Accurate determination of the boundary layer mixing height (MH) is critical to simulate pollutant levels using photochemical models. The daytime mixing heights derived from Parcel and Richardson number methods using hourly profile data from the Penn State/NCAR Mesoscale Model Version 3.3 (MM5V3) output are compared with the estimates obtained directly from the meteorological model with two PBL schemes (Blackadar and Gayno-Seaman) during July of 1999 over Philadelphia, PA to evaluate the uncertainties in the evolution of the MH. An analysis of the diurnal evolution of the urban PBL and its relationship to ground-level ozone (O3) levels is presented, along with differences in MH between the 15-19 July 1999 high O3 episode and the non-episode days.

Our results indicate that on the average the Richardson number method estimated larger MHs than the Parcel method. The MH from MM5V3 output is much smaller than those derived from the Parcel and Richardson number, especially for the Gayno-Seaman scheme (turbulent kinetic energy based). The daily maximum MH and ground-level O3 concentration were both twice as much on episode days than on non-episode days. The average hourly growth rates of MH and O3 were largest during the morning hours (0700 to 1000 EST), indicating that both entrainment and photochemical production are significant contributions to the accumulation of ground-level O3 in the urban areas.