As discussed above, the concentration of ozone is determined by transport from other locations and by local production and loss. Production and destruction rates of ozone are strongly influenced by the concentration of the free-radical catalysts NO and NO2 (NOx) and OH and HO2 (HOx). In the presence of NO, ozone is produced as a by-product when CO, CH4, and other hydrocarbons are oxidized by OH. NOx also influences the destruction rate of ozone directly as a catalyst (in the stratosphere) and indirectly as a result of reactions that couple NOx to other reactive species such as the odd hydrogen radicals, OH and HO2, and, in the stratosphere, the halogen free radicals chlorine monoxide (ClO) and bromine monoxide (BrO).
The production of ozone in the stratosphere is dominated by the photolysis of oxygen (O2) by sunlight. Because radiation with short wavelengths (less than 242 nm) is screened out by O2 and ozone in the upper atmosphere, this process is not very important in the troposphere. In the region where commercial aircraft fly, ozone (O3) is produced mainly from the oxidation of CO:
|OH + CO H + CO2||(1)|
|H + O2 + M HO2 + M||(2)|
|HO2 + NO NO2 + OH||(3)|
|NO2 + sunlight NO + O||(4)|
|O + O2 + M O3 + M||(5)|
|Net: CO + 2 O2
CO2 + O3
(where M represents a gaseous third body such as N2 or O2). The oxidation of CH4 also contributes to ozone formation but is less important in the UT than reaction 1.
The rates of these reactions depend directly on the concentrations of NOx and HOx. Increases in the concentration of NOx from aircraft generally will increase the rate of ozone production by speeding the oxidation of CO and CH4. Other aircraft emissions, such as H2O, CO, and NMHCs, are not expected to significantly affect this background chemistry because natural sources of these compounds far exceed the perturbation from aviation (Friedl, 1997; Brasseur et al., 1998). Within the plume, however, the production of particulate and subsequent contrail and cloud formation may influence NOx and HOx and therefore the production rate of ozone (see Section 2.1.3).
Aircraft emissions are one of many sources of NOx in the troposphere and the stratosphere. In the stratosphere, NOx is produced primarily from the oxidation of nitrous oxide (N2O). N2O is produced by numerous sources, and its concentration has been increasing at a rate of 0.5-0.8 ppbv yr-1 (IPCC, 1996). As a result of this source, NOx concentrations in the LS are quite large, increasing from about 100 pptv at the tropopause to as much as 3000 pptv at an altitude of 20 km. Aircraft exhaust provides an additional source of NOx, but there is no evidence that this source has appreciably altered the concentration of stratospheric NOx.
The primary sources of NOx in the troposphere are fossil fuel combustion, biomass burning, soil emissions, lightning, transport from the stratosphere, ammonia oxidation, and aircraft exhaust. The largest source is fossil fuel combustion; 95% of its emissions are in the Northern Hemisphere (Lee et al., 1997). Biomass burning occurs primarily in the continental tropics. Soil emissions come from microbial denitrification and nitrification processes, the rate depends on soil type and temperature, ecosystem type, water content, and several other variables (Matthews, 1983; M�ller, 1992; Williams et al., 1992; Yienger and Levy, 1995). The contribution of NOx produced by lightning is very uncertain because it is extremely difficult to measure directly. The distribution around thunderstorms is highly variable because NO production differs for cloud-to-cloud and cloud-to-ground strikes. Oxidation of ammonia (NH3) of NO occurs at the surface, mainly in the tropics. Although the sources of NH3 are fairly well known, the rates of reactions that result in NO are uncertain (Lee et al., 1997), and, under some conditions, ammonia oxidation can be a sink for NOx.
Various studies have estimated the emission of NOx from these sources (M�ller, 1992; Atherton et al., 1996; Lee et al., 1997). Additionally, estimates for individual sources have been produced: fossil fuel combustion (Dignon, 1992; Benkovitz et al., 1996); biomass burning (Crutzen and Andreae, 1990; M�ller, 1992; Atherton, 1995); soil emissions (Matthews, 1983; Williams et al., 1992; Yienger and Levy, 1995); lightning (Price and Rind, 1992; Levy et al., 1996; Ridley et al., 1996); aircraft exhaust (Wuebbles et al., 1993; Metwally, 1995; Baughcum et al., 1996; Gardner et al., 1997); and transport from the stratosphere (Kasibhatla et al., 1991; Wang et al., 1998a).
