The atmosphere is an oxidizing medium. Many environmentally important trace gases are removed from the atmosphere mainly by oxidation: greenhouse gases such as CH4, toxic combustion gases such as CO, agents for stratospheric O3 depletion such as HCFCs, and others. Oxidation in the troposphere is of key importance because the troposphere contains the bulk of atmospheric mass (85%, see section 2.3 ) and because gases are generally emitted at the surface.
The most abundant oxidants in the Earth's atmosphere are O2 and O3. These oxidants have large bond energies and are hence relatively unreactive except toward radicals (O2 only toward highly unstable radicals). With a few exceptions, oxidation of non-radical atmospheric species by O2 or O3 is negligibly slow. Work in the 1950s first identified the OH radical as a strong oxidant in the stratosphere. OH reacts rapidly with most reduced non-radical species, and is particularly reactive toward H-containing molecules due to H-abstraction reactions converting OH to H2O. Production of OH is by reaction of water vapor with O(1D) ( section 10.2.1 ):
We saw in chapter 10 how OH oxidizes a number of trace gases in the stratosphere. A simple expression for the source POH of OH from reactions (R1) - (R3) can be obtained by assuming steady state for O(1D). Laboratory studies show that (R2) is much faster than (R3) at the H2O mixing ratios found in the atmosphere, allowing for simplification:
Critical to the generation of OH is the production of O(1D) atoms by (R1) . Until 1970 it was assumed that production of O(1D) would be negligible in the troposphere because of near-total absorption of UV radiation by the O3 column overhead. It was thought that oxidation of species emitted from the Earth's surface, such as CO and CH4, required transport to the stratosphere followed by reaction with OH in the stratosphere:
This mechanism implied long atmospheric lifetimes for CO and CH4 because air takes on average 5-10 years to travel from the troposphere to the stratosphere ( section 4.4.4 ) and the stratosphere accounts for only 15% of total atmospheric mass. In the 1960s, concern emerged that accumulation of CO emitted by fossil fuel combustion would soon represent a global air pollution problem.
A major discovery in the early 1970s was that sufficient OH is in fact produced in the troposphere by reactions (R1) - (R3) to allow for oxidation of species such as CO and CH4 within the troposphere. A calculation of the rate constant for (R1) at sea level is shown in Figure 11-1 as the product of the solar actinic flux, the absorption cross-section for O3, and the O(1D) quantum yield. Tropospheric production of O(1D) takes place in a narrow wavelength band between 300 and 320 nm; radiation of shorter wavelengths does not penetrate into the troposphere, while radiation of longer wavelengths is not absorbed by O3. Although the production of O(1D) in the troposphere is considerably slower than in the stratosphere, this is compensated in terms of OH production by the larger H2O mixing ratios in the troposphere (102-103 times higher than in the stratosphere). Model calculations in the 1970s accounting for the penetration of UV radiation at 300-320 nm found tropospheric OH concentrations of the order of 106 molecules cm-3, resulting in a tropospheric lifetime for CO of only a few months and allaying concerns that CO could accumulate to toxic levels. Crude measurements of OH concentrations in the 1970s confirmed this order of magnitude and hence the importance of OH as an oxidant in the troposphere; further confirmation came from long-lived proxies ( section 11.1.2 ). The accurate measurement of OH turned out to be an extremely difficult problem because of the low concentrations, and only in the past decade have instruments been developed that can claim an accuracy of better than 50%.
Figure 11-1 . Computation of the rate constant k1 for photolysis of O3 to O(1D) in the troposphere as a function of wavelength. (1) Solar actinic flux at sea level for 30o solar zenith angle and a typical O3 column overhead; (2) Absorption cross-section of O3 at 273 K; (3) O(1D) quantum yield at 273 K; and (4) rate constant k1 calculated as the product of (1), (2), and (3).
where ni is the number density of species i reacting with OH, ki is the corresponding rate constant, and the sum is over all reactants in the air parcel. One finds that CO is the dominant sink of OH in most of the troposphere, and that CH4 is next in importance. The resulting OH lifetime is of the order of one second. Because of this short lifetime, atmospheric concentrations of OH are highly variable; they respond rapidly to changes in the sources or sinks.
Calculating the atmospheric lifetimes of gases against oxidation by OH requires a knowledge of OH concentrations averaged appropriately over time and space. This averaging cannot be done from direct OH measurements because OH concentrations are so variable. An impossibly dense measurement network would be required.
