Two of the DC-8 flights during the 1991--1992 second Airborne Arctic Stratospheric Expedition (AASE 2) were between California and Tahiti. Extremely aburpt changes in O3 and NOy were observed on both flights as the aircraft crossed the subtropical jet. They indicate that the width of the transition from midlatitude to tropical air in the troposphere can be as short as 1 km. The NOy/O3 ratio was remarkably stable across the transition. We discuss some of the dynamical features associated with the transitions and speculate on the reasons for their abruptness. They occurred south of the subtropical frontal zone and were accompanied by changes in humidity, NOx/NOy, and modest changes in CO, CH4, and CO2. In addition, a chemical model constrained by measurements of the long-lived species is used to simulate the variation of NOx/NOy along the two flight tracks.

Although this model is quite successful at simulating observed NOx/NOy in midlatitude air, it drastically overestimates NOx/NOy in tropical air. The rate at which the model converts NOx to HNO3 via the NO2+OH reaction is very slow in the upper tropical troposphere because the low O3 concentrations and cold temperatures force most of the NOx to be in the form of NO during the day. We argue that there is an important NO to NO2 pathway in this region not presently included in models, the much of the NOy is in a stable (possibly aerosol) form that is not readily converted to NOx, or that there has been insufficient time since convection for NOx to be released from other more stable forms of NOy. It is important to resolve this discrepancy because present models which have the correct O3 and NOy may overestimate O3 production rates and OH concentrations in the upper tropical troposphere. ¿ American Geophysical Union 1995.