Transport and diffusion of ozone in the nocturnal and morning planetary boundary layer of the Phoenix valley

The evolution of ozone (03) in the nocturnal and morning-transitional planetary boundary layer (PBL) of the Phoenix valley was measured as a part of the 'Phoenix Sunrise Experiment 2001' of the U.S. Department of Energy conducted in June 2001. The goal of the field program was to study the transport. distribution and storage of ozone and its precursors in the urban boundary layer over a diurnal cycle. The ground level O-3 as well as mean meteorological variables and turbulence were measured over the entire period, and vertical profiling (using a tethered balloon) was made during the morning transition period. Approximately half of the observational days showed the usual diurnal cycle of high O-3 during the day and low O-3 at night, with nitrogen oxides (NOx = NO2 + NO) showing an out of phase relationship with O-3. The rest of the days were signified by an anomalous increase of O-3 in the late evening (similar to 2200 LST), concomitant with a sudden drop of temperature, an enhancement of wind speed and Reynolds stresses, a positive heat flux and a change of wind direction. NOx measurements indicated the simultaneous arrival of an 'aged' air mass, which was corroborated by the wind predictions of a mesoscale numerical model. In all, the results indicate that the recirculation of O-3 rich air masses is responsible for the said high-O-3 events. Such air masses are produced during the transport of O-3 precursors by the upslope flow toward mountainous suburbs during the day, and they return back to the city at night via downslope winds (i.e. mountain breeze). The corresponding flow patterns, and hence the high-O-3 events, are determined by back-ground meteorological conditions. The vertical profiling of O-3 and flow variables during the morning transition points to a myriad of transport, mixing and chemical processes that determine the fate of tropospheric O-3. How well such processes are incorporated and resolved in predictive O-3 models should determine the accuracy of their predictions.