Résumé
Although it is today accepted that population viability analyses are needed at a meta-population level for most species, usually only single populations are monitored in the context of management and conservation programmes. This paper outlines a fairly general and easy-to-implement approach based on counts and capture-recapture data that allow the dynamics of single populations to be assessed even when they are highly connected to other populations. This approach was motivated by a study of the French population of the slender-billed gull Larus genei, which experienced a sharp population increase in the 1980s and 1990s, suggesting that the species was not at risk of extinction. However, several recently raised concerns indicate that the observed population increase is unlikely to have been achieved uniquely by an intrinsic growth rate. We estimated local adult survival probability at 0.81 (0.79-0.83), which is considerably lower than that of other gull species of comparable size. Moreover, local fecundity observed in slender-billed gulls [0.66 (0.47-0.85)] is lower than that observed in similar species. Massive reproduction failures and the low demographic parameters observed could be caused by chick exposure to aerial and terrestrial predation, leading to permanent emigration. Unrealistically high demographic parameter values would be needed to generate the observed local population increase. The results of our study indicate that connections with other neighbouring populations are responsible for the local population dynamics, and that about 10% of the individuals may be immigrants into this local population annually. However, our results suggest that the population of the slender-billed gull may be stable at the west Mediterranean scale. The high annual fluctuations of population size observed at a larger scale also highlight the necessity for coordinated international action to protect a maximum of potential breeding sites in order to protect the species.