Designing ow regimes to support entire river 1 ecosystems 2

10 Overcoming the physical limits of dam operations under nonstationarity will require cre11 ative approaches to ow management and modeling approaches that forecast the e ects of 12 management actions on multiple ecosystem components simultaneously. Using a novel mul13 tispecies modeling approach, we investigated the cross-ecosystem e ects of environmental 14 ow regimes designed for speci c ecosystem outcomes. We reveal tradeo s associated with 15 ow regimes targeting riparian vegetation, shes, and invertebrates. The di erent frequen16 cies associated with each ow regime in some cases caused non-target ecosystem components 17 to become locally extirpated within 50 years. By incorporating multiple ow frequencies, the 18 natural ow regime enabled a balanced but sub-optimal response of the three ecosystem com19 ponents (mean 72% of designer ow). Although returning to a natural ow regime may not be 20 possible in highly managed rivers, novel ow regimes must incorporate diverse frequencies 21 inherent to such a regime and accomodate the sometimes con icting requirements of di erent 22 taxa at di erent times. 23