Hatfield J.L., Gandini A. (eds.) Nitrogen in the Environment

Подождите немного. Документ загружается.

344 Nitrogen in the Environment

leading to the formation of NH



aerosol are therefore important to understand the

long-range influence of NH

emissions. These have been summarized by Seinfield

and Pandis (1998) .

Ammonia is rapidly transformed to NH



aerosol in the atmosphere by reac-

tion with acidic compounds, including H

(sulfuric acid), HNO

(nitric acid)

and HCl (hydrochloric acid) according to the following reactions:

NH NH NO

3g g() ()

 

343



(1)

NH Cl NH Cl

3g g() ()

 

(2)

324 424

NH SO NH ) SO

gg() ()

( 

(3)

Direct emissions of H

to the atmosphere occur due to the combustion of sul-

fur rich fuels. However, a more significant source of H

is due to emissions of

and subsequent oxidation by a variety of reactions. In the presence of sulfuric

acid, a rapid non-reversible reaction occurs with NH

to form ammonium sulfate

(Equation 3).

Fine ammonium nitrate (NH

) aerosol is also formed via a reversible gas

phase reaction of NH

with HNO

. At low relative humidities, the rate of produc-

tion or destruction of NH

aerosol is dependent on the equilibrium coefficient

, which is equal to the sum of the partial vapor pressures of HNO

and NH

. K

a strong function of temperature, with lower temperatures shifting the equilibrium

towards an increased mass of NH

. At higher relative humidities, NH

found in the aqueous state, with increasing humidity moving the equilibrium fur-

ther to the aerosol phase. Small changes in relative humidity and temperature will

therefore shift this equilibrium and lead to evaporation/condensation of the aerosol.

Most of the mass of NH



aerosol occurs in the fine “ accumulation ” mode in the

size range 0.1–1  m.

The mechanisms by which both gaseous NH

and particulate NH



are

removed from the atmosphere and deposited to the surface can be broadly divided

into two processes: “ dry deposition ” and “ wet deposition. ”

Dry deposition occurs due to the turbulent mixing of air which leads to the ver-

tical transport of material and impaction of gases and particulates to surface vegeta-

tion. Due to the combination of a low height of emissions (e.g., from livestock or

fertilizer application) and the efficiency with which NH

is deposited to vegetation,

the ground-level concentrations of NH

and deposition to surfaces are usually high-

est in the immediate vicinity of a source. A significant proportion of the emitted gas

can be deposited to vegetation within the nearest 1 km of the source, depending on

the type of land cover and on meteorological conditions. Fowler et al. (1998) found

that 3–10% of the NH

emitted from a poultry unit was deposited within a range

CH11-P374347.indd 344CH11-P374347.indd 344 5/31/2008 6:11:07 PM5/31/2008 6:11:07 PM

Sources, Dispersion and Fate of Atmospheric Ammonia 345

of 300 m from the source. Beyond this distance, the concentration of NH

became

harder to distinguish from background levels. A modeling study by Asman (1998)

found that up to 60% of the NH

emitted from a 3 m high point source could be

deposited to woodland within 2 km of the source. Factors affecting the rate of dry

deposition of NH

to vegetation are considered in detail in Section 3.2.

Wet deposition refers to the natural process by which hydrometeors (cloud and

fog droplets, rain and snow) scavenge material in the atmosphere and deposit it to

the Earth ’ s surface. The terms “ wet deposition, ” “ precipitation scavenging, ” and

“ washout ” are commonly used in the same context. A number of different physical

atmospheric processes contribute to wet deposition. A full description of the micro-

physical processes influencing wet deposition through the formation of clouds and

precipitation is contained in Pruppacher and Klett (1996) .

The processes which occur in clouds include nucleation scavenging (the

mechanism by which a cloud droplet is formed in supersaturated air on an aero-

sol particle), the growth of cloud droplets and conversion to raindrops (by convec-

tive or orographic ascent and cooling of air) and dissolution of gases into cloud

droplets. An overview of in-cloud removal mechanisms is given by Asman (1995) .

The rate of transfer of a gas to the surface of a cloud droplet is determined by the

mass transfer coefficient of the chemical species, the relative aqueous phase and

gas phase concentrations of the species and the Henry ’ s law coefficient (which

relates the aqueous phase concentration to the concentration of the species at

the droplet surface). Due to the high solubility of NH

, it is rapidly absorbed by

cloud droplets and, in equilibrium conditions, most of the NH

in the atmosphere

is contained within the cloud droplets as compared with the within-cloud inters-

titial air.

