Ozone

Ozone in the troposphere can damage plants and materials and act as a respiratory irritant. In the upper troposphere it is also an important greenhouse gas. The two main sources of O are transport from the stratosphere and in-situ photochemical production, which occurs when CO, CH, and NMHCs react in the presence of NO and sunlight. Trends in tropospheric O have been noted, but no global universal trend exists because of the importance of regional processes in determining the abundance of O. Surface O at the South Pole, for example, decreased 17% over 1976-1990, perhaps due to a combination of increased transport of marine air from lower latitudes and an increase in the destruction of O by photolysis (Schnell et al., 1991). O photolysis has increased as a result of depletion of stratospheric O allowing more ultraviolet radiation to enter the troposphere (see also Madronich and Granier, 1992). For other locations, trends vary. Surface O has increased slightly at Mauna Loa (0.37 ± 0.26%/yr), and at Barrow (0.67 ± 0.30%/yr), probably due to photochemical processes, but has shown no trend at America Samoa (Oltmans and Levy, 1994). Mid-tropospheric (500 mb) concentrations at 10 ozonesonde sites (9 in the northern hemisphere) have increased 1.2 ± 0.5%/yr since the 1960s and 1970s (London and Liu, 1992). Lefohn et al. (1992) note the difficulties in discerning trends.

Several summer aircraft campaigns have characterized regionally important sources of tropospheric ozone. The ABLE-3A campaign characterized Alaskan polar air prior to transport across North America while ABLE-3B characterized the sources of ozone in eastern and central Canada, sampling air that was more impacted by pollutants. In ABLE-3A, correlations of ozone and aerosols showed that in summer ``clean'' polar air, natural processes controlled tropospheric O. Stratospheric intrusion was a significant source of O (Gregory et al., 1992). For 0 -- 2 km, stratospheric intrusions contributed 12% of the O budget, but this increased to 50% at 4 -- 6 km (Browell et al., 1992). The main sinks were photochemical destruction (because of low NOx concentrations) and deposition. However, Jacob et al. (1992) point out that even at very low NO concentrations, in situ production increases O by a factor of 2 over levels expected in the absence of NO. The NOx budget (unlike O) was not dominated by transport of NO from the stratosphere, but possibly by transport of NO and NO (from biomass burning, fossil fuel combustion) from lower mid-latitudes, decomposition of PAN, and the possible decomposition of other nitrate species (Jacob et al., 1992; Singh et al., 1992).

Over eastern and central Canada in ABLE-3B, stratospheric intrusions of O were still important, and accounted for 33% of tropospheric O (Browell et al., 1994) and 40% of upper tropospheric O (Bachmeier et al., 1994). However, the O budget was also affected on the regional scale by sources of pollution (urban, biomass burning) and long range transport. Evidence of biomass burning was present in two-thirds of the flights and it affected 25% of the troposphere below 4km (Browell et al., 1994). Boreal forest fires were a major contributor to PAN and hence to NO (Sandholm et al., 1994). Urban pollution increased the lowest 5km tropospheric column O by 40% when present (Anderson et al., 1994).

Measurements at Mauna Loa (MLOPEX, May-June, 1988) showed that the budget of O was most strongly affected by photochemical production and destruction, with transport being secondary. During periods of downslope flow (representative of the free troposphere), O ranged from 16 - 75 ppbv, with an average concentration of 43±12 ppbv (Walega et al., 1992). The ozone budgets of Ridley et al. (1992) and Liu et al. (1992) showed a small net destruction of ozone, which was the difference between two larger numbers representing the photochemical production and photochemical destruction. Theoretically, transport from other regions (including the stratosphere) should balance these two terms. The small net destruction of O indicates that the stratosphere is not as important in the budget of O at this location as it is the summer Arctic and sub-arctic.

The Southern Oxidant Study (S.O.S.) characterized summer photochemistry at several southeastern U.S. sites. Ozone is substantially affected by local in-situ chemistry, but entrainment and transport from other regions also play a role in determining its abundance (Kleinman et al., 1994).

