This compensation mechanism may have caused the small net deposition rates of 0.8 ± 2.0 kg N ha -1 y-I ( ± standard error) measured during June (lOP-l). High rates of NH3 deposition of about 17 ± 9 kg N ha -1 y-1 were measured during IOP-2 when vegetation was tall (up to 1.40 m) and solar radiation was lower• In the spring (IOP-4), when vegetation was short again, a lower deposition of 6.5 + 3.0 kg N ha -1 y-i was measured•

Figure 5 shows the mean diurnal pattern of NH3 concentration, NH3 flux calculated by linear regression and mean concentration, and deposition velocity Vd °. In IOP-1 and IOP-4 diurnal variations in NH3 concentra-

tion showed a strong decrease around noon while in IOP-2 this decrease was overlapped by a peak stemming from a single event when in the afternoon NH3 concentration increased rapidly from 10 ppbv to 100 ppbv then decreased again down to 10 ppbv in 4 h. Highest concentrations were observed in the evening, at night or in the early morning. The diurnal variation of the calculated fluxes was characterized by largest deposition rates in nights and by a lower deposition around noon which even may turn to an emission for mid-summer conditions. Diurnal variations in NH3 deposition velocity observed during IOP-2 and IOP-4 revealed an influence of stomatal uptake during daylight hours. During IOP-1, no valid data were available between 8:00 a.m. and 8:00 p.m. because of frequent periods with low NH3 concentrations and small gradients• In such conditions, the concentration differences are often smaller than the accuracy of the measuring system•

Moni tor ing o f N H 3 concentrat ions with passive samplers Weekly average NH3 concentrations were obtained with passive samplers (Fig. 6). Extreme values of up to 50 ppbv NH3 were observed when manure was applied to neighbouring fields. This occurred at sowing time and after harvests when the fields were ploughed (e.g. in


Fig. 6.

3O

2O w

" 10 Z ~ 0

- 1 0

6(I

. ~ 4 0 -

~ 2 0 -

0

4

a)

b)

i 'l I ! -V I I I I I I I J J A S O N D J F M A M

1992 1993

Weekly variation in (a) flux (pos. = deposition) and (b) concentration of NH3 measured with passive samplers.

July, late November, mid March). The annual mean concentrations at 0.5 and 5 m above ground were about 6 and 16 ppbv, respectively.

To estimate deposition, weekly average values of K were calculated by linear combination of daily mean global radiation, wind speed and temperature. The parameters used were determined from measurements during the IOPs (Hesterberg, 1994). Diurnal variations in the concentration gradient and in K were corrected for by multiplying depositions rate by the time-averaged factor K.(6c/Sz)/(K.Sc/Sz). This factor was 0.32 during IOP-1, 0.91 during IOP-2, and 0.45 during IOP-4. Hence, this approach resulted in an uncertainty in the estimated dry deposition rate of NH3 of about 60 %.

A disadvantage of the determination of NH3 deposition with passive samplers is that no screening for erro- neous data can be applied, e.g. for data influenced by wind from an unsuitable sector. With a fetch of only about 30 m with winds from SW or NE, the measuring height should not have exceeded 30 cm, according to the fetch criterion suggested by Businger (1986): In order to minimize this error, the NH3 flux was calculated from concentrations measured at the lowest two levels between 0.5 and 1.5 m. The resulting average NH3 deposition was 16 + 9 kg N ha -1 y-l, with highest deposition rates of 30 kg N ha -1 y-1 and one emission of 1.5 kg N ha -1 y-i (at 3.3 ppbv NH3). By linear regression analysis of the relationship between Fc and c, it was found that Co = 4.6 + 1.3 ppbv NH3 and Vd* = 1.40 + 0.13 cm s -1 (r E : 0.51, 4- standard error). The value of Co was in agreement with the value observed during the IOPs, whereas va* was comparatively high.

HNOs deposition During IOP-2, an attempt was made to determine the HNO3 deposition rate. The mean concentration of 0.23 (0.14 median) ppbv was very low. The precision of the system (uncertainty of 0.1 ppbv) was not sufficiently high to allow the measurement of gradients (Meixner & Wyers, 1990). An additional problem was the high ratio between NH3 and HNO3. The product of NH3 and HNOs concentrations was lower than the product of vapour pressures of NHs and HNO3 over solid ammonium nitrate (NH4NO3) only during daytime (Seinfeld, 1986). Hence, the gradient method yielded unrealistic emission rates.

As an alternative, HNO3 fluxes were estimated by the inferential method using eqn (6). The canopy resistance Rc for extensively managed land given by Wesely (1989) was used. HNO3 deposition was estimated to be about 0.91 kg HNO3-N ha -1 y-1. Assuming a NO3-/(HNO3 + NO3-) ratio of 0.45 according to Lindberg et aL (1990) and Meixner et aL (1985), HNO3 concentrations were estimated to be 1.1 ppbv HNO3 using the data for total nitrate measured during the IOP-4 and IOP-3. This resulted in a deposition rate of 2.0 kg HNO3-N ha -1 y-l. Assuming identical concentrations during IOP-I and using the measured concentrations during IOP-2, an annual deposition of about 1.1 kg N ha -I y-~ was derived.

