Abstract. We use ozonesondes launched from Samoa (14øS) during the Pacific Exploratory Mission (PEM) Tropics A to show that O3 mixing ratios usually start increasing toward stratospheric values near 14 km. This is well below the tropical tropopause (as defined either in terms of lapse rate or cold point), which usually occurs between 16 and 17 km. We argue that the main reason for this discrepancy in height between the chemopause and tropopause is that there is very little convective detrainment of ozone-depleted marine boundary layer air above 14 km. We conjecture that the top of the Hadley circulation occurs at roughly 14 km, that convective penetration above this altitude is rare, and that air that is injected above this height subsequently participates in a slow vertical ascent into the stratosphere. The observed dependence of ozone on potential temperature in the transitional zone between the 14-km chemopause and the tropical tropopause is consistent with what would be expected from this hypothesis given calculated clear-sky heating rates and typical in situ ozone production rates in this region. An observed anticorrelation between ozone and equivalent potential temperature below 14 km is consistent with what would be expected from an overturning Hadley circulation, with some transport of high O3/low 0• air from midlatitudes. We also argue that the positive correlations between 03 and N20 in the transitional zone obtained during the 1994 Airborne Southern Hemisphere Ozone Experiment/Measurements for Assessing the Effects of Stratospheric Aircraft) (ASHOE/MAESA) campaign support the notion that air in this region does have trace elements of stratospheric air (as conjectured previously), so that some of the ozone in the transitional zone does originate from the stratosphere rather than being entirely produced in situ.

The concentrations of the hydrogen radicals OH and HO2 in the middle and upper troposphere were measured simultaneously with those of NO, O3, CO, H2O, CH4, non-methane hydrocarbons, and with the ultraviolet and visible radiation field. The data allow a direct examination of the processes that produce O3 in this region of the atmosphere. Comparison of the measured concentrations of OH and HO2 with calculations based on their production from water vapor, ozone, and methane demonstrate that these sources are insufficient to explain the observed radical concentrations in the upper troposphere. The photolysis of carbonyl and peroxide compounds transported to this region from the lower troposphere may provide the source of HOx required to sustain the measured abundances of these radical species. The mechanism by which NO affects the production of O3 is also illustrated by the measurements. In the upper tropospheric air masses sampled, the production rate for ozone (determined from the measured concentrations of HO2 and NO) is calculated to be about 1 part per billion by volume each day. This production rate is faster than previously thought and implies that anthropogenic activities that add NO to the upper troposphere, such as biomass burning and aviation, will lead to production of more O3 than expected.

Simultaneous in situ measurements of the concentrations of OH, HO(2), ClO, BrO, NO, and NO(2) demonstrate the predominance of odd-hydrogen and halogen free-radical catalysis in determining the rate of removal of ozone in the lower stratosphere during May 1993. A single catalytic cycle, in which the rate-limiting step is the reaction of HO(2) with ozone, accounted for nearly one-half of the total O(3) removal in this region of the atmosphere. Halogen-radical chemistry was responsible for approximately one-third of the photochemical removal of O(3); reactions involving BrO account for one-half of this loss. Catalytic destruction by NO(2), which for two decades was considered to be the predominant loss process, accounted for less than 20 percent of the O(3) removal. The measurements demonstrate quantitatively the coupling that exists between the radical families. The concentrations of HO(2) and ClO are inversely correlated with those of NO and NO(2). The direct determination of the relative importance of the catalytic loss processes, combined with a demonstration of the reactions linking the hydrogen, halogen, and nitrogen radical concentrations, shows that in the air sampled the rate of O(3) removal was inversely correlated with total NOx, loading.

No abstract