Urban areas located in Tornado Alley are consistently threatened by severe weather. Tornado sirens are commonly utilized by local governments to inform citizens of approaching danger. Geographic Information System (GIS)-based analyses have provided a practical means by which to assess the spatial extent and coverage of siren networks. Commonly, these studies apply arbitrary distance buffers to siren locations to determine areas where people can presumably hear them. This approach, however, does not account for the complexities of sound propagation (e.g., environmental factors such as weather conditions and topography). This study proposes the application of the SPreAD-GIS toolset to model the propagation of tornado siren sound. We model siren sound propagation in the city of Stillwater, Oklahoma (USA) for three scenarios, which align with the last three occasions when tornado warnings were issued in the city and the sirens were activated. We calculate mean siren decibel levels for all buildings in the city and relate this to population information to obtain a more comprehensive understanding of the performance of the siren network. Across all scenarios, findings indicate that the siren network performs well with less than 3% of the population residing in locations with a mean value of 80 dB or less.