Résumé
The dynamics of spatially structured populations (SSPs) depend on dispersal movements between habitat patches. Individual‐level studies show that dispersal can be informed by social cues about habitat quality, such as breeding success or the number of conspecifics in a patch. However, informed dispersal has rarely been considered in studies investigating SSP dynamics. In particular, little is known about how patch dynamics are influenced by characteristics of neighboring patches, and the spatial scale of this influence. Here, we used occupancy and growth rate models to investigate the effect of breeding success and numbers in all patches of an SSP of black‐headed gulls ( Chroicocephalus ridibundus ) on patch colonization, persistence and growth rates. The spatial scale of influence of neighboring patches was estimated using a distance decay function weighting the effect of patch covariates. Our results showed a strong influence of breeding success and breeding numbers, both in the focal patch and in neighboring patches, on patch dynamics. Persistence was higher for patches occupied by large and successful colonies, suggesting high attractiveness of such colonies, although their growth rate was reduced, probably due to increased costs of group‐living. Patch colonization rates were higher within a few kilometers from patches occupied by large and successful colonies, in accordance with an attraction of dispersers to areas showing favorable breeding conditions. The breeding success of neighboring patches from the entire SSP reduced patch persistence, but only for patches that failed to produce chicks, and negatively affected the growth rate of persisting colonies, consistent with individuals moving away towards higher quality habitats. These results advocate for the use of social information in black‐headed gull dispersal decisions. However, we were unable to distinguish between the use of breeding success (public information) and breeding numbers (conspecific attraction), which were highly correlated. Our findings emphasize the need to consider neighboring patches as dynamic components of the landscape that actively shape dispersal decisions, with consequences on SSP dynamics.