Sasha Madronich, National Center for Atmospheric Research Boulder, Colorado USA
ABSTRACT. The photochemistry of the troposphere is highly non-linear, and may be changing due to emissions of gases related to human activities. Increases in tropospheric ultraviolet (UV) radiation, due to stratospheric ozone depletion, may also perturb the troposphere.
The concern that air pollution may grow to the global scale (4 6371[2] km[2]) is real. There is good evidence that many tropospheric gases, active either chemically or radiatively, have grown in abundance since the pre-industrial era. Among the chemically reactive species, long term increases in CH4 and tropospheric O3 are well known (IPCC, 1990). For NOX, current athropogenic emissions are thought to be comparable to or larger than global natural sources (IPCC, 1990; WMO, 1991). The first major global concern is that an increase in these gases, some of which are strong infrared absorbers, may change the radiative balance of the atmosphere, possibly resulting in climate change. The combined radiative effect of all of these gases is about double that of carbon dioxide (CO2) alone. Additional climate perturbations may come from aerosols, formed when some compounds such as sulfur dioxide (SO2) and dimethyl sulfide (CH3-S-CH3, or DMS) react in the troposphere to form H2SO4, which can lead to the formation of sulfate aerosols. These aerosols can absorb and scatter light, and, if appropriate in size, can also function as cloud condensation nuclei (CCN), changing the size distribution of cloud droplets and affecting cloud reflectivities.
The second major global issue is the tropospheric "oxidizing" or "self-cleaning" capacity. Many compounds in the atmosphere are removed by reaction with the hydroxyl radical (OH) (Levy, 1971; Thompson, 1992 and refs. therein). OH is not available in infinite supply, and can be lost or at least reduced, if emissions associated with human activities exceed certain values. Should this happen, the lifetime of many compounds will increase, and ultimately their atmospheric concentrations will rise to higher levels. Tropospheric OH is made mostly from tropospheric O3, and both are sensitive to ultraviolet (UV) radiation, and therefore to the overhead ozone abundance. Thus, one consequence of stratospheric ozone depletion is the alteration of tropospheric chemistry. Ultimately, the chemistries of the troposphere and the stratosphere must be viewed as fully coupled, because the troposphere is also the first chemical filter for surface-emitted compounds on their way to the stratosphere.
Tropospheric chemistry is non-linear, it involves a large number of compounds emitted at the surface, and is complicated by interactions between different phases including gas, liquid, aerosol, and various surfaces. Large uncertainties remain that go beyond simple computing resource limitations. Among all the complications, a few of the basic features do appear to be understood, particularly with regards to the roles of OH, O3, CH4, and NOx, and are discussed below.
However, the source of the O atoms is different than in the stratosphere where the O atoms are made by the photodissociation of O2 at UV wavelengths < 240 nm. In the troposphere, only UV radiation with > 290 nm is available, due to essentially complete absorption of shorter wavelengths by O2 and O3 above the tropopause. Thus, an alternate source of O atoms is necessary to explain tropospheric ozone formation. Exhaustive search of plausible tropospheric reactions reveals only a few possibilities which yield O atoms, and among these only one appears to be important: the photodissociation of NO2,
This reaction however cannot be the complete story. While the O([3]P) produced by NO2 photolysis does indeed form ozone, the NO produced in the same reaction reacts rapidly with ozone,
Thus, the net ozone production is strongly limited. This is illustrated in Figure 1. Starting with an initial NO2 amount [NO2]o and zero nitric oxide and ozone, the ozone concentration at time t is
and is therefore always smaller than the initial NO2. In practice, however, observations show nearly always [O3] >> [NOx], so that an alternate way of producing the ozone is indicated. The alternative must still use NO2 photolysis as the primary source of O3, but the NO2 must be created without the loss of ozone that occurred in reaction 3. The necessary reactions are supplied by the oxidation of CO and various hydrocarbons.
