Using a Canadian weather surveillance radar (CWSR), we assessed the relationship between aural passerine counts and radar reflectivity during autumn migration on 16 nights. Reflectivity was positively correlated on all but 1 night with the number of birds detected aurally, but the correlation strength varied between -0.58 and 0.93 among nights (mean ± SD = 0.69 ± 0.42). Using linear mixed-effects models with aural counts nested within nights, we found that the number of birds detected by observers increased with reflectivity. The slope of this relationship did not vary between observers, nor was it affected by time since sunset, but the number of birds detected aurally tended to be lower when ambient noise levels were high. We know that the radar was relatively sensitive to low bird densities, because the intercept was slightly positive and its 95% confidence interval marginally included zero. However, the relationship between the number of birds detected aurally and reflectivity varied significantly among nights. Such variation was likely caused by a combination of (interacting) factors, including bird species and behavior (e.g., calling rate, flight altitude), influencing bird detectability by the observers and the radar. The weather radar network of the United States (NEXRAD) is already used for bird migration studies, and we conclude that the use of CWSR can extend NEXRAD's coverage farther north by hundreds of kilometers, thereby increasing our understanding of how birds use the North American landscapes during migration.