Although measurements of individual exposure remain the "gold standard" for evaluating health risks associated with air pollution, atmospheric models have become important tools for understanding the liniks between energy use, air pollution, and public health. The spatial and temporal coveragae of model data typpically far exceeds that of measurement data; the costs of simulating air pollution concentrations with a model is low compared with the equipment and personnel costs to operate a network of measurement stations; and models permit analysis of future projections and implications of air quality management policy. Still, important limitations remain. No model is perfect, and data quality is limited by model accuracy, meteorological input, and quantification of emissions. Exploiting the strengths of models in public health anlaysis, while appropriately dealing with model uncertainty, poses challenges that span public health and atmospheric science disciplines. Here we review past studies in which models have been employed in air pollution health analyses, and we discuss new directions in public health that capitalize on atmospheric model strengths.