H O N O deposition HONO concentrations measured during IOP-2 showed a peak during the early morning hours, and values were generally below 0.1 ppbv during daytime. Because of these low values, fluxes based on 30-min values showed a large scatter. Hence, HONO fluxes could only be estimated by the inferential method (eqn (6)). Calculated annual HONO deposition was 0.16 kg N ha -1 y-~ which was a negligible contribution to the total N input ( < 1%).

Deposition of aerosol-bound NOa-, NO2" and NH4 + Measurements with the MCS provided the sum of the gas phase and aerosol-bound portions of NH 3, HNO3 and HONO, or the corresponding ions. Due to the lower deposition efficiency of aerosols, their gradient is small, and the precision of the MCS was too low for flux measurements. Therefore, deposition was estimated by the inferential method. The deposition velocity of aerosol-bound NO3-, NO2- and NH4 + was assumed to be 0.12 cm s -1 (van Aalst, 1983; Asman & Maas, 1986; Asman & Jaarsveld, 1990; Keuken et al., 1990). Calcu- lated annual depositions were 0.3 kg NO3--N ha -1 y-l, less than 0.01 kg NO2--N ha -1 y-1 and 1.1 kg NH4÷-N ha-I y-t.

Wet deposition of nitrogen Wet deposition was measured on a weekly basis from July 1992 to May 1993 (Fig. 7). During winter, small amounts of precipitation were observed leading to small rates of N deposition of 0.4 kg NO3--N ha -1 y-i and 1.3 kg NH4+-N ha 1 y-1 (January 1993). In summer, the


4 0

- ~ 30

'~ 2 0 Z I0

0 I0.0

~ 7.5

E 5.0 E 2.5

0.0

Fig. 7.

J J A S O N D 1992

Weekly variation of (a) wet N deposition

a)

J F M A M 1993

( - -NO3- ; - -NH4 +) and (b) precipitation.

precipitation of up to 7 mm d -~ delivered high deposition rates corresponding to 31 kg NO3--N ha 1 y-1 and 33 kg NHa+-N ha -1 y-l. The annual averages were 3.4 kg NO3--N ha -l y-i and 5.6 kg NHa+-N ha -1 y-1. These rates were in quite good agreement with those pre- viously observed in Switzerland (Fuhrer, 1986a; G/illi Purghart, 1989) and lay within the range of variation between different years.

DISCUSSION

Annual input o f nitrogen and speeiation The arithmetic mean NH3 deposition measured during three lOPs of 8.2 + 3.4 (+ standard error) kg NH3-N ha d yd was corrected according to the annual concentration in order to obtain the annual deposition rate. Measurements with passive samplers yielded an annual mean NH 3 concentration at 2 m above ground of 13.1 ppbv, which was slightly higher than the mean for the three lOPs (11.2 ppbv). The correction leads to a NH3 flux of 9.2 ± 3.8 kg NH3-N ha -1 y-l which differs from the mean flux of 16 + 9 kg NH3-N ha -~ y-1 determined with passive samplers. For further discussions, the average of both values, i.e. 13 + 5 kg NH3-N ha -l y-1 will be used. For the calculation of the annual mean NO2 deposition, a simple interpolation of the fluxes observed during the four IOPs was 2.8 + 0.6 kg N ha -~ y-i (+ standard deviation) (Eugster, 1994). The sum of dry deposition of total nitrate, total nitrite and NH4 + was 2.7 kg N ha -~ y-~. Rates of deposition of NO, PAN

and N205 of 0.8, 0.1 and 0.2 kg N ha -1 y-l, respectively, were estimated from concentrations reported for rural regions and deposition velocities measured over grass (Hesterberg, 1994). They contribute only a few percent to the total input. Table 4 lists all estimated and measured annual N inputs. The sum of all fluxes was domi- nated by the dry and wet deposition of NH3-N and NH4 +-N. followed by wet nitrate deposition and by dry NO2 deposition (Fig. 8). Total N input was about 28 kg N ha -1 y-l of which 32 % was due to wet deposition. Possible occult N deposition is not taken into account. This input could contribute significantly to the total N- input in hilly and forested regions (Fabian, 1987; Dash, 1988; Grosch and Schmitt, 1988; Kroll & Winkler, 1988; Bobbink et al., 1992), but reported values for flat terrain

[] 20.1% NH4÷wet

[] 12.2 % NO 3" wet

[] 46.6 % NH 3

[] 10.0% NO 2

[] 5.0% ( HNO3 + NO3 )

[] 3.9 % NH4 +

[] 2.2% others

Fig. 8. Relative contribution of different N-containing compounds to the total annual N deposition (100% = 28 kg N ha- 1 y- 1) over an extensively managed litter meadow. Possible emission is included for NO and NH3. Occult deposition was

neglected.