In the late 1970s it was discovered that the industrial solvent methylchloroform (CH3CCl3) could be used to estimate the global mean OH concentration. The source of CH3CCl3 to the atmosphere is exclusively anthropogenic. The main sink is oxidation by OH in the troposphere (oxidation and photolysis in the stratosphere, and uptake by the oceans, provide small additional sinks). The concentration of CH3CCl3 in surface air has been measured continuously since 1978 at a worldwide network of sites ( Figure 11-2 ). Rapid increase of CH3CCl3 was observed in the 1970s and 1980s due to rising industrial emissions, but concentrations began to decline in the 1990s because CH3CCl3 was one of the gases banned by the Montreal protocol to protect the O3 layer. Although only a small fraction of CH3CCl3 is oxidized or photolyzed in the stratosphere, the resulting Cl radical source was sufficient to motivate the ban.
Industry statistics provide a reliable historical record of the global production rate P (moles yr-1) of CH3CCl3, and it is well-established that essentially all of this production is volatilized to the atmosphere within a few years. The global mass balance equation for CH3CCl3 in the troposphere is:
where N is the number of moles of CH3CCl3 in the troposphere, Ltrop is the loss rate of CH3CCl3 in the troposphere, and Lstrat and Locean are the minor loss rates of CH3CCl3 in the stratosphere and to the ocean. We calculate Ltrop as
where k(T) is the temperature-dependent rate constant for the oxidation of CH3CCl3 by OH, C is the mixing ratio of CH3CCl3, na is the air density, and the integral is over the tropospheric volume. We define the global mean OH concentration in the troposphere as
where we have assumed C to be uniform in the troposphere ( Figure 11-2 ) and neglected the minor terms Lstrat and Locean. All terms on the right-hand side of (11.6) are known. The values of C and dN/dt can be inferred from atmospheric observations ( Figure 11-2 ). The integral can be calculated from laboratory measurements of k(T) and climatological data for tropospheric temperatures. Substituting numerical values we obtain [OH] = 1.2x106 molecules cm-3.
This empirical estimate of [OH] is useful because it can be used to estimate the lifetime ti = 1/(ki[OH]) of any long-lived gas i against oxidation by OH in the troposphere. For example, one infers a lifetime of 9 years for CH4 and a lifetime of 2.0 years for CH3Br ( See Methyl bromide ). One can also determine the atmospheric lifetimes of different hydrochlorofluorocarbon (HCFC) species and hence the fractions of these species that penetrate into the stratosphere to destroy O3 ( problem 3. 3 ).
Carbon monoxide and methane are the principal sinks for OH in most of the troposphere. These two gases play therefore a critical role in controlling OH concentrations and more generally in driving radical chemistry in the troposphere.
See Present-day global budget of CO gives a global budget of CO for the present-day atmosphere. Fossil fuel combustion and biomass burning (principally associated with tropical agriculture) are large anthropogenic sources, and oxidation of CH4 is another major source ( See Sources of CO ). Most of the CO in the present-day troposphere is anthropogenic. The main sink of CO is oxidation by OH and results in a 2-month mean lifetime; because of this relatively short lifetime, CO is not well-mixed in the troposphere. Concentrations are 50-150 ppbv in remote parts of the world, 100-300 ppbv in rural regions of the United States, and up to several ppmv in urban areas where CO is considered a hazard to human health.
Atmospheric concentrations of CH4 have increased from 800 to 1700 ppbv since preindustrial times ( Figure 7-1 ). The reasons are not well understood. A present-day global budget for CH4 is given in See Present-day global budget of CH4 . There are a number of anthropogenic sources, some combination of which could have accounted for the observed CH4 increase. One must also consider the possible role of changing OH concentrations. Oxidation by OH in the troposphere provides 85% of the global CH4 sink (uptake by soils and oxidation in the stratosphere provide small additional sinks; see problem 4. 8 ). A decrease in OH concentrations since pre-industrial times would also have caused CH4 concentrations to increase. Long-term trends in OH concentrations will be discussed in section 11.5 .
