Vertical diffusion is a critical factor in the photochemical modeling of ozone and particulate matter (PM). Improper specification of vertical diffusivity in the model may permit excessive vertical mixing of ozone to offset the ozone lost through titration in the nighttime stable boundary layer, or abnormal accumulation of ozone and PM concentrations during the daytime when peaks usually occur. In this study, the EPA Community Multiscale Air Quality (CMAQ) model is used to simulate the August, 2002 high ozone and PM episode. The meteorological and emission inputs to CMAQ were prepared using the Penn State/NCAR Mesoscale Model version 5 (MM5) and the Sparse Matrix Operator Kernel Emissions (SMOKE) emission process. An approach for modifying the vertical diffusion coefficients in CMAQ to better represent the variation of the daytime and nighttime diffusion in the surface layers will be explored. We also investigate the effect of the vertical grid resolution to the CMAQ predictions on ozone and PM concentrations. The hourly ozone measurements from the AQS network and PM hourly measurements from Special Trend Network (STN) will be used to evaluate the performance of the model. The impact of the model predictions due to the differences in the treatment of the vertical diffusion and the model vertical resolution will be discussed.