Environmental disturbances, both natural and anthropogenic, have the capacity to substantially impact animal behavior and abundance, which can in turn influence patterns of genetic diversity and gene flow. However, little empirical information is available on the nature and degree of such changes due to the relative rarity of longitudinal genetic sampling of wild populations at appropriate intervals. Addressing this knowledge gap is therefore of interest to evolutionary biologists, policy makers, and managers. In the past half century, populations of the brown pelican (Pelecanus occidentalis) in the southeastern United States have been exposed to regional extirpations, translocations, colony losses, and oil spills, but potential impacts on genetic diversity and population structure remain unknown. To investigate the cumulative impacts of recent disturbances and management actions, we analyzed seven microsatellite loci using genetic samples collected from 540 nestlings across twelve pelican colonies from two time periods, corresponding to before (n = 305) and after (n = 235) the 2010 Deepwater Horizon oil spill. Pre-2010 populations in Texas were significantly differentiated from Louisiana, Alabama, and Florida populations to the east, with reintroduced populations in southeastern Louisiana having less genetic diversity than sites in Texas, consistent with a recent bottleneck. In contrast, there was no evidence of a geographic component to genetic structure among colonies sampled after the spill, consistent with increased dispersal among sites following the event. This pattern may be associated with reduced philopatry in response to colony abandonment in the areas most heavily impacted by the Deepwater Horizon event, though other factors (e.g., rehabilitation and translocation of oiled birds or colony loss due to erosion and tropical storms) were likely also involved. Future monitoring is necessary to determine if bottlenecks and loss of genetic variation are associated with the oil spill over time, and is recommended for other systems in which disturbance effects may be inferred via repeated genetic sampling.