Ozon ydroxyl Radical,and Oxidativ pacity · 4.01 Ozone, Hydroxyl Radical,and OxidativeCapacity...

4.01Ozone, Hydroxyl Radical, andOxidative CapacityR. G. Prinn

Massachusetts Institute of Technology, Cambridge, MA, USA

4.01.1 INTRODUCTION 1

4.01.2 EVOLUTION OF OXIDIZING CAPABILITY 34.01.2.1 Prebiotic Atmosphere 44.01.2.2 Pre-industrial Atmosphere 5

4.01.3 FUNDAMENTAL REACTIONS 54.01.3.1 Troposphere 54.01.3.2 Stratosphere 7

4.01.4 METEOROLOGICAL INFLUENCES 8

4.01.5 HUMAN INFLUENCES 94.01.5.1 Industrial Revolution 94.01.5.2 Future Projections 10

4.01.6 MEASURING OXIDATION RATES 114.01.6.1 Direct Measurement 114.01.6.2 Indirect Measurement 13

4.01.7 ATMOSPHERIC MODELS AND OBSERVATIONS 16

4.01.8 CONCLUSIONS 16

REFERENCES 17

4.01.1 INTRODUCTION

The atmosphere is a chemically complex anddynamic system interacting in significant wayswith the oceans, land, and living organisms. Akey process proceeding in the atmosphere isoxidation of a wide variety of relatively reducedchemical compounds produced largely by thebiosphere. These compounds include hydrocar-bons (RH), carbon monoxide (CO), sulfur dioxide(SO2), nitrogen oxides (NOx), and ammonia(NH3) among others. They also include gasesassociated with advanced technologies such ashydrofluoro carbons and hydrochlorofluoro-carbons. A summary of the composition of theatmosphere at the start of the twenty-first centuryis given in Table 1. Due to their key role,oxidation reactions are sometimes referred to asnature’s atmospheric “cleansing” process, and the

overall rate of this process as the “oxidationcapacity” of the atmosphere. Without this effi-cient cleansing process, the levels of manyemitted gases could rise so high that they wouldradically change the chemical nature of ouratmosphere and biosphere and, through thegreenhouse effect, our climate.Oxidation became an important atmospheric

reaction on Earth once molecular oxygen (O2)from photosynthesis had reached sufficientlyhigh levels. This O2 could then photodissociatein the atmosphere to give oxygen atoms, whichcombine with O2 to form ozone (O3). When O3

absorbs UV light at wavelengths less than310 nm, it produces excited oxygen atoms(O(1D)) which can attack water vapor toproduce the hydroxyl free radical (OH). It isthe hydroxyl radical that, above all, definesthe oxidative capacity of our O2- and

1

H2O-rich atmosphere (Levy, 1971; Chameidesand Walker, 1973; Ehhalt, 1999; Lelieveld et al.,1999; Prather et al., 2001). As of early 2000s,its global average concentration is only ,106radicals cm23 or ,6 parts in 1014 by mole inthe troposphere (Prinn et al., 2001). But itsinfluence is enormous and its life cyclecomplex. For example, it reacts with CO,usually within ,1 s, to produce CO2 (and alsoa hydrogen atom, which quickly combines withO2 to form hydroperoxy free radicals, HO2).Another key player is nitric oxide (NO), whichcan be produced by lightning, combustion, andnitrogen microbes, and undoubtedly becamepresent at significant levels in the Earth’satmosphere as soon as O2 and N2 became thedominant atmospheric constituents. NO canreact with either O3 or HO2 to form NO2.The reaction of NO with HO2 is special becauseit regenerates OH giving two major sources(Oð1DÞþ H2O! 2OH and NOþ HO2! NO2þOH) for this key cleansing radical. The NO2

is also important because it photodissociateseven in violet wavelengths to produce oxygenatoms and thus O3. Hence, there are also twomajor sources of O3 (photodissociation of bothO2 and NO2).

The importance of OH as a removal mechanismfor atmospheric trace gases is illustrated inTable 2,which shows the global emissions of each gas andthe approximate percentage of each of theseemitted gases which is destroyed by reaction withOH (Ehhalt, 1999).Evidently, OH removes a total of ,3:65 £

1015 g of the listed gases each year, or an

Table 1 Gaseous chemical composition of the atmosphere (1 ppt ¼ 10212; 1 ppb ¼ 1029; 1 ppm ¼ 1026).

Constituent Chemical formula Mole fraction in dry air Major sources

Nitrogen N2 78.084% BiologicalOxygen O2 20.948% BiologicalArgon Ar 0.934% InertCarbon dioxide CO2 360 ppm Combustion, ocean, biosphereNeon Ne 18.18 ppm InertHelium He 5.24 ppm InertMethane CH4 1.7 ppm Biogenic, anthropogenicHydrogen H2 0.55 ppm Biogenic, anthropogenic, photochemicalNitrous oxide N2O 0.31 ppm Biogenic, anthropogenicCarbon monoxide CO 50–200 ppb Photochemical, anthropogenicOzone (troposphere) O3 10–500 ppb PhotochemicalOzone (stratosphere) O3 0.5–10 ppm PhotochemicalNMHC CxHy 5–20 ppb Biogenic, anthropogenicChlorofluorocarbon 12 CF2Cl2 540 ppt AnthropogenicChlorofluorocarbon 11 CFCl3 265 ppt AnthropogenicMethylchloroform CH3CCl3 65 ppt AnthropogenicCarbon tetrachloride CCl4 98 ppt AnthropogenicNitrogen oxides NOx 10 ppt–1 ppm Soils, lightning, anthropogenicAmmonia NH3 10 ppt–1 ppb BiogenicHydroxyl radical OH 0.05 ppt PhotochemicalHydroperoxyl radical HO2 2 ppt PhotochemicalHydrogen peroxide H2O2 0.1–10 ppb PhotochemicalFormaldehyde CH2O 0.1–1 ppb PhotochemicalSulfur dioxide SO2 10 ppt–1 ppb Photochemical, volcanic, anthropogenicDimethyl sulfide CH3SCH3 10–100 ppt BiogenicCarbon disulfide CS2 1–300 ppt Biogenic, anthropogenicCarbonyl sulfide OCS 500 ppt Biogenic, volcanic, anthropogenicHydrogen sulfide H2S 5–500 ppt Biogenic, volcanic

Source: Brasseur et al. (1999) and Prinn et al. (2000).

Table 2 Global turnover of tropospheric trace gasesand the fraction removed by reaction with OH accordingto Ehhalt (1999). The mean global OH concentrationwas taken as 1 £ 106 cm23 (Prinn et al., 1995)

ð1 Tg ¼ 1012gÞ.Trace gas Global

emission rate(Tg yr21)

Removalby OH(%)

Removalby OH

(Tg yr21)

CO 2,800 85 2,380CH4 530 90 477C2H6 20 90 18Isoprene 570 90 513Terpenes 140 50 70NO2 150 50 75SO2 300 30 90(CH3)2S 30 90 27

Ozone, Hydroxyl Radical, and Oxidative Capacity2

amount equal to the total mass of theatmosphere every 1:4 £ 106 yr: Without OHour atmosphere would have a very differentcomposition, as it would be dominated by thosegases which are presently trace gases.Because both UV radiation fluxes and water

vapor concentrations are largest in the tropicsand in the southern hemisphere, the levels ofOH generally have their maxima in the lowertroposphere in these regions and seasons(Figure 1). Also, because much O3 is producedin and exported from polluted areas, itsconcentrations generally have maxima in thenorthern hemisphere mid-latitude summer(Figure 1).Even from the very brief discussion above, it is

clear that the levels of OH in the atmosphere arenot expected to remain steady but to changerapidly on a wide variety of space and timescales.In the lower atmosphere, OH sources can bedecreased or turned off by lowering UV radiation(night time, winter, increasing cloudiness,thickening stratospheric O3 layer), lowering NOemissions, and lowering H2O (e.g., in a coolingclimate). Its sinks can be increased by increasingemissions of reduced gases (CO, RH, SO2, etc.).Hence, OH levels are expected to have changed

substantially over geological and even recenttimes. But because natural or anthropogeniccombustion of biomass or fossil fuel is a primarysource of gases driving the oxidation processes,and because CO and NO are both products ofcombustion with (usually) opposite effects on OHlevels, the system is not as unstable as it couldotherwise be.Because O3 is a powerful infrared absorber and

emitter (a “greenhouse” gas), and because manyother greenhouse gases, notably CH4, aredestroyed principally by reaction with OH, thereare significant connections between atmosphericoxidation processes and climate. Added to thesegreenhouse gas related connections is the fact thatthe first step in the production of aerosols fromgaseous SO2, NO2, and hydrocarbons is almostalways the reaction with OH. Aerosols play akey role in climate through absorption and/orreflection of solar radiation.These oxidation processes also dominate in the

chemistry of air pollution,where the occurrence ofharmful levels of O3 and acidic aerosols dependson the relative and absolute levels of urbanemissions of NO, CO, RH, and SO2. In the marineand polar troposphere, reactive compounds ofchlorine, bromine, and iodine provide a small butsignificant additional pathway for oxidationbesides OH, as discussed by von Glasow andCrutzen (see Chapter 4.02).This chapter reviews the possible evolution of

oxidizing capacity over geologic time, the funda-mental chemical reactions involved, and theinfluences of meteorology and human activity onatmospheric oxidation. The measurement ofoxidation rates (i.e., oxidative capacity), and thecomplex interactive models of chemistry, trans-port, and human and other biospheric activitiesthat are now being developed to help understandatmospheric oxidation as a key process in theEarth system are also discussed.

