Quantification of the spatial and temporal variations of outdoor air pollutant concentrations provides important information for epidemiological and other air-pollution studies, many of which have relied in the past on data from a single, centrally-located air pollution monitoring site. A method is developed for combining air pollution measurements from multiple monitors and monitoring networks to generate daily air pollution concentration fields representing spatial variations over distances of approximately 1–10 km. Meteorological and co-pollutant data are used to estimate missing site measurements, yielding more realistic concentration fields as the number of monitoring locations with available data increases. Monitoring data are interpolated with weights computed from intersite pollutant correlations, which decay with distance, so distances between interpolation points and monitoring sites are factored into the interpolation weights. The approach minimizes the influence of source-oriented sites that represent limited areas, because data from such sites exhibit low intersite correlations and yield interpolation weights that decay rapidly to zero. Interpolated values represent pollutant concentrations averaged over spatial scales that depend on intersite distances and the interpolation grid, and do not delineate sharp spatial gradients associated with roadside or near-source conditions. The approach yields quantified interpolation errors the values of which depend on measurement uncertainties, intersite distances, and the representativeness of monitoring site locations. The method is illustrated using an 11-year period of measurements of ozone, PM2.5, and PM10 concentrations from Jefferson County, Alabama. The principal city is Birmingham, which is influenced by regional-scale air pollution and by local emissions from mobile sources, industrial facilities, and residential communities. Emission sources are not distributed uniformly throughout Birmingham, the ridge-and-valley topography complicates dispersion of local emissions, and monitoring data indicate that air pollutant concentrations vary spatially as well as temporally. No single monitor represents air quality across the entire study area.