Compilations and evaluations of emission rates (Friedl, 1997; Lee et al., 1997) have emphasized that the contribution from lightning remains highly uncertain. The relative impact of NOx produced from aviation is critically dependent on the strength of this source. Despite this uncertainty, it is likely that averaged over the globe, the lightning source of NOx in the UT is 2 to 8 times as large as the aircraft source. Although lightning, aircraft emissions, and transport of NOx from the boundary layer during deep convection clearly provide significant sources of NOx to the UT, the contribution of each in determining the total NOx concentration is not well understood.
In situ observations have been made in the UT to help identify and quantify sources of NOx. The concentration of NO measured in these field campaigns is extremely variable. Based on a number of in situ aircraft missions, however, typical NO concentrations in the UT are in the range 50-200 pptv (Emmons et al., 1997; Schlager et al., 1997; Tremmel et al., 1998). Because of the large variability in the atmosphere, these relatively sparse observations cannot quantitatively define the strengths of the various NOx sources in the UT. Recent aircraft campaigns-for example, SONEX, POLINAT-2, NOXAR-should, however, provide important constraints on the role of aircraft in perturbing NOx. For example, a full year of observations of NO in the UT and lowermost stratosphere obtained onboard a commercial airplane have demonstrated the importance of NOx sources associated with convection (Brunner et al., 1998).
The NOx species belong to a family of chemicals collectively called reactive nitrogen (NOy). This family includes species that are coupled to NOx on time scales of one day to several weeks, such as HNO4, nitrogen pentoxide (N2O5), nitric acid (HNO3), and, in the troposphere, peroxyacetylnitrate (PAN). The primary form of reactive nitrogen emitted into the atmosphere is NO. Once NO is in the atmosphere, it undergoes rapid interconversion with NO2 and then slower conversion to other reactive nitrogen species such as HNO3.
Figure 2-1: Net ozone production (24-hr average) as a function of NOx in the upper troposphere (adapted from Jaegl� et al., 1998).
Reactions occurring on aerosol surfaces also play an important role in the chemical partitioning of NOy. It is well-established that NOx concentrations are reduced and HNO3 concentrations are enhanced as a result of heterogeneous chemistry on aerosol surfaces. It is also possible that reactions occurring on soot may play the opposite role, converting HNO3 back to NOx (Section 2.1.3).
Observations of NOy species have demonstrated that our understanding of the sources of NOy and the partitioning of NOy in the stratosphere is relatively good (within � 30%) (Rinsland et al., 1996; Gao et al., 1997). Photochemical models of the partitioning and distribution of NOy in the UT, however, have typically demonstrated poor agreement with observations. Because the reactions that convert NOx, PAN, and HNO3 occur on time scales similar to transport, it is not clear whether the observed differences are caused by errors in the chemical description or by dynamic influences that are not captured in the models (Prather and Jacob, 1997).
HOx is produced in the UT and LS via a number of processes. One important primary
source of OH is the reaction of water vapor with O(1D) produced in the photolysis
|O3 + sunlight O(1D) + O2||(6a)|
|O(1D) + H2O OH + OH||(7)|
Additional sources of OH and HO2 in the troposphere include the photolysis of acetone (Singh et al., 1995), peroxides (Chatfield and Crutzen, 1984), and formaldehyde. In the stratosphere, the photolysis of HOBr and HNO3 produced in reactions occurring on aerosols may also be important HOx sources (Hanson et al., 1996). The existence of important primary HOx sources in addition to reactions 6a and 7 has been demonstrated by recent measurements of OH and HO2 in the UT and LS (McKeen et al., 1997; Jaegl� et al., 1997, 1998; Brune et al., 1998; Wennberg et al., 1998). These primary sources of OH can be further amplified during the oxidation of CH4 and other hydrocarbons.
