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

Подождите немного. Документ загружается.

334 Nitrogen in the Environment

including air chemistry ( Eriksson, 1952 ) and quantification of nitrogen balances in

agriculture ( Allison, 1955 ), which motivated the first measurements of ammonia

fluxes (e.g., Hanawalt, 1969 ; Denmead et al., 1974 ). However, there was already

an awareness of the ammonia problem at the start of the 19th century, even while

the principles of plant nitrogen nutrition remained uncertain. In the British Farmers

Cyclopedia ( Potts, 1807 ), the agricultural pioneer Arthur Young argued for the

benefits of immediate use of manures to avoid ammoniacal and other losses to the

atmosphere. As Young strikingly put it:

He who is within the sphere of the scent of a dunghill, smells that which his

crop would have eaten, if he would have permitted it. Instead of manuring his

land, he manures the atmosphere; and before his dunghill is finished turning,

he has manured another parish, perhaps another county.

Such issues of emission, dispersion, and deposition remain of central impor-

tance to the ammonia problem of today ( Asman et al., 1998 ; Sutton et al., 1998 ;

Dragosits et al., 2006 ).

In this chapter, the current state of knowledge on sources, dispersion, and fate

of atmospheric ammonia is reviewed. While the analysis is globally relevant, we

particularly select examples from our own UK-based work to illustrate issues.

This chapter covers the characteristics, spatial distribution, and temporal trends of

emission sources, dispersion, deposition and effects on the environment, at mul-

tiple scales from local to global. Modeling and measurement methodologies are

discussed, including combinations of these approaches for validation. Finally, we

address issues of emission abatement and policy strategies, taking account of links

with other forms of nitrogen and pollution swapping.

2 . AMMONIA SOURCES, SPATIAL DISTRIBUTION, AND TEMPORAL

TRENDS OF EMISSIONS

2.1 . Sources of Ammonia

Global NH

emissions are estimated at 58.2 Tg N/year for the mid-1990s

( Galloway et al., 2004 ; Galloway, 2005 ), including marine emissions. This estimate

is similar to earlier publications by Bouwman et al. (1997) and Olivier et al. (1998) at

54 Tg N/year for 1990, and Holland et al. (1999) at 46.2 Tg N/year for the 1980–1990

(with an uncertainty range of 45–83 Tg N/year).

The main anthropogenic sources of NH

in Europe and North America are, in

decreasing order of total emissions: agricultural livestock manures and slurries, appli-

cation of N fertilizer to crops and grassland, and other miscellaneous non-agricultural

sources. In addition, on a global scale, biomass burning and the oceans may be

important sources.

Emissions of ammonia from agricultural livestock are associated with all

stages of manure management: housing of animals, manure and slurry storage

and application to agricultural land (both crops and improved grassland), as well

CH11-P374347.indd 334CH11-P374347.indd 334 5/31/2008 6:10:59 PM5/31/2008 6:10:59 PM

Sources, Dispersion and Fate of Atmospheric Ammonia 335

as with grazing livestock on pastures. Livestock convert only a small amount of

the nitrogen in their feed to products such as milk, eggs, meat, and wool. Van der

Hoek (1998) estimates that global N use efficiency of livestock ranges from 4% for

goats, 8% for cattle, and 21% for pigs to 34% for poultry, with an average value

of 11%. The remainder of the ingested N is excreted in urine and dung (manure).

For instance, estimates of N excretion for high-yield dairy cows in Western Europe,

New Zealand, and North America are between 100 and 160 kg N/year ( Oenema and

Tamminga, 2005 ). These estimates are approximately three times higher than in

developing countries (60 kg N/year – Bouwman et al., 1997 ; 45 kg N/year – Smil,

1999 ; and 60–70 kg N/year – Mosier et al., 1998 ).

There are, however, large global variations and uncertainties associated with

these estimates, and differences within one animal type can be as large as differ-

ences between different animal types, depending on the genetic potential of the ani-

mal breed, feed, and management systems ( Oenema and Tamminga, 2005 ). When

the excreta come into contact with air, ammonia easily volatilizes (e.g., Monteny

and Erisman, 1998 ).

Most emissions of reduced nitrogen occur in the form of NH

, when the sur-

face concentration is larger than that of the surrounding air ( Sutton et al., 1993b ;

Asman et al., 1998 ). This is usually the case in animal manures and slurries, as

well in N-containing mineral fertilizers. For plants or oceans, where NH

concen-

trations in equilibrium with the canopy or water surface may be more similar to

air concentrations, deposition as well as emission can occur ( Sutton et al., 1993b ).

