Abstract Analyses of observations point to a more detectable change signal in the frequency than magnitude of flood events across the conterminous United States (CONUS); however, little is known about how the frequency of these events is projected to change under different scenarios. Here we apply a statistical attribution-and-projection approach to thousands of streamgages across CONUS and assess how the frequency of flood events is expected to change under multiple scenarios. We find that more frequent flood events are projected for the eastern United States, whereas decreases are slightly more common than increases in the U.S. Southwest and Great Plains. Additionally, the seasonality of flood events is projected to shift, with distinct regional patterns based on geography, topography, and season. These changes are expected to manifest themselves by the middle of this century, highlighting the pressing needs for flood mitigation and adaptation efforts.