Abstract
Forecasting population responses to rapidly changing marine ecosystems requires mechanistic models integrating complex demographic processes, fitted to long time series, across large spatial scales. We used a Bayesian metapopulation model fit to colony census data and climatic covariates spanning 1900-2100 for all Northeast Atlantic colonies of an exemplar seabird, the Northern gannet (Morus bassanus) to investigate metapopulation dynamics under two climate scenarios. Fecundity varied non-linearly with near-surface air temperature and recruitment was depressed by sea surface temperature. We predict regime changes in density dependence as marine carrying capacities become constrained with increasing SST. Sensitivity to climate change varied across space and time, disadvantaging southwestern colonies whilst benefitting northern ones. Such sensitivity is noteworthy for a species previously assumed robust to climate change. We provide a spatial overview of climate sensitivities across a metapopulation to help with evidence-based conservation management and open the way for similar mechanistic explorations for other colonial species.