handbook of atmospheric science || atmospheric particulate matter

9.1 INTRODUCTION

In the past several decades, atmospheric aerosolshave increasingly been recognized as constitutingone of the major uncertainties in the current un-derstanding of climate change (IPCC 1996, 2001).The uncertainty is mainly due to the large variabil-ity in their physical and chemical properties, aswell as their temporal and spatial distributions.Atmospheric aerosols may originate from eithernaturally occurring or anthropogenic processes.Major natural aerosol sources include emissionsfrom the oceans such as sea spray, volcanoes, and mineral dust from arid regions, while majoranthropogenic sources include emissions from industry and combustion processes. Further classification within each category into so-calledprimary and secondary sources may also be made.Aerosols directly emitted into the atmosphereconstitute primary sources, while secondarysources arise from the gas-to-particle conversionof gaseous precursor compounds such as nitricoxide and nitrogen dioxide (collectively known asNOx), sulfur dioxide (SO2), and hydrocarbons.

Characterization of the life cycle of atmos-pheric aerosols is a complex and many-facetedissue. A schematic is illustrated in Fig. 9.1, whichsummarizes: (i) aerosol sources; (ii) transformationmechanisms in the atmosphere; and (iii) aerosolsink processes. As aerosol size is one of the mostimportant parameters in describing aerosol prop-erties and their interaction with the atmosphere,its determination and use is of fundamental impor-

tance. Aerosol size covers several decades in diam-eter and hence a variety of instruments are re-quired for its determination. This necessitatesseveral definitions of the diameter, the most com-mon being the geometric diameter d, which is usedhere. Another commonly used aerosol diameter isthe aerodynamic diameter, which normalizes fordensity and shape. A good discussion of these as-pects is given in Reist (1993) and Hinds (1999).With reference to Fig. 9.1, the size fraction with di-ameter d > 1–2mm is usually referred to as thecoarse mode, and the fraction d < 1–2mm is the finemode. The latter mode can be further divided intothe accumulation (d ~ 0.1–1mm), Aitken (d ~0.01–0.1mm) and nucleation (d < 0.01mm) modes.Due to the d3 dependence of aerosol volume (andmass), the coarse mode is typified by a maximumvolume concentration and, similarly, the accumu-lation mode by the surface area concentration andthe Aitken and nucleation modes by the numberconcentration.

Aerosol formation arises from heterogeneous orhomogeneous nucleation. The former refers tocondensation growth on existing nuclei, and thelatter to the formation of new nuclei through con-densation. Heterogeneous nucleation preferen-tially occurs on existing nuclei. Condensationonto a host surface occurs at a critical supersatura-tion, which is substantially lower (<1–2%) than forhomogeneous nucleation in the absence of impuri-ties (>300%). Examples of gas-to-particle conver-sion are combustion processes and the ambientformation of nuclei from gaseous organic emis-

9 Atmospheric Particulate Matter

URS BALTENSPERGER, STEFAN NYEKI, ANDMARKUS KALBERER

Handbook of Atmospheric Science: Principles and ApplicationsEdited by C.N. Hewitt, Andrea V. Jackson

Copyright © 2003 by Blackwell Publishing Ltd

Atmospheric Particulate Matter 229

sions. The high initial number concentration ofparticles with d < 0.1mm are rapidly reducedthrough coagulation, resulting in aerosol lifetimesof the order of minutes for these smallest particles.

Particles in the Aitken/accumulation modetypically arise from either: (i) the condensation oflow-volatility vapors; or (ii) coagulation. Atmos-pheric aerosols may be composed of a range ofchemical species. When particles are chemicallydistinct from one another, the aerosol is termed anexternal mixture. Alternatively, when particleshave a similar composition they are known as aninternal mixture. This may physically range fromcomplete mixing of chemical components (e.g. hy-

groscopic components in a cloud droplet) to vari-ous other structures, such as a coated aerosol (e.g.sulfate coating on soot aerosols). Particles in theaccumulation mode have a longer atmosphericlifetime than other modes, as there is a minimumefficiency in sink processes. Of these processes,wet deposition (in-cloud and below-cloud scaveng-ing) is the major sink process.

Particles in the coarse mode are usually pro-duced by weathering and wind erosion processes.Dry deposition (primarily sedimentation) is thedominant removal process. Chemically their com-position reflects their sources, as demonstrated bymineral dust from arid regions and sea salt from

Hotvapor

Lowvolatility

vapor

Condensation

Primary particles

Coagulation

Chain aggregates

Chemical conversionof gases to low

volatility vapors

Homogeneousnucleation

Condensation growthof nuclei

Droplets

Coagulation

Coagulation

Rainout&

washout Sedimentation

Wind blown dust+

Emissions+

Sea spray+

Volcanoes+

Plant particles

0.002 0.01

Transient nuclei orAitken nuclei range

Fine particles Coarse particles

Accumulationrange

Mechanically generatedaerosol range

0.1

Particle diameter (mm)

1 2 10 100

Coagulation

Fig. 9.1 Schematic of the aerosolsurface area concentrationillustrating the Aitken,accumulation and coarse modes.(Reprinted from AtmosphericEnvironment (Whitby & Sverdrup1980). Copyright 1980, withpermission from Elsevier Science.)

230 urs baltensperger E T A L .

oceans. Organic compounds such as biological(spores, pollens, and bacteria) and biogenic parti-cles resulting from direct emission of hydrocar-bons into the atmosphere may also be constituentsof the coarse mode. As the sources and sinks of thecoarse and fine modes are different, there is only aweak association of particles in both modes.

The above brief summary of atmosphericaerosols serves to illustrate the complex processesinvolved in modeling their behavior and assessingtheir influence on climate. It is mainly for the lat-ter reason that interest in aerosols has grown oflate and hence this chapter focuses on the climateeffect of aerosols. Aerosol types and compositionare considered first, followed by their interactionwith radiation through the so-called direct and in-direct aerosol effects. Further reading on aerosolfundamentals and atmospheric aerosols may befound in the following list of reprinted and new books: Levine (1991), Baron and Willeke(2001), Hobbs (1993), Reist (1993), Charlson andHeintzenberg (1995), Singh (1995), Pruppacher andKlett (1997), Seinfeld and Pandis (1998), Brasseur et al. (1999), Finlayson-Pitts and Pitts (1999), andHinds (1999). A list of acronyms used throughoutthe following discussion may be found at the end ofthis chapter.

9.2 SIZE DISTRIBUTION,COMPOSITION, AND

CONCENTRATION

Knowledge of the different moments of an aerosolsize distribution and its composition are of pri-mary importance, as most other parameters of in-terest may be deduced from this information. Theatmospheric aerosol may be classified into severalcategories, according to source and geographic location.

Figure 9.2 illustrates typical size distributionsfor number, surface, and volume concentration forvarious aerosol types. A more recent compilationof size distributions, mainly from the Atlantic andEuropean regions (Raes et al. 2000), exhibits simi-lar features to those in Fig. 9.2. The number sizedistribution of the atmospheric aerosol may be ap-

proximated by an empirical power law equationfor radii r > 0.1mm (e.g. Jaenicke 1988, and refer-ences therein):

(9.1)

where N is the aerosol number concentration, c is aconstant and v depends on the aerosol type. Thelarge range in magnitude of N is demonstrated byvalues <20cm-3 for the polar regions and >105 cm-3

for an urban aerosol. Not only is there a large geo-graphic variation in aerosol concentrations, butthe vertical extent also varies substantially. For instance, a background aerosol at 3km over theoceans or 5km over the continents is typified by N~ 150cm-3 and the stratospheric aerosol at 20kmby N ~ 10cm-3. Urban, remote continental, and re-mote marine aerosol models from Fig. 9.2 may bedescribed as a first approximation by typical chem-ical compositions appearing in Table 9.1. Sulfateand organics are the major aerosol components ofurban and remote continental regions and sodiumchloride of remote marine regions. The aerosolchemical composition and geographic type arethus recognized as being fairly specific to anaerosol source and are discussed below in greaterdetail.

9.3 AEROSOL SOURCES

The different types of aerosol sources mentionedabove, i.e. natural and anthropogenic, primary and secondary, are described in greater detail inthis section. An estimate of annual atmosphericcontributions to major aerosol sources is given inTable 9.2. Estimates suggest that anthropogenicemissions of SO4

2- are greater than from naturalsources. As a result of large spatial and temporalvariations in aerosol sources and sinks, the range ofvalues in Table 9.2 highlights the uncertainties involved in estimating emissions. Owing to theshort atmospheric lifetime of aerosols it should beremembered that globally averaged figures inTable 9.2 are not necessarily indicative of localburdens. For instance, natural sources will domi-nate on a global scale due to their large areal emis-

d dlogN r cr= -n


sion sources (e.g. deserts and oceans). In contrast,anthropogenic emissions from industrialized re-gions in Europe, the USA, and East Asia, which arerelatively smaller, are likely to exceed the contri-butions from natural sources.

9.3.1 Natural sources

Primary emissions

Sea-salt aerosol. Sea-salt aerosol results from thebursting of bubbles, formed by wave and wind ac-tion at the ocean surface (e.g. O’Dowd et al. 1997).As a result, sea-spray droplets are ejected and ei-ther return to the water surface or evaporate toform inorganic/organic aerosols, which may then

be entrained into the marine boundary layer (MBL)by wind turbulence. The wind speed is a major fac-tor in controlling the sea-salt aerosol concentra-tion and roughly exhibits a linear dependence. Thefairly global uniformity of sea water compositionis reflected in the composition of sea-salt aerosol,although enrichment of particular species duringformation or subsequent chemical transformationin the atmosphere may occur. The composition ofsea-salt aerosol is given in Table 9.1, and is mainlythe salts NaCl, KCl, CaSO4, and Na2SO4. Solubleand insoluble organic compounds may also be animportant component, the fraction depending on anumber of parameters, such as location and time ofseason. As a result of the uniform trace aerosol

Urban

Rural

Remotecontinental

Back-ground

Polar

Desertdust

storm

Stratosphere

10–3

10–8

10–7

10–6

10–5

10–4

10–3

10–2

10–1

102

103

104

105

106

1

10

10–2

10–1

Radius (mm)

Num

ber

per

cm

–3

102

1 10

Remotecontinental

Back-ground

Polar

Desertdust storm

Stratosphere

10–3

10–12

10–11

10–10

10–9

10–8

10–7

10–6

10–5

10–4

10–3

10–2

10–1

1

10

102

10–2

10–1

Radius (mm)

Surf

ace

are

a (

cm2)

per

cm

3

102

1 10

Urban

RuralBack-ground

Desertdust storm

Urban

Rural

Remotecontinental

Polar

Stratosphere

10–3

10–19

10–18

10–17

10–16

10–15

10–14

10–13

10–12

10–11

10–10

10–9

10–8

10–7

10–6

10–5

10–2

10–1

Radius (mm)Volu

me

(cm

3)

per

cm

310

21 10

(a) (b) (c)

Fig. 9.2 (a) Number, (b) surface, and (c) volume size distributions (dN/dlogr, dS/dlogr, and dV/dlogr, respectively) forvarious atmospheric aerosols. Aerosol types are self-explanatory apart from background aerosol, which refers to thetropospheric aerosol 5km above the continents and 3km above the oceans. (From Jaenicke 1988.)


chemical composition over the oceans, source regions cannot be easily identified.