The simple case of methane oxidation is illustrated in Table 1. Reactions 1-3 form ozone as discussed above. The real starting point is reaction 4, the UV photolysis of O3 leading to the formation of excited oxygen atoms, O([1]D). If these atoms are quenched (i .e., collisionally transfer their excitation and return to the ground state), as occurs with about 95% probability, ozone is simply generated through reaction 1 and no net change occurs. With some smaller probability, the O([1]D) atoms can encounter a water molecule (H2O) and produce two OH radicals. The OH radicals then initiate a sequence of reactions which oxidize methane first to formaldehyde (CH2O) (reactions 7-11), then to CO and ultimately CO2 (reactions 12-15). In the presence of NOx, the complete oxidation of CH4 to CO2 can lead to a total of five NO-->NO2 conversions, and also regenerates 2 OH radicals necessary to initiate the oxidation of other CH4 molecules. Each NO-->NO2 conversion which occurs without loss of ozone is effectively a net ozone production step.
The methane oxidation illustrated above is clearly a powerful agent for NO-->NO2 conversions, and may be expected to lead to extremely large tropospheric ozone amounts. However, other chemical reactions are in competition with the reactions listed in Table 1. Some of these may destroy ozone directly, i.e., the HOx (= OH + HO2) catalytic cycle,
Another important ozone loss is reaction 6, since it represents the net loss of ozone dissociated in reaction 4. Also, the effectiveness of the oxidation of CO and CH4 can be greatly reduced by termination reactions in which two radicals are lost,
and by reactions in which both radicals and NOx are lost,
The products H2O2, CH3O2H, and HNO3 should be viewed as reservoirs, rather than terminal products, since they can lead to radical (or NOx) regeneration, e.g.,
if they are not removed from the atmosphere by rainout or surface deposition.
The effects of CO and hydrocarbon oxidation on ozone production may be summarized as
where R-OO is intended to represent organic peroxy radicals (roughly analogous to CH3O2) derived from non-methane hydrocarbons. In all cases, the ozone production is dependent on the NO levels. Surprisingly small values of NO can offset the ozone losses. Considering for simplicity only reactions 11 and 16, ozone production is greater than loss when
which is about 5 ppt (1ppt = 10[-12] molar ratio = 2.69x10[7] molec cm[-3] at STP) given the values of k16 and k11, and typical tropospheric ozone values. The full dependence of ozone production and loss on NOx levels is shown in Figure 2. The importance of NOx is clear, and will be discussed further in the next section.
These temperatures are reached during most combustion processes (including fossil fuel and biomass burning) and lightning. Additional NOx sources may be associated with bacterial processes in soils.
Once in the atmosphere, NO and NO2 partake in many chemical reactions. Some of these are simple NO - NO2 interconversions, while others are actual NOx sinks. In particular, the reaction
removes NOx quickly, with about 1 day lifetime for typical mid-latitude conditions. The short NOx lifetime has one important implication: If the sources of NOx are not geographically uniform (which is certainly the case), the global NOx distributions will be highly variable, being very sensitive to both chemical and meteorological (transport) processes. Typical NOx values observed in different non-urban regions of the northern hemisphere are shown in Figure 3. NOx levels are seen to span about 3 orders of magnitude, and can be on either side of the ozone net -production threshold shown in Figure 2.
While most of the reactive nitrogen is emitted as NO or NO2, field data suggest that atmospheric reactive nitrogen may be present also in different forms, such as HNO3, PAN, particulate nitrate (NO3-), and organic nitrates, among others. The ratio of NOx to the total reactive nitrogen (NOy, measured by instruments which are sensitive to the total reactive nitrogen rather than to individual components), is shown in Figure 4. The general trend is that in relatively polluted regions, NOx is a high fraction of NOy, but decreases toward more remote regions. This is consistent with the NOx short lifetime and with conversions to other nitrogen species during several days of atmospheric transport. However, the budget of the NOy species is still incomplete. This is shown in Figure 5, which compares simultaneous but independent measurements of total reactive nitrogen (NOy) with the individually measured species NO, NO2, HNO3, PAN, and NO3-. Clearly there is a shortfall in the NOy which cannot be explained by any measured species. The magnitude of the shortfall can be greater than the NOx amounts, a disturbing fact in view of the potential role of even small amounts of NOx in global ozone production. The identity of the "mystery" NOy species is still uncertain.