In the early 1970s when the importance of OH as a tropospheric oxidant was first realized, it was thought that the O3 molecules necessary for OH production would be supplied by transport from the stratosphere. As we saw in section 10.1.2 , the chemical lifetime of O3 in the lower stratosphere is several years, sufficiently long to allow transport of O3 to the troposphere. The transport rate F of O3 across the tropopause is estimated to be in the range 1-2x1013 moles yr-1 ( section 11.5 and problem 11. 2 ). One can make a simple argument that this supply of O3 from the stratosphere is in fact far from sufficient to maintain tropospheric OH levels. Each O3 molecule crossing the tropopause can yield at most two OH molecules in the troposphere by reactions (R1) + (R3) (some of the O3 is consumed by other reactions in the troposphere, and some is deposited at the Earth's surface). The resulting maximum source of OH is 2F = 2-4x1013 moles yr-1. In comparison, the global source of CO to the atmosphere is 6-10x1013 moles yr-1 ( See Present-day global budget of CO ) and the global source of CH4 is about 3x1013 moles yr-1 ( See Present-day global budget of CH4 ). There are therefore more molecules of CO and CH4 emitted to the atmosphere each year than can be oxidized by OH molecules originating from O3 transported across the tropopause. In the absence of additional sources OH would be titrated; CO, CH4, HCFCs, and other gases would accumulate to very high levels in the troposphere, with catastrophic environmental implications.
A key factor preventing this catastrophe is the presence in the troposphere of trace levels of NOx (NOx NO + NO2) originating from combustion, lightning, and soils. The sources and sinks of NOx will be discussed in Section 11.4 . As we first show here, the presence of NOx allows the regeneration of OH consumed in the oxidation of CO and hydrocarbons, and concurrently provides a major source of O3 in the troposphere to generate additional OH.
which regenerates OH, and also produces NO2 which goes on to photolyze as we have already seen for the stratosphere ( section 10.2.2 ):
Reaction (R11) regenerates NO and produces an O3 molecule, which can then go on to photolyze by reactions (R1) - (R3) to produce two additional OH molecules. Reaction (R10) thus yields up to three OH molecules, boosting the oxidizing power of the atmosphere. The sequence of reactions (R4) + (R6) + (R10) + (R11) is a chain mechanism for O3 production in which the oxidation of CO by O2 is catalyzed by the HOx chemical family (HOx H + OH + HO2) and by NOx:
The chain is initiated by the source of HOx from reaction (R3) , and is terminated by the loss of the HOx radicals through (R7) . The propagation efficiency of the chain (chain length) is determined by the abundance of NOx. A diagram of the mechanism emphasizing the coupling between the O3, HOx, and NOx cycles is shown in Figure 11-3 .
Remarkably, HOx and NOx catalyze O3 production in the troposphere and O3 destruction in the stratosphere. Recall the catalytic HOx and NOx cycles for O3 loss in the stratosphere ( section 10.2 ):
The key difference between the troposphere and the stratosphere is that O3 and O concentrations are much lower in the troposphere. The difference is particularly large for O atom, whose concentrations vary as [O3]/na2 (equation (10.4) ). In the troposphere, reaction (R12) is much slower than reaction (R4) , and reaction (R15) is negligibly slow. Ozone loss by the HOx-catalyzed mechanism (R12) - (R13) can still be important in remote regions of the troposphere where NO concentrations are sufficiently low for (R13) to compete with (R10) . Ozone loss by the NOx-catalyzed mechanism (R14) - (R15) is of no importance anywhere in the troposphere.
Methylhydroperoxide (CH3OOH) may either react with OH or photolyze. The reaction with OH has two branches because the H-abstraction can take place either at the methyl or at the hydroperoxy group. The CH2OOH radical produced in the first branch decomposes rapidly to formaldehyde (CH2O) and OH:
and HO2 reacts further as described in section 11.3.2 .
Formaldehyde produced by (R22) can either react with OH or photolyze (two photolysis branches):
CO is then oxidized to CO2 by the mechanism described in section 11.3.2 .
In this overall reaction sequence the C(-IV) atom in CH4 (the lowest oxidation state for carbon) is successively oxidized to C(-II) in CH3OOH, C(0) in CH2O, C(+II) in CO, and C(+IV) in CO2 (highest oxidation state for carbon). Ozone production takes place by NO2 photolysis following the peroxy + NO reactions (R10) and (R18) , where the peroxy radicals are generated by reactions (R5) + (R16) , (R20) , (R22) , (R24) , (R26) , and (R4) + (R6) .
Let us calculate the O3 and HOx yields from the oxidation of CH4 in two extreme cases. Consider first a situation where CH3O2 and HO2 react only by (R18) and (R10) respectively (high-NOx regime) and CH2O is removed solely by (R24) . By summing all reactions in the mechanism we arrive at the following net reaction for conversion of CH4 to CO2:
with an overall yield of five O3 molecules and two HOx molecules per molecule of CH4 oxidized. Similarly to CO, the oxidation of CH4 in this high-NOx case is a chain mechanism for O3 production where HOx and NOx serve as catalysts. Reaction (R24) , which provides the extra source of HOx as part of the propagation sequence, branches the chain ( section 9.4 ).
