A new meteorological-chemical model is used to determine the relative contribution of regional-scale transport and local photochemical production on air quality over Philadelphia. The model performance is evaluated using surface and airborne meteorological and chemical measurements made during a 30-day period in July and August of 1999 as part of the Northeast Oxidant and Particulate Study (NE-OPS). Good agreement between the simulations and observations was obtained. The bias in the vicinity of Philadelphia over the simulation period was -5.8 ppb for the peak ozone mixing ratio during the day and 2.0 ppb for the minimum ozone mixing ratio at night. Layers of ozone above the convective boundary layer were measured by both research aircraft and ozonesondes during the morning between 0900 and 1100 LT. The model demonstrates that upwind vertical mixing processes the previous afternoon, subsequent horizontal transport aloft, and depletion of ozone by NO titration within the stable boundary layer at night lead to the development of these layers. Ozone aloft was then entrained into the growing convective boundary, contributing to surface ozone concentrations. Through a series of sensitivity studies, we find that most of the ozone is the result of emissions in the vicinity of Philadelphia and Chesapeake Bay area, but up to 30--40% of the ozone during high ozone episodes was due to transport from upwind sources. Local emissions and meteorological conditions were largely responsible for one high ozone episode because of light winds.