This dataset is part of a multi-scale assessment and classification for segments of 15 rivers of the Upper Midwest United States that meet various criteria for largeness. All rivers are tributary to the Mississippi River system. The 11,600 kilometers (km) that qualified as large were parsed into 10-kilometer-long segments and classified by major alterations (free-flowing, navigation pools, storage reservoir). The dataset also includes a statistically based, component classification based on the 10-km segments. Cluster analysis of hydrologic variables from 66 streamflow gaging stations yielded 5 clusters calculated from 5 ecohydrological metrics related to lateral connectivity with the floodplain. A separate cluster analysis of 5 geomorphologic variables associated with each of 1,172 river segments also yielded 5 clusters. When the hydrologic variables were associated with corresponding segments, the cluster analysis yielded 8 hydrogeomorphic clusters that can be explained in terms of their contribution to floodplain connectivity. Although the clusters overlap considerably in principal component space, the resulting hydrogeomorphic classification leads to a physically reasonable distribution of classes.