The mass median diameter of sea-salt aerosolnear the sea surface is about 8mm, which results in a short lifetime and inefficient light-scatteringproperties. However, smaller sea-salt aerosols maybe transported over the ocean surface and a globalannual emission of 1300Tgyr-1 (Table 9.2) is con-sidered as representative of the fraction present inthe lower MBL.

Mineral dust. Wind-blown mineral dust fromdesert and semiarid regions is an important sourceof tropospheric aerosols (Duce 1995) and of partic-ular interest in paleoclimatological studies due toits inert properties (Leinen & Sarnthein 1989).Mineral dust arises from the physical and chemi-cal weathering of rock and soils. The wind speed is

the main controlling factor in entraining particlesinto the atmosphere, followed by other factors,such as soil moisture and surface composition. At-mospheric size distributions in the vicinity of soilsources are generally bimodal, in which the range d ~ 10–200mm consists mainly of quartz grains,and the range d < 10mm of clay particles. Quartzgrains will preferentially sediment close to theirsource, resulting in a fractionation process from a quartz/clay to a clay aerosol with increasing distance. The size distribution will also changewith downwind distance as a result of sedimenta-tion. Measurements have indicated that the modalvolume diameter changes from d ~ 60–100mm to d ~ 2mm at a distance of 5000km, where it ap-pears to stabilize.

The principal elemental constituents of min-eral dust (oxides and carbonates of Si, Al, Ca, Fe)

Table 9.1 Typical mass composition (mgm-3) of various chemical species in urban, remote continental, and remotemarine aerosol types (adapted from Pueschel 1995).

Element or Urban aerosol — photochemical Remote continental Remote marinecompound smog aerosol aerosol

SO42- 16.5 0.5–5 2.6

NO3- 10 0.4–1.4 0.05

Cl- 0.7 0.08–0.14 4.6

Br- 0.5 – 0.02

NH4+ 6.9 0.4–2.0 0.16

Na+ 3.1 0.02–0.08 2.9

K+ 0.9 0.03–0.01 0.1

Ca2+ 1.9 0.04–0.3 0.2

Mg2+ 1.4 – 0.4

Al2O3 6.4 0.08–0.4 —

SiO2 21.1 0.2–1.3 —

Fe2O3 3.8 0.04–0.4 0.07

CaO — 0.06–0.18 —

Organics 30.4 1.1 0.9

EC 9.3 0.04 0.04

Total 112.9 2.99–12.44 12.04

Selected mass fractions and

molar ratios

SO42- (%) 15.9 30.2–45.7 22.6

NO3- (%) 9.6 13.3–22.7 0.44

NH4+/SO4

2- 2.2 2.1–3.4 0.47


bear a close resemblance to the average crustalcomposition and may be used to identify source regions. Due to the inert nature of mineral dust, chemical transformation processes in the at-mosphere are thus considered minor, althoughsurface chemical reactions may be important, e.g. reactions with gaseous HNO3 and SO2,hence enhancing coarse mode sulfate and nitrate, respectively.

Mineral dust is estimated to contribute 1500Tgyr-1 to global atmospheric emissions (Table 9.2),which originates from areas totalling about 10% ofthe Earth’s surface. This compares to a similar sea-

salt emission from oceans covering an area largerthan 70% of the Earth’s surface. Principal sourceregions of mineral dust cover about one-third ofthe land surface and include the Saudi Arabianpeninsula, the US Southwest, and the Sahara andGobi deserts. Only the latter two regions are signif-icant sources of long-range transported dust, oc-curring mainly westward over the tropical NorthAtlantic and eastward over the North Pacific, re-spectively. Although mineral dust is considered anatural emission, recent work has suggested thatbetween 20 and 50% of the current atmosphericburden may arise from disturbed soils through

Table 9.2 Global emission source strengths for atmospheric aerosols (Tgyr-1).

Aerosol component d’Almeida et al. (1991) Pueschel (1995) IPCC (1995)* Aerosol size mode

Natural

Primary

Sea salt 1,000–10,000 300–2,000 1,300 Coarse

Mineral dust 500–2,000 100–500 1,500 Mainly coarse

Primary organic 80 3–150 50 Coarse

aerosols/biological debris

Volcanic ash 25–250 25–300 33 Coarse

Secondary 345–1,100

Sulfate from biogenic gases 121–452 90 Fine

Sulfate from volcanic SO2 9 12 Fine

Nitrate from NOx 75–700 22 Fine/coarse†

Organics from biogenic VOC 15–200 55 Fine

Natural total 1,950–13,430 648–4,311 3,062

Anthropogenic

Primary

Industrial dust 10–90 167 100 Coarse/fine

Biomass burning 3–150 29–72 80 Fine

Soot (all sources) 24 10 Mainly fine

Secondary 175–325

Sulfate from SO2 70–220 140 Fine

Nitrate from NOx 23–40 40 Fine/coarse

Ammonium from NH3 269

Organics from VOC 15–90 10 Fine

Anthropogenic total 188–565 597–882 380

Overall total 2,138–13,995 1,245–5,193 3,442

*“Best” estimate.

†Relative fractions uncertain.

Numbers have changed slightly in IPCC (2001), which appeared after finalization of this chapter.


human activity (Sokolik & Toon 1996; Tegen et al.1996).

Primary organic aerosols/biological debris.These categories of natural emissions from thebiosphere have only recently been investigated inany detail. Continental sources mainly arise fromvegetation (plant waxes and fragments, pollen,spores, fungi, and decaying material), while ma-rine sources consist of organic surfactants formedvia bubble bursting. Typical size distributions aredominated by the coarse mode. Continental andmarine source burdens are similar, with a totalglobal emission of 50Tgyr-1, although the scarcityof data makes such an estimate very uncertain.

Volcanic emissions. Although volcanic activityoccurs on a sporadic basis and is mainly located in the northern hemisphere, the more recent outbreaks of Mount St Helens (USA, 1980), ElChichón (Mexico, 1982), and Mount Pinatubo(Philippines, 1991) have highlighted the impor-tance of volcanic emissions to the atmosphere.These emissions are composed of ash (principallySiO2, Al2O3, and Fe2O3), gases (SO2, H2S, CO2, HCl,HF), and water vapor. Mount Pinatubo was esti-mated to have emitted 9Tg sulfur (S) compared to3.5Tg(S) for El Chichón, of which a large fraction inthe form of SO2 was buoyantly injected into thestratosphere. Table 9.2 gives an estimated emis-sion of 33Tgyr-1 for ash particles in the coarsemode, which mainly limit their impact to the regional scale and to the troposphere. Of greaterlong-term importance is the emission of SO2 intothe stratosphere and the subsequent formation ofH2SO4 aerosol droplets, which generally exhibit abimodal structure in the Aitken accumulationmode size range. Whereas aerosols in the tropo-sphere have an average lifetime of 5–7 days, duemainly to wet scavenging by clouds, stratosphericaerosols have a 6–9-month lifetime, due to thethermal stratification of the stratosphere and theabsence of wet scavenging. Stratospheric aerosolremoval mechanisms are primarily sedimenta-tion, subsidence, and exchange with the upper troposphere. Although Table 9.2 indicates that explosive volcanic emissions may only contribute

up to 10–20% of the total natural sulfur emissionto the atmosphere, the radiative impact of stratos-pheric aerosols may be quite significant. Figure 9.3illustrates the effect that El Chichón and MountPinatubo had on stratospheric optical properties.Peak values of the integrated aerosol backscatter,as measured with a ground-based lidar (Osbornet al. 1995), occurred about 6 months after eacheruption. During this time, H2SO4 droplets of asufficient size to interact with radiation wereformed. A nonvolcanic background aerosol is alsoevident in Fig. 9.3 and is attributed to the forma-tion of H2SO4 from the upward flux of COS (car-bonyl sulfide), emitted from the oceans. Theclimatic effect of stratospheric aerosols is furtherdescribed in Section 9.5.2, while a good review oftheir formation, properties, and effects is given byPueschel (1996).

Secondary emissions

Secondary natural aerosols may be formed from anumber of natural precursor gas sources, contain-ing sulfur, nitrogen, and hydrocarbons. The mainnatural source is the release of gaseous dimethylsulfide, or DMS (CH3SCH3), from the oceans. DMSis formed from the biological activity of phyto-plankton and eventually forms aerosol sulfate viathe photooxidation to methanesulfonic acid andSO2. The contribution to sulfate from DMS andother sources, except sea water, is known as nss (nosea salt) sulfate to differentiate it from sea water as a source. Emissions of nss sulfate are estimatedat 90Tgyr-1, toward which other marine sulfursources, such as H2S, CS2, and COS contribute lessthan 10% of total sulfur emissions. The seasonalvariation of DMS follows the ocean productivitycycle and may be a magnitude higher in the sum-mer than in the winter season. Concentrations ofnss sulfate in the remote MBL generally range from20–800ngm-3 for the southern hemisphere oceansto 400–3000ngm-3 for the northern Atlantic andillustrate the enhanced contribution from anthro-pogenic sources in the northern hemisphere(Heintzenberg et al. 2000).

Typical global concentrations above the MBLrapidly decline with altitude to several ngm-3 in


the troposphere. In comparison, typical valuesover the continents are <100ngm-3. These concen-trations may be put into context by consideringthat <50% of sulfate in the MBL is of marine origin,while the rest may be attributed to soil dust and anthropogenic sulfate.

The formation of nitrate aerosols from nitrogenprecursor gases has two main natural sources: (i)NOx from lightning and soils; and (ii) N2O frombacterial activity in soils and the oceans. An emis-sion rate of 22Tgyr-1 is estimated in Table 9.2.

Secondary organic aerosols or SOAs are formedby oxidation products of volatile organic com-pounds (VOCs). These reactions are mainly initiat-ed by reactions with ozone and OH and NO3radicals. Some products of these oxidation reac-tions have a low enough vapor pressure for parti-tioning between the gas phase and the aerosolphase to become significant. The equilibrium con-stant K, describing the partitioning of a compoundbetween the gas and the particle phase, is mainly afunction of its vapor pressure and of its activity inthe particulate phase.

(9.2)

where RG is the gas constant, T is the temperature,MOM is the mean molecular weight of the organic

K R T pMG OM= g

particulate phase into which the compound parti-tions, g is the activity coefficient of the compoundin the particulate phase, and p is its vapor pressure.Thus the formation of SOA is described by K for allcompounds that are found in the organic aerosolphase. This, however, is a challenging task, sincegenerally only a small fraction of the organicaerosol mass can be analyzed on a molecular level.However, laboratory experiments have shownthat for most cases the overall aerosol yield of acompound can be parameterized assuming two hypothetical compounds with different K. This allows the SOA mass in modeling studies to be estimated without knowing the actual molecularcomposition of the organic particulate phase. Theaerosol yield Y of a single compound is defined as:

(9.3)

where DM0 is the organic mass formed, DHC is thefraction of hydrocarbon reacted, and ai is a stochio-metric coefficient. The aerosol yield of an organiccompound was shown to depend on its structure(e.g. number and location of double bonds) and onthe gaseous reaction partners available (Griffin et al. 1999a). In general, compounds with a larger

Y M HC Ma K

Kj i

ii

= =+ÂD D0 0 1

1974 1976 1978

Inte

gra

ted b

ack

scatt

er (

1/

sr)

Fueg

o

August

ine

Sier

ra N

egra

St. H

elen

sU

law

un

Ala

id P

agan

Nya

mura

gir

aEl C

hic

hon

Una U

na

Ruiz

Pin

atu

bo

Spurr

1980 1982 1984 1986 1988 1990 1992 1994

10–6

10–5

10–4

10–3

10–2

Fig. 9.3 The integrated aerosolbackscatter (a measure of the totalstratospheric column of aerosol)using a ground-based lidar at l =0.694mm. Injections of aerosolsassociated with El Chichón andMount Pinatubo are seen to besuperimposed on a naturalbackground. (From Osborn et al.1995.)


all biomass emissions. Table 9.2 gives an anthro-pogenic component of 80Tgyr-1, which consists ofsoot, sulfate, nitrate, and incomplete combustionproducts containing carbonaceous compounds.Release of biomass products to the atmosphere oc-curs mainly in the Tropics during the dry seasons,i.e. December to March in the northern hemi-sphere and June to September in the southernhemisphere. Although Table 9.2 only considersparticle emissions, gaseous emissions of CO2, CO,CH4, and VOCs are also important, and the lattermay result in SOA formation.