The O budget over the Atlantic Ocean may be divided into three regions. Over the North Atlantic, August and September GTE-CITE3 flights and 5 day back trajectories show that eastward transport of air parcels from the North American continent increase the tropospheric column O more than 10% (Anderson et al., 1993b). As part of the NARE 1991 summer campaign, Parrish et al. (1993) measured surface CO and O at three North Atlantic locations (north of 42ûN at approximately 500 km intervals downwind from the northeastern U.S.). Transport from North America is an important source of these species, as reflected in the strong correlation between CO and O (and slope = 0.22 -- 0.30 for all three sites). The ozone source from North American pollutants was significant relative to its cross tropospause flux. Chin et al. (1994) used correlations of O with CO and with NO -- NO to constrain estimates of ozone production from pollution sources.

Over the tropical Atlantic, Bermuda (32ûN, 65ûW) and Barbados (13ûN, 60ûW) appear to be affected more by transport from the stratosphere and net photochemical destruction (due to low NO levels) than by transport of polluted air from North America (Oltmans and Levy, 1992). Back trajectories indicate that the spring-time maximum at Bermuda is due to the transport of upper tropospheric air from Canada, where injection by stratospheric intrusion is at a maximum at this time of year. The August minimum in O concentrations is associated with prolonged transport over the subtropical ocean, driven by the circulation associated with the Bermuda High. At Barbados, the maximum was in January, and periods of high O were characterized by transport from higher altitudes and latitudes (Savoie et al., 1992). The Barbados concentrations were higher in magnitude, but similar in seasonal cycle to American Samoa. Samoa is far removed from anthropogenic sources. The main source of O in the equatorial Pacific Ocean is believed to be transport from the stratosphere, and the main sink photochemical destruction (due to low NO) (Thompson et al., 1993b), similar to the tropical Atlantic. However, Samoa may also be affected by large-scale meridional and zonal circulations that lead to O minima over the equatorial Pacific Ocean (Piotrowicz et al., 1991).

Satellite analysis shows enhanced tropospheric O levels over southern Africa and the southeastern Atlantic (Fishman et al., 1991) during September-November. Ozone concentrations may increase due to transport of biomass burning products (Richardson et al., 1991). At lower levels of NO the net number of O molecules produced per molecule of NO consumed is higher that at high levels of NO, hence, the ozone production potential of these products can increase if associated with dilution and convection (Pickering et al., 1992). The observed ozone maximum appears to arise from a combination of biomass burning over both Africa and South America and downward transport from the upper troposphere/lower stratosphere (Anderson et al., 1993a), which can result from both synoptic scale and planetary scale circulations (Krishnamurti et al., 1993). Further work investigating the role of southern hemisphere biomass burning should be forthcoming from the TRACE-A (Transport and Atmospheric Chemistry near the Equator-Atlantic) experiment, conducted during September and October, 1992 (Fishman, 1994).

The correlation of O with other species has yielded additional information. At six eastern North American sites in summer, O and the products of photooxidation of NO (NOy -- NO) were strongly correlated in photochemically aged air (Trainer et al., 1993). The correlation may be applicable to other air mass types, as well, and the effectiveness of NO and/or NMHC emission reductions may be indicated by NO levels (Milford et al., 1994a). In the upper troposphere, O and NOy are weakly correlated (as opposed to the lower stratosphere, where they are much more strongly correlated), indicating the importance of the upper tropospheric lightning source of NO (Murphy et al., 1993). Correlations of NO and O in the stratosphere can be used to deduce a stratospheric O flux of 3.5 x 1010 cm-2 s-1 (= 450 Tg/yr = 19 x 1028 molecules/yr) (Murphy and Fahey, 1994), which is smaller than a recent estimate of the tropospheric photochemical production (96 x 1028 molecules/yr) (Fehsenfeld and Liu, 1993).

Modeling studies have shown that since preindustrial times, concentrations of both O and HO have increased while OH has decreased (Thompson, 1992; Thompson et al., 1993a). In the future, as precursor concentrations increase, O levels are predicted to increase (Prather et al., 1994). Regional modeling studies show that reducing NO emissions (either alone, or in conjunction with reductions in NMHCs) is far more effective in reducing O over large regions of the eastern U.S. than strictly reducing NMHC emissions (Milford et al., 1994b; Mathur et al., 1994). This is true because over much of the region, O production is NO limited; the exceptions are regions such as New York City, with strong NO sources and high NO to hydrocarbon emission ratios. A continental scale model showed that if NOx emissions are reduced by 50%, over much of the rural eastern U.S., O is predicted to decrease by 15% (Jacob et al., 1993). It also showed that as much as 70% of the O photochemically produced over the U.S. is exported. Penner et al. (1994) demonstrate the global importance of NOx emissions for O production.