In contrast, consider an atmosphere devoid of NOx so that CH3O2 reacts by (R17) ; further assume that CH3OOH reacts by (R19) and CH2O reacts by (R23) . Summing all reactions in the mechanism yields the net reaction:
Oxidation of larger hydrocarbons follows the same type of chain mechanism as for CH4. Complications arise over the multiple fates of the organic peroxy (RO2) and oxy (RO) radicals, as well as over the structure and fate of the carbonyl compounds and other oxygenated organics produced as intermediates in the oxidation chain. These larger hydrocarbons have smaller global sources than CH4 and are therefore less important than CH4 for global tropospheric chemistry. They are however critical for rapid production of O3 in polluted regions, as we will see in chapter 12, and play also an important role in the long-range transport of NOx, as discussed below.
We now turn to an analysis of the factors controlling NOx concentrations in the troposphere. Estimated tropospheric sources of NOx for present-day conditions are shown in See Estimated present-day sources of tropospheric NOx . Fossil fuel combustion accounts for about half of the global source. Biomass burning, mostly from tropical agriculture and deforestation, accounts for another 25%. Part of the combustion source is due to oxidation of the organic nitrogen present in the fuel. An additional source in combustion engines is the thermal decomposition of air supplied to the combustion chamber. At the high temperatures of the combustion chamber (~ 2000 K), oxygen thermolyzes and subsequent reaction of O with N2 produces NO:
The equilibria (R27) - (R29) are shifted to the right at high temperatures, promoting NO formation. The same thermal mechanism also leads to NO emission from lightning, as the air inside the lightning channel is heated to extremely high temperatures. Other minor sources of NOx in See Estimated present-day sources of tropospheric NOx include microbial nitrification and denitrification in soils ( Section 6.3 ), oxidation of NH3 emitted by the biosphere, and transport from the stratosphere of NOy produced by oxidation of N2O by O(1D). Oxidation of N2O does not take place in the troposphere itself because concentrations of O(1D) are too low.
Because of this rapid cycling, it is most appropriate to consider the budget of the NOx family as a whole, as in the stratosphere ( section 10.2.2 ). At night, NOx is present exclusively as NO2 as a result of (R14) .
Human activity is clearly a major source of NOx in the troposphere, but quantifying the global extent of human influence on NOx concentrations is difficult because the lifetime of NOx is short. The principal sink of NOx is oxidation to HNO3, as in the stratosphere; in the daytime,
The resulting lifetime of NOx is approximately one day. In the stratosphere, we saw that HNO3 is recycled back to NOx by photolysis and reaction with OH on a time scale of a few weeks. In the troposphere, however, HNO3 is scavenged by precipitation because of its high solubility in water. The lifetime of water-soluble species against deposition is typically a few days in the lower troposphere and a few weeks in the upper troposphere ( problem 8. 1 ). We conclude that HNO3 in the troposphere is removed principally by deposition and is not an effective reservoir for NOx.
Research over the past decade has shown that a more efficient mechanism for long-range transport of anthropogenic NOx to the global troposphere is through the formation of another reservoir species, peroxyacetylnitrate (CH3C(O)OONO2). Peroxyacetylnitrate (called PAN for short) is produced in the troposphere by photochemical oxidation of carbonyl compounds in the presence of NOx. These carbonyls are produced by photochemical oxidation of hydrocarbons emitted from a variety of biogenic and anthropogenic sources. In the simplest case of acetaldehyde (CH3CHO), the formation of PAN proceeds by:
Formation of PAN is generally less important as a sink for NOx than formation of HNO3. However, in contrast to HNO3, PAN is only sparingly soluble in water and is not removed by deposition. Its principal loss is by thermal decomposition, regenerating NOx:
The lifetime of PAN against (R37) is 1 hour at 295 K and several months at 250 K; note the strong dependence on temperature. In the lower troposphere, NOx and PAN are typically near chemical equilibrium. In the middle and upper troposphere, however, PAN can be transported over long distances and decompose to release NOx far from its source, as illustrated in Figure 11-4 .
Measurements of PAN and NOx concentrations in the remote troposphere over the past decade support the view that long-range transport of PAN at high altitude plays a critical role in allowing anthropogenic sources to affect tropospheric NOx (and hence O3 and OH) on a global scale. Although PAN is only one of many organic nitrates produced during the oxidation of hydrocarbons in the presence of NOx, it seems to be by far the most important as a NOx reservoir. Other organic nitrates either are not produced at sufficiently high rates or do not have sufficiently long lifetimes.