Industrial aerosols. Aerosol emissions from in-dustrial processes, estimated at 100Tgyr-1, have adiverse number of sources. Major sources in indus-trialized countries include: coal and mineral dustfrom mining, aerosols formed from incombustibleinorganic compounds in oil and coal fuels, stone-crushing, cement manufacture, metal foundries,and grain elevators. However, decreasing emis-sions in industrialized nations are occurringagainst rapidly rising emissions in emerging na-tions, where the implementation of modern tech-nologies is not keeping pace with rapid economicand industrial development.

Soot aerosols. Soot is a ubiquitous component ofthe atmospheric aerosol. As a result of it beingchemically relatively inert and having poor hygro-scopic properties, it may be used as an anthro-pogenic tracer. For instance, a recent study of sootin an alpine ice-core exhibited enhanced soot con-centrations since the beginning of industrializa-tion in the mid-nineteenth century and has beenattributed to anthropogenic activity in Europe(Lavanchy et al. 1999). Soot aerosols are generatedin incomplete combustion processes and consistmainly of an organic fraction and an inorganicgraphite-like carbon fraction. These two fractionsare often not clearly distinguished in the litera-ture, which is in large part due to technical diffi-culties in distinguishing them. The graphite-likecarbon fraction is also often known as black carbon(BC) or elemental carbon (EC). Depending on thesource and burning conditions, the amount of sootis highly variable. For instance, biomass soot has a

number of carbon atoms have a larger aerosol yield.However, compounds with fewer than six carbonatoms are mostly unable to form SOAs, becausetheir oxidation products are too small and thus toovolatile to partition into the particle phase.

Terpenes emitted from plants are considered tobe the largest class of natural VOCs able to formSOAs. Guenther et al. (1995) estimated the globalemission of terpenes and other reactive organiccompounds at 127Tgyr-1, which is equally dis-tributed between the northern and southern hemispheres. From emission data and laboratorymeasurements of aerosol formation yields (Griffinet al. 1999a), the global formation rate of SOA hasbeen estimated at 13–24Tgyr-1 from this source(Griffin et al. 1999b).

Information on the molecular composition ofbiogenic SOA is still largely unavailable. In gener-al, the small amounts of sample available foranalysis and the highly oxidized compounds makechemical analysis difficult. Studies concerning thecomposition of laboratory-generated aerosols atthe molecular level have only recently begun to re-solve an appreciable amount of the organic particlemass. In ambient aerosol samples only 10–20% ofthe organic aerosol mass can usually be analyzedon a molecular level. In ambient samples it is diffi-cult to distinguish between primary and secondaryorganic mass, as no simple experimental methodsexist to separate these two fractions.

It is assumed that SOA mass grows mainly onexisting aerosol particles (such as salt aerosols orsulfuric acid nuclei). However, some events havebeen observed in field measurements where nucle-ation of new particles was attributed to low-volatility organic compounds (Marti et al. 1997). Atpresent it is unclear how important this process is.

9.3.2 Anthropogenic sources

Primary emissions

Biomass burning. Natural wild fires and anthro-pogenic fires (e.g for agricultural clearing) are com-monly termed biomass burning (Levine 1991). Thelatter source has grown so quickly in the past twodecades that it is estimated to account for 95% of


high organic carbon content, in contrast to dieselsoot with a high EC content. Emissions of anthro-pogenic soot aerosols from fossil fuel combustionare estimated at ~10Tgyr-1, which include contri-butions from biomass sources.

Secondary emissions

Sulfate and nitrate aerosols. The main atmos-pheric source of secondary particles is the oxida-tion of SO2 and NOx. It is estimated that about 50%of SO2 and NOx is oxidized before being deposited(Langner & Rodhe 1991), where oxidation mayoccur in either the gas or condensed phases. Gas-phase oxidation to both H2SO4 and HNO3 is domi-nated by OH. For the condensed phase, about 50%of NOx and ≥80% of SO2 is oxidized to HNO3 andH2SO4, respectively, by heterogeneous reactions.Oxidation of SO2 always results in the formationof aerosol mass, due to the low H2SO4 vapor pres-sure, and is in contrast to HNO3, which is distrib-uted between the gas and aerosol phases. Thechemical transformation of gases into particles de-pends on many factors, including chemical reac-tion kinetics and physical factors such as plumemixing and dispersion, oxidant concentration,sunlight, and catalytic aerosol surfaces. Despitethis, the conversion rates of SO2 are generallyaround 1–2% per hour and somewhat higher for ni-trate. Table 9.2 indicates that current sulfate emis-sions from anthropogenic sources exceed naturalsources, while different estimates exist in the caseof nitrate.

The largest source of secondary anthropogenicaerosol comes from fossil fuel emissions of SO2and subsequent conversion to H2SO4. Over conti-nental surfaces in the PBL or above, where gaseousammonia is present, H2SO4 forms NH4HSO4 and(NH4)2SO4. These components may exist simulta-neously and are illustrated by varying molar ratiosof NH4

+/SO42- according to the atmospheric

aerosol type, as in Table 9.1. In regions of the strat-osphere and upper troposphere, H2SO4 is found tobe the major aerosol component.

Atmospheric emissions from fossil fuel com-bustion have been increasing since the beginningof industrialization in about 1850. The current in-

dustrial SO2 emission of about 70–90TgSyr-1 ac-counts for about 80–85% of the total SO2 annualflux in the northern and 30% in the southernhemisphere (see Berresheim et al. 1995; Möller1995). Ninety percent of these emissions arise inindustrialized regions of the northern hemisphere.Little mixing occurs into the southern hemispheredue to a long inter-hemispheric mixing time, of ~1 year, which compare to aerosol lifetimes of ~1 week.

Current estimates of NOx emissions from bio-mass burning and NH3 oxidation (~17TgNyr-1)are lower than the anthropogenic value of ~32TgNyr-1 from fossil fuel combustion. The formation ofHNO3 from NOx is a major removal mechanismfor tropospheric NOx as most HNO3 is subse-quently lost through wet and dry deposition.

Sulfate aerosol concentrations are more stableto fluctuations in H2SO4 concentration, tempera-ture and humidity conditions, in contrast to ammonium nitrate and chloride aerosols. Forthese aerosols, the reversible reactions shownbelow occur to form the parent gaseous com-ponents under conditions of low atmospheric am-monia concentration, high temperature and lowhumidity:

where (g) and (s) denote the gaseous and solid phases, respectively. The main source of atmos-pheric HCl is from refuse incineration and coalcombustion.

Ammonia plays an important role in the neutralization of acid species, as it is the mostcommon atmospheric alkaline gas. Conversion toammonium salts is a function of not only altitude,but also of temperature and humidity. Major natural and anthropogenic sources respectively include: (i) soils and organic decomposition, (ii) fertilizers and animal farming, and (iii) catalyzedvehicle. The annual emission of ammonia is esti-mated at 269Tgyr-1.

The aqueous-phase production of aerosol mate-rial on cloud droplets is an important mechanism

NH Cl s NH g HCl g4 3( ) ´ ( ) + ( )

NH NO s NH g HNO g4 3 3 3( ) ´ ( ) + ( )


in nonprecipitating clouds. Cloud droplets, whichform on CCN, may undergo on average 10 evapora-tion/condensation cycles before precipitabledroplets are formed. During this process, gaseousspecies are scavenged and undergo chemical trans-formation, while aerosols and other droplets arescavenged by coagulation and phoresis mecha-nisms. As a result, the aerosol mass and hygro-scopicity increases, in turn increasing the CCN activity. The conversion of SO2 and NH3, whichare dissolved in droplets, appears to be an efficientprocess for the production of NH4HSO4 and(NH4)2SO4. Such processes have been postulated asresponsible for the rather uniform composition of the background tropospheric aerosol (e.g. Raeset al. 2000).

Secondary organic aerosol. Fossil fuel combus-tion and biomass burning, caused by human activ-ities, are the main sources of anthropogenic VOCsthat can lead to SOAs. VOC emissions due to fossilfuel combustion are estimated at ~65Tgyr-1, i.e.60% of total anthropogenic emissions (Piccot et al.1992). The same principles apply to anthropogenicSOA formation as for SOA formed from biogenicprecursors compounds.

Aromatic compounds are the main class of an-thropogenic VOCs that lead to significant aerosolformation. Pandis et al. (1992) estimated that aro-matics of anthropogenic emissions can be respon-sible for about two-thirds of the total SOAformation in an urban atmosphere. Aerosol yieldsof different single aromatic compounds, whencompared to the SOA yield of whole gasolinevapor, show that aromatics are mainly responsiblefor SOA from fossil fuel VOC emissions (Odum et al. 1997). As in the case of biogenic aerosol particles, little is known about the molecular composition of anthropogenic SOA.

9.4 HETEROGENEOUS CHEMISTRY

In a general sense, heterogeneous reactions are reac-tions involving more than one phase, such as reac-tions of gaseous molecules with compounds on asolid or liquid surface or in bulk liquids. Some reac-

tions that are unfavorable in the gas phase are able tooccur on the surfaces or in the bulk of atmosphericaerosol particles. Many authors (e.g. Ravishankara1997) further split these heterogeneous reactionsinto heterogeneous reactions in a narrower sense,i.e. on solid surfaces, and multiphase reactions inliquids. In the following, the term “heterogeneousreactions” is used in the broad sense.

There are two different implications when considering such reactions. First, they modify theaerosol composition; second, they influence gasphase chemistry. Although the importance of het-erogeneous reactions in tropospheric and especial-ly in stratospheric chemistry (e.g. ozone depletionin polar regions) has been realized, only few exper-imental data exist compared to gaseous reactions,which are relatively well characterized. One of themajor difficulties in investigating heterogeneousreactions in laboratory experiments is to accurate-ly simulate relevant aerosols, as ambient particlesare often complex mixtures of organic and inorgan-ic compounds, and water.