Ammonia volatilization processes vary for different forms of agricultural N, such

as urea (contained in urine and mineral urea fertilizers), uric acid (a main constitu-

ent of poultry manures), and mineral fertilizers such as urea and ammonium nitrate

(e.g., Hutchinson, 1950 ; Jarvis and Pain, 1990 ). For all these sources, ammonia is

emitted by volatilization from ammonium in aqueous solution or from volatile salts

attached to surfaces. As a result, the principles that regulate solubility and disso-

lution are those that primarily drive the magnitude of these ammonia emissions.

Thus emissions are largely promoted in warm drying conditions, while the small-

est emissions generally occur in cool wet conditions. There are nevertheless many

exceptions. For example, water is needed for the mineralization of ammonium from

organic matter and the hydrolysis of urea, and can promote ammonia emission from

these sources, while very wet soils can limit infiltration of surface-spread livestock

slurries, which can also increase emissions.

Losses of NH

by volatilization from the application of N fertilizers range from

negligible amounts to  50% of the applied fertilizer N, depending on fertilizer/

manure type, application practice (e.g., injection, surface application), and envi-

ronmental conditions ( Peoples et al., 1995 ; Freney, 2005 ). Roelcke et al. (2002)

report NH

emissions of up to 50% of the applied fertilizer N on calcareous soils

on the loess plateau of China (in this case for ammonium bicarbonate and urea),

and volatilization rates of 20% to  80% have been observed for flooded rice pad-

dies typically fertilized with urea ( DeDatta et al., 1989 ; Mosier et al., 1989 ; Freney

et al., 1990 , in Freney, 2005 ). On average, volatilization rates from N fertilizer

CH11-P374347.indd 335CH11-P374347.indd 335 5/31/2008 6:10:59 PM5/31/2008 6:10:59 PM

336 Nitrogen in the Environment

application are highest from urea (global average: 21%), ammonium sulfate (16%)

and ammonium bicarbonate (this being mainly used in China), according to a

review of 149 research papers by Bouwman et al. (2002) . The same study reports

global average NH

volatilization rates of 14% for ammonium nitrate fertilizers and

of 23% for livestock manures.

Other agriculture-related emissions of ammonia occur from, for example, bio-

mass burning or fertilizer manufacture. Ammonia is also emitted from a range of

non-agricultural sources, such as catalytic converters in petrol cars, landfill sites,

sewage works, composting of organic materials, human breath and sweat, babies ’

nappies, cigarette smoking, combustion, industry, domestic pets, wild mammals,

wild birds, etc. ( Sutton et al., 2000 ; Wilson et al., 2004 ). Ammonia emissions from

these sources together are, however, relatively small compared with agricultural

sources. In the UK, they are estimated to contribute about 10–15% of the total NH

emissions. By contrast, at a global-scale Bouwman et al. (1997) estimate that 36%

is from sources other than livestock and fertilizers. Of the global total of 54 Tg NH

emission, they estimate that 7.2 Tg (or 13%) originate from natural terrestrial emis-

sions (soils, human excreta), 8.2 Tg (15%) from oceans, 5.9 Tg (11%) from biomass

burning, and  1% from fossil fuel combustion and industrial processes.

2.2 . Estimating NH

Emission Source Strength and Emission Inventories

emissions from livestock (and other sources) may be calculated using

either simple “ emission factors ” or by estimating nitrogen flows:

● In the emission factor approach, volatilization losses are calculated for

each of the stages of livestock manure management (housing, grazing, etc.)

under average conditions, using the latest available scientific publications

on measured emission fluxes and adding these individual losses together to

achieve an emission “ factor ” per animal ( Sutton et al., 1995a ). Losses dur-

ing livestock housing, manure storage, manure spreading and grazing may be

expressed as component emission factors, which are very useful for examin-

ing the differences between source strength estimates.

● More recently, Asman (1992) and Sutton et al. (1995a) , for instance, started

analyzing the flow of nitrogen through the livestock husbandry system fol-

lowing excretion. With this approach, the contribution of the component

emissions to the total are linked to the nitrogen available for volatilization at

each stage, taking account of any losses at previous management stages. This

approach has been used by TFEI (1996) and Cowell (1998) .

