Understanding how fish respond to drought and streamflow intermittency is important for designing effective conservation actions. Fish could move to refuge areas prior to intermittency, or be trapped in nearby habitats at the onset. We used mechanical pumps to simulate streamflow intermittency in the Rio Grande, New Mexico. We conducted a 5‐week before–after impact‐only experiment. We seined at 10 sites each week to observe changes in Red Shiner Cyprinella lutrensis abundance and distribution in a 35‐km reach. We conducted two surveys prior to any flow reduction. Next, we reduced flows each week to simulate streamflow intermittency and repeated surveys to identify changes in distribution that would suggest large‐scale movement of Red Shiner to refuge areas. During the first survey, we marked Red Shiner with elastomeric tags to observe movement. In a separate effort, we surveyed all isolated pools that formed after flow reductions for tagged Red Shiner. We observed no patterns that suggested large‐scale movement to refuge areas. Instead, numbers remained steady or increased at all sites in surveys 1 through 4 and declined sharply when sites dried. Tagged fish also supported this conclusion: dispersal models estimated 56% remained within ~0.5 km of the tagging location. We suggest Red Shiner is resilient to drought because of their opportunistic life history and potential for rapid recolonisation, rather than resistant to drought through population‐scale movements to refuge areas. Conservation actions to benefit species threatened by drought must consider the mechanisms of persistence, including both behaviour and life‐history strategy, to be effective.