Below-cloud processes include the washout of both particles and gases by fall-

ing precipitation (rain or snow). The physical mechanism for the washout of gases

by raindrops in the atmosphere is similar to that for in-cloud scavenging of gases by

cloud droplets. However, the mass transfer coefficient of gas to solution is depend-

ent on the ratio of the sizes of the raindrop and the gas molecule. Calculations with

a realistic raindrop size distribution, dependent on precipitation rate ( Marshall and

Palmer, 1948 ), lead to the conclusion that droplets with a diameter smaller than

2 mm are responsible for most below-cloud gas scavenging. Below-cloud wash-

out of aerosols by raindrops is also highly dependent on the sizes of both raindrop

and aerosol particle. For very large aerosol particles, with a diameter above 10  m,

nearly all of the aerosol is scavenged. For ammonium nitrate and ammonium sulfate

aerosol particles, with sizes typically in the range of 0.1–1  m, the scavenging effi-

ciency is at a minimum. In practice, these particles can only be scavenged if they

are close to the center of the volume passed through by a raindrop. This slow rate

of washout of NH



aerosol is the main reason for the long-range transport of these

species. The evaporation of both cloud droplets and precipitation in subsaturated air

leads to the formation of aerosol particles from the aqueous phase. In these conditions,

out-gassing of NH

from solution can also occur ( Bower et al., 1995 ; Dore et al., 2000 ).

CH11-P374347.indd 345CH11-P374347.indd 345 5/31/2008 6:11:07 PM5/31/2008 6:11:07 PM

346 Nitrogen in the Environment

Cloud chemical processing can result in a re-partitioning of the chemical composi-

tion of different aerosol sizes ( Osborne et al., 2001 ).

Low-level orographic clouds occur due to the forced ascent of moist air over

hills. Such clouds are often short lived and do not form into rain. However, the

cloud droplets in these hill clouds form in air close to the ground, which is often

polluted, and the concentrations of ions in hill clouds have been observed to be sig-

nificantly higher than in rain water samples ( Fowler et al., 1988 ; Dore et al., 2001 ).

The washout of cloud droplets by precipitation from above, known as the seeder–

feeder effect, was found to lead to enhanced wet deposition of NH



in upland

areas.

3.2 . Dry Deposition and Bi-directional Exchange Processes

Dry deposition of NH

to a surface involves three main processes: (1) move-

ment from the “ free air ” to the vicinity of the surface; (2) crossing the laminar

boundary layer surrounding the surface; and (3) depositing to the surface at a

molecular level. It is useful to think of these processes using the electrical analogy

of resistances, where each process is assigned a resistance that controls the flow

of NH

through that process. The resistances of the three processes listed above

are called the atmospheric surface layer resistance ( R

), molecular sublayer resist-

ance ( R

) and surface resistance ( R

), which is dependent on surface characteristics

( Sutton et al., 1994 ). With analogy to the calculation of current in electrical circuits

using Ohm ’ s law, the deposition flux is calculated as shown in Eq. 4 and Figure 4a ,

where  is the atmospheric concentration (analogous to the potential difference in

an electrical circuit). The reciprocal of the sum of R

, R

, and R

is also known as

the deposition velocity V

Flux NH

abc

()





RRR

V

(4)

The resistance R

accounts for deposition both to the leaf surface (cuticle) and

the stomata. Although widely used for many trace gases, the approach of Eq. 4 is

often an inadequate simplification for ammonia. The reason is that the R

approach

is constructed on the assumption that the trace gas concentration at the absorb-

ing surface is zero, which for ammonia is often not the case. As a result both dry

deposition and emission of ammonia can occur with both land and water surfaces,

depending on the relative magnitude of the surface and air concentrations. Since

these conditions vary with time as well as in space, “ dry deposition ” of ammonia

is typically seen more as a “ bi-directional ” process. In the case of plant canopies,

Sutton et al. (1995b) showed how plant cuticles may act as a sink for ammonia

at the same time as plant stomata may act as either a sink or source. In this case,

the atmospheric ammonia concentration in equilibrium with substomatal fluids is

referred to as the “ stomatal compensation point ” ( 

). If ammonia concentrations in

CH11-P374347.indd 346CH11-P374347.indd 346 5/31/2008 6:11:07 PM5/31/2008 6:11:07 PM

Sources, Dispersion and Fate of Atmospheric Ammonia 347

the vicinity of stomata are larger than 

, then absorption will occur, whereas if they

are larger than 

then emission will occur from the stomata. However, looking at

the canopy as a whole, ammonia emitted from stomata may subsequently be recap-

tured by deposition to the plant cuticles, (regulated by a cuticular resistance R

This reduces the net NH

emission to the atmosphere and can even switch the flux

to net deposition. The modified resistance analog of Sutton et al. (1995b) is repre-

sented by the model in Figure 4b . In this approach the net flux with the atmosphere

depends on the relative magnitude of the air concentration above the canopy ( 

)

and 

, referred to as the “ canopy compensation point ” ammonia concentration.

