Wild bees are declining, mainly due to the expansion of urban habitats that have led to land-use changes. Effects of urbanization on wild bee communities are still unclear, as shown by contrasting reports on their species and functional diversities in urban habitats. To address this current controversy, we built a large dataset, merging 16 surveys carried out in 3 countries of Western Europe during the past decades, and tested whether urbanization influences local wild bee taxonomic and functional community composition. These surveys encompassed a range of urbanization levels, that were quantified using two complementary metrics: the proportion of impervious surfaces and the human population density. Urban expansion, when measured as a proportion of impervious surfaces, but not as human population density, was significantly and negatively correlated with wild bee community species richness. Taxonomic dissimilarity of the bee community was independent of both urbanization metrics. However, occurrence rates of functional traits revealed significant differences between lightly and highly urbanized communities, for both urbanization metrics. With higher human population density, probabilities of occurrence of above-ground nesters, generalist and small species increased. With higher soil sealing, probabilities of occurrence of above-ground nesters, generalists and social bees increased as well. Overall, these results, based on a large European dataset, suggest that urbanization can have negative impacts on wild bee diversity. They further identify some traits favored in urban environments, showing that several wild bee species can thrive in cities.

The response of insect pollinator communities to increasing urbanization is shaped by landscape and local factors. But what about habitats that are already highly artificial? We investigated the drivers of pollinator diversity in a dense urban matrix, the city of Paris. We monitored insect pollinator communities monthly (March-October) for two consecutive years in 12 green spaces that differed in their management practices, focusing on four insect orders (Hymenoptera, Diptera, Lepidoptera, Coleoptera). Pollinator abundance and species richness were both positively tied to green space size and flowering plant species richness, but negatively linked to surrounding impervious surfaces. In addition, environmental features at both the local and landscape scales influenced the composition and functional diversity of wild bee communities. Indeed, small and large bees responded differently, with the occurrence of large-bodied species being impaired by the proportion of impervious surfaces but strongly enhanced by plant species richness. Also, sites with a majority of spontaneous plant species had more functionally diverse bee communities, with oligolectic species more likely to be found.These results, consistent with the literature, can guide the design and management practices of urban green spaces to promote pollinator diversity and pollination function, even in dense urban environments.