Dispersal and shared environmental conditions can both synchronize the dynamics of local populations, but disentangling their relative influence on dynamics is challenging. We used a Bayesian approach to estimate the synchrony of a metapopulation of Atlantic salmon (Salmo salar) composed of 18 populations in Brittany, France, including a 24-year time series of the abundances of juveniles. We estimated the spatial synchrony at a regional and local spatial scale over the study period. We found a strong regional synchrony despite spatiotemporal variability of local synchrony in the abundance of juveniles. We then explored the drivers of synchrony, including environmental conditions (aspects of river flow) and abundance of adult breeders. This revealed that summer low-flow conditions seemed to synchronize the abundances of juveniles more than the synchrony in the abundance of adult breeders, suggesting a Moran effect. Given that drought conditions are expected to become more common with climate change, our work highlights the potentially strong synchronizing effect of summer low flow on the dynamics of local salmon populations and the benefits of considering synchrony at multiple scales.