4.01.2 EVOLUTION OF OXIDIZINGCAPABILITY

A discussion of the evolution of O3 and OHover geologic time needs to be placed in thecontext of the evolution of the Earth system(see Table 3). The Earth accreted between 4.5 and4.6 billion years (Gyr) before present (BP).Enormous inputs of kinetic energy, ice, andhydrated minerals during accretion would haveled to a massive H2O (steam) atmosphere withextremely high surface temperatures (Lange andAhrens, 1982). This earliest atmosphere may alsohave been partly or largely removed by thesubsequent Moon-forming giant impact. Due tothe apparent obliteration of rocks much olderthan 4 Gyr on Earth, very little is known directly

200

400

600

800

1,00090˚ S 60˚ S 30˚ S

30

40 40

30˚ N 60˚ N 90˚ N0˚

12080

60

60 40

200

400

600

800

1,00090˚ S 60˚ S 30˚ S 30˚ N 60˚ N 90˚ N0˚

Latitude

5 1015

20

10 5

(a)

(b)

Pres

sure

(hPa

)Pr

essu

re(h

Pa)

Figure 1 (a) Ozone mole fractions (in ppb) and(b) hydroxyl radical concentrations (in 105 radicalscm23) as functions of latitude and pressure (in hPa)from a model of the atmosphere in the 1990s (source

Wang and Jacob, 1998).

Evolution of Oxidizing Capability 3

about the Earth’s earliest atmosphere. It is hypo-thesized that, as the Earth cooled, the massiveH2O atmosphere condensed forming oceansand a much less massive CO2/CO/N2/H2O/H2

atmosphere then evolved.The next big evolutionary step, involving the

aqueous weathering of rocks and removal ofCO2 as carbonates, had occurred ,3.8 Gyr ago(Holland, 1984). Thus, the early Archaean atmos-phere was probably largely dominated by N2 withsmaller amounts of CO2, CO, H2O, H2, CH4,HCN, and NOx (from lightning). The evolution ofthe first life forms (procaryotes), and specificallyphotosynthetic cyanobacteria, would haveallowed the slow accumulation of biologicallyproduced O2 in the atmosphere in the lateArchaean period. Similarly, the early evolutionof methane and nitrogen bacteria along with theevolution of the aerobic biosphere led, apparentlyearly in the Proterozoic, to an N2/O2/H2O/CO2/CH4/N2O atmosphere with all of these gases, withthe exception of H2O, being largely controlled bybiological sources and biological or atmosphericchemical sinks.

4.01.2.1 Prebiotic Atmosphere

Before the evolution of the photosyntheticsource of O2, the atmosphere was relativelychemically neutral or even reducing. Its reductionstate would have depended significantly on thequantities of reduced gases produced from volca-nic sources (e.g., sources in which primordial iron,sulfur, carbon, and organic compounds werethermally oxidized by H2O and CO2 producingH2, CO, and CH4 (Prinn, 1982; Lewis and Prinn,1984)). In these prebiotic atmospheres, smallamounts of O2 can be produced from H2Ophotodissociation (Levine, 1985), and smallquantities of NO can be formed by shock heating,by lightning, and large collisions (Chameidesand Walker, 1981; Prinn and Fegley, 1987;Fegley et al., 1986).However, it is reasonable to conclude that the

evolution of the powerful oxidizing processesinvolving O3 and OH in the present atmospherehad to await the evolution of photosynthetic O2.Before that, atmospheric chemical transformationswere probably driven by simple photodissociation

Table 3 Schematic showing time in billions of years before present, major eras and associated biosphericand geospheric evolutionary events, and hypothesized atmospheric composition. Gases are listed in order oftheir hypothesized concentrations and those whose sources and/or sinks are dominated by biological processes aredenoted with an asterisk. Note the transition from an abiotic to an essentially biologically controlled atmosphere over

geologic time.

Time Major evolutionary events Hypothesized atmospheric compositions(Gyr BP)

Earth accretes and massivesteam atmosphere forms

[H2O, CO2, CO, N2, H2]

Surface cools, volcanism begins?Ocean forms, continental weathering,and carbonate precipitation begin

[CO2, CO, N2, H2O, H2]

4 Archean era begins, oldest rocksLate heavy bombardmentSediment subduction and volatile

recycling begins?[N2, CO2, CO, H2O, H2, CH4, HCN, NOx]

First living organisms evolve,prokaryotes

3 First protosynthetic organisms evolve:cyanobacteria, stromatolites

[N2, H2O, CO2, *O2, CH4]

Proterozoic era begins: scarce fossils2 Aerobic biosphere with nitrogen

and methane microbes evolves[*N2, *O2, H2O, *CO2, *CH4, *N2O]

1 Eucaryotes evolveMetazoans appearInvertebrates appearPaleozoic era begins, ubiquitous fossilsFishes appear [*N2, *O2, H2O, *CO2, *CH4, *N2O]Land plants appearAmphibians appearMesozoic (reptilian) era beginsFission of Pangean supercontinentCretaceous–Tertiary collisionCenozoic (mammalian) era begins

0 Present [*N2, *O2, H2O, *CO2, *CH4, *N2O, *CF2Cl2, etc.]


and dissolution reactions and by episodic pro-cesses like lightning and collisions. The trans-formation from reducing/neutral to oxidizingatmosphere was itself a long-term competitionbetween abiotic sources of reduced gases, anaero-bic microbial processes, and finally the aerobicbiological processes which came to dominate thebiosphere. Once the land biosphere evolved, theburning of vegetation ignited by lightningprovided an additional source of trace gases(see Chapter 4.05).

4.01.2.2 Pre-industrial Atmosphere

We expect that OH and O3 changed even underthe approximately constant levels of atmosphericO2 in the Cenozoic era. For example, the glacial–interglacial cycles should have significantlychanged atmospheric H2O (inferred to be low inice ages) and CH4 (observed to be low in iceages). But these particular changes would havepartially offsetting effects on OH (by loweringboth OH production and OH removal rates in iceages). And unfortunately, the lack of directinformation about the abundance of the keyshort-lived gases (CO, NOx, nonmethane hydro-carbons (NMHC), and stratospheric O3) inpaleoenvironments make it difficult to be quan-titative about past OH changes (Thompson,1992). One observational record over the pastmillennium which is useful in this respect isthe analysis of hydrogen peroxide (H2O2) inGreenland permafrost, which indicates stableH2O2 from the years 1300 to 1700 followed bya 50% increase through to 1989 (however, mostof the increase has occurred since 1980 and is notseen in Antarctic permafrost (Sigg and Neftel,1991; Anklin and Bales, 1997)).Increases in H2O2 during industrialization

could signal a decrease in OH due to netconversion of HO2 to H2O2 rather than recyclingof HO2 back to OH. Staffelbach et al. (1991)have used CH2O measurements in Greenland icecores to suggest that pre-industrial OH levelswere 30% higher than the levels in the latetwentieth century. Indeed, a variety of modelstudies which assume low pre-industrial emis-sions of CO, NOx, and NMHC compared to thelevels in the twentieth century suggest that the OHlevels in the years 1200–1800 were ,10–30%higher than the levels at the end of the twentiethcentury (Thompson, 1992). Ozone levels inthe 1880s can be inferred from iodometricmeasurements (Volz and Kley, 1988; Marencoet al., 1994) indicating mole fractions as low asð7–12Þ £ 1029; which are lower by a factor of 3or more from those observed in the 1990s. Suchlow levels of O3 are difficult to reconcile with COmeasurements in ice cores (Haan et al., 1996)

indicating CO mole fractions of ,60 £ 1029

(Antarctica) and ,90 £ 1029 (Greenland), whichare not much less than those observed today. Aswe will discuss in Section 4.01.3.1, oxidation ofeach CO molecule ought to produce one O3

molecule, so that presuming OH levels in the1880s to be similar to those in the 1990s impliessimilar O3 production rates from CO.

4.01.3 FUNDAMENTAL REACTIONS

4.01.3.1 Troposphere

As we noted in Section 4.01.1, the ability of thetroposphere to chemically transform and removetrace gases depends on complex chemistry drivenby the relatively small flux of energetic solar UVradiation that penetrates through the stratosphericO3 layer (Levy, 1971; Chameides and Walker,1973; Crutzen, 1979; Ehhalt et al., 1991;Logan et al., 1981; Ehhalt, 1999; Crutzen andZimmerman, 1991). This chemistry is also drivenby emissions of NO, CO, and hydrocarbons andleads to the production of O3, which is one of theimportant indicators of the oxidizing power ofthe atmosphere. But themost important oxidizer isthe hydroxyl free radical (OH), and a key measureof the capacity of the atmosphere to oxidize tracegases injected into it is the local concentration ofhydroxyl radicals.Figure 2 summarizes, withmuch simplification,

the chemistry involved in production and removalof O3 and OH (Prinn, 1994). This chemistry isremarkable because it involves compoundsthat also influence climate, air pollution, andacid rain.The greenhouse gases involved here include

H2O, CH4, and O3. Primary pollutants emittedmainly as a result of human activity (includingfossil-fuel combustion, biomass burning, andland-use) include RH, CO, and nitrogen oxides(NO, NO2). Reactive free radicals or atoms are intwo categories: the very reactive species, such asO(1D) and OH, and the less reactive ones, such asHO2, O(3P), NO, and NO2.The atmospheric cleansing role of OH is very

evident in Figure 2. When OH reacts with an RH,the latter is oxidized in a series of steps mostly toCO. These steps consume OH but may alsoproduce HO2. In turn, OH oxidizes CO to CO2,and it also oxidizes the gases NO2 and SO2 tonitric (HNO3) and sulfuric (H2SO4) acids, respect-ively. The primary OH source involves watervapor which reacts with the very reactive singletoxygen atom, O(1D), that comes from photo-dissociation of O3 by solar UV radiation atwavelengths less than 310 nm. Typically, OHwithin a second of its formation oxidizes othercompounds either by donating oxygen or by