Until recently, the “ emission factor ” method has been used for calculating the offi-

cial UK NH

inventory ( Pain et al., 1998 ; Misselbrook et al., 2000 ). However since

2005 (for the inventory year 2004), the “ N flow ” method ( Webb et al., 2002 ; Webb

and Misselbrook, 2004 ; Misselbrook et al., 2006 ) has been adopted.

CH11-P374347.indd 336CH11-P374347.indd 336 5/31/2008 6:10:59 PM5/31/2008 6:10:59 PM

Sources, Dispersion and Fate of Atmospheric Ammonia 337

Ammonia emission inventories are commonly calculated at national levels (e.g.,

UK – Misselbrook et al. (2006) , Ireland – Hyde et al. (2003) , South Korea – Lee

and Park (2002) , Denmark – Hutchings et al. (2001) , Canada – Kurvits and Marta

(1998) , USA – Anderson et al. (2003) , Pinder et al. (2004) ), and are often updated

annually for policy purposes. Emission inventories have become increasingly impor-

tant as a tool for policy makers, informing them of the current status and for pre-

dicting scenarios for the future, as well as for submission to international bodies,

such as the United Nations Economic Commission for Europe (UNECE). They thus

become essential for reporting emissions in relation to national abatement targets,

for example, under the Gothenburg Protocol ( UNECE, 1999 ), which came into force

in 2005. Many countries have developed their own inventory methodology using

national emission factors best suited for their agricultural systems, and guidebooks

such as the “ Atmospheric Emission Inventory Guidebook ” (UNECE Task force on

Emission Inventories and Projections ( TFEIP, 2004 )) have been published to pro-

vide assistance to countries in the form of guidelines and default methodologies.

Ammonia emission inventories have also been compiled at continental scales (e.g.,

Europe – Erisman et al., 2003 ; East Asia – Klimont et al., 2001 ), and global scales

(e.g., Bouwman et al., 1997 ; Olivier et al., 1998 ; Holland et al., 1999 ; Bouwman et

al., 2002 ; Galloway et al., 2004 ; Galloway, 2005 ). More recently, local-scale ammo-

nia emission inventories have been developed at a “ landscape ” scale (e.g., Dragosits

et al., 2002, 2005, 2006 ; Theobald et al., 2004 ; see Section 2.4).

2.3 . Temporal Trends

During the early history of humans, the nitrogen cycle was not changed signifi-

cantly, neither by hunter gatherers, nor by early agriculture ~ 10,000 years ago, apart

from mobilizing existing reactive nitrogen (N

) by, for example, biomass burning or

local use of livestock manures as fertilizer ( Galloway, 2005 ). This resulted in only

very small local NH

emissions. From 5000 years BP (Before Present) the amount

of N

began to increase with the cultivation of legumes, with the amount of biologi-

cally fixed N increasing slowly over the next few millennia. But essentially, rotation

systems with fallow periods dominated and restricted agricultural production, and

livestock and crop production were closely coupled. Ju et al. (2005) illustrate how

recycling of organic manures, legume rotations, and green manures have all played

a part in maintaining and increasing yields slowly over the last two millennia in

China (until the advent of inorganic fertilizers in the 20th century).

With mining and transportation of natural N-rich deposits such as guano from

South America to Europe, larger amounts of additional nitrogen started to become

available and were applied in areas far from the source. However, it was not until

the large-scale application of the Haber–Bosch process, (i.e., fertilizer production

by synthesis of NH

from hydrogen and nitrogen, invented and commercialised in

the early 20th century by F. Haber and K. Bosch, respectively), that the amount of N

added into the global N cycle started to increase dramatically. Through this develop-

ment, global NH

emissions started to increase from an estimated 21 Tg N/year in 1860

CH11-P374347.indd 337CH11-P374347.indd 337 5/31/2008 6:10:59 PM5/31/2008 6:10:59 PM

338 Nitrogen in the Environment

to 58 Tg/year in the early 1990s, speeding up rapidly from the 1960s ( Galloway,

2005 ), with these figures including a fixed marine NH

emission of 5.6 Tg/year for

both periods. Galloway et al. (2004) estimate that NH

emissions may increase to

119 Tg N/year by 2050, with large amounts of additional (fertilizer) nitrogen needed

to feed the growing global population.

Between 1960 and 2000, the human population roughly doubled, while the

number of domestic animals tripled (FAO, 2003). This reflects a global increase in

the consumption of animal protein per capita ( Smil, 2002 ). Van Drecht et al. (2003)

estimate that fertilizer N consumption increased by 600% over the same period.

