Floods, with their inherent spatiotemporal variability, drive floodplain physical and ecological processes. This research identifies a flood regime typology and approach for flood regime characterization, using unsupervised cluster analysis of flood events defined by ecologically meaningful metrics, including magnitude, timing, duration, and rate of change as applied to the unregulated lowland alluvial Cosumnes River of California, United States. Flood events, isolated from the 107-year daily flow record, account for approximately two-thirds of the annual flow volume. Our analysis suggests six flood types best capture the range of flood event variability. Two types are distinguished primarily by high peak flows, another by later season timing and long duration, two by small magnitudes separated by timing, and the last by later peak flow within the flood event. The flood regime was also evaluated through inter- and intra-annual frequency of the identified flood types, their relationship to water year conditions, and their long-term trends. This revealed, for example, year-to-year variability in flood types, associations between wet years and high peak magnitude types and between dry years and the low magnitude, late season flood type, and increasing and decreasing contribution to total annual flow in the highest two peak magnitude classes, respectively. This research focuses needed attention on floodplains, flood hydrology, ecological implications, and the utility of extending flow regime classification typically used for environmental flow targets. The approach is broadly applicable and extensible to other systems, where findings can be used to understand physical processes, assess change, and improve management strategies.