Figure 4 . (a) Canopy resistance model; (b) the single-layer canopy compensation point

model according to Sutton et al. (1995b) . R, resistance; Subscripts: a  aerodynamic;

b  boundary layer; w  leaf cuticle; and s  stomata. 

, NH

concentration in air,



and 

stomatal and canopy compensation points, respectively.

Net flux

(a)

Stomata

flux

Net flux

Leaf

cuticle

flux

(b)

The value of 

is strongly dependent on the NH



concentration within the

apoplast and the apoplastic pH ( Farquhar et al., 1980 ), and is a strong function of

temperature, through the Henry and dissociation equilibria. Hence for practical

purposes, 

can be estimated as  · f ( T ) where  is the ratio [NH



]/[H



] of the

apoplast and T is temperature in Kelvin ( Nemitz et al., 2001 ). The canopy compen-

sation point ( 

) ( Sutton and Fowler, 1993 ; Sutton et al., 1995b ; Nemitz et al., 2000 )

can be expressed as shown in Eq. 5.

CH11-P374347.indd 347CH11-P374347.indd 347 5/31/2008 6:11:07 PM5/31/2008 6:11:07 PM

348 Nitrogen in the Environment





aa b ss

ab w s











()

RR R

RR R R

111

(5)

Compensation points are useful quantities for the analysis of ammonia fluxes.

However, it is important to note that 

is a real physical quantity compared with 

which is a calculated concentration representing the balance of several processes,

and therefore it must be clear which compensation point is being referred to. In this

system, R

and R

are controlled principally by meteorological conditions and the

turbulence created by the surface. The stomatal resistance ( R

) is strongly depend-

ent on water vapor pressure and leaf temperature and the cuticular resistance ( R

)

is primarily a function of relative humidity. These complex interactions result in an

exchange flux that is dependent on plant species, meteorology, and the pre-existing

chemical state of the environment.

In fact, Eq.5 gives the simplest possible formulation of the canopy compensa-

tion point. In practice other sources and sinks of ammonia can occur in plant cano-

pies, including the ground surface (from decomposing litter and fertilizer; Nemitz

et al., 2001 ; Riedo et al., 2002 ), while the exchange with the cuticle can itself be

bi-directional depending on adsorption/desorption processes and interactions with

air chemistry ( Sutton et al., 1998 ; Flechard et al., 1999 ).

These complex interactions explain the wide diversity of measured ammonia flu-

xes. In general, deposition dominates to semi-natural ecosystems, while bi-directional

fluxes and emissions predominate over agricultural ecosystems. Fluxes of many

trace gases are typically analyzed in relation to a deposition velocity ( V

), being the

flux divided by 

, which provides a normalized representation of the deposi-

tion rate. Where 

is non-zero, this concept loses its usefulness, since V

itself

becomes by definition a function of 

However, in many semi-natural ecosys-

tems where deposition dominates, analysis of fluxes according to V

has remained

instructive.

Sutton et al. (1993a) measured deposition velocities in the range of 15–20 m m / s

over acid grassland, 1–11 mm/s over calcareous grassland, and 66 mm/s over

coniferous woodland. For heathland, values of 19 m/s, 8 mm/s, and 12 mm/s were

reported by Duyzer (1994) , Hansen (1999) and Flechard and Fowler (1998) , respec-

tively. Over coniferous woodland, values of 20–30 mm/s were measured by Duyzer

et al. (1994) , 32 mm/s by Wyers et al. (1992) , and 14–200 mm/s by Andersen et al.

(1993) . In many cases, these values are toward the upper limits permitted by turbu-

lence ( V

max

 1/( R

 R

), showing that R

is small and that any stomatal/soil sur-

face emissions are fully recaptured within the canopy. As a result, cuticular deposition

is the dominant pathway for NH

deposition to semi-natural ecosystems for most

atmospheric concentrations in temperate and cold localities, represented by the

European studies mentioned. By contrast, there are insufficient data to make gen-

eral statements for tropical and subtropical climates, and detailed measurements

are needed. Exceptions still apply for cool temperate conditions; in the example

CH11-P374347.indd 348CH11-P374347.indd 348 5/31/2008 6:11:08 PM5/31/2008 6:11:08 PM

Sources, Dispersion and Fate of Atmospheric Ammonia 349

of Sutton et al. (1993a) , deposition rates over calcareous grassland were found to

be very small, and this was attributed to the high pH of soil particles which had

been distributed on the surfaces (reducing the net solubility of ammonia). A further

exception occurs adjacent to strong point sources of ammonia, where at very large

concentrations R

increases due to a tendency to saturation of the uptake sites

on the plant cuticle ( Sutton et al., 1993c, 2004a ; Jones et al., 2007 ).