The most basic example of a heterogeneous re-action, the adsorption of a gaseous molecule onto asolid particle surface, can be described with the following equation:

where species in { } denote surface-bound com-pounds. The rate with which gaseous moleculesadsorb to aerosol particles is described by:

(9.4)

where J is the flux, Nmol is the number concentra-tion of the molecule of interest, d is the aerosol di-ameter, Ni(d) is the particle number concentrationwith particle diameter d and B(d) is the attachmentcoefficient. A variety of different equations for B(d)exist in the literature, among which a simplifiedformula gives B(d) as:

(9.5)B d

DdD

cd dMFP

( ) =+

+

28 1

1 2

p

g l

J N N d B dmol ii

= ( ) ( )Â

A g B AB( ) + { } Æ { }


where D, c, and lMFP are the diffusion coeffi-cient, the mean thermal velocity, and the meanfree path of the diffusing molecule, respectively.The parameter g is the dimensionless sticking coefficient, defined as the ratio of reactive gas molecule-particle collisions to total gas molecule-particle collisions, i.e. an indication of adsorptionefficiency.

When g = 1, eqn 9.5 illustrates that for smallparticles (i.e. d << gMFP) B(d) μ d2, while for largeparticles (i.e. d >> lMFP) B(d) μ d. However, for g £0.01, eqn 9.5 may be simplified to the following forall particle diameters of interest (d £ 10mm):

(9.6)

Equations 9.4 to 9.6 describe an overall adsorption,but do not take the chemical nature of this processinto account. Several different categories may bedistinguished, which is sometimes a reason forconfusion in the literature. A gaseous moleculecan be physically (or reversibly) adsorbed, i.e. nocovalent bonding is formed between the adsorbentand the adsorbate. In many cases this is the firststep in a heterogeneous reaction. The reversiblyadsorbed molecule may desorb again from its physisorbed state, hence resulting in equilibriumbetween the gaseous and the particulate phases.However, the adsorption process can also result incovalent bonding of the gas molecule and a particlesurface compound (chemisorption). If the com-pound remains irreversibly adsorbed on the parti-cle (i.e. the residence time on the particle is longerthan the lifetime of the aerosol in the atmosphere),adsorption then represents a permanent sink forthe gaseous compound in its atmospheric lifecycle.

The adsorbed molecule may also undergo achemical reaction on/in the aerosol, resulting ineither an altered aerosol composition or a release ofa reaction product into the gas phase, or both. De-pending on the aerosol chemical and physical (i.e.solid or liquid) composition, heterogeneous reac-tions may be quite diverse, and are governed bymany factors.

B d cd( ) =14

2p g

9.4.1 Dry aerosols

Reactions of gaseous compounds with dry, solidaerosols have remained relatively uncharacter-ized. It has been known for quite a while that thereaction between sea-salt aerosols and gaseous ni-tric acid results in the release of HCl into the gasphase:

Another example is the reduction NO2 undergoeson the surface of soot aerosol particles, where ni-trous acid (HONO) is the main gaseous reactionproduct.

The products {Cred}(s) and {Cox}(s) denote a reducedand an oxidized compound on the soot particle sur-face, respectively (Ammann et al. 1998). Nitrousacid is an important precursor of the most impor-tant oxidant in the lower troposphere, the OH rad-ical, and is easily photolyzed in the atmosphere.The reaction efficiency appears to largely dependon the chemical characteristics of soot, renderingthe applicability to ambient atmospheric condi-tions difficult. However, on other chemically inertsurfaces, NO2 appears to react with absorbed waterto nitrous and nitric acids.

where HONO is released into the gas phase andHNO3 remains predominantly on the surface. Several studies on the interaction of gaseousHNO3 with crustal aerosols exist (e.g. Phadnis & Carmichael 2000), while other reactions such as the interaction of OH radicals are mostly spe-culative (Saylor 1997).

9.4.2 Liquid aerosols

When considering liquid aerosols (i.e. cloud, fog, orhaze droplet), eqns 9.4 to 9.6 are also applicable,where g is then referred to as the uptake coeffi-cient. After accommodation, the molecule dis-

2NO g H O ads HONO g HNO ads2 2 3( ) + ( ) Æ ( ) + ( )

NO g C s HONO g C s2 red ox( ) + { }( ) Æ ( ) + { }( )

HNO g NaCl s HCl g NaNO s3 3( ) + ( ) Æ ( ) + ( )


solves into the bulk solution of a droplet. Chemi-cal reactions occur mainly in the bulk, although ithas been found that reactions may occur on thesurface of a liquid aerosol in some systems. Onecan envisage the overall interaction of a gas mole-cule with a liquid aerosol droplet using an electri-cal circuit analogy (Davidovits et al. 1995). In thismodel the overall measured uptake of a gas mole-cule into a liquid aerosol (gmeas) can be described asthe sum of dimensionless coefficients:

(9.7)

The first two terms represent processes in the gas phase (diffusion to the droplet, gdiff) and at the gas/liquid interface (accommodation on thedroplet surface, a), respectively. The third term de-scribes processes in the droplet, i.e. the chemicalreaction (grn) and the solubility of the gas moleculein the liquid (gsol). Chemical reactions in the liquiddepend heavily on the chemical composition ofthe liquid aerosol, i.e. parameters such as the ionicstrength and redox potential play an importantrole. Thus different reactions occur in cloud orhaze droplets, as a result of differing water content.

An important example of heterogeneous reac-tions in cloud droplets is the aqueous oxidation ofSO2. The main oxidation paths are reactions withaqueous H2O2 and ozone:

accommodation, solution

in liquid

in liquid

Another important example of a heterogeneous re-action scheme in liquid aerosols is the nighttimeformation of nitric acid :

in gas phase

in gas phase

in liquidN O aq H O 2 HNO2 5 2 3( ) + Æ

NO NO M N O M3 2 2 5+ + Æ +

NO O O NO2 3 2 3+ Æ +

SO aq O aq H O H SO O2 3 2 2 4 2( ) + ( ) + Æ +

SO aq H O aq H SO2 2 2 2 4( ) + ( ) Æ

SO g SO aq2 2( ) Æ ( )

1 1 1 1g g a g gmeas diff m sol

= + ++

During the day, these reactions are of minor impor-tance due to the rapid photolysis of NO3 radicals.Model calculations have shown that with these re-actions, the yearly average global NOx burden de-creases by 50%, due to a decreased residence timein the atmosphere (Dentener & Crutzen 1993). Ob-served nitrate wet deposition patterns in NorthAmerica and Europe are better simulated if theseaerosol reactions are included.

In conclusion, heterogeneous reactions are ofpotential importance in atmospheric chemistry,but experimental studies remain sparse.

9.5 CLIMATE FORCING

Aerosols are now recognized as being a major un-certainty in studies of global climate change (IPCC2001). Emissions of anthropogenic aerosols to the atmosphere may explain the lower observedtemperature increase than is otherwise predictedfor greenhouse gas emissions. Aerosols are consid-ered to be responsible for a negative forcing or cooling of the Earth-atmosphere system, in con-trast to a positive forcing, i.e. “warming,” fromgreenhouse gases. In this definition, a forcingrefers to a natural or anthropogenic perturbation inthe radiative energy budget of the Earth’s climatesystem.

Aerosols may influence the atmosphere in two important ways, by direct and indirect effects (Charlson et al. 1987, 1992; Charlson &Heintzenberg 1995; Andreae & Crutzen 1997;Baker 1997). Direct effects refer to the scatteringand absorption of shortwave radiation and the sub-sequent influence on the climate system and plan-etary albedo (see Schwartz 1996). Indirect effectsrefer to a complex positive feedback system,whereby an increase in CCN arises from an in-crease in anthropogenic aerosol concentration. Asa consequence, cloud droplets become smaller fora given cloud liquid water content, thereby in-creasing cloud albedo and resulting in a negativeforcing. The principal aerosol–cloud interactionsare schematically summarized in Fig. 9.5, whichillustrates the life cycle in marine and continentalair respectively. Many of the physical and chemi-


cal aspects of the life cycle have already been con-sidered in Sections 9.1 to 9.4.

The climate effects of aerosols are still poorlyquantified and it is at present still unclear whatmagnitudes are involved. Figure 9.4 illustrates recent estimates of the mean annual radiative forcing for various climate change mechanisms,averaged globally for the period 1750–2000 (IPCC2001). Current estimates suggest that positivegreenhouse forcing is dominated by anthropogenicemissions of CO2, CH4, N2O, and halocarbons. Additional mechanisms, such as the depletion ofstratospheric ozone, the formation of troposphericozone by photochemical smog, and variation inthe solar irradiance, are believed to contributesmaller forcings. When viewed on a global scale,shortwave radiative forcing from anthropogenicaerosols is considered to offset part of the long-wave radiative forcing due to greenhouse gases.However, on a more regional scale, the large uncer-tainty in aerosol radiative forcing compared to thatfor greenhouse gases does not allow a meaningfulnet value from all mechanisms to be defined.

As already mentioned, aerosols may be respon-sible for the lower than predicted temperature in-

crease in the northern hemisphere. While this mayapply on a global basis, the same argument cannotbe applied on a regional basis for a number of rea-sons. First, greenhouse gases have lifetimes meas-ured in decades to centuries and are globally wellmixed, in contrast to aerosols, which have life-times of less than one week on average and are geo-graphically variable in extent. Second, aerosolforcing responds more rapidly to changes inaerosol emissions than greenhouse forcing, whichis still influenced by accumulated past emissions.Third, aerosol radiative forcing is more restrictedto source and downwind regions, which may thenexperience an overall negative forcing. Lastly,greenhouse gas forcing is not as diurnally and sea-sonally variable as aerosol forcing, which has agreater influence during: (i) daylight, (ii) cloudlessconditions, and (iii) summer. Such difficultiesgreatly hinder the modeling of aerosol direct andindirect effects.

It is important to note that in the above defini-tion of climate forcing, complex feedback systems,such as the hydrological cycle, are not considered,as they are in the overall discussion of climatechange. Although this simplifies the considera-

Fig. 9.4 Estimates of globally andannually averaged anthropogenicradiative forcing from pre-industrial times to the present.Confidence levels (from high tovery low) in each estimate arereflected in the uncertainty rangeof the error bars. Due to theepisodic and sporadic nature ofvolcanoes, they are omitted fromthe above estimates, althoughtheir effects on time scales of theorder of decades may besignificant. (From IPCC 2001.)

SO2

SO2

H2SO4

H2SO4

Oxidation(a)

(b)

NucleationCondensational

growthSmall SO4aerosol

2–

SO4, NO3production

2– –

Small SO4aerosol

2–

Transportof DMSto upper

troposphereby

convectiveclouds

Detrain tofree troposphere

Detrain tofree troposphere

Sedimentationfrom free troposphere

Cuclouds

Condensationalgrowth

Absorption

Aerosolremoval in

precipitation

Aerosolremoval in

precipitation

Activation

Activation

Activation

Activation

Aerosol producedin and aroundclouds

Aerosol producedin and aroundclouds

Aerosolproducedaroundclouds

Sea SaltOrganicsCondensation

Oxidation

Oxidation

Oxidation Nucleation

Combustion

Combustion

Biomassburning

Industry Automobiles Ground

Oxidation

SO2, NOxorganics

Nucleation

DMS

DMS

Activation

Activation

AqueousSO4

production

2–

CCN

MSA

Condensationalgrowth

Soil dustsaltation

Absorption

Cuclouds

LargeCCN

Aerosolremoval in

precipitation

Small, combustionnuclei

H2SO4, HNO3,organic acids

Ocean

St–Scclouds

St–Scclouds

Aerosolremoval in

precipitation

Fig. 9.5 Schematic of aerosol–cloud interactions for (a) marine air and (b) continental air. Cloud types: Cu, cumulonimbus; St, stratus; Sc, stratocumulus. (From Hobbs 1993.)


tion of aerosols, it does not fully consider theiroverall impact. In order to assess the direct aerosoleffect on climate change, the following aspects oftropospheric aerosols are further considered: (i) theradiative properties of aerosols; and (ii) estimatesof aerosol direct radiative forcing.