Large increases are expected to continue in developing countries, while slow

increases in, for example, North America or (further) decreases are expected in,

for example, Western Europe ( Van Drecht et al., 2003 ; Bouwman et al., 2005 ). In

the UK, agricultural NH

emissions peaked in the late 1980s/early 1990s, and have

decreased since then, mainly through (ongoing) decreases in livestock populations

rather than the implementation of abatement measures. By contrast, other European

countries, such as the Netherlands or Denmark, have enforced extensive environ-

mental regulations to increase N use efficiency and reduce NH

emissions. Some

of the decrease in agricultural production in developed countries may be attributed

to the globalization of markets, but also to concerns about animal welfare and envi-

ronmental damage caused by intensive agriculture.

In 2000, Asia (especially the south and east) consumed  50% of the world ’ s N

fertilizer production, and by 2030 Asian NH

emissions are expected to increase by

 30% through increasing intensification of animal production and large increases

in fertilizer consumption ( Zhu et al., 2005 ). The increases in fertilizer use are

mainly expected in the form of urea, which has one of the highest NH

volatiliza-

tion rates of mineral N fertilizers. Global fertilizer consumption statistics show that

in 2001, Asia accounted for 70% of the world urea consumption and that this is a

rising trend, as urea is inexpensive to produce and has a relatively high N content

( Prud ’ homme, 2005 ).

2.4 . Spatial Distribution of Ammonia Emissions

Unlike other pollutants that are mainly associated with combustion processes,

industry, and transport (e.g., SO

, NO

, heavy metals), ammonia is emitted mainly

from agricultural/rural sources, therefore the spatial distribution patterns are signifi-

cantly different. Due to the reactive nature of NH

gas, a considerable proportion of

the NH

emitted from a source does not travel further than about 5 km. Singles et

al. (1998) and Sutton et al. (1998) estimated that approximately 5–50% of all NH

emissions are deposited within a 5 km radius from their source. The deposition of

and its effects in receptor ecosystems are intrinsically linked with the sources

through atmospheric transport mechanisms, and these are simulated by a wide range

of atmospheric chemistry and transport models. Spatially disaggregated invento-

ries are the primary input into such models, and are hence a fundamental require-

ment to simulate atmospheric NH

concentrations, deposition, and impacts, as

well as to develop abatement strategies. Achieving a reliable spatial representation

CH11-P374347.indd 338CH11-P374347.indd 338 5/31/2008 6:10:59 PM5/31/2008 6:10:59 PM

Sources, Dispersion and Fate of Atmospheric Ammonia 339

of NH

emissions is essential if the models are to be used to assess the location

of areas affected by ammonia and nitrogen deposition. This is especially important

in landscapes where intensive agricultural production and semi-natural ecosystems

exist in close proximity, which is the case for large parts of the UK (e.g., Dragosits

et al., 2002, 2006 ) and other European countries.

National NH

emission in the UK are mapped at a 5-km grid resolution, using

the AENEID model ( Dragosits et al., 1998 ) for agricultural sources, and at a 1- or

5-km grid resolution for non-agricultural sources, and are freely available at www.

naei.org (National Atmospheric Emission Inventory). An example map of ammonia

Figure 1 . The spatial distribution of UK NH

emissions: (a) emissions from all

agricultural and non-agricultural sources and (b) dominant NH

sources.

(a)

Kg NH

-N / ha / year

0–1

 50

0 100 200 km

1–2.5

2.5–5

5–10

10–25

25–50

CH11-P374347.indd 339CH11-P374347.indd 339 5/31/2008 6:10:59 PM5/31/2008 6:10:59 PM

340 Nitrogen in the Environment

emissions from the UK is shown in Figure 1a for 2004, demonstrating a very high

spatial variability in emissions. Accompanying this image, Figure 1b shows the main

ammonia emission sources in the different parts of the country. High emission areas

with intensive dairy farming can be distinguished from low emission areas with

extensive sheep and beef farming, or the “ hot-spot ” patterns associated with inten-

sive pig and poultry farming. Some non-agricultural emission sources (e.g., seabird

colonies) contribute only small amounts to the overall NH

emissions in the UK, but

are – due to their location – often the dominant emission source in remote and other-

wise “ clean ” areas. Larger seabird colonies have been shown to emit similar amounts

of NH

to large intensive poultry farms ( Sutton et al., 2000 ; Wilson et al., 2004 ).