The same principles apply over water surfaces ( Asman et al., 1994 ; Sutton

et al., 1994 ; Bouwman et al., 1997 ). Experimental studies show that atmospheric

fluxes over marine waters may be upward or downward ( Lee et al., 1998 ; Sorensen

et al., 2003 ) depending on the relative concentrations of NH

in air and NH



and



at the water surface. Cool waters in coastal regions are expected to be a net sink

( 

small as cold; 

larger due to adjacent terrestrial sources), while warmer waters

in the remote oceanic environment are expected to be sources ( 

large due to warm

conditions; 

small as remote from terrestrial NH

sources). However, these val-

ues remain extremely uncertain, especially regarding the actual magnitude of 

for ocean surfaces, and the likelihood that  will often be larger in more polluted

coastal waters.

3.3 . Measurements of Concentration and Fluxes

3.3.1 . Spatial monitoring at local and national scales

Monitoring concentrations and the deposition/emission of NH

is important

for assessing the effectiveness of any current and future policies to abate ammonia

emissions, and to improve the understanding of the processes involved. The spatial

variability of NH

concentrations near ground level (1–2 m) is very large. Therefore,

a very dense monitoring network would be required to provide data at a resolution

matching the spatial variability. Particulate NH



, as a secondary pollutant with a

low spatial variability and low formation rates, requires fewer sites.

To assess spatial patterns and temporal trends, basic monitoring with low cost

methods can be implemented at many sites in a network with a low temporal fre-

quency. For example, the UK National Ammonia Monitoring Network (NAMN;

Sutton et al., 2001a ) operates with 95 sites to quantify the spatial distribution and

long-term trends of atmospheric concentrations of NH

and aerosol NH



, using

monthly sampling. At 59 of these sites, a diffusion denuder methodology (DELTA

system: described in detail by Sutton et al., 2001b ) provides the spatial and tempo-

ral patterns of NH

(and NH



at a subset of 43 sites) across the UK. This is com-

plemented by passive diffusion sampling ( Tang et al., 2001 ) at a further 50 sites.

The DELTA method uses a small air pump to sample air and a gas meter to record

sampled volume. Under laminar flow, the selective removal of NH

gas (due to high

diffusion coefficient) onto acid-impregnated glass tubes ( “ denuders ” ) is achieved

( Ferm, 1979 ). Particulate ammonium passes through, which may be collected on

a downstream filter pack. Passive diffusion methods, on the other hand, operate on

the principle of diffusion of gases from the atmosphere along a sampler of defined

CH11-P374347.indd 349CH11-P374347.indd 349 5/31/2008 6:11:08 PM5/31/2008 6:11:08 PM

350 Nitrogen in the Environment

dimensions onto an absorbing medium, according to Fick ’ s law. The theoretical uptake

rate can be calculated, which is a function of the length, L (m), and the cross sectional

area, A (m

), of the stationary air layer within the sampler ( Brown, 2000 ).

The UK NAMN has been operating for over 10 years. Measurements confirm

the high spatial variability of NH

(0.05–15  g/m

), consistent with it being a pri-

mary pollutant emitted from ground-level sources ( Figure 5a ). Ammonium aerosol,

with its longer atmospheric lifetime, has a much smoother spatial distribution than

( Figure 5b ).

Figure 5 . Interpolated annual atmospheric concentrations of (a) NH

and (b) aerosol



for 2004, from the UK NAMN at a 10 k m  10 km grid resolution.

2004

(µg/m

3

)

 0.25

0.25–1

1–2.5

2.5–5

5

100 Kilometers

(a)

CH11-P374347.indd 350CH11-P374347.indd 350 5/31/2008 6:11:08 PM5/31/2008 6:11:08 PM

Sources, Dispersion and Fate of Atmospheric Ammonia 351

Atmospheric NH

concentrations reflect variations in local emission source

types (e.g., sheep, cattle, pigs, etc.) and strengths and are influenced by changes

in meteorological conditions, with warm, dry weather in the summer favoring

increased volatilization. This is illustrated for Inverpolly, a remote site in NW

Scotland, where very low ammonia concentrations occur, influenced by a combina-

tion of extremely low-density grazing emissions, the potential for occasional veg-

etation emissions in summer, and the partitioning between aerosol and NH

from

long-range transported NH



. Figure 6 shows the warmer, drier conditions of summer

2004



(µg/m

)

 0.25

0.25 – 0.5

0.5–1

1–1.5

 1.5

0 50 100 Kilometers

(b)

Figure 5 . ( Continued )

CH11-P374347.indd 351CH11-P374347.indd 351 5/31/2008 6:11:09 PM5/31/2008 6:11:09 PM

352 Nitrogen in the Environment

favor release of NH

(which would occur for each of the three sources mentioned),

leading to peak NH

concentrations.