Table 4. Measured concentrations, estimated annual N depositions, measured deposition velocities, and fraction of different species of dry, wet and total deposition to an extensively managed litter meadow. Possible emission is included for NO and NH3, but not for the

other species (e.g. HONO, HNO3, N2Os). Occult deposition was neglected

Species Concentration Dry and wet Deposition Fraction of dry Fraction of wet Fraction of total (ppbv) deposition (kg N ha -1 y-l) velocity (cm s l ) deposition (%) deposition (%) deposition (%)

NH3 NO2 HNO3 + NO3- NH4 + NO, N205, HONO, NO2- , PAN NH4 + wet NO 3 wet Total

13.1 13 0.13-1.40 68.8 4.6 9.4 2.8 0.11-0.24 14.8 10.0 1.65 1.4 7.4 5.0 4.8 1.1 5.8 3.9

0.6 3.2 2.2

5.6 62.2 20.1 3.4 37.8 12.2

27.9

Deposition o f N-containing compounds 31

are low (0.1 kg NO3--N ha -1 y-l and 0.4 kg NH4+-N ha-l y-l Fuhrer, 1986b).

Our observations are in agreement with results from other studies. Stocker et al. (1993) reported a NOy (= reactive N oxides) dry deposition rate between 2.1 and 6.9 kg N ha -I y-l, which is similar to our estimate of 4.4- 7.3 kg N ha -1 y-1. The deposition velocities for NO2 (between 0.11 and 0.26 cm s -1) are similar to the observations made by Dollard et al. (1987) (0.25 cm s-l), Erisman et al. (1989) (0.33 cm s -l) and L6vblad et al., 1993 (0.2q).5 cm s-l). NH3 deposition of 10 kg N ha -l y-l determined by Erisman et al. (1989) was similar to the 13 kg N ha -1 y-i measured here. In the framework of the Convention on Long Range Transboundary Air Pollution coordinated by the United Nations Economic Commission for Europe, a map of the exceedance of the critical loads has been produced for Switzerland based on a 1 km x 1 km grid (Buwal, 1994b). The procedure to determine the N input followed the guidelines produced by the Coordination Centre for Effects (CCE, 1993), and parametrization of deposition was simple, without using a source-receptor model. The emission of NH3 has been calculated by using widely accepted emission factors and statistical data for animal husbandry in individual communities. For the grid point near Merenschwand, the calculation differs by 24 % for NH3, by - 2 1 % for NOy and by 8 % for total N deposition (Buwal, 1994a).

Environmental consequences of large nitrogen inputs The total N input exceeds the critical load for litter meadows and moors given by Grennfelt and Thtrnel t f (1992), and it is likely that the N input exceeds the critical load for most natural ecosystems located on the Swiss plateau. Consequently, changes in biodiversity, eutrophication, and deficiencies due to enhanced leaching of nutrients such as potassium, calcium and magne- sium, could occur in these ecosystems, and are already detectable (Ulrich & Pankrath, 1983; Ellenberg, 1985, 1989; Fabian, 1987; Bobbink et al., 1992). For the experimental field used in the present study the critical load of nitrogen is about 15 kg N ha -1 y-l, and the actual load exceeds the critical load by a factor of nearly 2. Over forests, the deposition velocity for NH3 is larger by a factor of 2-3 (Sutton et al., 1992; Wyers et al., 1992; Andersen & Hovmand, 1993; Duyzer et al., 1994) and by a factor of about 1.5-5 for NO2 (Grosch & Schmitt, 1988; Hanson et al., 1989). Assuming the dry deposition velocity is doubled over forests as compared to our measurements and identical concentrations, a total net dry deposition of about 40 kg N ha -I y-l and a total deposition of up to 50 kg N ha -I y-1 can be estimated for forests. These values reach the upper limit of the range of critical load values, and it suggests that the critical load for nitrogen is exceeded in forests as well.

For the ecosystem studied in this work, N availability was not likely to be limiting because natural mineralization of organic material in the soil may have delivered up to 60 kg N ha -1 y-1 (A. Grub, personal communication). During the last 15 years, no effects on vegetation

from N overloading were observed (J. Fischer, personal communication). Hence, dry N deposition is not likely to have a direct effect on this meadow.

Atmospheric N input increases the N-pool in the soil. This leads to enhanced emissions of NO and nitrous oxide (N20) and increased NO3- leaching into the groundwater. Emission of NO stimulates formation of photooxidants during the summer, whereas N20 is a greenhouse gas and enhances 03 destruction in the stratosphere (Finlayson-Pitts & Pitts, 1986). The presence of a compensation point of NH3 decreases the dry deposition of NH3 considerably (ApSimon et al., 1994), and, in turn, NH3 is transported over larger distances and is deposited preferentially to surfaces of ecosystems with a lower compensation point. Often, such ecosystems are sensitive to additional N input (Ellenberg, 1985). With respect to possible strategies to reduce N loading of natural and semi-natural ecosystems, NH3 is particularly important when these systems are surrounded by intensively managed land. In urban areas, changes in ambient concentrations are likely to lead to a lower fraction of NH3 relative to the fraction of NOy (Jaeschke et al., 1992).

