Vertical profiles of ozone obtained from ozonesondes in Brazzaville, Congo (4 degrees S, 15 degrees E), and Ascension Island (8 degrees S, 15 degrees W) show that large quantities of tropospheric ozone are present over southern Africa and the adjacent eastern tropical South Atlantic Ocean. The origin of this pollution is widespread biomass burning in Africa. These measurements support satellite-derived tropospheric ozone data that demonstrate that ozone originating from this region is transported throughout most of the Southern Hemisphere. Seasonally high levels of carbon monoxide and methane observed at middle- and high-latitude stations in Africa, Australia, and Antarctica likely reflect the effects of this distant biomass burning. These data suggest that even the most remote regions on this planet may be significantly more polluted than previously believed.

The seasonal tropospheric ozone maximum in the tropical South Atlantic, first recognized from satellite observations [Fishman et al., 1986, 1991], gave rise to the IGAC/STARE/SAFARI 1992/TRACE A campaigns (International Global Atmospheric Chemistry/South Tropical Atlantic Regional Experiment/Southern African Atmospheric Research Initiative/Transport and Atmospheric Chemistry Near the Equator‐Atlantic) in September and October 1992. Along with a new TOMS‐based method for deriving tropospheric column ozone, we used the TRACE A/SAFARI 1992 data set to put together a regional picture of the O3 distribution during this period. Sondes and aircraft profiling showed a troposphere with layers of high O3 (≥90 ppbv) all the way to the tropopause. These features extend in a band from 0° to 25°S, over the SE Indian Ocean, Africa, the Atlantic, and eastern South America. A combination of trajectory and photochemical modeling (the Goddard (GSFC) isentropic trajectory and tropospheric point model, respectively) shows a strong connection between regions of high ozone and concentrated biomass burning, the latter identified using satellite‐derived fire counts [Justice et al., this issue]. Back trajectories from a high‐O3 tropical Atlantic region (column ozone at Ascension averaged 50 Dobson units (DU)) and forward trajectories from fire‐rich and convectively active areas show that the Atlantic and southern Africa are supplied with O3 and O3‐forming trace gases by midlevel easterlies and/or recirculating air from Africa, with lesser contributions from South American burning and urban pollution. Limited sampling in the mixed layer over Namibia shows possible biogenic sources of NO. High‐level westerlies from Brazil (following deep convective transport of ozone precursors to the upper troposphere) dominate the upper tropospheric O3 budget over Natal, Ascension, and Okaukuejo (Namibia), although most enhanced O3 (75% or more) equatorward of 10°S was from Africa. Deep convection may be responsible for the timing of the seasonal tropospheric O3 maximum: Natal and Ascension show a 1‐ to 2‐month lag relative to the period of maximum burning [cf. Baldy et al., this issue; Olson et al., this issue]. Photochemical model calculations constrained with TRACE A and SAFARI airborne observations of O3 and O3 precursors (NOx, CO, hydrocarbons) show robust ozone formation (up to 15 ppbv O3/d or several DU/d) in a widespread, persistent, and well‐mixed layer to 4 km. Slower but still positive net O3 formation took place throughout the tropical upper troposphere [cf. Pickering et al., this issue (a); Jacob et al., this issue]. Thus whether it is faster rates of O3 formation in source regions with higher turnover rates or slower O3 production in long‐lived stable layers ubiquitous in the TRACE A region, 10–30 DU tropospheric O3 above a ∼25‐DU background can be accounted for. In summary, the O3 maximum studied in October 1992 was caused by a coincidence of abundant O3 precursors from biomass fires, a long residence time of stable air parcels over the eas...