9.5.1 Direct aerosol effects: radiativeproperties of aerosols

The interaction of aerosols with radiation dependsprimarily on the aerosol diameter and the wave-length g of radiation. The largest interaction occurs when the aerosol size parameter pd/l ~ 1.As a consequence, the longer-lived accumula-tion mode interacts with the shortwave solar radiation spectrum to a greater extent than withthe longwave infrared spectrum emitted by theEarth’s atmosphere and surface, which peaks at l ~ 7mm.

Optical properties of aerosols may be describedusing a number of parameters. Scattered or ab-sorbed radiation from an incident beam is definedby the total scattering (sSP) and absorption coeffi-cients (sAP), which are a measure of the fractionalchange in beam intensity per meter. Radiationscattered into the backward hemisphere is correspondingly described by the hemisphericbackscattering coefficient (sBSP). The sum of sSPand sAP is the extinction coefficient sEXT and whenintegrated over the beam path length dl gives theaerosol optical depth (AOD: W).

(9.8)

(9.9)

An important parameter in global aerosol modelsis the ratio of scattering to extinction, otherwiseknown as the single scattering albedo wo:

(9.10)

The extinction coefficient and its components areoften approximated as being proportional to l-å

where å is the Ångström exponent and K is a con-

w s so SP EXT=

W = Ús EXTdl

s s sSP AP EXT+ =

stant (eqn. 9.11). Furthermore, å is related to theJunge parameter v by eqn 9.12.

(9.11)

(9.12)

Typical values of å range from ~4 for gases, ~2 forurban aerosols, and ~1–2 for rural haze, to ~0 forcoarse aerosols. Hence if å can be measured theninformation on the number size distribution isgained and vice versa.

The asymmetry factor g is defined as the cosineweighted mean of the angular scattering phasefunction b(f), where b(f) describes the amount oflight scattered through an angle f:

(9.13)

The value of g ranges from -1 for complete backscattering to +1 for complete forward scattering.The asymmetry factor is important in radiativemodels of the atmosphere and takes the angularscattering of radiation into account. As directmeasurement is not possible, it may be parameter-ized by the hemispheric backscattering ratiosBSP/sSP.

The radiative effect of aerosols on global cli-mate may be predicted using two computer modeltypes: (i) a general circulation model (GCM) providing information on atmospheric state parameters (wind and temperature fields, etc.) andradiative properties; and (ii) a chemical transportmodel (CTM) to provide information on aerosolemission, transport, transformation, and removal.Early models used prescribed databases of meas-ured aerosol parameters (e.g. d’Almeida et al. 1991;Koepke et al. 1997), while recent progress has fo-cused on developing algorithms to dynamicallymodel many complex aspects of atmosphericaerosols, such as the interaction with radiation orclouds. Global distributions of sulfate aerosol havebeen widely modeled. A good review and compari-son of currently used sulfate GCM/CTM models isgiven by Barrie et al. (2001) and IPCC (2001). More

gd

d=

( ) ( )( ) ( )

ÚÚ

b f f f

b f f

cos cos

cos

å = -n 2

s lEXTåK= -


recent model studies have been widened to includeglobal distributions of soot (e.g. Cooke et al. 1999),mineral dust (e.g. Miller & Tegen 1998), and seasalt (e.g. Gong et al. 1997).

GCM/CTM models approach global direct ra-diative forcing by considering column-integratedand local aerosol optical properties. Both ap-proaches are necessary to determine the verticalstructure of the atmosphere, and to validateground-based against aircraft and satellite observa-tions. As a number of assumptions are inherent in such calculations, measurements of column-integrated and local properties are ideally con-ducted simultaneously (Quinn et al. 1996). Muchprogress has been made in this respect through re-cent column-closure experiments in internation-ally coordinated field campaigns such as TARFOXand ACE II (Russell et al. 1999; Russell &Heintzenberg 2000).

Column-integrated properties refer to aerosolproperties averaged throughout the atmosphericcolumn. Such measurements have been com-monly conducted using Sun photometers by sub-tracting the attenuation from atmospheric gasesand water vapor to give the AOD. Despite numer-ous in situ Sun photometer measurements, satel-lites are gaining in importance, due mainly to their large spatial coverage of the Earth’s surface.These aspects are discussed in Section 9.6.

Local measurements involve directly measur-ing the above-defined aerosol parameters. Aerosoloptical properties may be derived from in situmeasurement or by calculating their propertiesthrough knowledge of the aerosol chemical com-position, size distribution, and refractive index. As

many CTMs calculate global aerosol mass concen-tration distributions, the ability to simultaneous-ly derive distributions of scattering and absorptionby aerosols is advantageous. Parameters that relateaerosol scattering and absorption coefficients tomeasurements of the aerosol concentration andcomposition are the specific scattering aSP andspecific absorption aAP efficiencies.

(9.14)

(9.15)

These parameters are given for either a particularchemical species or a mass concentration M for aparticular aerosol mode. Typical values are givenin Table 9.3 and illustrate a mean value of aSP ~5m2 g-1 for sulfate and aAP ~ 10m2 g-1 for soot. Thelatter high value emphasizes the important role ofsoot aerosols in atmospheric extinction despitethe relatively small atmospheric burden. Specificefficiencies are dependent on a number of parame-ters: the size distribution, the wavelength of inci-dent light, and the ambient relative humidity(RH). The latter is of importance due to the uptakeof water, resulting in an increase in d, and hence insSP. For a pure component aerosol (e.g. inorganicsalts), growth into an aqueous droplet will abrupt-ly occur at the deliquescence point. The phasetransition is characteristic of the chemical species(e.g. (NH4)2SO4, NaCl, and NH4HSO4 deliquesce atRH = 80, 75, and 39%, respectively), but becomespoorly defined for mixed aerosols. Upon reductionof RH, recrystallization (or efflorescence) of puresalts occurs at a lower RH than the deliquescence

a sAP AP M=

a sSP SP M=

Table 9.3 Specific scattering (aSP) and absorption (aAP) coefficients for the indicated aerosol species. Also indicated isf(RH = 80%), the fractional increase in sSP at RH = 80%. All measurements are for l = 0.525mm. Figures in parenthesesrepresent approximate standard deviations (adapted from IPCC 1995).

Optical property Sulfate Organic carbon Mineral dust Soot Fine aerosol

aSP (m2 g-1) 5 (3.6–7) 5 (3.0–7) aEXT ~ 0.73 3 (2–4)

aAP (m2 g-1) 0 0 10 (8–12) (0–10)

f(RH = 80%) 1.7 (1.4–4) 1.7 (1.4–4) < 0.05 (0–1.7) 1.7 (0–4)

�


point, due to the droplet remaining in a supersatu-rated state. For atmospheric aerosols, no effect isgenerally observed under RH = 40%. Table 9.3 alsodemonstrates the fractional increase in sSP at RH =80% with respect to a “dry” value at RH < 30%,referred to as f(RH = 80%). Values indicate a factor1.7 increase in most cases, although the standarddeviations in parentheses demonstrate the de-pendence on chemical composition. Soot does notalways exhibit an increase below RH = 80% and avalue is therefore only presented as a standard deviation. Apart from the large increase in sEXT,enhanced absorption in the polluted marine andurban models results from increased absorption inan internal mixture containing soot. These resultsdemonstrate the significance of water uptake onoptical properties, especially as such measure-ments are still sparse for ground locations and almost nonexistent for the free troposphere. Whenthe RH increases beyond RH = 100%, aerosol–cloud interactions become important, and theseare considered in Section 9.5.3. Table 9.4 summa-rizes the range of optical properties representativeof polluted continental, clean continental, andclean marine aerosol types. Polluted continentalaerosol concentrations are nearly a factor 10 largerthan for remote regions, while a lower value of woindicates a higher proportion of absorbing species.Large efforts are presently under way to obtain detailed aerosol databases for various representa-tive locations around the globe. The Global

Atmosphere Watch (GAW) program of the WMOhas, for instance, been established to collect dataover decadal time scales to monitor long-termchanges in atmospheric composition.

9.5.2 Direct aerosol effects: estimates of direct forcing

The application of GCMs/CTMs in assessing theinfluence of atmospheric aerosols is made difficultby the large spatial and temporal variation in theirproperties. This is illustrated by the fact that pres-ent assessments of the role of aerosols have largelycome from models as opposed to observations.Most models predict a regional offset of green-house forcing by sulfate aerosols in the industrial-ized regions of the eastern USA, central Europe andeastern China. However, as mentioned above, theregional forcing is not expected to be indicative ofregional climate response, as atmospheric circula-tion may result in a nonlocal response to local forcing.

Direct radiative forcing is estimated to be -0.4Wm-2 for sulfate, -0.2Wm-2 for biomassburning aerosols, -0.1Wm-2 for fossil fuel organiccarbon, and +0.2Wm-2 for fossil fuel black carbonaerosols (IPCC 2001). Uncertainties are given as afactor of 2 for sulfate and fossil fuel black carbonaerosols, and as a factor of 3 for biomass burningaerosols and fossil fuel organic carbon. For mineraldust aerosols, a range of -0.6 to +0.4Wm-2 is indi-

Table 9.4 Representative values of observed aerosol optical properties in the lower troposphere for l = 0.5–0.55mm andRH < 60% (adapted from IPCC 1995).

Parameter Polluted continental Clean continental Clean marine

Optical depth (W) 0.2–0.8 0.02–0.1 0.05–0.1

Single scattering albedo (wo) 0.8–0.95 0.9–0.95 ~1

Back/total scattering ratio (R) 0.1–0.2 0.13–0.21 0.15

Total scattering coefficient (sSP; m-1) 50–300 ¥ 10-6 5–30 ¥ 10-6 5–20 ¥ 10-6

Absorption coefficient (sAP; m-1) 5–50 ¥ 10-6 1–10 ¥ 10-6 <0.05 ¥ 10-6

Fine mass concentration (mg m-3) 5–50 1–10 1–5

CN number concentration (cm-3) 103–105 102–103 <102

CCN number concentration (cm-3; 0.7–1% supersaturation) 1000–5000 100–1000 10–200

Ångström exponent (å) 1–2 1–2 1.5–2.1


cated. The northern hemisphere to southern hemi-sphere is forcing ratio >>1 except for biomass burn-ing aerosols (<1), while no estimate is given formineral dust aerosols (IPCC 2001). In general, thevariation in the above estimates is mainly due tothe differing sophistication of GCMs/CTMs,while the uncertainty depends on assumed opticalproperties and the global modeled distribution ofaerosol species. Sulfate and biomass aerosols havebeen most widely modeled to date, but additionalaerosol types, such as volcanic emission, mineraldust, and carbonaceous aerosols, are being increas-ingly included in GCM/CTM simulations.

