Hourly fluctuations in flow from Glen Canyon Dam were increased in an attempt to limit the population of nonnative rainbow trout Oncorhynchus mykiss in the Colorado River, Arizona, due to concerns about negative effects of nonnative trout on endangered native fishes. Controlled floods have also been conducted to enhance native fish habitat. We estimated that rainbow trout incubation mortality rates resulting from greater fluctuations in flow were 23-49% (2003 and 2004) compared with 5-11% under normal flow fluctuations (2006-2010). Effects of this mortality were apparent in redd excavations but were not seen in hatch date distributions or in the abundance of the age-0 population. Multiple lines of evidence indicated that a controlled flood in March 2008, which was intended to enhance native fish habitat, resulted in a large increase in early survival rates of age-0 rainbow trout. Age-0 abundance in July 2008 was over fourfold higher than expected given the number of viable eggs that produced these fish. A hatch date analysis indicated that early survival rates were much higher for cohorts that hatched about 1 month after the controlled flood (~April 15) relative to those that hatched before this date. The cohorts that were fertilized after the flood were not exposed to high flows and emerged into better-quality habitat with elevated food availability. Interannual differences in age-0 rainbow trout growth based on otolith microstructure supported this hypothesis. It is likely that strong compensation in survival rates shortly after emergence mitigated the impact of incubation losses caused by increases in flow fluctuations. Control of nonnative fish populations will be most effective when additional mortality is applied to older life stages after the majority of density-dependent mortality has occurred. Our study highlights the need to rigorously assess instream flow decisions through the evaluation of population-level responses.