Figure 1 . ( Continued )

Dominant NH

source

Background ( 1kg/ha/year)

Cattle

Sheep, goats and horses

Pigs and poultry

Fertilisers and crops

Non-agricultural

Mixed

100

200 km

(b)

CH11-P374347.indd 340CH11-P374347.indd 340 5/31/2008 6:11:00 PM5/31/2008 6:11:00 PM

Sources, Dispersion and Fate of Atmospheric Ammonia 341

Effects of N deposition on individual semi-natural areas can best be assessed

by detailed modeling and measurements, aided by a landscape/local-scale emission

inventory, at a resolution of fields, woodlands, and farmsteads (e.g., Dragosits et al.,

2002, 2006 ; Theobald et al., 2004 ). Local-scale inventories allow emissions to be

estimated at the spatial scale of environmental effects, and these might be used in

short-range atmospheric transport models (ATMs) to address questions of spatial

variability of NH

deposition and impacts. This is illustrated in Figure 2 , which

shows an area of approximately 5 k m  5 km in central England, the equivalent of

a single grid square in the national inventory. This figure shows the detailed mix

that occurs between source fields and farms and ecosystems which are receiving

0–2.5

2.5–5

5–10

10–30

25

Emission (kg NH

-N / ha / year)

1 2 km

Nature reserves

Semi-natural vegetation

Figure 2 . Example local-/landscape-scale emission inventory: the spatial distribu-

tion of emissions in a study area in central England of approximately 5 k m  5 km.

F1–F4 represent four livestock farms and R1–R3 represent three nature reserves

(adapted from Dragosits et al., 2006).

CH11-P374347.indd 341CH11-P374347.indd 341 5/31/2008 6:11:01 PM5/31/2008 6:11:01 PM

342 Nitrogen in the Environment

and sensitive to ammonia deposition. The study area represents a real landscape in

central England, and Figure 2 shows a land-use scenario with three areas designated

for nature protection (R1–R3), four livestock farms (F1–F4), and mixed grazed and

fertilized fields ( Dragosits et al., 2006 ).

As discussed in Section 2.3, global emissions of NH

are predicted to increase

considerably over the next few decades, especially in southern and eastern Asia.

Figure 3 . The spatial distribution of global NH

emissions estimated for 1970, 2000,

and predicted for 2030 ( courtesy of Lex Bouwman, RIVM, The Netherlands ).

0–4

4–8

8–16

16–32

32

N ha

1

(a)

(b)

CH11-P374347.indd 342CH11-P374347.indd 342 5/31/2008 6:11:04 PM5/31/2008 6:11:04 PM

Sources, Dispersion and Fate of Atmospheric Ammonia 343

This will polarize current spatial patterns of NH

emissions even further, with parts

of Asia providing the largest “ hot-spots ” globally by far ( Figure 3 ). Bouwman et al.

(2005) estimate that the contribution of industrialized countries to global NH

emis-

sions, which decreased from 33% in 1970 to 22% in 1995, will decrease to 18% by

2030 for intensive agriculture (i.e., excluding pastoral systems). This projection is

based on assumptions of improving agricultural practice and technology, leading to

increased N efficiency.

3 . ATMOSPHERIC TRANSPORT, TRANSFORMATION, AND

DEPOSITION

3.1 . Transport and Transformation

Ammonia is a soluble and reactive gas which is readily deposited to vegetation.

These properties of NH

contribute to a relatively short lifetime in the atmosphere,

typically of a few hours following emission ( Flechard and Fowler, 1998 ). Following

emission to the atmosphere, the transport of NH

is by meteorological processes

such as advection and convection. Ammonia is readily absorbed by cloud droplets

where an equilibrium reaction in solution leads to the formation of the ammonium

(NH



) ion. In normal atmospheric conditions, for a cloud water pH below 8, prac-

tically all the dissolved NH

in clouds is in the form of the NH



ion. Although NH

gas has a relatively short atmospheric lifetime, NH



is an important component of

atmospheric aerosol (which has a longer lifetime) and is associated with interna-

tional and even intercontinental scale transport. The atmospheric chemical reactions

(c)

Figure 3 . ( Continued )

CH11-P374347.indd 343CH11-P374347.indd 343 5/31/2008 6:11:06 PM5/31/2008 6:11:06 PM