Fundamental Reactions 5

removing hydrogen leaving an H atom or organicfree radical (R). Then R and H attach rapidly tomolecular oxygen to form hydroperoxy radicals(HO2) or organoperoxy radicals (RO2), which arerelatively unreactive. If there is no efficient way torapidly recycle HO2 back to OH, then levels ofOH are kept low.Adding the nitrogen oxides (NO and NO2),

which have many human-related sources, sig-nificantly changes this picture. Specifically, NOreacts with the HO2 to form NO2, and reformOH. UV radiation then decomposes NO2 atwavelengths less than 430 nm to form O3 andreform NO, so the nitrogen oxides are notconsumed in this reaction. In a study ofrelatively polluted air in Germany, Ehhalt(1999) estimated that the production rate ofOH by this secondary path involving NO is 5.3times faster than the above primary pathinvolving O(1D). Note that the reaction of NOwith HO2 does not act as a sink for the so-calledodd hydrogen (the sum of the H, OH, and HO2

concentrations). Rather, it determines the ratio ofOH to HO2. In the Ehhalt (1999) study, the[HO2] : [OH] ratio was 44 : 1. This is due to, andindicative of, the much greater reactivity of OHcompared to HO2. To summarize, the principal

catalyzed reactions creating OH and O3 in thetroposphere are shown in Equations (1)–(3),(4)–(9), and (10)–(15):

O3 þ UV! O2 þ Oð1DÞ ð1Þ

Oð1DÞþ H2O! 2OH ð2Þ––––––––––––––––––––––––––––––––––––––––

Net effect : O3 þ H2O! O2 þ 2OH ð3Þand

OHþ CO! Hþ CO2 ð4Þ

Hþ O2! HO2 ð5Þ

HO2 þ NO! OHþ NO2 ð6Þ

NO2 þ UV! NOþ O ð7Þ

Oþ O2! O3 ð8Þ––––––––––––––––––––––––––––––––––––––––

Net effect : COþ 2O2! CO2 þ O3 ð9Þ

Stratosphere

NO2

HO2

HNO3 H2SO4

N2O

CFCs

SO2 (CH3)2S

NO

O2

O2

O2

O3

CO2 CO2CO

OH

OH

O(1D)

H2O RHRH

UV

UV

Greenhouse gases

Primary pollutants

Natural biogenic species

Reactive free radical/atom

Less reactive radicals

Reflective aerosols

Hydrosphere HydrosphereBiosphere

OH

OH

OH

Lightning

Figure 2 Schematic showing principle oxidation processes in the troposphere in NOx-rich air (after Prinn,1994). In NOx-poor air (e.g., remote marine air), recycling of HO2 to OH is achieved by reactions of O3

with HO2 or by conversion of 2HO2 to H2O2 followed by photodissociation of H2O2. In a more completeschematic, nonmethane hydrocarbons (RH) would also react with OH to form acids, aldehydes and ketones in

addition to CO.


and

OHþ RH! Rþ H2O ð10Þ

Rþ O2! RO2 ð11Þ

RO2 þ NO! ROþ NO2 ð12Þ

NO2 þ UV! NOþ O ð13Þ

Oþ O2! O3 ð14Þ––––––––––––––––––––––––––––––––––––––––

Net effect : OHþ RHþ 2O2! ROþ H2Oþ O3

ð15ÞIn theoretical models, the global production of

tropospheric O3 by the above pathway beginningwith CO is typically about twice that beginningwith RH. For most environments, it is theseoverall catalytic processes that pump themajority of the OH and O3 into the system. Ifthe concentration of NO2 becomes too high,however, its reaction with OH to form HNO3

ultimately limits the OH concentration (seeSection 4.01.4). The global impact of theabove tropospheric OH production mechanismshas been summarized in Table 2. Hough andDerwent (1990) have estimated the globalproduction rate of tropospheric O3 by theabove reactions to be 2,440 Tg yr21 in pre-industrial times and 5,130 Tg yr21 in the 1980swith the increase caused by man-made NOx,hydrocarbons, and CO.An update to the above 1980s calculation

(Ehhalt, 1999) gives a production rate of4,580 Tg yr21. A comparison of several tropo-spheric O3 models showed a chemical productionrate range for O3 of 3,425–4,550 Tg yr21

(Lelieveld et al., 1999). Since the tropospherecontains ,350 Tg of O3 (Ehhalt, 1999), thechemical replenishment time for O3 (defined asthe tropospheric content divided by the tropo-spheric chemical production rate) is 28–37 d. Thisis much longer than the replenishment time forOH,,1 s, and indicates the enormous reactivity ofOH relative to O3.

How stable is the oxidizing capability of thetroposphere? How might it have changed overgeologic time even after oxygen reached neartoday’s levels? If emissions of gases that reactwith OH such as CH4, CO, and SO2 are increasingthen, keeping everything else constant, OHlevels should decrease. Conversely, increasingNOx -emissions from combustion (of biomass inthe past (see Chapter 4.05) augmented by fossilfuel today) should increase tropospheric O3 (andthus the primary source of OH), as well as increasethe recycling rate of HO2 to OH (the secondarysource of OH).

Also, if the temperatures of oceans are increas-ing, we expect increased water vapor in the lowertroposphere. Because water vapor is part of theprimary source of OH, climate warming alsoincreases OH. This increase could be lowered orraised due to changes in cloud cover accompanyingthe warming leading to more or less reflection ofUV back toward space. Rising temperature alsoincreases the rate of reaction of CH4 with OH,thus lowering the lifetimes of both chemicals.The opposite conclusions apply if trace gasemissions increase or the climate cools. Finally,decreasing stratospheric O3 can also increasetropospheric OH.

4.01.3.2 Stratosphere

The O3 in the stratosphere is maintained by adistinctively different set of chemical reactionsthan the troposphere. The flow diagram inFigure 2 shows how the driving chemicals likewater vapor, chlorofluorocarbons, and nitrousoxide, are transported up from the troposphereto the stratosphere, where they produce chlorine-,nitrogen-, and hydrogen-carrying free radicals,which destroy O3. There is an interestingparadox here: the nitrogen oxide free radicals,which are responsible for destroying a significantfraction of the O3 in the stratosphere, are thesame free radicals that are producing O3 in thetroposphere through the chemistry outlinedearlier (Figure 3).When the ratio of nitrogen oxide free radicals to

O3 is low, as it is in the stratosphere, the net effectis for O3 to be destroyed overall, while if thisratio is high, as it is in much of the troposphere,then O3 is produced overall (Crutzen, 1979).The stratosphere is chemically very active due

to its high O3 and UV levels. But the densitiesand usually temperatures are much lower thanin the lower troposphere so that only a smallfraction of the destruction of the gases listedin Table 2 occurs in the stratosphere. Thestratosphere contains ,90% of the world’s O3

and exports ,400–846 Tg yr21 of O3 to thetroposphere to augment the chemical productiondiscussed in Section 4.01.3.1 (Lelieveld et al.,1999).There is also an important link between strato-

spheric chemistry and climate. The precursorgases for the destructive free radicals in thestratosphere are themselves greenhouse gases, asis the stratospheric O3 itself. Therefore, whileincreases in the concentrations of the sourcegases will increase the radiative forcing ofwarming, these increases will, at the sametime, lead to decreases in stratospheric O3, thuslowering the radiative forcing. Once again, as inthe troposphere, there are important feedbacks

Fundamental Reactions 7

between chemical processes and climate throughthe various greenhouse gases. It is not just man-made chemicals that can change the stratosphericO3. Changes in natural emissions of N2O and CH4,and changes in climate that alter H2O flows intothe stratosphere, undoubtedly led to changes in thethickness of the O3 layer over geologic time evenafter O2 levels reached levels near those in theearly 2000s.There is also a strong link between the thickness

of the stratospheric O3 layer and concentrations ofOH in the troposphere. Because O(1D) can only beproduced from O3 at wavelengths less than310 nm, and the current O3 layer effectivelyabsorbs all incoming UV at less than 290 nm,there is only a very narrow window of 290–310 nm radiation that is driving the majoroxidation processes in the troposphere. Decreasein the amount of stratospheric O3, which hasoccurred since about 1980, can therefore increasethe tropospheric OH by increasing the flux ofradiation less than 310 nm which reaches thetroposphere to produce O(1D) (Madronich andGranier, 1992). By the same arguments, if thestratospheric O3 layer was very much thicker inthe past, it would have removed a great deal of the“cleansing” capacity of the Earth’s troposphereleading to much higher levels of gases like COand CH4.

4.01.4 METEOROLOGICAL INFLUENCES

The oxidative capability of the atmosphere isnot simply a function of chemistry. Convectivestorms can carry short-lived trace chemicals fromthe planetary boundary layer (the first few hundredto few thousand meters) to the middle and uppertroposphere in only a few to several hours. Thiscan influence the chemistry of these upper layersin significant ways by delivering, e.g., reactivehydrocarbons to high altitudes. Conversely, theoccurrence of very stable conditions in theboundary layer can effectively trap chemicalsnear the surface for many days, leading to pollutedair. Larger-scale circulations serve to carry gasesaround latitude circles on timescales of a fewweeks, between the hemispheres on timescales ofa year, and between the troposphere and strato-sphere on timescales of a few years.A convenient measure of the stability of a

trace chemical is its local chemical lifetime,which is defined as its local concentration (e.g., inmol L21) divided by its local removal rate (e.g., inmol L21 s21). Concentrations of chemicals whoselifetimes (due to chemical destruction or depo-sition) are comparable to, or longer than, the abovetransport times will be profoundly affected bytransport. As a simple example, in a steady(constant concentration) state with a horizontal

O2

N2O

CH4

CFCl3CF2Cl2H2O

H2O

O2

O2

O2

O3 OUV

UVetc.

