Air quality (AQ) simulation models for PM_2.5 (by mass, constituent speciation, as well as size distribution and number) and for many individual species and classes of air toxics pollutants provide a basis for implementation of National Ambient Air Quality Standards (NAAQS) and a tool for performing risk based assessments and developing environmental management strategies. These (and other) air pollutants exhibit different degrees of spatial and temporal variability especially in urban areas and in different geographical-climatic regimes. In this study, we explore the specific role of AQ models to drive human exposure models especially in situations where the pollutants exhibit high spatial and temporal variability. We seek a capability to capture both resolved scale of these concentration fields as well as providing measures of sub-grid scale variability such as peak-to-mean ratios (and other statistical measures of variability) in concentration distributions that impact human exposures. We investigate the use of a high resolution emissions-based modeling approach to enhance and complement the more limited data from central site monitoring networks to provide the concentration fields at high temporal and spatial resolutions. By providing further information of concentration variability at sub-grid scales, we complete the requirements needed for exposure assessments.

In our implementation of these conceptual requirements, the gridded air quality fields are produced using the Models-3/Community Multiscale Air Quality (CMAQ) modeling system. We describe and show results of refinements that allows CMAQ predictions of the concentration fields to 1.3 km grid resolution. At this resolution, we are testing and incorporating urban scale parameterizations for urban building morphologies and land use patterns for improved modeling of the transport and dispersion fields. The pollutants variable chemical and photochemical reactivity in atmospheric transported mixtures are modeled using the CMAQ chemical-transport modeling system. For the sub-grid variability, we identify and explore two contributing sources, including: (1) those that arise from turbulence induced concentration fluctuation (using Large-Eddy Simulation (LES) techniques) and (2) dispersion of point street and area sources from street canyon flows using a combination of Computational Fluid Dynamics (CFD) and wind tunnel modeling techniques. This variability information, in the form of probability density functions (PDFs), provide the basis for various statistical information such as peak to mean ratios, and spatial structure functions of the variability as a function of the building morphology.

Preliminary results for a case study for Philadelphia show that the constituent species of PM_2.5 exhibits greater spatial variability than the total mass of fine particulate. In addition the extent of the resolved scale spatial variability is seen to vary with each pollutant and also, the grid resolved variability does not necessarily increase monotonically with increased grid resolution. This means that the grid resolution selected for use in exposure modeling may need to be ascertained by numerical experiments. Results to date for the sub-grid modeling are from the contributions from turbulence-induced concentration fluctuations. We show that sub-grid variability from this source exhibits a wide range (e.g., formaldehyde, acetaldehyde/large; CO/small), both at the surface as well as aloft in the mixed layer due to degree of photochemical reactivity in atmospheric mixtures.