Air dispersion modeling is a promising approach for assessing the odor impact of confined animal feeding operations (CAFOs). This paper shows that predicted ambient odor levels from AERMOD, when compared to measured odor intensities from CAFOs, need scaling to adjust short-time interval measurements in the odor plume to predictions from the relatively long intervals used by the model. Odor footprints based on ten-year odor annoyance-free frequency isopleths produced through AERMOD indicate substantial differences in necessary separation distances depending on the odor tolerance of the community, time of year and location. Calm conditions must also be accounted for to accurately assess odor impacts in communities surrounding CAFOs