Tropospheric ozone is the precursor of OH by (R1) - (R3) and plays therefore a key role in maintaining the oxidizing power of the troposphere. It is also of environmental importance as a greenhouse gas (chapter 7) and as a toxic pollutant in surface air (chapter 12). We saw in section 11.3 that O3 is supplied to the troposphere by transport from the stratosphere, and is also produced within the troposphere by cycling of NOx involving reactions of peroxy radicals with NO:
The reactions of NO with peroxy radicals (R10) - (R18) , driving O3 production, compete with the reaction of NO with O3, driving the null cycle (R14) - (R11) . Reactions (R10) - (R18) represent therefore the rate-limiting step for O3 production, and the O3 production rate PO3 is given by
Other organic peroxy radicals RO2 produced from the oxidation of nonmethane hydrocarbons also contribute to O3 production but are less important than HO2 and CH3O2 except in continental regions with high hydrocarbon emissions (chapter 12).
Ozone is also consumed by reactions with HO2 and OH in remote regions of the troposphere ( section 11.3.2 ):
Global models of tropospheric chemistry which integrate HOx-NOx-CO-hydrocarbon chemical mechanisms in a 3-dimensional framework (chapter 5) have been used to estimate the importance of these different sources and sinks in the tropospheric O3 budget. See Present-day global budget of tropospheric ozone gives the range of results from the current generation of models. It is now fairly well established that the abundance of tropospheric O3 is largely controlled by chemical production and loss within the troposphere. Transport from the stratosphere and dry deposition are relatively minor terms.
Figure 11-5 shows the global mean distributions of NOx, CO, O3, and OH simulated with a 3-dimensional model of tropospheric chemistry for present-day conditions.
Figure 11-5 . Longitudinally averaged concentrations of NOx, CO, O3, and OH as a function of latitude and pressure computed with a global 3-dimensional model for the present-day atmosphere. Values are annual averages. Adapted from Wang, Y., and D.J. Jacob, J. Geophys. Res., in press.
Concentrations of NOx and CO are highest in the lower troposphere at northern midlatitudes, reflecting the large source from fossil fuel combustion. Lightning is also a major source of NOx in the upper troposphere. Recycling of NOx through PAN maintains NOx concentrations in the range of 10-50 pptv throughout the remore troposphere. Ozone concentrations generally increase with altitude, mainly because of the lack of chemical loss in the upper troposphere (water vapor and hence HOx concentrations are low). Higher O3 concentrations are found in the northern than in the southern hemisphere, reflecting the abundance of NOx. Concentrations of OH are highest in the tropics where water vapor and UV radiation are high, and peak in the middle troposphere because of opposite vertical trends of water vapor (decreasing with altitude) and UV radiation (increasing with altitude). Concentrations of OH tend to be higher in the northern than in the southern hemisphere because of higher O3 and NOx, stimulating OH production; this effect compensates for the faster loss of OH in the northern hemisphere due to elevated CO.
Figure 11-6 Relative enhancements of NOx, CO, O3, and OH concentrations from preindustrial times to present, as computed with a global model of tropospheric chemistry. Values are annual longitudinal averages plotted as a function of latitude and pressure. Adapted from Wang, Y., and D.J. Jacob, J. Geophys. Res., in press.
Figure 11-6 shows the relative enhancements of NOx, CO, O3, and OH computed with the same model from preindustrial times to today. The preindustrial simulation assumes no emission from fossil fuel combustion and a much reduced emission from biomass burning. Results suggest that anthropogenic emissions have increased NOx and CO concentrations in most of the troposphere by factors of 2-8 (NOx) and 3-4 (CO). Ozone concentrations have increased by 50-100% in most of the troposphere, the largest increases being at low altitudes in the northern hemisphere.
The anthropogenic influence on OH is more complicated. Increasing NOx and O3 act to increase OH, while increasing CO and hydrocarbons act to deplete OH ( section 11.3 ). Because CO and CH4 have longer lifetimes than NOx and O3, their anthropogenic enhancements are more evenly distributed in the troposphere. It is thus found in the model that the net effect of human activity is to increase OH in most of the lower troposphere and to decrease OH in the upper troposphere and in the remote southern hemisphere ( Figure 11-6 ). There is compensation on the global scale so that the global mean OH concentration as defined by (11.5) decreases by only 7% since preindustrial times (other models find decreases in the range 5-20%). The relative constancy of OH since preindustrial times is remarkable in view of the several-fold increases of NOx, CO, and CH4. There remain large uncertainties in these model analyses. From the CH3CCl3 observational record, which started in 1978, we do know that there has been no significant global change in OH concentrations for the past 20 years.