Characteristics of total O3 in southern Africa and over the adjacent Atlantic during the IGAC/STARE/SAFARI‐92/TRACE A (International Global Atmospheric Chemistry/South Tropical Atlantic Regional Experiment/Southern African Fire Atmospheric Research Initiative/Transport and Atmospheric Chemistry near the Equator‐Atlantic) field experiments are described. Most of the analysis is based on data from the Nimbus 7/total ozone mapping spectrometer (TOMS) gridded O3 data archive (version 6.0), which is used to examine O3 in terms of seasonal and interannual variability. Total O3 column variability is compared to the tropospheric O3 column derived from balloonborne ozonesondes at four fixed SAFARI‐92/TRACE A sites (Ascension Island, Brazzaville, Okaukuejo, and Irene) from September 1 to October 23, 1992. All of these sites except Okaukuejo had regular ozonesonde launches from 1990 to 1992. Total O3 and integrated tropospheric O3 at the sounding sites showed the expected September–October maxima over southern Africa and the adjacent Atlantic Ocean. Statistical analysis of the TOMS record for 1979–1992 allows disaggregation of components contributing to total O3 variability: Signals due to semiannual and annual cycles and the quasi‐biennial oscillation are identified at the sounding sites. The tropospheric O3 column estimated from integrated sondes (to ∼16 km) at the four sites ranged from 24 to 62 Dobson units (DU) (mean, 45 DU) and averaged 15% of total O3 at Irene (14 launches) and 19% of total O3 at Ascension (20 launches). Tropospheric O3 was higher at Ascension and Brazzaville than at the sites south of 15°S because transport from biomass burning regions was more direct at these sites. This transport is seen in Hovmöller (time‐longitude) plots of total O3. A comparison of 1990–1992 integrated tropospheric O3 amounts with the annual total ozone cycle shows that tropospheric ozone variations may account for all of the annual signal at Ascension (8°S) and Brazzaville (4°S) but only 30–40% of the seasonal total O3 variation at Irene (26°S). Hovmöller plots of daily TOMS O3 over southern Africa and the Atlantic show easterly flow of local O3 maxima at 0°–10°S and westerly movement from 30°–40°S. At 0°–10°S the continent‐ocean total O3 gradient and Ascension and Brazzaville O3 soundings are used to estimate a photochemical O3 formation rate of 1–2 ppbv O3/d over the Atlantic. This agrees with model calculations of moderately aged biomass burning emissions from SAFARI‐92/TRACE A [Jacob et al., this issue; Thompson et al., 1996, this issue].

A set of four ozonesonde stations located at Ascension Island, Brazzaville, Okaukuejo, and Irene, operational during the TRACE A and SAFARI‐92 experiments has provided an opportunity to investigate the vertical distribution of ozone over southern Africa and adjacent oceans. All stations display a springtime maximum in tropospheric ozone. Enhanced tropospheric ozone, which occurs between June and September at Brazzaville and between July and October at Ascension Island, is linked to dry season biomass burning. The influence of tropical biomass burning is delayed until September at Okaukuejo when a sharp increase in tropospheric ozone is experienced. The biomass burning influence at Irene is less because of its more southerly location. A general tropospheric enhancement is observed at all stations. It is manifest as an enriched layer in the upper troposphere at Okaukuejo (9–12 km) and Brazzaville (11–14 km) and in the lower troposphere (2–8 km) at Ascension Island. At Ascension Island lower tropospheric ozone values are about 20 parts per billion by volume greater than elsewhere and the tropospheric component here accounts for about 18% of the total column ozone. A series of tethersonde soundings conducted at hourly intervals at Okaukuejo revealed ozone to be well mixed in the lower boundary layer during the day, but to display marked vertical stratification at night.

The large amount of sulfuric acid aerosol formed in the stratosphere by conversion of sulfur dioxide emitted by the eruption of Mount Pinatubo (15.14°N, 120.35°E) in the Philippines around June 15, 1991, has had a pronounced effect on lower stratospheric ozone in the tropics. Measurements of stratospheric ozone in the tropics using electrochemical concentration cell (ECC) sondes before and after the eruption and the airborne UV differential absorption lidar (DIAL) system after the eruption are compared with Stratospheric Aerosol and Gas Experiment II (SAGE II) measurements from several years before the eruption and ECC sonde measurements from the year prior to the eruption to determine the resulting changes. Ozone decreases of up to 33% compared with SAGE II climatological values were found to be directly correlated with altitude regions of enhanced aerosol loading in the 16‐ to 28‐km range. A maximum partial‐column decrease of 29±9 Dobson units (DU) was found over the 16‐ to 28‐km range in September 1991 along with small increases (to 5.9±2 DU) from 28 to 31.5 km. A large decrease of ozone was also found at 4° to 8°S from May to August 1992, with a maximum decrease of 33±7 DU found above Brazzaville in July. Aerosol data from the visible channel of the advanced very high resolution radiometer (AVHRR) and the visible wavelength of the UV DIAL system were used to examine the relationship between aerosol (surface area) densities and ozone changes. The tropical stratospheric ozone changes we observed in 1991 and 1992 are likely be explained by a combination of dynamical (vertical transport) perturbations, radiative perturbations on ozone photochemistry, and heterogeneous chemistry.