In Figure 7 , the average seasonal cycles in NH

concentration at NAMN sites

from four different emission source categories are compared. There are a number of

broad patterns in these temporal trends that are related to the emission source types.

Background sites have a strong seasonal cycle with summer maxima and winter

minima, as at the Inverpolly site.

This type of profile is also seen for sites dominated by emission from sheep

grazing, although the summer maximum in NH

emission is larger than for back-

ground sites, because grazing emissions are larger. It is notable that the peak NH

concentration occurs earlier for sheep areas (June–August) than for background

areas (July–September). This may be related to the seasonal presence of lambs,

which are in the fields from spring until the autumn.

By contrast to sheep and background sites, areas with intensive livestock farm-

ing (cattle, pigs, and poultry), show the largest concentrations in spring and autumn,

corresponding to periods of manure application to land. For cattle areas, similarly

sized peaks occur in March and September, while in pig/poultry areas the autumn

0.5

0.4

0.3

0.2

0.1

0.0

(g/m

)

Temperature (°C) / Rainfall (mm  25)

Sep.1996

Jan.1997

May 1997

Sep.1997

Jan.1998

May 1998

Sep.1998

Jan.1999

May 1999

Sep.1999

Jan.2000

May 2000

Sep.2000

Jan.2001

May 2001

Sep.2001

Jan.2002

May 2002

Sep.2002

Jan.2003

May 2003

S3: Inverpolly

(g/m

)

Temperature (°C)

Rainfall (mm  25)

Figure 6 . Relationship between temporal trends in NH

concentrations at Inverpolly

(UK NAMN Site) and meteorological conditions (temperature and rainfall); peaks

in NH

concentrations coincide with hot, dry conditions of summer, whilst minima

occur in cold, wet conditions of winter.

CH11-P374347.indd 352CH11-P374347.indd 352 5/31/2008 6:11:10 PM5/31/2008 6:11:10 PM

Sources, Dispersion and Fate of Atmospheric Ammonia 353

peak is larger, since this is the main period for manure application in arable areas

(i.e., before the sowing of autumn crops). Ammonia concentrations in these areas

are also larger in summer than winter, due to warmer conditions promoting volatili-

zation, adding to the complex temporal profile.

At a local scale, NH

concentrations also show a large degree of spatial vari-

ability, influenced by emission source strength and type. In the vicinity of an emis-

sion source, NH

concentrations generally exhibit an exponential decay away from

the source, with concentrations reaching background levels within 1–2 km of a large

source (e.g., a poultry farm) ( Dragosits et al., 1998, 2006 ; Pitcairn et al., 1998 ). The

simplest approach is to monitor along a transect downwind of the source to measure

the horizontal gradient in concentrations. Depending on resources, passive samplers

(which can only measure NH

) or low-cost denuder systems (e.g., DELTA systems,

which require electricity) that can provide an integrated quantitative measure of both

and NH



concentrations, can be used for these measurements. Figure 8 shows

typical concentration profiles measured downwind of a poultry farm. It should be

noted that the lowest concentrations in Figure 8 , at cf. 8–10  g/m

, remain rather

large, suggesting that this is a residual effect of the farm  320 m, and/or a rather

high regional background in this area, which is an intensive dairy farming area of

Northern Ireland ( Tang et al., 2005 ).

3.3.2 . Detailed process measurements

High time-resolution data are important for identifying peaks in concentration

and the duration of “ acute ” exposures, and to link concentrations with environmental

10.00

1.00

0.10

concentration (µg/m

)

Jan.

Feb.

Mar.

Apr.

May

Jun.

Jul.

Aug.

Sep.

Oct.

Nov.

Dec.

Cattle Pig and Poultry Sheep Background

Figure 7 . Seasonal profiles of NH

concentrations from the mean data (1996–2005)

of sites in the NAMN classified by different dominant emission categories: cattle,

sheep, pig and poultry and background.

CH11-P374347.indd 353CH11-P374347.indd 353 5/31/2008 6:11:11 PM5/31/2008 6:11:11 PM