Volcanic emissions to the stratosphere influ-ence the climate by warming the lower strato-sphere through aerosol absorption and by reducingthe net radiation transmitted to the troposphereand surface. However, as stratospheric aerosol res-idence times are of the order of 1 year, the radiativeinfluence is also restricted to similar time scales.The El Chichón (1982) and Mount Pinatubo (1991)eruptions allowed the climate effect of a large tran-sient forcing to be studied. GCM temperature pre-dictions over short time scales were found to be inreasonable agreement with observations. MountPinatubo is estimated to have contributed a maxi-mum forcing of -4Wm-2 and about -1Wm-2 up to2 years later (Hansen et al. 1992; McCormick et al.1995), which illustrates the large transient coolingeffect when compared to other radiative forcingmechanisms.

The organic aerosol fraction has only recentlybeen considered. Field studies have observed simi-lar values of aSP for organics and sulfate (e.g. Hegget al. 1997). In contrast, the EC fraction of sootaerosols absorbs light and may contribute to a pos-itive radiative forcing (Penner 1995; Schult et al.1997) when wo is less than ~0.85. Current esti-mates for various aerosol types (fossil fuel EC, biomass EC) and various mixing states (internal,external, core treatment) place the globally aver-aged forcing in the range +0.2 to +0.5Wm-2 (seeJacobson 2000, 2001, and references therein). Min-eral dust is also receiving more attention (Sokolik& Toon 1996; Tegen et al. 1996; Miller & Tegen1998; IPCC 2001). The influence of sea salt on

aerosol radiative properties in the MBL has untilrecently been considered negligible. However,new findings indicate that the contribution of seasalt to a negative radiative aerosol forcing may belarger than previously considered (Murphy et al.1998; Haywood et al. 1999).

9.5.3 Indirect aerosol effects: aerosol effects on clouds

The formation of nss sulfate via the emission ofDMS has been proposed as a cloud–climate feed-back mechanism. The present greenhouse warm-ing of the Earth’s surface/atmosphere is consideredto warm the ocean surface waters, which in turnlead to an increase in phytoplankton activity andhence DMS emissions. As a result of increasedcloud condensation nuclei (CCN) concentrationdue to aerosol formation from DMS, the albedo ofmarine stratiform clouds may increase (Twomey1977), thereby offsetting a global temperature rise(Charlson et al. 1987). While the mechanism anddifferent pathways are complex, the large areal extent of stratiform clouds, covering 25% of theoceans, renders such a feedback mechanism of po-tential importance. Cloud lifetimes and precipita-tion frequencies are also thought to be affected.However, the aerosol–climate–DMS feedback the-ory has not been conclusively proven to date. Inorder to assess the indirect effects of aerosols cloudproperties are considered further.

The presence of CCN in the atmosphere allowscloud droplet formation to occur at supersatura-tions below 1–2% (i.e. RH = 101–2%). Withouttheir presence, supersaturations of several hun-dred percent would be required. The physicalmechanisms of these processes have long been investigated (Pruppacher & Klett 1997), but atmospheric observations and secular studies remain sparse due to the overall complexity of the aerosol–cloud effect on climate. Not only areaerosol–cloud interactions difficult to model, butcloud parameterizations themselves are still anoutstanding issue.

The critical supersaturation at which CCN be-come activated depends on aerosol composition,


size, and age, and may be described by the Köhlertheory (e.g. Pruppacher & Klett, 1997). Growth ofaerosols to RH = 100% was described in Section9.5.1. As RH increases beyond 100% the aerosoldroplets continue to grow until a critical supersat-uration SC is reached, corresponding to a critical diameter dCRIT at which the aerosol becomes “activated.” At this point droplets are in an unsta-ble equilibrium with their environment and maygrow uncontrollably or return to a stable equili-brium, where they will exist as unactivateddroplets or haze.

The value of SC depends on the solubility ofchemical components and dry diameter of theaerosol. The larger both parameters are, the lowerthe critical supersaturation required to activatethe aerosol. As a result the larger, hygroscopicaerosols tend to form CCN first. Typical supersat-urations in marine stratus clouds, estimated at0.1%, imply that dry aerosol diameters with d >0.1mm will be activated to produce cloud droplets,whereas for d < 0.1mm the aerosol will remain interstitial to cloud droplets.

Effective CCN sources are, in general, second-ary aerosols of either natural or anthropogenic ori-gin. Non-sea-salt sulfate is considered to be themajor natural source, while biomass/organic andsulfate aerosols are major anthropogenic sources.The ability of pure inorganic salts and acids such asnitrates, sulfates, or sulfuric acid to act as CCN isrelatively well investigated in experimental andmodeling work. The role of organic aerosols or themixture of organic and inorganic particles, how-ever, is less characterized and largely dependent on the hygroscopic behavior of the organic com-pounds. Studies have shown that smoke particlesfrom biomass burning exhibit a CCN activation ofup to 100%, in contrast to an activation of less thanseveral percent for fresh soot from petroleum fuel.Other studies suggest that organics mixed with in-organic salt aerosols alter their hygroscopic behav-ior. Organics were found to reduce the criticalsupersaturation needed to activate inorganicaerosols (Shulman et al. 1996). Others found thateven large amounts of hydrophobic organics didnot affect the hygroscopic growth of inorganic par-

ticles below 100% humidity. Soot aerosols frombiomass burning containing large parts of organicshave been shown to act as CCN. Field measure-ments, e.g. in the Amazon region, found signifi-cant influence of biomass burning events andCCN concentrations (Kaufman et al. 1998). Therole of organics seems to be an important but as yetunresolved aspect in the activation of ambientaerosols to cloud droplets.

Cloud physical properties over continental re-gions differ from those over oceans. Over land,larger CCN concentrations result in increasedcloud droplet concentrations (NCLOUD), and sincethe LWC of both cloud types are similar, continen-tal clouds have a smaller average droplet size thanmarine clouds. Observations indicate that marinecumulus clouds have a median value NCLOUD ~ 45cm-3 and a broad droplet size spectrum with medi-an at d ~ 30mm, while continental cumuli have amedian value NCLOUD ~ 230cm-3 and a narrowersize spectrum with median ~10mm. Parameteriza-tions of clouds use the cloud optical thickness d,which is defined as:

(9.16)

where L is the cloud LWC and re is the effective radius of cloud droplets. L may be approximatedfrom eqn 9.16 and, when solving for re and differen-tiating, eqn 9.18 results.

(9.17)

(9.18)

Hence for a constant LWC, an increase in N as a re-sult of anthropogenic sources will result in a linearincrease in d. This simple approach is, however,complicated by many other factors apart from theinherent assumptions in the above derivation. A further cloud parameter of importance is thealbedo Ao, which may be approximated by (Hobbs1993):

D Ddd

=13

NN

L r Ne CLOUDμ43

3p

d μ L re


processes involved are as yet still very poorly un-derstood.

9.5.4 Indirect aerosol effects: estimates ofindirect forcing

As a consequence of the large uncertainties in theinteraction of aerosols with clouds, an estimate of the sign and magnitude of indirect forcing isfraught with assumptions. Present estimates ofaerosol indirect forcing in Fig. 9.4 give a range ofuncertainty from 0 to -2.0Wm-2, which indicate asimilar sign and magnitude to direct forcing (IPCC2001). This report summarizes the rapid advancesthat have been made in understanding and assess-ing the indirect aerosol effect since the last reportwas published in 1996. Among the numerous newuncertainties and findings that have been high-lighted in recent studies, several investigationshave provided evidence that the incorporation ofsoot aerosols from biomass burning may be influ-encing cloud albedo. A satellite study of cumulusand stratocumulus clouds over the Amazon basinduring the burning season showed that the averagedroplet diameter decreased from 14 to 9mm, with acorresponding increase in cloud reflectance from0.35 to 0.45 when smoke aerosols were present(Kaufman & Fraser 1997).

9.6 TROPOSPHERIC ANDSTRATOSPHERIC AEROSOLS:

REMOTE SENSING

In order to reduce the uncertainty in our under-standing of climate change, satellite measure-ments of atmospheric gases and aerosols arerapidly gaining in importance (Browell et al. 1998;King et al. 1999). Due to the spatial and temporalvariability of aerosol sources and sinks, obtainingglobal distributions of tropospheric aerosols frompoint measurements is an arduous task. Even air-craft studies are limited by a necessarily large andcomplex infrastructure, which is especially true ofstratospheric measurements. The main advantageof satellites therefore lies in their ability to measure global distributions of atmospheric

(9.19)

For scattering of solar radiation by clouds, g ~ 0.85such that:

(9.20)

Hence a greater change in albedo occurs when d islow. Using the above equations, and assuming thatthe cloud LWC and depth remain constant then a parameter known as the susceptibility may be derived:

(9.21)

The susceptibility DAo/DN is most sensitive tochange when N is small and when Ao lies between~0.25 and 0.75, conditions that are both typical ofmarine clouds. Hence, small increases in anthro-pogenic CCN concentrations are likely to have agreater influence in marine than continental re-gions, i.e. in the southern hemisphere. As cloudsare already optically thick to longwave radiation,only shortwave radiation is influenced by cloudproperties. As a consequence, enhancement of theshortwave albedo is considered to result in in-creased reflection of solar radiation back to spaceand a cooling of the Earth’s surface.

The susceptibility of marine stratiform cloudsto increased CCN concentrations has been observed in satellite images where ship stack exhausts have resulted in an increase in cloud albe-do. The increase in droplet concentration and reduction in size have been confirmed simultane-ously by in situ and remote sensing measure-ments. Observational evidence suggests that for afactor 10 increase in the aerosol size range d ~0.1–0.3mm, a 2–5-fold increase in droplet concen-tration results. An additional important effect, as a result of the above processes, is a reduction in precipitation efficiency and increased cloud lifetime (Albrecht 1989). Such observations are at present difficult to quantify and the complex

DDAN

A AN

o o o=-( )1

3

Ao ª+d

d 6 7.

Ag

go =-( )

+ -( )1

1 1d

d


constituents. In addition, subtle changes in at-mospheric composition may be monitored usingaccurate calibration standards, and is necessary ifanthropogenic-induced climate change is to be de-tected. As a multitude of satellite sensors withvarying capabilities exist and are planned for thefuture, the discussion below can give only a briefoverview of some highlights. Further details ingood review articles can be found in both of theabove-mentioned references.

Although satellites with instrument packagesdedicated to aerosol measurements have only beenlaunched in recent years, the retrieval of AOD hasbeen possible since the late 1970s from two nadir-viewing satellites. The advanced very high resolu-tion radiometer (AVHRR) is a five-band sensor,aboard the National Oceanic and Atmospheric Ad-ministration’s series of satellites, and is dedicatedto weather and ocean observations. The total ozonemapping spectrometer (TOMS) is an ultravioletscanning monochromator, aboard various satel-lites, dedicated to measuring the total ozone col-umn depth. Retrieval of the AOD over clear-skyregions of the ocean has proven to be accurate (e.g.Husaret al. 1997), due to the low and well defined re-flectance of the ocean surface at solar wavelengths.In comparison, retrieval over land surfaces has re-mained difficult, and this stems from several prob-lems: the complex variation in surface reflectancewith viewing angle, wavelength, season of year(summer/winter biological cover), and surface features (agricultural land, mountainous regions,deserts, etc.). Nevertheless, AOD has been re-trieved over land surfaces (e.g. AVHRR and TOMSdata), but requires a number of assumptions, morespecifically on the aerosol size distribution, surfacereflectance, and wo. Progress has been made in several areas by using novel retrieval algorithms(Kaufman et al. 1997; Herman et al. 1997) and dedi-cated aerosol sensors (see King et al. 1999), althougha comprehensive network of ground-based Sunphotometers and sky radiometers are still presentlyrequired to provide ground-truth verification. Sev-eral of the above-mentioned aerosol parameters cannow be retrieved with less ambiguity using severaltechniques aboard recent satellite sensors: (i) atmultiple wavelengths and angles (moderate resolu-

tion imaging spectrometer (MODIS) and multian-gle imaging spectroradiometer (MISR) sensorsaboard the Terra satellite); and (ii) using polarizedradiance (the POLDER (polarization and direction-ality of the Earth’s reflectances) sensor aboard theADEOS I and II satellites).

