A regional-scale Observation System Simulation Experiment is used to examine how changes in the horizontal covariance localization radius employed during the assimilation of infrared brightness temperature observations in an ensemble Kalman filter assimilation system impacts the accuracy of atmospheric analyses and short-range model forecasts. The case study tracks the evolution of several extratropical weather systems that occurred across the contiguous U.S. during 07-08 Jan 2008. Overall, the results indicate that assimilating 8.5 μm brightness temperatures improves the analysis and forecast accuracy for cloud-sensitive variables but has the tendency to degrade the moisture and thermodynamic fields unless a small localization radius is used. Vertical cross-sections showed that varying the localization radius had a minimal impact on the shape of the analysis increments; however, their magnitude consistently increased with increasing localization radius. By the end of the assimilation period, the moisture, temperature, cloud, and wind errors generally decreased with decreasing localization radius and became similar to the Control case only when the shortest localization radius was used. Short-range ensemble forecasts showed that the large positive impact of the infrared observations on the final cloud analysis diminished rapidly during the forecast period, which indicates that it is difficult to maintain beneficial changes to the cloud analysis if the moisture and thermodynamic forcing controlling the cloud evolution are not simultaneously improved. These results show that although assimilation of infrared observations consistently improves the cloud field regardless of the length of the localization radius, that it may be necessary to use a smaller radius to also improve the accuracy of the moisture and thermodynamic fields.