OH is chemically coupled to HO2 by reactions such as 1-3 that convert these
species on time scales of seconds to minutes. Removal of HOx (leading to the
formation of H2O) occurs on a longer time scale (5-30 min) and is dominated
by processes such as:
|OH + HO2 H2O + O2||(8)|
|OH + NO2 + M HNO3 + M||(9)|
|OH + HNO3 H2O + NO3||(10)|
|Net: OH + OH + NO2 H2O + NO3|
|HO2 + NO2 + M HNO4 +M||(10)|
|OH + HNO4 H2O + NO2 + O2||(12)|
|Net: OH + HO2 H2O + O2|
|HO2 + HO2 + M H2O2 + O2 + M||(13)|
|OH + H2O2 H2O + HO2||(14)|
|Net: OH + HO2 H2O + O2|
NOx and HOx are therefore linked by a number of important reactions (3,9,10-12); the concentration of each depends on the concentration of the other.
Figure 2-1 illustrates how the background photochemistry
works to change ozone in the UT in response to variation in the concentration
of NOx (Jaegl� et al., 1998). The points are calculated from in situ measurements
of HOx and NO obtained during the SUCCESS campaign. The lines show how the calculated
response varies with assumptions about HOx source strength. The response includes
ozone production via reactions 4 and 5 and ozone destruction primarily via:
|O3 + HO2 OH + 2 O2||(15)|
In all models, the net ozone production rate (production rate - loss rate) increases rapidly with NOx until a maximum is reached. At NOx concentrations larger than 500 pptv, the net rate of ozone production is expected to decrease with increasing NOx. Depending on the background concentration of NOx, additions of NOx from aviation can increase or decrease the net ozone production rate. Thus, the background concentration of NOx determines both the magnitude and the sign of the perturbation. Field measurements of NOx in the middle and upper troposphere typically have found NOx to be 50-200 pptv (Emmons et al., 1997; Schlager et al., 1997; Tremmel et al., 1998). At these concentrations, the rate of net ozone production increases almost linearly with NOx.
Figure 2-1 also illustrates that the rate of ozone production depends critically on HOx source strength. In this figure, the calculation in Case 0 assumes that the primary source of HOx is limited to the reaction of O1D with H2O and CH4. This assumption is made in many (but not all) of the chemical transport models used to assess the effect of aviation in Section 2.2.1 and Chapter 4. Case 1 includes an additional source of HOx from acetone photolysis (Singh et al., 1995), assuming acetone is present at a concentration roughly consistent with recent measurements in the UT (Singh et al., 1995; Arnold et al., 1997). Case 2 assumes that in addition to acetone, peroxides and formaldehyde are transported to the UT by convection (Chatfield and Crutzen, 1984; Jaegl� et al., 1997). It is clear from observations of HOx obtained from the DC-8 (Brune et al., 1998) and the ER-2 (Wennberg et al., 1998) that HOx sources in addition to O1D + H2O are needed to explain measured concentrations of OH and HO2. The effect of these HOx sources is most pronounced in UT air when the water vapor mixing ratio is less than 100 ppmv. At median NOx concentrations observed during these campaigns-50-100 pptv, typical of the UT (Brunner et al., 1998)-the net ozone production rate is calculated to be 1-2 ppbv per day. This rate is significantly faster than would be calculated assuming only the simple O1D HOx chemistry.
The sensitivity of the net ozone production rate to assumptions about the sources of odd hydrogen is high and remains an area of significant uncertainty. The budget of acetone, for example, is poorly understood, and relatively few measurements of its concentration have been made in the UT. Observational constraints on the HOx chemistry of the UT are just now becoming available; the number of measurements is expected to increase greatly over the next few years.
The concentration of NOx also influences the rate at which ozone is destroyed
in the atmosphere, particularly in the stratosphere and in the lower troposphere.
In the stratosphere, because of the abundance of ozone and lower pressures,
the concentration of atomic oxygen is sufficiently large that NOx can catalytically
|O3 + sunlight O + O2||(6b)|
|O + NO2 NO + O2||(16)|
|NO + O3 NO2 + O2||(17)|
|Net: 2 O3 3 O2|
Reactions of the HOx family also destroy ozone. In particular, reaction 15 leads to significant ozone loss in the LS and in the troposphere. Reaction 3 is in competition with reaction 15, so the rate of ozone loss by HOx decreases with increasing NO.
Figure 2-2: A calculation of the rate of ozone loss in the lower stratosphere for springtime mid-latitude conditions during March.
Finally, ozone loss by halogen chemistry is important in the stratosphere.