UV

Source/Greenhouse gas Destructive atom/free radical Benign reservoir gas

OH NOCl

ClO

(ClO)2HO2 NO2

HNO3

HNO3

ClONO2

HOCl

HClHCl

N2O5

H2O2

TROPOSPHERE

Figure 3 Schematic showing major chemical processes forming and removing O3 in the stratosphere (after Prinn,1994). UV photodissociaton of largely natural (N2O, CH4, H2O) and exclusively man-made (CFCl3, CF2Cl2)source gases leads to reactive HOx, NOx, and ClOx free radicals which catalytically destroy O3. The catalystsreversibly form reservoir compounds (HNO3, etc.), some of which live long enough to be transported down to

the troposphere.


wind u; the concentration ½i* of chemical i whoselifetime is ti decreasesdownwind of a source region(located at horizontal position x ¼ 0) according to

½i*ðxÞ ¼ ½i*ð0Þexp32x

ðutiÞ4

Specifically, the concentration of i decreasesby a factor of e over the distance scale uti: Notealso that, if the transport time x=u is much lessthan ti; the concentration remains essentiallyconstant and the species can be considered tobe almost inert. For oxidation by OH of NO2 andSO2 to form the acids HNO3 and H2SO4, the timeti is typically 3 d ð2:6 £ 105 sÞ; so the e-foldingdistance uti is only ,800 km for a typical u of3 m s21. Alternatively, for the oxidation of COby OH to form CO2, ti ¼ 3 months ð7:9 £ 106 sÞ;so uti ¼ 2:4 £ 104 km for the same u. In a simpleway this illustrates why NO2 and SO2 are local orregional pollutants, whereas CO is a hemisphericpollutant.In the above example, if the chemical lifetime

ti is much less than the transport time x=u then wecould ignore transport and simply considerchemical sources and sinks to determine concen-trations. However, for the chemicals controllingoxidation reactions, care must be taken indefining ti: Specifically, sometimes members ofa “chemical family” are converted from one toanother on very short timescales relative totransport, whereas the total population of thefamily is produced and removed on longertimescales.For example, in the family composed of NO and

NO2 (called the NOx family) the reactions shownin Equations (16)–(18) rapidly interconvert NOand NO2 on timescales much less than typicaltransport times and do so without affecting thetotal ½NOx* ¼ ½NO*þ ½NO2*:

NOþ O3! NO2 þ O2 ð16Þ

NO2 þ hn! NOþ O ð17Þ

Oþ O2 þM! O3 þM ð18Þ

NO2 þ OHþM! HNO3 þM ð19ÞHowever, the removal of NOx by the reaction

shown in Equation (19) is much slower (e.g.,,3 d). Thus, we can use the chemical steady-state approximation (which equates chemicalproduction to chemical loss only) to define theratio of NO to NO2. We must however take fullaccount of transport in calculating ½NOx*:Similar arguments hold for the “odd oxygen”ð½Ox* ¼ ½O*þ ½O3*Þ and “odd hydrogen”ð½HOx* ¼ ½H*þ ½OH*þ ½HO2*Þ families.Besides being a vigorous vertical transport

mechanism, moist convection also directly affects

the oxidizing reactions in the atmosphere. First,lightning and thundershocks serve to thermallyconvert atmospheric N2 and O2 to 2NO, leading toa very important natural source of NOx. Second,the formation of cloud droplets is aided by (andconsumes) aerosols (as cloud condensationnuclei), and the cloud droplets then serve assinks for soluble species like H2O2, HNO3, NOx,SO2, and H2SO4. For some oxidation reactions,e.g., oxidation of SO2 to H2SO4, this dissolutionleads to an additional nongaseous pathway foroxidation generally considered comparable tothe gas-phase processes for SO2. Finally, theformation of convective clouds, and particularlythe extensive anvils, leads to reflection ofUV radiation enhancing the photo-oxidationprocesses above the cloud and inhibiting thembelow (e.g., Wang and Prinn, 2000).

4.01.5 HUMAN INFLUENCES

4.01.5.1 Industrial Revolution

The advent of industrialization since 1850 hasled to significant increases in the global emis-sions of NOx, CO, NMHC, and CH4 (Wang andJacob, 1998; Prather et al., 2001). Biomassburning associated with human land use providesan additional source of these trace gases (seeChapter 4.05). Also, the observed temperatureincrease of 0.6 ^ 0.2 8C in this time period(Folland et al., 2001) should have led to increasesin tropospheric H2O and possibly changes inglobal cloud cover (of unknown sign). Anthro-pogenic aerosols, both scattering and absorbing,have also increased due to human activity(Penner et al., 2001). These aerosol increasesshould decrease UV fluxes below the cloud topsdue to both their direct optical effects (absorbingor reflecting) and their indirect effects ascloud condensation nuclei, which should increasethe reflectivity of water clouds. Finally, depletionof stratospheric O3, which has occurred due toman-made halocarbons, will increase tropo-spheric UV irradiation (Madronich and Granier,1992).All of these changes should have affected O3

and OH concentrations, sometimes offsetting andsometimes augmenting each other, as discussedearlier. Table 4 summarizes the expectedchanges in OH and O3 relative to pre-industrialtimes. Where quantitative estimates are notavailable, only the signs of expected changesare indicated. Note in particular that for OH thepositive effects of NOx increases have been morethan offset by the negative effects of CO,NMHC, and CH4 increases, so the net changein OH is a modest 10% decrease (Wangand Jacob, 1998; see also Levy et al., 1997).

Human Influences 9

Since combustion is a common, and (except forCH4) a dominant source for all these gasesaffecting OH, then the offsetting increases ofNOx, CO, and hydrocarbons mean that OH levelsare not very sensitive overall to past increases incombustion. Note, however, that if future airpollution controls lower NOx emissions morethan CO and hydrocarbon emissions, then OHdecreases should result.For O3, in contrast to OH, the effects of NOx,

CO, and RH emission increases are additive andin sum are calculated to have caused ,63%increase in O3 (Wang and Jacob, 1998). Theactual increase may be even larger, sinceobservations (albeit not very accurate) of O3

during the 1800s (Marenco et al., 1994) suggesteven lower values than those computed by Wangand Jacob (1998).

4.01.5.2 Future Projections

The importance of oxidation processes inclimate led the Intergovernmental Panel onClimate Change (IPCC) to investigate the effectson O3 and OH of 11 IPCC scenarios for futureanthropogenic emissions of the relevant O3 andOH precursors (Prather et al., 2001). In all of thescenarios, except one, OH was projected todecrease between 2000 and 2100 by 3–18%,while O3 was projected to increase by 2.6–46%(Table 5). The smallest decrease in OH andsmallest increase inO3 was seen as expected in thescenario in which emissions of NOx, CO, NMHCs,and CH4 were projected to change by onlyþ3.4%, 28.3%, þ2.0%, and þ9.2%, respect-ively, between 2000 and 2100. Also as expected,the greatest OH decreases were seen in two

Table 4 Effects of anthropogenically driven or observed changes in trace gas emissions, concentrations, ormeteorological variables between pre-industrial and present times on the oxidizing capability of the atmosphere

(expressed where available as percentage changes (D) in [OH] or [O3] from pre-industrial to present).

Present/pre-industrial D[OH]/[OH] D[O3]/[O3]

NOxa

4.7 þ36% þ30%CO

a3.2 232%

b þ10%b

NMHCa

1.6CH4

a2.1 217% þ13%

Combineda

210%c þ63%c

H2O(T) .1 .0 .0Clouds

d.1, ,1 ,0, .0 ,0, .0

Scat. aeroe

.1 ,0 ,0Abs. aero

f.1 ,0 ,0

Strat. O3g

,1 .2% .0

a Wang and Jacob (1998). Individual effect of each specific chemical computed in sensitivity runs with pre-industrial conditions except forthe specific chemical which is at present levels. b CO and NMHC effects combined. c CO, NMHC, CH4, and NOx effects combined.Total effect not equal to sum of individual effects due to interactions. d Cloud cover changes unknown. Effects shown for either an increase or adecrease. e Specifically aerosols with high single scattering albedos (e.g., sulfates) which reflect UV radiation back to space. f Specifically aerosolswith low single scattering albedos (e.g., black carbon) which absorb UV radiation. g Madronich and Granier (1992) and Krol et al. (1998).

Table 5 Percentages by which either emissions of NOx, CO, NMHC (plus oxidized NMHC), and CH4, orconcentrations of tropospheric OH, or column amounts of tropospheric O3, are projected to change between 2000 and

2100 for various IPCC scenarios.

Scenario NOx

(%)CO(%)

NMHC(%)

CH4

(%)OH(%)

O3

(%)a

A1B þ26 þ90 þ37 211 210 þ7.7A1T 212 þ137 29.2 215 218 þ5.5A1FI þ243 þ193 þ198 þ128 214 þ47A2 þ241 þ165 þ143 þ175 212 þ46B1 242 259 238 227 þ5 28.6B2 þ91 þ128 þ21 þ85 216 þ23A1p þ27 þ138 þ15 213 218 þ11A2p þ238 þ140 þ133 þ163 212 þ46B1p þ3.4 28.3 þ2.0 þ9.2 23 þ2.6B2p þ86 þ100 þ13 þ46 211 þ19IS92a þ124 þ64 þ125 þ95 211 þ29

Source: Prather et al. (2001).aOzone changes decreased by 25% to account for modeling errors as recommended by IPCC.


scenarios in which small increases or evendecreases in NOx emissions (OH source) wereaccompanied by relatively very large increases inCO emissions (OH sink). The greatest O3

increases were seen in two scenarios whereemissions of all four O3 precursors increasedvery substantially. Finally, the one anomalousscenario (in which OH increased by 5% and O3

decreased by 8.6%) involved significant decreasesin emissions of all four precursors (which is veryunlikely).The above results make it very clear that

forecasts of the future oxidation capacity of theatmosphere depend critically on the assumedemissions. The IPCC did not assign probabilitiesto its emission scenarios but it is apparent thatsome of these scenarios are highly improbable foroxidant precursor emissions. Integrated assess-ment models, which couple global economic andtechnological development models with naturalEarth system models provide an alternativeapproach to the IPCC scenario approach with theadded advantage that objective estimates ofindividual model uncertainties can be combinedwith Monte Carlo approaches to provide moreobjective ways of defining means and errors in

forecasts.We will use here one such model (Prinnet al., 1999) that integrates a computable generalequilibrium economics model with a coupledchemistry and climate model, and an ecosystemmodel. The economics model automaticallyaccounts for the strong correlations betweenemissions of air pollutants (gases and aerosols)and greenhouse gases due to their commonproduction processes (e.g., combustion of coal,oil, gasoline, gas, biomass).Results from this model for (i) a high-

probability reference emissions forecast and(ii) a lower-probability high and low emissionsforecasts are shown in Figure 4 (Wang and Prinn,1999). Emissions driving these projections aresummarized in Prinn et al. (1999); referenceanthropogenic CO, NOx, and CH4 emissionsincrease by ,18%, ,95%, and ,33%, respect-ively, between 2000 and 2100 in these projections.Evidently, these emission projections from theeconomics model (which consistently models allof the relevant industrial and agricultural sectorsproducing these gases) lead to projections ofsignificant depletion of OH and enhancement ofO3 between 2000 and 2100 with magnitudescomparable to the larger of the changespredicted from the IPCC scenarios (Table 5).

