Piovesan, A.; Praz, C.; Voelkl, B.; Lanz, S.; Neumann, P.; Beaureapaire, A.

Pollinator populations are facing worldwide declines, underscoring conservation needs. Yet, conservation assessments still mostly rely on occurrence data, often derived from heterogeneous and opportunistic observations. While such data can inform on species presence and distribution, they may overlook important markers of population declines. This is particularly problematic for social species such as bumble bees, which typically exhibit low effective population sizes despite high abundance of workers observed in the field. Despite these putative pitfalls, the relationship between occurrence-based and genetic-based estimates remains largely unexplored in social bees. We here investigated spatio-temporal genetic patterns in five Swiss <Bombus> species representing contrasting population trajectories over the last century: <B. humilis> and <B. sylvarum> (stable), <B. ruderatus> (increasing), <B. pomorum> (regionally extinct), and <B. veteranus> (declining). Museum specimens collected between 1929 and 2023 were genotyped at 11 microsatellite loci to compare spatio-temporal fluctuations in genetic diversity and population structure with occurrence data. Overall, multilocus heterozygosity and allelic richness remained stable in all species during the time period investigated, indicating that the diverging population trends did not result in substantial variation of genetic diversity. In contrast, strong and significant shifts in allelic frequencies between time periods were detected in three species, suggesting recent immigration events. Isolation by distance was detected in the cold-adapted <B. veteranus>, while the extant warm-adapted species (<B. humilis>, <B. sylvarum>, <B. ruderatus>) showed high levels of gene flow between locations. In <B. pomorum>, increasing genetic homogenization was observed before extinction. Altogether, these findings show that genetic diversity indexes are not the most adapted tools to monitor conservation status of social bee populations, and that estimates of population structure such as allelic shifts may be more informative. Moreover, these results highlight the importance of monitoring metapopulation dynamics and ensuring connectivity among populations to facilitate gene flow and enable demographic rescue processes.