Spatial pollutant patterns and peak concentrations are strongly influenced by meteorological parameters. Therefore, accurate hourly, gridded meteorological data sets are crucial inputs for photochemical modeling. An effort has been underway to apply both diagnostic and dynamic meteorological models in order to generate inputs needed in photochemical grid model simulations. The model being employed is a modified version of the Urban Airshed Model (UAM), which was designed to accept input files generated from both meteorological approaches. In this effort, both meteorological models were exercised in two different urban domains situated next to water bodies and with significant terrain features (i.e., greater metropolitan NYC and LA basin). A historical high ozone day in the NYC domain was simulated which exhibited a strong large scale flow pattern conducive to interurban transport along the northeastern coast. Results from simulations of an ozone episode from the 1987 Southern California Air Quality Study (SCAQS) indicated that the simulations using dynamic model winds with data assimilation displayed less absolute error than simulations using diagnostically or objectively-derived winds.