The NMHCs confound tropospheric chemistry in several ways. First, near their sources, their strong reactivity may have an immediate effect on local O3, OH, and NOx levels. Second, because of their short lifetimes, the geographical distribution of the NMHCs will be highly variable and determined by complex couplings between chemistry and transport. Third, and perhaps most important for global considerations, the NMHCs are not converted immediately to their ultimate oxidation products CO2 and H2O; instead, intermediate compounds are produced after the initial OH attack, and may persist much longer in the atmosphere than the initial NMHC, possibly being transported over global distances. It has been suggested, for example, that these intermediate compounds may tie up at least some of the missing NOy species discussed in the previous section. These organic compounds, which may collectively be called Cy by analogy with NOy, may pervade the troposphere but, as of today, remain largely undetected.
Consider, specifically, a compound X whose atmospheric burden is determined by the balance between its emission fluxes (Fx), and its sinks which are inversely proportional to its atmospheric lifetime x. Its atmospheric concentration, [X], can rise for two reasons: (a) in response to increased fluxes, and (b) in response to an increase in x, which for many gases is given by the reaction with OH radicals. If the OH concentration, [OH], is in turn dependent on [X], a potentially powerful chemical feedback is possible: increases in Fx may increase the atmospheric concentrations of X, which in turn decrease [OH] and increase the lifetime of X, leading to further increases in [X]. The response of [OH] to current and future emissions is thus a major issue in tropospheric chemistry.
Current OH levels are illustrated in Figure 6, together with some other species of interest to the oxidizing capacity. OH is produced primarily from photolysis reactions, and is therefore nearly negligible at night, while noon-time values reach about 4x10[6] molec cm[-3] for the conditions shown in the Figure. Table 4 gives a detailed breakdown of the chemical reactions which contribute to the formation and destruction of OH in the remote free troposphere. It is seen that (for the 24-hr average), primary photolytic production accounts for about 60% of the OH radical production, while conversion of HO2 accounts for the remainder. OH loss is dominated by reaction with CO (49%) and CH4 (18%), with another 12% arising from intermediates formed by methane oxidation (CH3OOH, CH2O). The dependence of OH on NOx levels is fairly small for the low-NOx conditions shown in Table 4 (5-20 ppt NOx), but increases in more polluted regions. This is illustrated in Figure 7, which shows both OH and HO2 radicals as a function of NOx concentrations. If NOx is increased, OH is seen to first increases due to the NO + HO2 --> NO2 + OH reaction, but ultimately falls due to the radical termination reaction OH+NO2+M-->HNO3+M.
The possibility of serious OH changes due to anthropogenic emissions is illustrated in Figure 8. Here, reasonably high steady state OH concentrations (~10[6] molec cm[-3]) are found for low CO and NOx fluxes. But as FCO and FNO are increased beyond certain threshold values, [OH] can decrease catastrophically, although some specific combinations of high NO and CO inputs can still result in acceptable oxidizing capacity.
The chemical system is clearly complex, and may even exhibit some multiple solutions, that is, different possible atmospheric composition states for identical input conditions. This is illustrated in Figure 9, where the high-NO collapse of Figure 8 is shown in cross sectional detail. Starting at PO, increases in FNO lead first to more OH (to P1), but subsequent increases in FNO lead to much lower OH levels (to P2). Suppose now that, after the environmental seriousness of the OH collapse is realized, attempts are made to restore oxidizing capacity by reducing the FNO Reversing the FNO trend will not result in the same OH solution until the point P3 is reached, where FNO is considerably lower than the FNO which was capable of sustaining the highest OH levels at P1.
A most important question relates to how today's fluxes compare with the threshold fluxes at which the oxidizing capacity fails. Rough estimates suggest that we are within an order of magnitude of the critical values, but such estimates should be viewed in the context of large uncertainties in both the estimation of fluxes and in the model calculations of the global oxidizing capacity.
The lifetime of a molecules X with respect to photodissociation is given by the inverse of the photolysis rate coefficient (J, also termed the photodissociation frequency),
where is wavelength (nm), F() is the spectral actinic flux (quanta cm[-2] s[-1] nm[-1]), X() is the absorption cross section (cm[2] molec[-1]) and Xi() is the quantum yield (molec quantum[-1]) for the dissociation of X into one of several possible product sets, say channel i. Changes in photolysis rates are likely to have a profound impact on various aspect of tropospheric composition, including the formation and destruction of O3 and OH radicals, and the lifetime of various reduced species.