ACKNOWLEDGEMENTS

We thank J. Fischer for giving important information about the measuring field, A. Grub for estimating natural mineralization, P. K. Dasgupta for helpful discussions concerning the NH3-HNO3-HONO measuring system, R. Perler and E. Kratochvil for NH 3 measurements with passive samplers and for wet N deposition measurements. C. Fuhrer is acknowledged for a careful language editing of the manuscript. This work was financially supported by the Swiss Federal Office for Environment, Forests and Landscape (BUWAL).

REFERENCES

Andersen, H. V. & Hovmand, M. F. (1993). Measurements of ammonia flux to a spruce stand in Denmark. Atmos. Environ., 27A, (2), 189-202.

Andersen, H. V., ApSimon, H., Asman, W., Barret, K., Kul- weit, J., Schjrring, J. & Sutton, M. (1993). Dry deposition of reduced nitrogen. Working group report published in 'Models and Methods for the Quantification of Atmo- spheric Input to Ecosystems', report of the International Workshop on the Deposition of Acidifying Substances in Gtteborg, 3-6 November 1992, ed. G. Ltvblad, J. W. Erisman & D. Fowler. ISBN 92-9120-253-3, Nordic Council of Ministers, Copenhagen.

Anderson, I. C. & Levine, J. S. (1987). Simultaneous field measurements of biogenic emissions of nitric oxide and nitrous oxide. J. Geophys. Res., 92, 965-76.

ApSimon, H. M., Barker, B. M. & Kayin, S. (1994). Model- ling studies of the atmospheric release and transport of ammonia in anticyclonic episodes. Atmos. Environ., 28, 665-78.

Asman, W. A. H. & Jaarsveld, van H. A. (1990). Regionale und europaweite Emission und Verfrachtung von NHx- Verbindungen. In Ammoniak in der Umwelt, Kreisldufe, Wirkungen, Minderungen. Hrsg. v. KTBL und VDI.


Asman, W. A. H. & Maas, J. F. M. (1986). Sch/itzung der Ammoniak- und Ammonium Deposition in den Nie- derlanden. Bericht R-86-8, Instituut voor Meteorologic en Oceanografie, Rjiksuniversiteit Utrecht, Niederlande.

Betlach, M. R. & Tiedje, J. M. (1982). Kinetic explanation for accumulation of nitrite, nitric oxide and nitrous oxide during bacterial nitrification. Appl. Env. Microbiol., 42, 1074-84.

Blatter, A., Fahrni, M. & Neftel, A. (1992). A new generation of NH3 passive samplers. Air Pollution Report, Develope- ment of analytical techniques for atmospheric pollutants. Commission of the European Communities, 41, 171-6.

Blatter, A., Neftel, A., Dasgupta, P. K. & Simon, P. K. (1993). A combined wet effluent denuder and mist chamber system for deposition measurements of NH3, NH4 +, HNO3 and NO3-. Presented at the Sixth European Symposium of Physico-Chemical Behaviour of Atmospheric Pollutants, 18-22 October, Varese, Italy.

Bobbink, R., Bik, L. & Willems, J. H. (1988). Effects of nitrogen fertilization on vegetation structure and domi- nance of Brachypodium pinnatum (L.) Beauv. in chalk grassland. Acta Botanica Neerlandica, 37, 231-42.

Bobbink, R., Heil, G. W. & Raessen, M. B. A. (1992). Atmo- spheric deposition and canopy exchange processes in heathland ecosystems. Environ. Pollut., 75, 29-37.

Businger, J. A. (1986). Evaluation of the accuracy with which dry deposition can be measured with current micro- meteorological techniques. J. Climate Appl. Meteorol., 25, 1100-24.

Businger, J. A., Wyngaard, J. C., Izumi, Y. & Bradley, E. F. (1971). Flux profile relationships in the atmospheric surface layer. J. Atmos. Sci., 28, 181-9.

Buwal (1994a). Der Stickstoffeintrag in ein Naturschutzgebiet im unteren Reusstal. Buwal-Materialien Nr. Swiss Federal Office for Environment Forests and Landscape, PO Box, CH-3000 Bern, 28.

Buwal (1994b). Mapping critical loads of acidity for Swiss forest soils and alpine lakes steady state balance method. Environmental Series. No. report prepared by METEOTEST on the request of the Swiss Federal Office for Environment, Forests and Landscape, PO Box CH- 3000 Bern, 234.

CCE (1993). Calculation and Mapping of Critical Loads in Europe. Status report (1993). UN-ECE Coordination Center for Effects, National Institute of Public Health and Environmental Protection Bilthoven, The Nether- lands, RIVM Report No. 259101003.