During winter, particularly in polar regions, it can dominate all other chemical
destruction mechanisms. As with HOx chemistry, NOx interferes with this chemistry
by binding to the reactive chlorine radical ClO:
|ClO + NO2 + M ClONO2 + M.||(18)|
As a result of these coupling reactions, changes in the concentration of NOx can lead to increased or decreased rates of stratospheric ozone destruction. When NOx is low-as it is in most of the LS during winter, fall, and spring-most of the ozone loss occurs through HOx and halogen chemistry. Under these conditions, enhancements of NOx will decrease ozone destruction. On the other hand, at higher altitudes and during summer, NOx-catalyzed ozone loss (reactions 16-17) can dominate the removal of lower stratospheric ozone, so enhancements in NOx will speed ozone loss (Br�hl et al., 1998). These effects have been demonstrated by direct measurements of free radicals in the stratosphere (Wennberg et al., 1994; Jucks et al., 1997).
This chemistry is illustrated in Figure 2-2. A calculation is shown for typical mid-latitude springtime conditions. This entire profile is within the stratosphere, where catalytic ozone loss competes with and can exceed photochemical production. The left panel shows the fraction of ozone destroyed during the month of March as a result of catalysis by NOx (squares), halogens (circles), and HOx (crosses). For this latitude and season, the loss is dominated by halogen and hydrogen oxides below 20 km, whereas above 25 km, nitrogen oxides are most important. To illustrate how changes in NOx perturb this chemistry, the right panel shows the effect of a uniform 20% increase in the concentration of NOx. In regions where NOx is high, ozone destruction increases. On the other hand, the opposite occurs in the LS because the increased NOx decreases the loss of ozone by hydrogen and halogen radicals. Thus, as with the production rate of ozone in the troposphere, the response of ozone destruction with changes in NOx is highly nonlinear. Because the photochemical lifetime of ozone in the LS is very long, the concentration of ozone in this region of the atmosphere is strongly influenced by transport. The change in ozone loss rates illustrated in Figure 2-2 does not translate directly into a change in ozone. For example, for a uniform 20% increase in NOx, enhanced loss rates at high altitudes will reduce the transport of ozone to the LS. As a result, ozone concentrations in the LS can decrease even when the local ozone loss rate slows. Thus, the change in the ozone column with added NOx is very sensitive to the altitude distribution of the perturbation. The subsonic aircraft fleet adds NOx only to the lowermost stratosphere (< 13 km), where large-scale dynamics tend to prevent advection to higher altitude. As a result, injection of NOx by the present fleet is thought to increase ozone in the LS.
If the major direct impact of aircraft on the chemistry of the UT and lowermost stratosphere (below approximately 16 km) is an increase in the concentration of NOx, we can say with high confidence that the ozone concentrations in this region will be higher than they would be in the absence of aviation. This increase occurs because NO speeds the catalytic oxidation rate of CO and reduces the destruction rate of ozone by HOx and halogens (primarily in the stratosphere). In this context, it is important to note that the conventional troposphere-stratosphere boundary (i.e., the tropopause), reflecting important changes in atmospheric dynamics, does not coincide with the separation between net positive and negative NOx-induced changes in ozone (see right panel of Figure 2-2). This distinction is important when considering future aviation scenarios that include a significant supersonic component.
The actual quantitative change in ozone from present aircraft operation is very sensitive to the meteorology of this region of the atmosphere; longer residence times will lead to larger NOx increases and therefore higher ozone. As a result, only coupled chemical and dynamic models can estimate how large an increase is expected. To accurately predict the perturbation, these models must accurately describe the background NOx concentration, the magnitude of the aircraft-induced NOx perturbation, the sources of HOx in this region of the atmosphere, and the meteorology. Finally, because ozone itself is relatively long-lived in the UT and LS, its concentration is strongly influenced by transport (as discussed in Section 18.104.22.168.3). The transport of ozone between different regions of the atmosphere significantly confounds attempts to assign causality to local ozone trends. As described in Section 2.1.3, the production of particulates and aircraft-induced formation of clouds may be important for ozone. Chemical processes occurring in aerosols and on ice clouds may lead to increases in chlorine radical abundance in the stratosphere and suppress NOx throughout the region.
Other reports in this collection