The calculated variation of tropospheric O3 with changing stratospheric O3 levels (and therefore UV levels) is shown in Figure 10. The net effect on ozone production depends, once again, on NOx levels. At low NOx, the UV increase reduces surface O3 levels, while at high NOx the net effect is an increase in O3. The high-NOx regime is of particular interest in cities where photochemical smog and ozone formation are of concern, because higher UV levels will make it more difficult to attain air quality standards, as illustrated in Table 5. The low-NOx ozone decrease may already have been observed, as shown in Figure 11, under the extreme conditions of UV increase experienced under the Antarctic ozone hole.
OH concentrations are generally expected to increase with increasing UV levels, as shown in Figure 12. The calculated relative increases in [OH] are greatest at low NOx, (about 50% increase for a 20% ozone column reduction), with decreasing sensitivity at high NOx. An analysis of recent total ozone column data from 1979 to 1989 (TOMS version 6, as reported by Stolarski et al. 1991 ) suggests an increase in the J value for the reaction
which is the primary OH source over most of the troposphere. The globally averaged J trend is about 4% per decade, with the latitudinal distribution shown in Table 6. Because this trend translates almost directly into OH increase at low NOx, the globally averaged OH concentration should also have increased by about 4% per decade. This OH increase has significant implications for the global concentrations of methane, which had been increasing by about 1% per year until the early 1980s, but recent observations indicate a slower increase of about 0.6% per year. This slowing of the methane trend is consistent with the J trend estimated from the TOMS ozone data. While other factors may have been contributing to the changing CH4 trends (i.e., changes in CH4 emissions, and factors which affect OH such as NOx, etc.), the UV-induced component is obviously significant.
Carroll MA, Ridley BA, Montzka DD, Hubler G, Walega JG, Norton RB, Huebert BJ, and Grahek FE (1992) Measurement of nitric oxide and nitrogen dioxide during the Mauna Loa Observatory photochemistry experiment. J Geophys Res 97:10361-10374
Gery MW (1989) Tropospheric air quality. In: UNEP Environmental Effects Panel Report. J. C. van der Leun, M. Tevini, and R. C. Worrest (eds) United Nations Environmental Programme, Nairobi Kenya
IPCC (1990) Climate Change. In: Intergovernmental Panel on Climate Change: Report by Working Group 1. Cambridge University Press, Cambridge
Liu SC and Trainer M (1988) Responses of tropospheric ozone and odd hydrogen radicals to column ozone changes. J Atmos Chem 6:221-233
Levy II, H (1971) Normal atmosphere: large radical and formaldehyde concentrations predicted. Science 173:141-143
Logan JA, Prather MJ, Wofsy SC, and McElroy MB (1981) Tropospheric chemistry: a global perspective. J Geophys Res 86:7210-7254
Madronich S and Granier C (1992) Impact of recent total ozone changes on tropospheric ozone photodissociation, hydroxyl radicals, and methane trends. Geophys Res Lett 19:465-467
Ridley BA, Madronich S, Chatfield RB, Walega JG, Shetter RE, Carroll MA, and Montzka DD (1992) Measurements and model simulations of the photostationary state during MLOPEX: Implications for radical concentrations and ozone production and loss rates. J Geophys Res 97: 10375-10388
Schnell RC, Liu SC, Oltmans SJ, Stone RS, Hoffman DJ, Dutton EG, Deshler T, Sturges WT, Harder JW, Sewell SD, Trainer M and Harris JM (1991) Decreases of summer tropospheric ozone concentrations in Antarctica. Nature 351:726-729
Stolarski RS, Bloomfield P, McPeters RD, and Herman JR (1991) Total ozone trends deduced from Nimbus 7 TOMS data. Geophys Res Lett 18:1015-1018
Thompson AM (1992) The oxidizing capacity of the earth's atmosphere: Probable past and future changes. Science 256:1157-1165
WMO (1991) Scientific Assessment of Ozone Depletion: 1991. World Meteorological Organization Global Ozone Research and Monitoring Project Report No 25, Geneva