Density-dependent feedback is recognized as important regulatory mechanisms of population size. Considering the spatial scales over which such feedback operates has advanced our theoretical understanding of metapopulation dynamics. Yet, metapopulation models are rarely fit to time-series data and tend to omit details of the natural history and behavior of long-lived, highly mobile species such as colonial mammals and birds. Seabird metapopulations consist of breeding colonies that are connected across large spatial scales, within a heterogeneous marine environment that is increasingly affected by anthropogenic disturbance. Currently, we know little about the strength and spatial scale of density-dependent regulation and connectivity between colonies. Thus, many important seabird conservation and management decisions rely on outdated assumptions of closed populations that lack density-dependent regulation. We investigated metapopulation dynamics and connectivity in an exemplar seabird species, the Northern gannet (Morus bassanus), using more than a century of census data of breeding colonies distributed across the Northeast Atlantic. We developed and fitted these data to a novel hierarchical Bayesian state-space model, to compare increasingly complex scenarios of metapopulation regulation through lagged, local, regional, and global density dependence, as well as different mechanisms for immigration. Models with conspecific attraction fit the data better than the equipartitioning of immigrants. Considering local and regional density dependence jointly improved model fit slightly, but importantly, future colony size projections based on different mechanistic regulatory scenarios varied widely: a model with local and regional dynamics estimated a lower metapopulation capacity (645,655 Apparently Occupied Site [AOS]) and consequently higher present saturation (63%) than a model with local density dependence (1,367,352 AOS, 34%). Our findings suggest that metapopulation regulation in the gannet is more complex than traditionally assumed, and highlight the importance of using models that consider colony connectivity and regional dynamics for conservation management applications guided by precautionary principles. Our study advances our understanding of metapopulation dynamics in long-lived colonial species and our approach provides a template for the development of metapopulation models for colonially living birds and mammals.

Colonial breeding is widespread among animals. Some, such as eusocial insects, may use agonistic behavior to partition available foraging habitat into mutually exclusive territories; others, such as breeding seabirds, do not. We found that northern gannets, satellite-tracked from 12 neighboring colonies, nonetheless forage in largely mutually exclusive areas and that these colony-specific home ranges are determined by density-dependent competition. This segregation may be enhanced by individual-level public information transfer, leading to cultural evolution and divergence among colonies.

Species with breeding distributions spanning a broad latitudinal range typically experience a correspondingly wide range of environmental conditions, and may also be subject to temporal changes in conditions operating either across their range or more locally. Phenological records are potentially useful for elucidating how species adjust to this spatial and temporal variation, particularly in the context of responses to climate change, and have been widely used in studies of terrestrial birds. In contrast, despite the fact that many marine environments are also changing markedly as a result of climate warming, studies of the phenology of seabirds are comparatively rare. We used data from a wide range of sources to examine trends in breeding time in relation to latitude (49 to 62 degrees N) and year (1980 to 2007) in the northern gannet Morus bassanus at 17 colonies in the eastern Atlantic. We found significant spatial and temporal effects but no significant interaction between latitude and year, indicating a consistent shift in breeding time at gannetries in the central and southern parts of the breeding range between 1980 and 2007. On average, median hatch date became 1.98 d later for every 10 latitude shift north, such that breeding in the Faeroe Islands was 26 d later than in northern France, and breeding was delayed by 0.21 d yr(-1), i.e. 6 d later in 2007 compared to 1980. There was no evidence that the trend towards later breeding was related to temporal changes in either local (late-winter sea surface temperature) or large-scale (winter North Atlantic Oscillation) ocean climate. We believe that this is the first evidence of a trend towards later breeding in a marine predator across a major part of its breeding range. More generally, our findings accord with other recent studies suggesting contrasting phenological responses between marine and terrestrial north temperate birds.