Dash, J. M. (1988). Hydrological and chemical inputs to fir trees from rain and clouds during a 1-month study at Clingmans Peak, N. C. Atmos. Environ., 22, 2255-62.

Dollard, G. J., Davies, T. J. & Lindstrom, J. P. C. (1987). Measurements of dry deposition of some trace gases. In Physico-Chemical Behaviour of Atmospheric Pollutants, ed. G. Angeletti and G. Restelli) Reidel Dordrecht, 470-9.

Drummond, J. W., Schiff, H. I., Karecki, D. R. & Mackay, G. I. (1989a). Measurements of NO2, NOx, 03, PAN, HNO3, H202 and H2CO during the Southern California air quality study. Presented at APCA 1982, Anaheim, CA.

Drummond, J. W., Castledine, C., Green, J., Denno, R., Mackay, G. I. & Schiff, H. I. (1989b). New technologies for use in acid deposition networks. In Monitoring Meth- ods for Toxics in the Atmosphere, Spec. Tech. Publ. 1052, ed. W. L. Zielinsky, Jr. & D. D. William. American Society for Testing and Materials, Philadelphia 133-49.

Drummond, J. W., Topham, L. A., Mackay, G. I. & Schiff, H. I. (1991). Use of chemiluminescence techniques in portable, lightweight, highly sensitive instruments for measuring NO2, NOx and O3. Measurements of Atmospheric Gases, ed. H. I. Schiff. Proc. SPIE, 1433, 224-31.

Duyzer, J. H., Verhagen, H. L. M., Weststrate, J. H., Bosveld, F. C. & Vermetten, A. W. M. (1994). The dry deposition

of ammonia onto a Douglas fir forest in The Netherlands. Atmos. Environ., 28(7), 1241-53.

Dyer, A. J., Garrat, J. R., Francey, R. J., Mcllroy, I. C., Bacon, N. E., Hyson, P., Bradley, E. F., Denmead, O. T., Tsvang, L. R., Volkov, Y. A., Koprov, B. M., Elagina, L. G., Sahashi, K., Monjis, N., Hanafusa, T., Tsukamato, O., Frenzen, P., Hicks, B. B., Wesely, M., Miyake, M. & Shaw, W. (1982). An international turbulence comparison experiment (ITCE 1976). Boundary Layer Meteorol., 24, 181-209.

Ellenberg, H. (1985). Ver~inderungen der Flora Mitteleuropas unter dem Einfluss von Diingung und Immissionen. Schweiz Z. Forstwes., 136, 19-39.

Ellenberg, H. (1989). Eutrophierung - - das gravierendste Problem im Naturschutz? Zur Einfiihrung. NNA-Berichte, 2/1, 4-13.

Erisman, J. W., de Leeuw, F. A. A. M. & van Aalst, R. M. (1989). Deposition of the most acidifying components in the Netherlands during the period 1980-1986. Atmos. Environ., 23(5), 1051-62.

Erisman, J. W., van Pul, A. & Wyers, P. (1994). Parametri- zation of surface resistance for the quantification of atmospheric deposition of acidifying pollutants and Ozone. Atmos. Environ., 28, 2595-607.

Eugster, W., Neftel, A. & Hesterberg, R. (1993). A criterion derived from eddy correlation measurements for rejecting Monin-Obukhov scaling over complex terrain. Analyt. Geophys., 11 (Suppl. II, Part 11), 268.

Eugster, W. (1994). Mikrometeorolgische Bestimmung des NO2-Flusses an der Grenzfl/iche Boden/Luft. Geo- graphica Bernenksia, G37, 164 pp, ISBN3-906290-90-5.

Eugster, W. & Senn, W. (1995). A cospectral correction model for measurement of turbulent NO2 flux. Boundary Layer Meteorol., 74, 321-40.

Fabian, P. (1987). Photochemischer Smog und seine Einwir- kung auf die Biosphare. Forstw. Cbl., 106, 223-5.

Fahey, D. W., Hubler, G, Parrish, D. D., Williams, E. J., Norton, R. B., Ridley, B. A., Singh, H. B., Liu, S. C. & Fehsenfeld, F. C. (1986). Reactive Nitrogen in the tropo- sphere. Measurements of NO, NO2, HNO3, particulate nitrate, peroxyacetyl nitrate (PAN) O3 and total reactive odd Nitrogen (NOy) at Niwot Ridge, Colorado. J. Geo- phys. Res., 91 (D9), 9781-93.

Fehsenfeld, F. C., Drummond, J. W., Roychowdhury, U. K., Galvin, P. J., Williams, E. J., Buhr, M. J., Parrish, D. D., Hubler, G., Langford, A. O., Calvert, J. G., Ridley, B. A., Grahek, F., Heikes, B. G., Kok, G. L., Shetter, J. D., Walega, J. G., Elsworth, C. M., Norton, R. B., Fahey, D. W., Murphy, P. C., Hovermale, C., Mohnen, V. A., Demerjian, K. L., Mackay, G. I. & Schiff, H. I. (1990). Intercomparison of NO2, measurements techniques. J. Geophys. Res., 95 (D4), 3579-97.

