Under a warming climate, it is unclear how environments associated with US tornado outbreaks are changing. This work narrows this gap by using maximum covariance analysis (MCA) to find multivariate relationships between 500-hPa geopotential height anomalies and a severe thunderstorm proxy – WMAXSHEAR (a product of convective available potential energy and 0–6 km vertical wind shear) associated with past (1950-2019) May major tornado outbreaks. Results highlight three main patterns that explain the majority of covariance between tornado outbreaks and the large-scale atmospheric environment. Tornado outbreaks occurring under the dominant pattern (MCA1) initiate at different hours of the day and tend to last for many hours. Tornado outbreaks associated with the second (MCA2) and third (MCA3) patterns are shorter in duration and tend to initiate during the warmest daytime hours. Moreover, an increase in the magnitude and the spatial extent of the patterns conducive to tornado outbreaks was observed after 1980.