4.01.6 MEASURING OXIDATION RATES

Given its importance as the primary oxidizingchemical in the atmosphere, a great deal of efforthas been given to measuring OH concentrationsand trends. Two approaches, direct and indirect,have been taken to address this measurement need.

4.01.6.1 Direct Measurement

The direct accurate measurement of local OHconcentrations has been one of themajor technicalchallenges in atmospheric chemistry since theearly 1980s. This goal was first achieved in thestratosphere (e.g., Stimpfle and Anderson, 1988),but the troposphere provedmore difficult (Crosley,1995). Nevertheless, early long-baseline absorp-tion methods for OH were adequate to test somebasic theory (e.g., Poppe et al., 1994). Currentsuccessful direct methods include differentialoptical absorption near-UV spectroscopy withlong baselines (e.g., Mount, 1992; Dorn et al.,1995; Brandenburger et al., 1998), laser-inducedfluorescence after expansion of air samples(e.g., Hard et al., 1984, 1995; Holland et al.,1995), and a variety of chemical conversiontechniques (Felton et al., 1990; Chen and Mopper,2000; Tanner et al., 1997).The long-baseline spectroscopic techniques

require path lengths in air of 3–20 km, which

1980 2000 2020 2040 2060 2080 2100

Year

12

11

10

9

8

20001980 2020 2040 2060 2080 2100

Year

40

35

30

25

OH

conc

entr

atio

n(i

n10

5ra

dica

lscm

–3)

O3

mol

efr

actio

n(i

npp

b)

(a)

(b)

highreferencelow

highreferencelow

Figure 4 (a) Predicted future tropospheric averageconcentrations of OH and (b) mole fractions of O3 forthree scenarios of future emissions of O3 and OHprecursors (as well as greenhouse gases) predicted in aneconomics model (Prinn et al., 1999). The naturalsystem model receiving these emissions includessimultaneous calculations of atmospheric chemistry

and climate (Wang and Prinn, 1999).

Measuring Oxidation Rates 11

can be achieved by a single pass through air(e.g., Mount and Harder (1995) used a 10.3 kmpath to achieve a sensitivity of 5 £ 105

radicals cm23), or multiple reflections in longcells (e.g., Dorn et al. (1995) used 144 passesthrough a 20 m cell to observe six OH absorptionlines around 308 nm). The laser-induced fluor-escence methods use either 282 nm or 308 nmphotons to excite OH, which then subsequentlyemits at 308–310 nm (e.g., Brune et al. (1995);Holland et al. (1995) use 308 nm laser lightand achieve detection limits of ð1–3Þ £ 105

radicals cm23).The common chemical conversion techniques

use a flow reactor in which ambient OH reactswith isotopically labeled SO2 or CO to yieldobservable products (e.g., Felton et al. (1990) used14CO with radioactive 14CO2 detection, whileEisele and Tanner (1991) used 34SO2 withH2

34SO4 detection by ion-assisted mass spec-trometry).These various direct OH measurement tech-

niques enable critical tests of current theoreticalmodels of fast photochemistry. This is achieved bysimultaneously measuring NO, NO2, CO, H2O,O3, RH, and other relevant trace species, UVfluxes, the frequency of photodissociation of O3 toproduce O(1D), and the concentrations of OH (andsometimes HO2). A model incorporating the bestestimates of the relevant kinetic rate constants and

absorption coefficients is then used, along with thetrace gas and photodissociation rate measure-ments, to predict OH concentrations for compari-son with observations. Examples of results fromthe Mauna Loa Observatory PhotochemistryExperiment (MLOPEX) and the Photochemistryof Plant-emitted Compounds and OH Radicals inNorth Eastern Germany Experiment (POPCORN)are shown in Figures 5 and 6, respectively.The 1992 MLOPEX experiment was able to

sample both free tropospheric air (associated withdownslope winds at this high mountain station)and boundary layer air (upslope winds). It isevident from Figure 5 that agreement betweenobserved (using the 34SO2 method) and calculatedOH concentrations was always good in the freetroposphere, but in the summer the calculatedboundary layer OH was about twice that observed(Eisele et al., 1996). Since the measurementaccuracy was argued to be much better than factorof 2, this suggests a missing OH sink in theirmodel, which they suggest might be due tounmeasured hydrocarbons from vegetation orundetected oxidation products of anthropogeniccompounds.The 1994 POPCORN experiment in rural

Germany compared OH measurements usinglaser-induced fluorescence to calculations in aregional air chemistry model. Figure 6(a) showsthe measured OH concentrations as a function

107

106

107

106

107

106

[OH

](m

olec

ules

cm–3

)

6 7 8 9 10 11 12 13 14 15 16 17 18

Time of day (h)

Measuredw/o May 7Calculated

(a)

(b)

(c)

107

106

107

106

107

106

1056 7 8 9 10 11 12 13 14 15 16 17 18

Time of day (h)

MeasuredCalculated

(d)

(e)

(f)

Figure 5 Hourly average measured and calculated OH concentrations during spring (a)–(c) and summer (d)–(f)of 1992 in MLOPEX: (a), (d) all time; (b), (e) free tropospheric air only; (c), (f ) boundary layer air only (after

Eisele et al., 1996). May 7 measurements were anomalous.


of time. Note the expected strong diurnalvariation in OH due to the daily cycle in UVradiation. Also shown in Figure 6(b) is ademonstration of the expected strong correlationbetween the measured frequency of photoproduc-tion of O(1D) from O3(J(O

1D)) and the measuredOH (see Section 4.01.3.1). The strong high-frequency variations in OH evident in Figure 6(a)are real and are due to strongly correlatedvariations in solar UV fluxes caused by transientcloud cover. The agreement between observedand calculated OH was quite good in POPCORNbut, as in MLOPEX, there was also a tendencyfor the model to overpredict OH (Hofzumahauset al., 1996; Ehhalt, 1999). Part of the discre-pancy may be attributable to the combined effectsin the modeling of rate constant errors andmeasurement errors in the observed RH, CO,NO, NO2, O3, and H2O used in the model. Poppeet al. (1995) have estimated this modeling errorto be ^30%.Overall, considering the difficulties in both

the measurements and the models, the agree-ments between observation and theory areencouraging at least for the relatively clean and

clear air environments investigated in theseexperiments.The power and great importance of the in situ

direct measurement is that in conjunction withsimultaneous in situ observations of the othercomponents of the fast photo-oxidation cycles,they provide a fundamental test of the photo-chemical theory. However, the very short lifetimeand enormous temporal and spatial variability ofOH, already emphasized in Section 4.01.1, makeit impossible practically to use these directtechniques to determine regional to global scaleOH concentrations and trends. It is these large-scale averages that are essential to understandingthe regional to global chemical cycles of all of thelong-lived (more than a few weeks) trace gases inthe atmosphere which react with OH as theirprimary sink. For this purpose, an indirectintegrating method is needed.

4.01.6.2 Indirect Measurement

The large-scale concentrations and long-termtrends in OH can in principle be measuredindirectly using global measurements of trace

14

12

10

8

6

4

2

0

12

16

20

8

4

0

1.2

1.6

2.0

0.8

0.4

0.0

06:00 09:00 12:00

12:30 12:45 13:00 13:15 13:30 13:45 14:00

15:00 18:00

[OH

]/10

6cm

–3[O

H]/

106

cm–3

J(O

1 D)/

10–5

s–1

(a)

(b) Daytime / h(UT)

J(O1D)

[OH]

Figure 6 (a) Diurnal variation of OH concentrations during POPCORN (August 16, 1994) and (b) higher timeresolution version of (a) showing fast response of OH concentrations to J(O1D) variations (solid lines are three-point

running averages) (after Hofzumahaus et al., 1996).


gases whose emissions are well known and whoseprimary sink is OH. The best trace gas for thispurpose is the industrial chemical 1,1,1-trichloro-ethane (methylchloroform, CH3CCl3). First, thereare accurate long-term measurements of CH3CCl3beginning in 1978 in the ALE/GAGE/AGAGEnetwork (Prinn et al., 1983, 2000, 2001) andbeginning in 1992 in the NOAA/CMDL network(Montzka et al., 2000). Second,methylchloroformhas fairly simple end uses as a solvent, andvoluntary chemical industry reports since 1970,along with the national reporting proceduresunder the Montreal Protocol in more recentyears, have produced very accurate emissionsestimates for this chemical (McCulloch andMidgley, 2001).Other gases which are useful OH indicators

include 14CO, which is produced primarily bycosmic rays (Volz et al., 1981; Mak et al., 1994;Quay et al., 2000).While the accuracy of the 14COproduction estimates, and especially the frequencyand spatial coverage of its measurements, do notmatch those for CH3CCl3, its lifetime (2 months)is much shorter than for CH3CCl3 (4.9 yr), so itprovides estimates of average concentrations ofOH that aremore regional than CH3CCl3. Anotheruseful gas is the industrial chemical chlorodi-fluoromethane (HCFC-22, CHClF2). It yields OHconcentrations similar to those derived fromCH3CCl3 but with less accuracy due to greateruncertainties in emissions and less extensivemeasurements (Miller et al., 1998). The industrialgases CH2FCF3 (HFC-134a), CH3CCl2F (HCFC-141b), and CH3CClF2 (HCFC-142b) are poten-tially useful OH estimators but the accuracy oftheir emission estimates needs improvement(Simmonds et al., 1998; Huang and Prinn, 2002).The most powerful methods for determining