Finlayson-Pitts, B. J. & Pitts, J. N. Jr. (1986). Atmospheric Chemistry: Fundamentals and Experimental Techniques. John Wiley and Sons, New York.

Fuhrer, J. (1986a). Study of acid deposition in Switzerland. Temporal variation in the ionic composition of wet precipitation at rural sites during 1983-1984. Environ. Pollut. Ser. B, 12, 111-29.

Fuhrer, J. (1986b). Chemistry of fogwater and estimated rates of occult deposition in an agricultural area of Central Switzerland. Agric. Ecosystems Environ., 17, 153-64.

Galbally, I. E. (1989). Factors controlling NOx emissions from soils. In Exchange of Trace Gases between Terrestrial Ecosystems and the Atmosphere, ed. M. O. Andreae & O. S. Schimel. New York, John Wiley and Sons, pp. 23-37.

G/illi Purghart, B. C. (1989). Schwermetalle auf grossen frak- tioniertem Aerosol und in der Deposition: Untersuchun- gen an einem Hohenprofil im Kanton Bern. Diss. phil. nat., Ziirich, ADAG, p. 126.

Gehrig, R. & Baumann, R. (1992). Comparison of 4 different types of commercially available monitors for nitrogen


oxides with test gas mixtures of NH3, HNO3, PAN and VOC and in ambient air. EMEP - - Workshop on Mea- surements of Nitrogen-Containing Compounds, 30 June- 3 July 1992, Les Diablerets, Switzerland, ed. R. Ballaman, R. Gehrig, I. M. Kvalvhgnes & J. Schaug. Norwegian Institute for Air Research, Postboks 64, 2001 Lillestrom, Norway.

Genfa, Z. & Dasgupta, P. K. (1989). Fluorometric measurement of aqueous ammonium ion in a flow injection system. Analyt. Chem., 61,408-12.

Gregory, G. L., Hoell, J. M. Jr., Carroll, M. A., Ridley, B. A., Davis, D. D., Bradshaw, J., Rodgers, M. O., Sandholm, S. T., Schiff, H. I., Hastie, D. R., Karecki, D. R., Mackay, G. I., Harris, G. W., Torres, A. L. & Fried, A. (1990). An intercomparison of airborne nitrogen dioxide instruments. J. Geophys. Res., 95 (D7), 10103-27.

Grennfelt, P. & Th6rnel6f, E. (1992). Critical loads for nitrogen. Report from a workshop held at Lokeberg, Sweden, 6-10 April. Nordic Council of Ministers, Nord 1992:41, ISBN 92 9120 121 9.

Grosch, S. & Schmitt, G. (1988). Experimental investigations on the deposition of trace elements in forest areas. In International Symposium on Environmental Meteorology 1st, Wiirzburg, W. Germany, 29 September-1 October 1987. Proceedings, Dordrecht, Holland, Kluwer Academic Publishers, pp. 201-16.

Hanson, P. K., Rott, K., Taylor, G. E., Gunderson, C. A., Lindberg, S. E. & Ross-Todd, B. M. (1989). NO2 deposition to elements representative of a forest landscape. Atmos. Environ., 23, 1783-94.

Hesterberg, R. (1994). Die Stickoxide im schweizerischen Mittelland und der Stickstoffeintrag in ein Nat- urschutzgebiet. Dissertation, Physikalisches Institut der Universit~it Bern, Schweiz.

Hicks, B. B. & McMillan, R, T. (1988). On the measurement of dry deposition using imperfect sensors and in non-ideal terrain. Boundary Layer Meteorol., 42, 79-94.

Hutter, M. & Biirki, D. (1991). Application of the UK mesoscale model to the lower Reuss valley. General Energy Technology Newsletter Annex V, Annual Report. Department of General Energy Technology, Paul Scherrer Institute, Villigen, Switzerland.

Jaeschke, W., Grieser, J., Herrmann, U., Kessel, M., Kosiol, W., Nietzsche, I. & Sattler, T. (1992). MassenfliJsse stick- stoffhaltiger Schadstoffe zwischen Atmosph/ire und Wal- dekosystem. Berichte des Zentrums ffir Umweltforschung, 17, Johann Wolfgang Goethe-Universitht. Frankfurt am Main, Verhundsforschung Frankfurter Stadtwald -III.

Johansson, C. & Galbaily, I. E. (1984). The production of nitric oxide in a loam under aerobic and anaerobic conditions. Appl. Env. Microbiol., 47, 1284-98.

Kaimal, J. C., Wyngaard, J. C., Izumi, Y. & Cote, O. R. (1972). Spectral characteristics of surface-layer turbulence. Quart. J. Roy. Meteorol. Soc., 98, 563-89.

