All of America's large rivers are regulated by dams. These rivers pose restoration problems that are fundamentally different than those for small streams: their scale of operation is orders of magnitude larger, physical reconstruction of the channel is usually impossible, and multi-objective management is a prerequisite. At a national scale, regulation by large dams on rivers has radically altered fluvial hydrologic regimes and geomorphology. Experience with four examples, the Klamath River in Oregon and California, the Rio Grande in New Mexico, the Platte River in Nebraska, and the Everglades in Florida shows that science and decision-making must effectively address three common themes in restoration of large rivers. First, the fundamental avenue to restoration is to at least partially restore the altered flows, requiring the use of dam operating rules and possibly re-engineering. Second, using regulated flows, restoration must reestablish physical integrity of the river before addressing biological integrity. Third, restoration planners must recognize targets and limits: it is likely that we can restore functionality but not at pre-dam scales, we are not likely to recreate a “natural” system, and the restored system must address historical and economic issues as well as ecological ones.

The shortcomings of science for restoration in each of the four examples have been the same: lack of adequate quality assurance and control, inability to ask the right science questions and be adaptive, failure to integrate efforts among the sciences, and disconnection between empirical (historical) research and modeling (future-based) investigations. Despite these issues, science and decision-making for restoration of regulated American rivers is rapidly improving, and the need for this restoration is likely to drive the next wave of major theoretical and applied advances in river science and management.