Article révisé par les pairs
Résumé : Recent reports of pollinator declines have stirred interest in investigating the impacts of habitat exploitation on the conservation of pollinator and plant communities. An important prerequisite to tailor conservation action is to understand the drivers and patterns of species-rich communities, and how they change in space and time during a whole season. To account for this, we surveyed wild bees and flowering plants using standardized transects in 11 natural habitat fragments of an IUCN important plant area along the coast of Israel. We used phylogeny- and taxon-based methods of community structure analyses to study the assembly processes of bee communities, and investigated the effects of several landscape parameters on bee diversity using generalized linear models (GLMs). Our results illustrate that natural habitat sites comprised significantly higher species richness compared to disturbed habitat sites, and show that even the smallest habitat fragments harbored unique bee assemblages, with significant species replacement (turnover) found in both space and time. Our GLMs indicated that flower diversity, and semi-natural habitat within 500 m of habitat fragments were important drivers of bee diversity, but we found no evidence for a species—area relationship among sites. Finally, we document a case of phylogenetic overdispersion despite low species richness, which highlights the importance of accounting for phylogenetic diversity rather than only species richness to reach a more fine-grained understanding of pollinator diversity. This, in turn, is pivotal to developing conservation actions to protect these essential pollinators and their interaction with rare and endemic plant species in this highly threatened ecosystem.