Kelly, Th. J., Spicer, Ch. W. & Ward, G. F. (1990). An assessment of the Luminol chemiluminescence technique for measurement of NO2, in ambient air. Atmos. Environ., 24A (9), 2397-403.

Keuken, M. P., Wayers-Ijpelaan, A., Otjes, R. P. & Slanina, J. (1990). Determination of NH2, HNO3 and NH2NO2 in ambient air by automated thermodenuder systems and a wet annular denuder system. In Physico-Chemical Behaviour of Atmospheric Pollutants. Publ. G. Restelli & G. Angeletti, Air Pollution Report Series of the CEC, 23, 6-11.

K16tzli, F. (1969). Die Grundwasserbeziehungen der Streu- und Moorwiesen im nrrdlichen Schweizer Mittelland. Beitr~ige zur geohotanischen Landesautnahme der Schweiz, Heft, 52, p. 296.

Kroll, G. & Winkler, P. (1988). Estimation of wet deposition via fog. In Environmental Meteorology, ed. K Grefen & J. L6bel, pp. 227-36.

Lindberg, S. E., Bredmeier, M., Schaefer, D. A. & Qi, L. (1990). Atmospheric concentrations and deposition of nitrogen and major ions in conifer forests in the United States and Federal Republic of Germany. Atmos. Environ., 24A (1), 00-14.

L6vblad, G., Erisman, J. W. & Fowler, D. (1993). Models and methods for the quantification of atmospheric input to ecosystems. An international workshop on the depoxition of acidifying substances in Grteborg, 3-6 November 1992, organized by the Nordic Council of Ministers in col- laboration with the BIATEX project and EMEP. Nordic Council of Ministers, Copenhagen 1993, ISBN 92-9120- 253-3.

Mandel, J. (1964). The Statistical Analysis of Experimental Data. John Wiley and Sons, New York.

Marquardt, D. W. (1963). An algorithm for least-squares estimation of non-linear parameters. J. Soc. lndust. Appl. Math., I1 (2), 431-41.

Meixner, F. X., M~iller, K. P., Aheimer, G. & Hrfken, K.-D. (1985). Measurements of gaseous nitric acid and particulate nitrate. In Physico-Chemical Behaviour of A tmospheric Pollutants. Publ. F. A. A. M. De Leeuw & N. D. Van Egmond, COST Aktion 611, Proc. Workshop Pollut. Cycles Transport-Modeling Field Experiments. Bilthoven, Niederlande, pp. 103-14.

Meixner, F. X. & Wyers, P. G. (1990). Messmethoden zur Quantifizierung der trockenen Deposition von gasfrr- migen Ammoniak. KTBL-VDI Symposium Ammoniak in der Umwelt, 10-20 October, Bundesforschungsanstalt fi~r Landlsnirtschaft, D-Braunschweig.

Mohr, H. (1993a). Die neuartigen Waldschaden in Mittel- europa. Forschung und Technik, Neue Z~'rcher Zeitung, 27 Jan 1993.

Mohr, H. (1993b). Waldschaden in Mitteleuropa - - wo liegen die Ursachen? In Verhandlungen der Gesellschaft deutscher Naturforscher und/t'rzte, 117, Versammlung, Aachen 1992, Wissenschaftliche Verlagsgesellschaft, Suttgart, 43 59.

Mohr, H. (1994). Stickstoffeintrag als Ursache neuartiger Waldsch~iden. Spektrum der Wissenschaft, January 1994.

Monin, A. S. & Obukhov, A. M. (1954). Basic laws of turbulent mixing in the ground layer of the atmosphere. Akademiia Nauk SSSR, Leningrad, Geofizicheskii Institut. Trudy, 151, pp. 163-87. AFCRL translation by American and Meteorological Society, reprinted in Aerophysics of Air Pollution, ed. J. A. Fay & D. P. Hoult, pp. 90-119. (AIAA selected reprints, Vol IX). American Institute of Aero- nautics and Astronautics, New York.

Monteith, J. L. & Unsworth, H. (1990). Principles of Environ- mental Physics (2nd edn). British Library Cataloguing in Publication Data, Edward Arnold, A Division of Hodder and Stoughton, London, ISBN 0-7131-2931 -X

Moore, C. J. (1986). Frequency response correction for eddy correlation systems. Boundary Layer Meterol., 37, 17-35.

Nilsson, S. & Grennfelt, P. (1988). Critical Loads for Sulphur and Nitrogen. Report from a workshop held at Skokloster, Sweden, 19-24 March 1988. NORD miljorapport, 15, Nordic Council of Ministers, Copenhagen.

Ohmura, A. (1982). Objective criteria for rejecting Bowen- ratio flux calculations. J. Appl. Meteorol., 21 (4), 595-8.

Panofsky, H. A. & Dutton, J. A. (1984). Atmospheric Turbu- lence. Models and Methods for Engineering Applications. John Wiley and Sons, New York, p. 397.

