Background As for other life history traits, variation occurs in movement patterns with important impacts on population demography and community interactions. Individuals can show variation in the extent of seasonal movement (or migration) or can change migratory routes among years. Internal factors, such as age or body condition, may strongly influence changes in movement patterns. Indeed, young individuals often tend to move across larger spatial scales compared to adults, but relatively few studies have investigated the proximate and ultimate factors driving such variation. This is particularly the case for seabirds in which the sub-adult period is long and difficult to follow. Here, we examine migration variation and the factors that affect it in a common Mediterranean seabird, the Yellowlegged gull (Larus michahellis). Methods The data include the encounter histories of 5158 birds marked as fledglings between 1999 and 2004 at 14 different colonies in southern France and resighted over 10 years. Using a multi-event mark-recapture modeling framework, we use these data to estimate the probability of movement and survival, taking into account recapture heterogeneity and age. Results In accordance with previous studies, we find that young individuals have greater mobility than older individuals. However, the spatial extent of juvenile movements depends on natal colony location, with a strong difference in the proportion of sedentary individuals among colonies less than 50 km apart. Colony quality or local population dynamics may explain these differences. Indeed, young birds from colonies with strong juvenile survival probabilities (~ 0.75) appear to be more sedentary than those from colonies with low survival probabilities (~ 0.36). Conclusions This study shows the importance of studying individuals of different ages and from different colonies when trying to understand seabird movement strategies. Local breeding success and the availability of food resources may explain part of the among colony differences we observe and require explicit testing. We discuss our results with respect to the feedback loop that may occur between breeding success and mobility, and its potential implications for population demography and the dissemination of avian disease at different spatial scales.