We discuss the results of multiscale model simulation study using the U.S. EPA Models 3 Community Multiscale Air Quality (CMAQ) modeling system (Byun and Ching, 1999) to provide the concentration information at fine or neighborhood scales (NS). For this study, our simulations incorporate urban canopy parameters in order to capture the influence of urban morphological structures on the subsequent air quality distributions. Fine grid size makes possible the modeling of spatial gradients in pollutant concentration fields arising from dispersion and photochemistry of large source emissions. Inherent spatial details at smaller than model grid sizes may be very large near sources, as a result, near-source peak-to-grid mean concentrations may be large. Thus, for urban areas, both grid and sub-grid concentration information are needed for the goal of driving risk assessment and exposure modeling. The urban canopy parameterizations are applied at smaller grid resolutions where we demonstrate that the details of urban structures have large impacts on the dynamics and thermodynamics of the meteorological fields and the subsequent air quality fields within and above the urban roughness sub layer (RSL). The meteorological fields for CMAQ are provided by the Penn State/NCAR Mesoscale Model (MM5). For high resolution modeling we modify the Gayno-Seaman planetary boundary layer (GSPBL) scheme of MM5 to incorporate urban canopy parameterization (UCP) based on two systems that apply a drag-force approach (DA). A model domain and study area centered on the Philadelphia Metropolitan area was set up and the CMAQ model was run for simulations at horizontal grid dimensions of 36 km, 12 km, 4 km, and 1.33 km, the latter representing the NS resolution.

To appreciate the characteristics of additional spatial variability at coarser grid resolutions from fine scale modeling, standard statistics such as variances, skewness, kurtoses as well as peak to mean and range of concentration variability is presented at coarse grid scales (4km and 12 km grid resolutions) derived from the fine scale simulations at 1.3 km grid resolution. These provide some limited guidance for the goal to prescribing characteristic probability density functions for representing within-grid fine scale concentration variation(s). Results will be described for pollutatns with differing chemical reactivities including and air toxics pollutant, formaldehyde, and photchemicical oxidant species including ozone and nitrogen oxides as well as for carbon monoxide.