Résumé
Evidence is accumulating on the crucial role of spatial demographic processes, particularly immigration, in driving the viability of wild animal populations. However, the contribution of immigration is rarely accounted for when performing population viability analysis and evaluating the efficiency of conservation actions, which can lead to spurious inferences. Here we analysed a 24-year monitoring scheme conducted on the French population of Bonelli's eagle, a population considered as vulnerable and targeted by two recent conservation plans. This population is located at the northern range-margin of the species' range and has suffered a long-standing decline before recovering recently thanks to efficient conservation actions (mainly retrofitting of power lines). We combined all available demographic information sources into an Integrated Population Model (IPM) for quantifying the relative contribution of immigration and local demography to the overall population dynamics. Population projections derived from local demography strikingly differed from the observed population trajectory. Throughout the study period, a sustained immigration rate (0.108 ± 0.03 immigrant females per occupied territory each year) first prevented the population to go extinct and then supported the recovery. The number of immigrants was about 1.0-1.5 times higher than the number of local recruits and we found only a weak evidence for a reduction in this ratio as survival rates increased. The fact that local conservation actions drove the recovery in mitigating anthropogenic perturbations revealed that range-margin populations should not be systematically considered as sink populations. Instead, demographic surveys should be promoted to assess the relative contribution of immigration and local demography alongside species' range (from centre to periphery) so as to improve our understanding of spatial demographic processes and ensure conservation actions are implemented at an appropriate spatial scale.