Rosset, M. O. (1990). Beziehungen zwischen Vegetation, Bodenwasser, Mikroklima und Energiehaushalt yon Feuchtwiesen unter besonderer BeriJcksichtigug der Evapotranspiration. Diss. Botanicae, 159, Berlin Stutt- gart, J. Cramer, p. 244.

Schiff, H. I., Mackay, G. I., Castledine, C., Harris G. W. & Tran, Q. (1986). Atmospheric measurements of nitrogen dioxide with a sensitive luminol instrument. Water, Air Soil Pollut., 311, 105-14.


Seinfeld, J. H. (1986). Atmospheric Chemistry and Physics of Air Pollution. John Wiley and Sons, New York.

Shepherd, M. F., Barzetti, S. & Hastie, D. R. (1991). The production of atmospheric NOx and N20 from a fertilized agricultural soil. Atmos. Environ., 25A (9), 1961-9.

Simon, P. K. & Dasgupta, P. K. (1995). Continuous automated measurement of the soluble fraction of atmospheric particulte matter. Environ. Sci. Technol., 29, 1534-41.

Stocker, D. W., Stedman, D. H., Zeller, K. F., Massman, W. J. & Fox, D. G. (1993). Fluxes of nitrogen oxides and ozone measured by eddy correlation over a shortgrass prairie. J. Geophys. Res., 98 (D7), 12619-30.

Stull, R. B. (1988). An Introduction to Boundary Layer Meteorology. Atmospheric Sciences Library, Kluwer Academic, Dordrecht, The Netherlands.

Stumm, W., Morgan, J. J. & Schnoor, J. L, (1982). Saurer Regen, eine Folge der Strrung hydrogeochemischer KreisMuse. Springer Verlag, New York, Nr. 1270, Mskr 1-19, pp. 1-8.

Sutton, M. A., Asman, W. A. H. & Schj~rring, J. K. (1992). Dry Deposition of Reduced Nitrogen. UV-ECE workshop on deposition of atmospheric pollutants. Grteborg, 3-6 November 1992.

Sutton, M. A., Fowler, D., Hargreaves, K. J. & Storeton- West, R. L. (1993). Interactions of NH3, and SO2, exchange inferred from simultaneous flux measurements over a wheat canopy. In General Assessment of Biogenic Emissions and Deposition of Nitrogen Compounds and Oxi- dants in Europe. Proceedings of the joint CEC/BIATEX workshop, Aveiro, May 1993, ed. J. Slanina, G. Angeletti & S. Beilke. Air Pollution Research Report 47, CEC, Brussels, pp. 125-39.

Sutton, M. A. & Fowler, D. (1993). A model for inferring bi-directional fluxes of ammonia over plant canopies. In WMO Region II1 Conference on the Measurement and

Modelling of Atmospheric Composition Changes including Pollution Transport, Sofia, Bulgaria, 4-8 October 1993. WMO/GAW, No. 91, World Meteorological Organiza- tion, Geneva, pp. 179-82.

Ulrich, B. & Pankrath, J. (1983). Effects of Accumulation of Air Pollutants in Forest Ecosystems. D. Reidel, Dordrecht.

van Aalst, R. M. (1983). Dry deposition of acid precursors in the Netherlands. VDI-Berichte Nr., 500, 97-102.

van Breemen, N., Burrough, P. A., Velthorst, E. J., van Dobben, F. F., de Wit, T., Ridder T. B. & Reijnders, H. F. (1982). Soil acidification from atmospheric ammonium sulphate in forest canopy throughfall. Nature, 299, 548-50.

Vecera, Z. & Dasgupta, P. K. (1991). Measurement of atmospheric nitric and nitrous acids with a wet effluent diffusion denuder and low-pressure ion chromatography- post column reaction detection. Anal Chem., 63, 2210-16.

Wesely, M. L. (1989). Parametrization of surface resistances to gaseous dry deposition in regional-scale numerical models. Atmos. Environ., 23 (6), 1293-304.

Wunderli, S. & Gehrig, R. (1991). Influence of temperature on formation and stability of surface PAN and ozone. A two year field study in Switzerland. Atmos. Environ., 25A (8), 1599-608.

Wyers, G. P., Vermeulen, A. T. & Slanina, J. (1992). Mea- surement of dry deposition on a forest. Environ. Pollut., 75, 25-8.

Zeller, K., Fox, D. & Massman, W. (1990), Simultaneous measurements of the eddy diffusivities and gradients of ozone, sensible heat and momentum. Presented at the Ninth Conference on Turbulence and Diffusion, AMS Roskilde, Denmark, Paper No. FA 5.4.

Zeman, O. & Jensen, N. O. (1987). Modification of turbulence characteristics in flow over hills. Quart. J. Roy. Meteorol. Soc., 113, 55-80.

deposition of nitrogen-containing compounds to an

Documents