OH concentrations from trace gas data involvesolution of an inverse problem in which theobservables are expressed as Lagrangian lineintegrals, and the unknown OH concentrations(expressed as functions of space and time) are theintegrands. The inverse problem of interestconsists of determining an “optimal” estimate inthe Bayesian sense of the unknown OH concen-trations from imperfect concentration measure-ments of say CH3CCl3 over space and time. Theunknown OH concentrations are arrayed in a“state” vector xt and the CH3CCl3 measurementerrors are arrayed in a “noise” vector. Approxi-mating the line integral by a summation leads tothe observed CH3CCl3 concentrations beingexpressed as the noise vector plus a matrix of“partial derivatives” ðHÞ multiplied by xt: Hexpresses the sensitivity of the chemical transportmodel (CTM) CH3CCl3 concentrations tochanges in the state vector elements (i.e., tochanges in OH). Given the discrete timeseries nature of CH3CCl3 measurements, it is

convenient (but not essential) to solve for xt

using a discrete recursive optimal linear filtersuch as the discrete Kalman filter (DKF). TheDKF has the specific useful property that itprovides an objective assessment of the uncer-tainty in estimates of xt as each CH3CCl3measurement is used and thus of the usefulnessof each measurement. Application of the DKFrequires a CTM to compute H:While the derivation of the measurement

equation uses Lagrangian concepts, H can beequally well derived using an Eulerian CTM.Several intuitive concepts exist regarding theimportant effects of observational errors andCTM errors on the value of the observations inimproving and lowering the errors in the OHestimates (Prinn, 2000). (The latter paper iscontained in a teaching monograph designed tointroduce researchers to these inverse techniquesapplied to the biogeochemical cycles.) Appli-cation of optimal linear filtering requires carefulattention to both the physics of the problemexpressed in the “measurement” and “system”(or “state-space”) equations or models, and thesources and nature of the errors in the observationsand CTMs. This approach was the one adopted foranalysis of CH3CCl3 data by the ALE/GAGE/AGAGE scientists from the inception of theexperiment.The use of CH3CCl3 has established that the

global weighted average OH concentration inthe troposphere is, over the 1978–2000 period,,106 radicals cm23 (Prinn et al., 1987, 1992,1995, 2001; Krol et al., 1998; Montzka et al.,2000). The weighting factor for this average isformally the rate constant k (which is temperaturedependent) of the reaction of OH with methyl-chloroform, multiplied by the CH3CCl3 concen-tration, [CH3CCl3]. Practically, this means thatthe average is weighted toward the tropicallower troposphere. A similar average concen-tration is derived from 14CO (Quay et al., 2000),although the weighting here is not temperaturedependent.While the average OH concentration appears to

be fairly well defined by this method, thetemporal trends in OH are more difficult todiscern since they require long-term measure-ments, and very accurate calibrations, modeltransports, and above all, a very accurate timeseries of CH3CCl3 emissions (Prinn et al., 1992).In the first attempt at trend determination,Prinn et al. (1992) derived an OH trend of1.0 ^ 0.8% yr21 between 1978 and 1990.Due primarily to a subsequent recalibration ofthe ALE/GAGE/AGAGE CH3CCl3 measure-ments, the estimated 1978–1994 trend waslowered to 0.0 ^ 0.2% yr21 (Prinn et al., 1995).In a subsequent reanalysis of the same measure-ments (but not with the same emissions),


Krol et al. (1998) derived a trend of 0.46 ^0.6% yr21. The reasons for this difference,including differences in models, data processing,emissions, and inverse methods, have beenintensely debated (Prinn and Huang, 2001; Krolet al., 2001). However, when the Prinn et al.(1995) method is used with the same emissions asKrol et al. (1998), it yields a trend of 0.3% yr21,in reasonable agreement with their value (Prinnand Huang, 2001).In the latest analysis applied to the entire 1978–

2000 ALE/GAGE/AGAGE CH3CCl3 measure-ments, Prinn et al. (2001) showed that global OHlevels grew by 15 ^ 22% between 1979 and 1989.But, the growth rate was decreasing at a statisti-cally significant rate of 0.23 ^ 0.18% yr21, so thatOH began slowly declining after 1989 to reachlevels in 2000 that were 10 ^ 24% lower than the1979 values (see Figure 7). Overall, the weightedglobal average OH concentration was ½9:4^1:3* £ 105 radicals cm23 with concentrations14 ^ 35% lower in the northern hemisphere thanin the southern hemisphere. The Prinn et al. (2001)analysis included, in addition to the measurementerrors, a full Monte Carlo treatment of modelerrors (transport, chemistry), emission uncertain-ties, and absolute calibration errors with the last

two being the dominant contributors to OH trenderrors.Because these substantial variations inOHwere

not expected from current theoretical models,these results have generated much attention andscrutiny. Recognizing the importance of theirassumed emissions estimates in their calculations,Prinn et al. (2001) also estimated the emissionsrequired to provide a zero trend in OH. Theserequired emissions differ by many standarddeviations from the best industry estimatesparticularly for 1996–2000 (McCulloch andMidgley, 2001). They are also at odds withestimates of European and Australian emissionsobtained from measurements of polluted air fromthese continents reaching the ALE/GAGE/AGAGE stations (Prinn et al., 2001). Indeed, ifthese required zero-trend emissions are actuallythe correct ones, then the phase-out of CH3CCl3consumption reported by the party nations to theMontreal Protocol must seriously be in error. Thistopic is expected to be studied intensively in thefuture. The good news is that as the emissions ofCH3CCl3 reduce to essentially zero (as theyalmost are in the early 2000s), the influence ofthese emission errors on OH determinationsbecomes negligible.

12

11

10

9

8

7

1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000

[OH

](1

05ra

dica

lscm

–3)

Globe

16

14

10

12

8

6

4

1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000

Southern hemisphere

Northern hemisphere

Year(b)

(a)

Figure 7 (a) Annual average global and (b) hemispheric OH concentrations estimated from CH3CCl3measurements (after Prinn et al., 2001). Uncertainties (one sigma) due to measurement errors are indicated bythick error bars and uncertainties due to both measurement, emission, and model errors are shown by thin error bars.Also shown are polynomial fits to these annual values (solid lines), and polynomials describing OH variations which

were estimated directly from CH3CCl3 measurements (dotted lines).


Presuming that they are correct, the abovepositive and negative OH trends are not simplyexplained by the measured trends in trace gasesinvolved in OH chemistry. There is no clearnegative or positive global-scale trend in themajor OH precursor gases O3 and NOx between1978 and 1999, while levels of the dominant OHsink CO decreased, rather than increased, in thenorthern hemisphere (Prather et al., 2001). At thesame time, concentrations of another OH sink,CH4, have risen significantly during 1978–1999.Prinn et al. (2001) have suggested that increases intropical and subtropical countries of hydrocarbonsand SO2 that react with OH, along with a decreaseof NOx emissions in developed countries, could beinvolved.Also, a growth in levels of anthropogenic

aerosols could increase heterogeneous OHdestruction. Such a growth in aerosols (bothabsorbing and reflecting) could also lower UVfluxes, through direct and indirect reflection andabsorption, thus lowering OH. Neglect of theseaerosol effects, as well as urban-scale chemistry,which can remove NOx locally, may also be thereason why the higher OH levels inferred in thesouthern hemisphere are not simulated in many ofthe models in use in the early 2000s.It is clear that further study on possible causes

of OH variations of the type suggested from theCH3CCl3 analysis is necessary. Toward this end,the lack of long-term global-covering measure-ment of O3, NOx, SO2, hydrocarbons, and aerosolsis a very serious impediment to our understandingof OH chemistry.

4.01.7 ATMOSPHERIC MODELS ANDOBSERVATIONS

A major approach for testing our knowledge ofatmospheric oxidation processes has involved acareful comparison between trace gas and freeradical observations and the predictions fromthree-dimensional (3D) CTMs that incorporatethe latest understanding of homogeneous andheterogeneous fast photochemistry. While these3D CTMs simulate qualitatively many featuresof the observations, there are discrepanciesthat point out the need for additional research.Some illustrative examples are briefly discussedhere.Berntsen et al. (2000) noted that their 3D CTMs

indicate increases in upper tropospheric O3 in the1980s, whereas observations show a leveling offor a decrease. They point to the possible role ofunmodeled stratospheric O3 changes whichmight explain this discrepancy. Houweling et al.(2000) compared their 3D CTM simulations ofCH4 and CH3CCl3 observations providing a testof their CH4 emission assumptions and OH

concentrations. They concluded that furtherwork is needed on simulating seasonal CH4

emissions from wetlands and rice paddies.Spivakovsky et al. (2000) extensively testedtheir OH calculations using CH3CCl3, CHClF2,and other trace gas data and concluded that their3D CTM gave good agreement except in the twotropical winters (possible OH underestimates) andsouthern extratropics (possible OH overestimate).They noted the greater usefulness of CH3CCl3 forregional OH tests, as its emissions are becomingnegligible.Models have also been tested using paleodata as

noted in Sections 4.01.2.2 and 4.01.5.1.Wang andJacob (1998) noted their significant overestima-tion of O3 levels in the 1800s and discussed thepossible effects of incorrectly modeled surfacedeposition and troposphere–stratosphere exchangein producing this discrepancy. Mesoscale 3Dmodels, which couple the dynamics of convectionwith cloud microphysics, and fast photochemistryhave also been tested against observations; e.g.,Wang and Prinn (2000) concluded that NOx

produced from lightning can deplete O3 in thetropical Pacific upper troposphere as is oftenobserved, but there were difficulties in simulatingthe observed high O3 concentrations in the middletroposphere in these regions.Underpinning these tests of our understanding

are a large number of observational experimentsusing surface, aircraft, balloon and satellite plat-forms, and ranging from short-term campaigns tomultidecadal measurement networks. Lelieveldet al. (1999) and Ehhalt (1999) have reviewedsome of the relevant experiments focused ontropospheric oxidants. They noted the advancesmade in direct NOx and HOx observations fortesting photochemical models. Thompson et al.(2003) have reviewed the gaps in tropospheric O3

observations and have presented an important newdata set for the southern hemisphere tropics thatwill provide new tests of chemical models.