An example of POLDER data is shown in Plate 9.1a, b (facing p. 180), which exhibits theglobal radiative forcing due to aerosols in winterand summer (Boucher & Tanré 2000). A negativeforcing is clearly seen over regions of enhancedaerosol concentrations, such as over the US eastcoast (industrial aerosols), the Mediterranean, andWest Africa (Saharan dust plume). A global meanclear-sky shortwave perturbation of -5 to -6Wm-2

was estimated from these measurements.Despite these advances, aerosol properties as

a function of altitude are also required forGCM/CTM models. Aerosol forcing depends notonly on the vertical distribution of aerosols butalso on their scattering and absorbing properties.Several remote sensing systems have allowed vertically resolved aerosol information to be re-trieved, such as the Stratospheric Aerosol and GasExperiment (SAGE) satellite and the Lidar in SpaceTechnology Experiment (LITE) aboard the spaceshuttle. The SAGE II was launched in 1984 and hasallowed the observation of aerosols and gases inthe stratosphere and upper troposphere (i.e. above6km). The sensor uses the solar occultation tech-nique to obtain 15 sunrise and sunset observationsper day (limb-viewing) with 1km vertical and 200km horizontal resolutions. As a result of interfer-ence from clouds, the limb-viewing technique isgenerally restricted to observations above 6km.SAGE II has a high sensitivity and is therefore able to measure aerosols in the stratosphere andupper troposphere, where concentrations aremuch lower than in the lower troposphere andPBL. For instance, typical AOD values (l = 1.0mm)are 0.02–0.5 in the troposphere, and may be compared with a 2 ¥ 10-4 to 2 ¥ 10-3 range for theunenhanced upper troposphere.

Two global maps of the aerosol extinction (l =1.02mm) at altitudes of 6.5 and 12.5km are shownin Plate 9.2a, b (facing p. 180), respectively, andhave been averaged over several years for the


September to November period (Kent et al. 1998a).The main features in Plate 9.2a are the high extinc-tion levels at high northern latitudes, attributedmainly to anthropogenic emissions, and distinctbands of rather uniform zonal distribution. En-hanced extinction over Brazil was attributed tobiomass burning, which peaks at this time of year.Plate 9.2b exhibits similar features, but the ele-vated extinction at high latitudes is due to the observation of stratospheric aerosols as a conse-quence of the tropopause lying below 12.5km. Anaerosol climatology of SAGE observations has indicated a higher aerosol concentration in thenorthern than in the southern hemisphere andmay be attributed to the larger number of land andanthropogenic sources (Kent et al. 1998b). An en-hancement in springtime concentrations above20° either side of the Equator and in each hemi-sphere has also been observed.

SAGE II has gathered a comprehensive data-base of globally averaged aerosol maps, which hashelped the understanding of atmospheric dynam-ics. In addition, it has been possible to conduct arobust comparison of in situ and remote measure-ments due to the long lifetime of stratosphericaerosols. The aerosol direct effect in the upper at-mosphere is minimal during volcanic quiescentyears; the utility of SAGE II in providing informa-tion to assess the indirect aerosol effect is perhapsof greater significance. Properties of cirrus cloudsand their extent, whether modified by natural or anthropogenic influences (see IPCC 1999) are a further major uncertainty in climate changestudies.

LITE was a nadir-pointing backscatter lidarflown aboard the Discovery space shuttle missionin September 1994. The objectives of LITE were to assess lidar space technology for use in futuresatellites, and to demonstrate how aerosol distri-butions and optical properties could be monitoredin the troposphere and stratosphere (e.g. Kent et al.1998a; Osborn et al. 1998). One of the main advan-tages demonstrated by LITE was the ability tomeasure vertical profiles down to the Earth’s sur-face through openings in the cloud-cover, hence al-lowing a larger database to be gathered than wouldbe possible with a limb-viewing system. An exam-

ple of stratospheric aerosol distributions is shownin Plate 9.3 (facing p. 180), where the aerosol scat-tering ratio (l = 0.532mm) on September 17, 1994 is shown (Osborn et al. 1998). The white line indicates the tropopause height, which forms thelower limit to the stratospheric aerosol layer.Stratospheric aerosols, mainly composed of sulfu-ric acid, appear to form a stable and globally dis-tributed layer above the tropopause. This so-calledJunge layer is long-lived due to the absence of wet-removal processes and the strong temperaturestratification of the stratosphere.

Although LITE was only flown for a 10-day peri-od, the potential to retrieve temporally and spatial-ly resolved aerosol parameters in three dimensionsproved to be very promising. Looking into the fu-ture, the satellite remote sensing capability will bestrengthened and widened with the launch of sev-eral new additional sensors dedicated to aerosolmeasurements. Major advances in the understand-ing of climate change are expected by utilizing bet-ter calibration techniques and advanced computeranalysis of multiple, cross-referenced satellitedatabases. Further progress is also foreseen in the refinement of column-closure experiments(ground-based and in situ aircraft measurements),which is essential if satellite measurements are tobe verified.

APPENDIX: NOMENCLATURE

å Ångström exponentAo cloud albedoa accommodation coefficientai stochiometric coefficientaAP aerosol specific absorption

efficiencyaSP aerosol specific scattering

efficiencyB(d) attachment coefficientb(f) angular scattering phase

functionc the mean thermal velocity of

a diffusing moleculed aerosol geometric

diameter


dCRIT aerosol critical diameter atwhich activation occurs

D diffusion coefficient of adiffusing molecule

d cloud optical thicknessf(RH = 80%) fractional increase in sSP at

RH = 80% with respect to RH= 30%

g asymmetry factor of ascattering aerosol

g dimensionless stickingcoefficient

gmeas measured uptake coefficientgdiff gas transport coefficientgrn reaction coefficientgsol solubility coefficientHC mass of hydrocarbonh viscosity of airJ flux of gaseous molecules

sticking to aerosol particlesK equilibrium constant for

partitioning of a compoundbetween the organic andgaseous phase

L cloud liquid water contentlMFP mean free path of a diffusing

moleculel wavelength of lightM fine mode aerosol mass

concentrationM0 organic aerosol mass formedMOM mean molecular weight of

organic matterN aerosol number concentrationNCLOUD cloud droplet concentrationNi(d) the particle number

concentration with particlediameter d

Nmol the molecular numberconcentration

wo aerosol single scatteringalbedo; sSP/sEXT

W aerosol optical depth; thevertical integral of sEXT overthe light beam path length dl

p vapor pressurev Junge parameter

R aerosol back/total scatteringratio; sBSP/sSP

RG gas constantr aerosol geometric radiusre effective cloud droplet radiusSC critical supersaturation at

which aerosol activation intoa cloud droplet occurs

sAP aerosol absorption coefficientsBSP aerosol hemispheric

backscatteringing coefficientsEXT aerosol extinction coefficient;

sSP + sAPsSP aerosol total scattering

coefficientT temperatureY aerosol yield from a single

compound

REFERENCES

Albrecht, B.A. (1989) Aerosols, cloud microphysics, andfractional cloudiness. Science 245, 1227–30.

Andreae, M.O. & Crutzen, P.J. (1997) Atmosphericaerosols: biogeochemical sources and role in atmos-pheric chemistry. Science 276, 1052–8.

Ammann, M., Kalberer, M., Jost, D.T. et al. (1998) Heterogeneous production of nitrous acid on soot inpolluted air masses. Nature 395, 157–60.

Baker, M.B. (1997) Cloud microphysics and climate. Science 276, 1072–8.

Baron, P.A. & Willeke, K. (eds) (2001) Aerosol Measure-ment. John Wiley, New York.

Barrie, L.A., Yi, Y., Leaitch, W.R. et al. (2001) A compari-son of large scale atmospheric sulfate aerosol models(COSAM): overview and highlights. Tellus, Series B 53,615–45.

Berresheim, H., Wine, P.H. & Davis, D.D. (1995) Sulfur inthe atmosphere. In Singh, H.B. (ed.), Composition,Chemistry and Climate of the Atmosphere. Van Nos-trand Reinhold, New York.

Boucher, O. & Tanré, D. (2000) Estimation of the aerosolperturbation to the Earth’s radiative budget overoceans using POLDER satellite aerosol retrievals. Geo-physics Research Letters 27, 1103–6.

Brasseur, G.P., Orlando, J.J. & Tyndall, G.S. (eds) (1999) Atmospheric Chemistry and Global Change. OxfordUniversity Press, Oxford.


Browell, E.V., Ismail, S. & Grant, W.B. (1998) Differentialabsorption lidar (DIAL) measurements from air andspace. Applied Physics B 67, 399–410.

Charlson, R.J. & Heintzenberg, J. (eds) (1995) AerosolForcing of Climate. Wiley, Chichester.

Charlson, R.J., Lovelock, J.E., Andreae, M.O. & Warren,S.G. (1987) Oceanic phytoplankton, atmospheric sulfur, cloud albedo and climate. Nature 326, 655–61.

Charlson, R.J., Schwartz, S.E., Hales, J.M. et al. (1992)Climate forcing by anthropogenic aerosols. Science255, 423–30.

Cooke, W.F., Liousse, C., Cachier, H. & Feichter, J. (1999) Construction of a 1° ¥ 1° degree fossil fuel emission data set for carbonaceous aerosol and imple-mentation and radiative impact in the ECHAM4model. Journal of Geophysical Research 104, 22,137–62.

d’Almeida, G.A., Koepke, P. & Shettle, E.P. (eds) (1991)Atmospheric Aerosols: Global Climatology and Radiative Characteristics. Deepak, Hampton.

Davidovits, P., Hu, Jh., Worsnop, D.R., Zahniser, M.S. &Kolb, C.E. (1995) Entry of gas molecules into liquids.Faraday Discussions 100, 65–82.

Dentener, F.J. & Crutzen, P.J. (1993) Reaction of N2O5 ontropospheric aerosols: Impact on the global distribu-tions of NOx, O3, and OH. Journal of Geophysical Research 98, 7149–63.

Duce, R.A. (1995) Sources, distributions and fluxes ofmineral aerosols and their relationship to climate. InCharlson, R.J. & Heintzenberg, J. (eds), Aerosol Forcingof Climate. Wiley, Chichester.

Finlayson-Pitts, B. & Pitts, J. (1999) Chemistry of theUpper and Lower Atmosphere. Acadamic Press, NewYork.

Gong, S.L., Barrie, L.A. & Blanchet, J.P. (1997) Modelingsea-salt aerosols in the atmosphere. Part 1: Model development. Journal of Geophysical Research 102,3805–18.

Griffin, R.J., Cocker, D.R., Flagan, R.C. & Seinfeld, J.H.(1999a) Organic aerosol formation from the oxidationof biogenic hydrocarbons. Journal of Geophysical Research 104, 3555–67.

