Barriers to migration can negatively affect population persistence. To explore how dams can influence the viability of a diadromous fish, we developed an empirically based stochastic model to estimate per-capita population growth rate (r) and probability of population decline (Pr(r < 0)). Our simulations incorporated life-history parameters common for many populations of Atlantic salmon (Salmo salar L., 1758), particularly in the southern part of the species range. Additionally, we explored the influence of individuals that reproduce more than once, i.e., “kelts”, on r and Pr(r < 0). For the life-history scenarios examined here, dams are forecast to negatively affect persistence, even at the comparatively high per-dam smolt survival rate of 90%. As the number of dams increases from one to four, the probability of negative population growth increases four-fold from 10% to 47%. Kelt survival rate, number of dams, and smolt dam-passage survival were all found to be significant factors in predicting population persistence. The present study suggests two primary conclusions: (1) dams are likely to have a negative influence on Atlantic salmon and (2) kelts can have considerable and positive influence on population viability. Our work provides compelling support for the hypothesis that mortality attributable to dam facilities can adversely affect survival, persistence, and recovery of depleted migratory fish populations.