4.01.8 CONCLUSIONS

Oxidation processes have played amajor role inthe evolution of Earth’s atmosphere. Observationsof trace gases and free radicals in the atmospherein 1978–2003, and of chemicals in ice coresrecording the composition of past atmospheres,are providing fundamental information aboutthese processes. Also, basic laboratory studies ofchemical kinetics, while not reviewed here, haveplayed an essential role in defining mechanismsand rates. Models have been developed for fastphotochemistry and for coupled chemical andtransport processes that encapsulate current labo-ratory and theoretical understanding and helpexplain some of these atmospheric observations.


But there are important discrepancies betweenmodels and observations for OH and O3 (bothlocally and globally) that still need to be resolved.Looking into the twenty-first century, it is clear

that there is urgent need to gain more observationsto improve our understanding. We need muchmore extensive global 3D distributions for manytrace gas species (O3, CO, NOx, reactive hydro-carbons, H2O2, etc.) and many more measure-ments of key reactive free radicals (OH, HO2, etc.)to improve and ultimately validate models of theatmosphere. More attention needs to be given, inboth modeling and observational programs, to therole of aerosols and clouds in governing UVfluxes and heterogeneous chemical reactions.Refinement of existing methods for determininglong-term trends in OH, along with developmentof new independent approaches is clearlyimportant.Innovative methods for discerning the compo-

sition of past atmospheres are needed, since suchinformation both tests current understanding andhelps elucidate the process of atmospheric evol-ution. Finally, human and natural biosphericactivities serve as the primary sources of bothoxidant precursors and oxidant sink molecules,and the past, present, and possible future trendsin these sources need to be much betterquantified if we are to fully understand theoxidation processes in our atmosphere.

REFERENCES

Anklin M. and Bales R. C. (1997) Recent increases in H2O2

concentrations at Summit, Greenland. J. Geophys. Res. 102,19099–19104.

Berntsen T. K., Myhre G., Stordal F., and Isaksen I. S. A.(2000) Time evolution of tropospheric ozone and itsradiative forcing. J. Geophys. Res. 105, 8915–8930.

Brandenburger U., Brauers T., Dorn H.-P., Hausmann M., andEhhalt D. H. (1998) In-situ measurement of tropospherichydroxyl radicals by folded long-path laser absorptionduring the field campaign POPCORN in 1994. J. Atmos.Chem. 31, 181–204.

Brasseur G. P., Orlando J. J., and Tyndall G. S. (1999)Atmospheric Chemistry and Global Change. Oxford Uni-versity Press, Oxford.

Brune W. H., Stevens P. S., and Mather J. H. (1995)MeasuringOH and HO2 in the troposphere by laser-induced fluor-escence at low pressure. J. Atmos. Sci. 52, 3328–3336.

ChameidesW. L. and Walker J. C. G. (1973) A photochemicaltheory of tropospheric ozone. J. Geophys. Res. 78,8751–8760.

Chameides W. L. and Walker J. C. G. (1981) Rates of fixationby lightning of carbon and nitrogen in possible primitiveatmospheres. Origins of Life 11, 291–302.

Chen X. and Mopper K. (2000) Determination of tropospherichydroxyl radical by liquidphase scrubbing and HPLC:preliminary results. J. Atmos. Chem. 36, 81–105.

Crosley D. R. (1995) The measurement of OH and HO2 in theatmosphere. J. Atmos. Sci. 52, 3299–3314.

Crutzen P. J. (1979) The role of NO and NO2 in the chemistryof the troposphere and stratosphere. Ann. Rev. Earth Planet.Sci. 7, 443–472.

Crutzen P. J. and Zimmerman P. H. (1991) The changingphotochemistry of the trophosphere. Tellus 43AB, 136–151.

Dorn H.-P., Brandenburger U., Brauers T., and Hausmann M.(1995) A new in situ long-path absorption instrument for themeasurement of tropospheric OH radicals. J. Atmos. Sci. 52,3373–3380.

Ehhalt D. H. (1999) Gas phase chemistry of the troposphere.In Global Aspects of Atmospheric Chemistry, Topics inPhysical Chemistry (ed. R. Zellner). Springer, Darmstadt,vol. 6, pp. 21–109.

Ehhalt D. H., Dorn H., and Poppe D. (1991) The chemistry ofthe hydroxyl radical in the troposphere. Proc. Roy. Soc.Edinburgh 97, 17–34.

Eisele F. L. and Tanner D. J. (1991) Ion-assisted troposphericOH measurements. J. Geophys. Res. 96, 9295–9308.

Eisele F. L., Tanner D. H., Cantrell C. A., and Calvert J. G.(1996) Measurements and steady state calculations of OHconcentrations at Mauna Loa Observatory. J. Geophys. Res.101, 14665–14679.

Fegley B., Prinn R. G., Hartman H., and Watkins G. H. (1986)Chemical effects of large impacts on the Earth’s primitiveatmosphere. Nature 319, 305–308.

Felton C. C., Sheppard J. C., and Campbell M. J. (1990) Theradiochemical hydroxyl radical measurement method.Environ. Sci. Technol. 24, 1841–1847.

Folland C. K., Karl T. R., Christy J. R., Clarke R. A., GruzaG. V., Jouzel J., Mann M. E., Oerlemans J., Salinger M. J.,and Wang S. W. (2001) Observed climate variabilityand change. In Climate Change 2001: The Scientific Basis(ed. J. T. Houghton), Cambridge University Press,New York, pp. 99–181.

Haan D., Martinerie P., and Raynaud D. (1996) Ice core data ofatmospheric carbonmonoxide overAntarctica andGreenlandduring the last 200 years.Geophys. Res. Lett. 23, 2235–2238.

Hard T.M., O’Brien R. J., Chan C. Y., and Mehrabzadeh A. A.(1984) Tropospheric free radical determination by FAGE.Environ. Sci. Technol. 18, 768–777.

Hard T. M., George L. A., and O’Brien R. J. (1995) FAGEDetermination of Tropospheric HO and HO2. J. Atmos. Sci.52, 3354–3372.

Hofzumahaus A., Aschmutat U., Hebling M., Holland F., andEhhalt D. (1996) The measurement of tropospheric OHradicals by laser-induced fluorescence spectroscopy duringthe POPCORN field campaign. Geophys. Res. Lett. 23,2541–2544.

Holland F., Hessling M., and Hofzumahaus A. (1995) In-situmeasurement of tropospheric OH radicals by laser-inducedfluorescence.A description of the KFA instrument. J. Atmos.Sci. 52, 3393–3401.

Holland H. D. (1984) The Chemical Evolution of theAtmosphere and Oceans. Princeton University Press,Princeton.

Hough A. M. and Derwent R. G. (1990) Changes in the globalconcentration of tropospheric ozone due to human activities.Nature 344, 645–648.

Houweling S., Dentener F., and Lelieveld J. (2000) Themodeling of tropospheric methane: how well can pointmeasurements be reproduced by a global model? J.Geophys.Res. 105, 8981–9002.

Huang J. and Prinn R. (2002) Critical evaluation of emissionsfor potential gases for OH estimation. J. Geophys. Res. 107,D24, 4784, doi: 10.1029/2002 JD002394.

KrolM., van Leeuwen P. J., and Leileveld J. (1998)Global OHtrend inferred from methyl chloroform measurements.J. Geophys. Res. 103, 10697–10711.

Krol M., van Leeuwen P. J., and Lelieveld J. (2001) Reply tocomment by Prinn and Huang (2001) on Krol et al. (1998)J. Geophys. Res. 106, 23158–23168.

Lange M. A. and Ahrens T. J. (1982) The evolution of animpact-generated atmosphere. Icarus 51, 96–120.

Lelieveld J., Thompson A. M., Diab R. D., Hov O., Kley D.,Logan J. A., Nielson O. J., Stockwell W. R., and Zhou X.(1999) Tropospheric ozone and related processes.

References 17

In Scientific Assessment of Ozone Depletion: 1998 (ed. D.Albritton). World Meteorological Organization, Geneva,pp. 8.1–8.42.

Levine J. S. (1985) The photochemistry of the earlyatmosphere. In The Photochemistry of Atmospheres (ed.J. S. Levine). Academic Press, Orlando, pp. 3–38.

Levy H. (1971) Normal atmosphere: large radical andformaldehyde concentrations predicted. Science 173,141–143.

LevyH., Kasibhatla P. S., MoximW. J., Klonecki A., HirschA.,Oltmans S., and Chameides W. L. (1997) The global impactof human activity on tropospheric ozone.Geophys. Res. Lett.24, 791–794.

Lewis J. and Prinn R. (1984) Planets and their Atmospheres:Origin and Evolution. Academic Press, Orlando.

Logan J., Prather M., Wofsy S., and McElroy M. (1981)Tropospheric chemistry: a global perspective. J. Geophys.Res. 86, 7210–7254.

Madronich S. and Granier C. (1992) Impact of recent totalozone changes on tropospheric ozone photodissociation,hydroxyl radicals, and methane trends. Geophys. Res. Lett.19, 465–467.

Mak J. E., Brenninkmeijer C. A. M., and Tamaresis J. (1994)Atmospheric 14CO observations and their use for estimatingcarbon monoxide removal rates. J. Geophys. Res. 99,22915–22922.