Griffin, R.J., Cocker, D.R., Seinfeld, J.H. & Dabdub, D.(1999b) Estimate of global atmospheric organic aerosolfrom oxidation of biogenic hydrocarbons. GeophysicalResearch Letters 26, 2721–4.

Guenther, A., Hewitt, C., Erickson, D. et al. (1995) Amodel of natural volatile organic compound emis-sions. Journal of Geophysical Research 100, 8873–92.

Hansen, J.E., Lacis, A., Ruedy, R. & Sato, M. (1992) Poten-

tial climate impact of Mount Pinatubo eruption. Geo-physical Research Letters 19, 215–218.

Haywood, J.M., Ramaswamy, V. & Soden, B.J. (1999) Tro-pospheric aerosol climate forcing in clear-sky satelliteobservations over the oceans. Science 283, 1299–303.

Hegg, D.A., Livingston, J., Hobbs, P.V., Novakov, T. &Russell, P. (1997) Chemical apportionment and aerosolcolumn optical depth off the mid-Atlantic coast of theUnited States. Journal of Geophysical Research 102,25,293–303

Heintzenberg, J., Covert, D.C. & Van Dingenen, R. (2000)Size distribution and chemical composition of marineaerosols: a compilation and review. Tellus 52B,1104–22.

Herman, M., Deuzé, J.L., Goloub, P., Bréon, F.M. & Tanré,D. (1997) Remote sensing of aerosols over land surfacesincluding polarization measurements and applicationto POLDER measurements. Journal of GeophysicalResearch 102, 17,039–49.

Hinds, W.C. (1999) Aerosol Technology: Properties, Behavior, and Measurements of Airborne Particles.Wiley, New York.

Hobbs, P.V. (ed.) (1993) Aerosol–Cloud–Climate Interac-tions. Academic Press, San Diego.

Husar, R.B., Prospero, J.M. & Stowe, L.L. (1997) Charac-terization of tropospheric aerosols over the oceanswith the NOAA Advanced Very High Resolution Radiometer optical thickness operational product.Journal of Geophysical Research 102, 16,889–909.

IPCC (1996) Climate Change 1995. The Science of Climate Change. Cambridge University Press,Cambridge.

IPCC (1999) Aviation and the Global Atmosphere.Cambridge University Press, Cambridge.

IPCC (2001) Climate Change 2001: The Scientific Basis.Cambridge University Press, New York.

Jacobson, M.Z. (2000) A physically-based treatment of el-emental carbon optics: implications for global directforcing of aerosols. Geophysical Research Letters 27,217–20.

Jacobson, M.Z. (2001) Strong radiative heating due to themixing state of black carbon in atmospheric aerosols.Nature 409, 695–7.

Jaenicke, R. (1988) Atmospheric physics and chemistry.In Fischer, G. (ed.), Meteorology: Physical and Chemi-cal Properties of Air. Springer-Verlag, Berlin.

Kaufman, Y.J. & Fraser, R.S. (1997) The effect of smokeparticles on clouds and climate forcing. Science 277,1636–9.

Kaufman,Y.J., Tanré, D., Gordon, H.R. et al. (1997) Passive remote sensing of tropospheric aerosol and


atmospheric correction for the aerosol effect. Journalof Geophysical Research 102, 16,815–30.

Kaufman, Y.J., Hobbs, P.V., Kirchhoff, V.W.J.H. et al.(1998) Smoke, Clouds, and Radiation —Brazil (SCAR-B) experiment. Journal of Geophysical Research 103,31,783–808.

Kent, G.S., Trepte, C.R., Skeens, K.M. & Winker, D.M.(1998a) LITE and SAGE II measurements of aerosols inthe southern hemisphere upper troposphere. Journal ofGeophysical Research 103, 19,111–27.

Kent, G.S., Trepte, C.R. & Lucker, P.L. (1998b) Long-termStratospheric Aerosol and Gas Experiment I and IImeasurements of upper tropospheric aerosol extinc-tion. Journal of Geophysical Research 103, 28,863–74.

King, M.D., Kaufman, Y.J., Tanré, D. & Nakajima, T.(1999) Remote sensing of tropospheric aerosols from space: past, present and future. Bulletin of theAmerican Meteorological Society 80, 2229–59.

Koepke, P., Hess, M., Schult, I. & Shettle, E.P. (1997)Global Aerosol Data Set. Report 243, Max Planck Institute for Meteorology, Hamburg.

Langner, J. & Rodhe, H. (1991) A global three-dimensional model of the tropospheric sulfur cycle.Journal of Atmospheric Chemistry 13, 255–63.

Lavanchy, V.M.H., Gäggeler, H.W., Schotterer, U.,Schwikowski, M. & Baltensperger, U. (1999) Historicalrecord of carbonaceous particle concentrations from a European high-alpine glacier (Colle Gnifetti,Switzerland). Journal of Geophysical Research 104,21,227–36.

Leinen, M. & Sarnthein, M. (eds) (1989) Paleoclimatol-ogy and Paleometeorology: Modern and Past Patternsof Global Atmospheric Transport. Kluwer Academic,Dordrecht.

Levine, J.S. (ed.) (1991) Global Biomass Burning: Atmos-pheric, Climatic and Biospheric Implications. MITPress, Cambridge, MA.

McCormick, M.P., Thomason, L.W. & Trepte, C.R.(1995) Atmospheric effects of the Mount Pinatuboeruption. Nature 373, 399–403.

Marti, J.J., Weber, R.J., McMurry, P.H., Eisele, F., Tanner,D. & Jefferson, A. (1997) New particle formation at aremote continental site: assessing the contributions ofSO2 and organic precursors. Journal of Geophysical Research 102, 6331–9.

Miller, R. & Tegen, I. (1998) Climate response to soil dustaerosols. Journal of Climate 11, 3247–67.

Möller, D. (1995) Sulfate aerosols and their atmos-pheric precursors. In Charlson, R.J. & Heintzenberg, J.(eds), Aerosol Forcing of Climate. John Wiley, Chich-ester.

Murphy, D.M., Anderson, J.R., Quinn, P.K. et al. (1998)Influence of sea salt on aerosol radiative properties inthe southern Ocean marine boundary layer. Nature392, 62–5.

O’Dowd, C., Smith, M.H., Consterdine, I.E. & Lowe, J.A.(1997) Marine aerosol, sea-salt, and the marine sulfurcycle: a short review. Atmospheric Environment 31,73–80.

Odum, J.R., Jungkamp, T.P.W., Griffin, R.J., Flagan, R.C.& Seinfeld, J.H. (1997) The atmospheric aerosol-forming potential of whole gasoline vapor. Science276, 96–9.

Osborn, M.T., DeCoursey, R.J., Trepte, C.R., Winker,D.M. & Woods, D.C. (1995) Evolution of the Pinatubovolcanic cloud over Hampton, Virginia. GeophysicalResearch Letters 22, 1101–4.

Osborn, M.T., Kent, G.S. & Trepte, C.R. (1998) Stratos-pheric aerosol measurements by the Lidar In-SpaceTechnology Experiment. Journal of Geophysical Research 103, 11,447–53.

Pandis, S.N., Harley, R.A., Cass, G.R. & Seinfeld, J.H.(1992) Secondary organic aerosol formation and trans-port. Atmospheric Environment 26, 2269–82.

Penner, J.E. (1995) Carbonaceous aerosols influencing atmospheric radiation: black and organic carbon. InCharlson, R.J. & Heintzenberg, J. (eds), Aerosol Forcingof Climate. Wiley, Chichester.

Phadnis, M.J. & Carmichael, G.R. (2000) Numerical investigation of the influence of mineral dust on thetropospheric chemistry of East Asia. Journal of Atmospheric Chemistry 36, 285–323.

Piccot, S.D., Watson, J.J. & Jones, J.W. (1992) A global inventory of volatile organic compound emissionsfrom anthropogenic sources. Journal of GeophysicalResearch 97, 9897–912.

Pruppacher, H.R. & Klett, J.D. (1997) Microphysicsof Clouds and Precipitation, 2nd edn. Reidel, Dordrecht.

Pueschel, R.F. (1995) Atmospheric aerosols. In Singh,H.B. (ed.), Composition, Chemistry and Climate of the Atmosphere. Van Nostrand Reinhold, NewYork.

Pueschel, R.F. (1996) Stratospheric aerosols: Formation,properties, effects. Journal of Aerosol Science 27,383–402.

Quinn, P.K., Anderson, T.L., Bates, T.S. et al. (1996) Clo-sure in tropospheric aerosol–climate research: a reviewand future needs for addressing aerosol direct short-wave radiative forcing. Contributions in AtmosphericPhysics 69, 547–77.

Raes, F., Van Dingenen, R., Vignatti, E. et al. (2000) For-


mation and cycling of aerosols in the global tropo-sphere. Atmospheric Environment 34, 4215–40.

Ravishankara, A.R. (1997) Heterogeneous and multi-phase chemistry in the troposphere. Science 276,1058–65.

Reist, P.C. (1993) Aerosol Science and Technology, 2ndedn McGraw-Hill, New York.

Russell, P.B. & Heintzenberg, J. (2000) An overview of the ACE-2 clear sky column closure experiment(CLEARCOLUMN). Tellus 52B, 463–83.

Russell, P.B., Hobbs, P.V. & Stowe, L.L. (1999) Aerosolproperties and radiative effects in the United StatesEast Coast haze plume: an overview of the Tropos-pheric Aerosol Radiative Forcing Observational Exper-iment (TARFOX). Journal of Geophysical Research104, 2213–22.

Saylor, R.D. (1997) An estimate of the potential signifi-cance of heterogeneous loss to aerosols as an addi-tional sink for hydroperoxy radicals in the troposphere.Atmospheric Environment 31, 3653–8.

Schult, I., Feichter, J. & Cooke, W.F. (1997) Effect of black carbon and sulfate aerosols on the global radia-tion budget. Journal of Geophysical Research 102,30,107–17.

Schwartz, S.E. (1996) The Whitehouse Effect —shortwave radiative forcing of climate by anthro-

pogenic aerosols: an overview. Journal of Aerosol Science 27, 359–82.

Shulman, M.L., Jacobson, M.C., Carlson, R.J., Synovec,R.E. & Young, T.E. (1996) Dissolution behavior andsurface tension effects of organic compounds in nucle-ating cloud droplets. Journal of Geophysical Research23, 277–80.

Seinfeld, J.H. & Pandis, S.N. (1998) Atmospheric Chem-istry and Physics: From Air Pollution to ClimateChange. John Wiley, New York.

Singh, H.B. (ed.) (1995) Composition, Chemistry and Cli-mate of the Atmosphere. Van Nostrand Reinhold,New York.

Sokolik, I. & Toon, O.B. (1996) Direct radiative forcing byanthropogenic airborne mineral aerosols. Nature 381,681–3.

Tegen, I., Lacis, A.A. & Fung, I. (1996) The influence onclimate forcing of mineral aerosols from disturbedsoils. Nature 380, 419–22.

Twomey, S.A. (1977) The influence of pollution on theshort-wave albedo of clouds. Journal of AtmosphericScience 34, 1149–52.

Whitby, K.T. & Sverdrup, G.M. (1980) Californiaaerosols: their physical and chemical characteristics.Advances in Environmental Science and Technology9, 477–517.

handbook of atmospheric science || atmospheric particulate matter

Documents