Marenco A., Gouget H., Nedelec P., and Pages J. P. (1994)Evidence of long-term increase in tropospheric ozonefrom Pic du Midi data series, consequences and positiveradiative forcing. J. Geophys. Res. 99, 16617–16632.

McCulloch A. and Midgley P. (2001) The history of methylchloroform emissions: 1951–2000. Atmos. Environ. 35,5311–5319.

Miller B. R., Weiss R. F., Prinn R.G., Huang J., and Fraser P. J.(1998) Atmospheric trend and lifetime of chlorodifluoro-methane (HCFC-22) and the global tropospheric OHconcentration. J. Geophys. Res. 103, 13237–13248.

Montzka S. A., Spivakovsky C. M., Butler J. H., Elkins J. W.,Lock L. T., and Mondecl D. J. (2000) New observationalconstraints on atmospheric hydroxyl on global and hemi-spheric scales. Science 288, 500–503.

Mount G. H. (1992) The measurement of tropospheric OH bylong-path absorption: 1. Instrumentation. J. Geophys. Res.97, 2427–2444.

Mount G. H. and Harder J. W. (1995) The measurement oftropospheric trace gases at Fritz Peak Observatory by long-path absorption: OH and ancillary gases. J. Atmos. Sci. 52,3342–3353.

Penner J., Andreae M., Annegarn H., Barrie L., Feichter J.,Hegg D., Jayaraman A., Leaitch R., Murphy D., Nganga J.,and Pitari G. (2001) Aerosols: their direct and indirecteffects. In Climate Change 2001: The Scientific Basis (ed.J. T. Houghton). Cambridge University Press, New York,pp. 289–348.

Poppe D., Zimmerman J., Bauer R., Brauers T., Bruning D.,Callies J., Dorn H. P., Hofzumahaus A., Johnen F. J.,KhedimA., Koch H., Koppman R., LondonH., MullerK. P.,Neuroth R., Plass-Dulmer C., Platt U., Rohrer F., Roth E. P.,Rudolph J., Schmidt U., Wallasch M., and Ehhalt D. H.(1994) Comparison of measured OH concentrations withmodel calculations. J. Geophys. Res. 99, 16633–16642.

Poppe D., Zimmerman J., and Dorn H. P. (1995) Field data andmodel calculations for the hydroxyl radical. J. Atmos. Sci.52, 3402–3407.

Prather M., Ehhalt D., Dentener F., Derwent R., DlugokenckyE., Holland E., Isaksen I., Katima J., Kirchhoff V., MatsonP., Midgley P., and Wang M. (2001) Atmospheric chemistryand greenhouse gases. In Climate Change 2001: TheScientific Basis (ed. J. T. Houghton). Cambridge UniversityPress, New York, pp. 239–288.

Prinn R. (1982)Origin and evolution of planetary atmospheres:an introduction to the problem. Planet. Space Sci. 30,741–753.

Prinn R. (1994) The interactive atmosphere: global atmos-pheric–biospheric chemistry. Ambio 23, 50–61.

Prinn R. and Fegley B. (1987) Bolide impacts, acid rain, andbiospheric traumas at the Cretaceous–Tertiary boundary.Earth Planet. Sci. Lett. 83, 1–15.

Prinn R. and Huang J. (2001) Comment on Krol et al. (1998).J. Geophys. Res. 106, 23151–23157.

Prinn R., Rasmussen R., Simmonds P., Alyea F., Cunnold D.,Lane B., Cardelino C., and Crawford A. (1983) Theatmospheric lifetime experiment 5. Results for CH3CCl3based on three years of data. J.Geophys. Res. 88, 8415–8426.

Prinn R., Cunnold D., Simmonds P., Alyea F., Boldi R.,Crawford A., Fraser P., GutzlerD., HartleyD., RosenR., andRasmussen R. (1992) Global average concentration andtrend for hydroxyl radicals deduced from ALE/GAGEtrichloroethane (methyl chloroform) data for 1978–1990.J. Geophys. Res. 97, 2445–2461.

Prinn R., Jacoby H., Sokolov A., Wang C., Xiao X., Yang Z.,EckhausR., Stone P., EllermanD., Melillo J., Fitzmaurice J.,Kicklighter D., Holian G., and Liu Y. (1999) Integratedglobal system model for climate policy assessment:feedbacks and sensitivity studies. Climat. Change 41,469–546.

Prinn R., Weiss R., Fraser P., Simmonds P., Cunnold D., AlyeaF., O’Doherty S., Salameh P., Miller B., Huang J., Wang R.,Hartley D., Harth C., Steele L., Sturrock G., Midgley P., andMcColloch A. (2000)A history of chemically and radiativelyimportant gases in air deduced from ALE/GAGE/AGAGE.J. Geophys. Res. 105, 17751–17792.

Prinn R. G. (2000) Measurement equation for trace chemicalsin fluids and solution of its inverse. In Inverse Methods inGlobal Biogeochemical Cycles Geophysical Monograph(ed. P. Kasibhatla). American Geophysical Union,Washington, DC, vol. 114, pp. 3–18.

Prinn R. G., Cunnold D. M., Rasmussen R., Simmonds P. G.,Alyea F. N., Crawford A., Fraser P. J., and Rosen R. (1987)Atmospheric trends in methylchloroform and the globalaverage for the hydroxyl radical. Science 238, 945–950.

Prinn R. G., Weiss R. F., Miller B. R., Huang J., Alyea F. N.,Cunnold D. M., Fraser P. J., Hartley D. E., and SimmondsP. G. (1995) Atmospheric trends and lifetime of CH3CCL3

and global OH concentrations. Science 269, 187–192.Prinn R. G., Huang J., Weiss R. F., Cunnold D.M., Fraser P. J.,Simmonds P. G., McCulloch A., Harth C., Salameh P.,O’Doherty S., Wang R. H. J., Porter L., and Miller B. R.(2001) Evidence for substantial variations of atmospherichydroxyl radicals in the past two decades. Science 292,1882–1888.

Quay P., King S., White D., Brockington M., Plotkin B.,Gammon R., Gerst S., and Stutsman J. (2000) Atmospheric14CO: a tracer of OH concentration and mixing rates.J. Geophys. Res. 105, 15147–15166.

Sigg A. and Neftel A. (1991) Evidence for a 50% increase inH2O2 over the past 200 years from a Greenland ice core.Nature 351, 557–559.

Simmonds P.G., O’Doherty S., Huang J., PrinnR., DerwentR.,Ryall D., Nickless G., and Cunnold D. (1998) Calculatedtrends and the atmospheric abundance of 1,1,1,2-tetrafluor-oethane, 1,1-dichloro-1-fluoroethane, and 1-chloro-1,1-difluorethane using automated in situ gas chromato-graphy-mass spectrometry measurements recorded at MaceHead, Ireland, from October 1994 to March 1997.J. Geophys. Res. 103, 16029–16037.

Spivakovsky C. M., Logan J. A., Montzka S. A., BalkanskiY. J., Foreman-Fowler M., Jones D. B. A., Horowitz L. W.,Fusco A. C., Brenninkmeijer C. A. M., PratherM. J., WofsyS. C., and McElroy M. B. (2000) Three dimensionalclimatological distribution of tropospheric OH: update andevaluation. J. Geophys. Res. 105, 8931–8980.

Staffelbach T. A., Neftel A., Stauffer B., and Jacob D. J. (1991)Formaldehyde in polar ice cores: a possibility to characterizethe atmospheric sink of methane in the past? Nature 349,603–605.


Stimpfle R. M. and Anderson J. G. (1988) In situ detection ofOH in the lower stratosphere with a balloon borne highrepetition rate laser system. Geophys. Res. Lett. 15,1503–1506.

Tanner D. J., Jefferson A., and Eisele F. L. (1997) Selected ionchemical ionization mass spectrometric measurement ofOH. J. Geophys. Res. 102, 6415–6425.

Thompson A. M. (1992) The oxidizing capacity of the Earth’satmosphere: probable past and future changes. Science 256,1157–1165.

Thompson A. M., Witte J. C., Mc Peters R. D., Oltmans S. J.,Schmidlin F. J., Logan J. A., Fujiwara M., Kirchoff V.,Posny F., Coetzee G., Hoegger B., Kawakami S., Ogawa T.,Johnson B., VomelH., and LabowG. (2003) The 1998–2000SHADOZ Tropical Ozone Climatology: comparison withTOMS and ground-based measurements. J. Geophys. Rev.108, D2, 8238, doi: 10.1029/2001 JD000967.

Volz A. and Kley D. (1988) Evaluation of the MontsourisSeries of ozone measurements made in the nineteenthcentury. Nature 332, 240–242.

Volz A., Ehhalt D. H., and Derwent R. G. (1981) Seasonaland latitudinal variation of 14CO and the troposphericconcentration of OH radicals. J. Geophys. Res. 86,5163–5171.

Wang C. and Prinn R. G. (1999) Impact of emissions,chemistry and climate on atmospheric carbon monoxide:100-year predictions from a global chemistry model.Chemosph. Global Change 1, 73–81.

Wang C. and Prinn R. G. (2000) On the roles of deepconvective clouds in tropospheric chemistry. J. Geophys.Res. 105, 22269–22297.

Wang Y. and Jacob D. J. (1998) Anthropogenic forcing ontropospheric ozone and OH since pre-industrial times.J. Geophys. Res. 103, 31123–31135.

q 2003, Elsevier Ltd. All rights reservedNo part of this publication may be reproduced, stored in a retrieval system ortransmitted in any form or by any means, electronic, mechanical, photocopying,recording or otherwise, without prior written permission of the Publisher.

Treatise On GeochemistryISBN (set): 0-08-043751-6

Volume 4; (ISBN: 0-08-044339-7); pp. 1–19

References 19

Ozon ydroxyl Radical,and Oxidativ pacity · 4.01 Ozone, Hydroxyl Radical,and OxidativeCapacity...

Documents

Transcript of Ozon ydroxyl Radical,and Oxidativ pacity · 4.01 Ozone, Hydroxyl Radical,and OxidativeCapacity...