Many studies have shown that the quality of biodiversity influences the well-being of citizens. Nevertheless, the drivers that shape biodiversity in urbanized zones are poorly understood. Although tree bases present reduced surface areas, they occur in great numbers in a deliberate spatial arrangement and may play an important ecological role in urban environments by offering limited favorable spaces for the development of spontaneous flora. The purpose of this study was to identify the factors that influence the composition of plant communities harbored by tree bases in an urban district. We analyzed floristic inventory data collected in 2014 about plants growing at the bases of the 1474 trees on the 26 streets of the Bercy district in Paris (France). Our results indicated that the plant communities growing in the urban tree bases varied according to different factors. The abundance and distribution of these species were dependent on their biological traits (seed longevity in the soil bank) as well as the tree base characteristics (tree trunk diameter, equipment type around the tree bases, soil compaction, animal excrement, solar radiation, and urban tree species), the street orientation according to the air flow following the Seine River, and the geographic structure of the district (the influence of the presence of green spaces). The results of this study showed that the tree bases could be considered favorable stepping-stone habitats for certain species between more important green spaces such as parks and gardens. Thus, these areas actively participate in the enhancement of urban biodiversity.

In cities, trees planted along streets could play an important ecological role for spontaneous plants growing at their bases. For example, these trees could represent corridors by potentially connecting large green spaces (e.g., parks, gardens), which allow species to move within the urban matrix. We considered sets of urban trees in 15 streets in Paris, France, as metapopulations for 15 plant species. Our objective was to determine the factors influencing the dynamics of colonization and extinction of populations based on the distance of the streets to green spaces and biological traits of each species. Plant species in 1,324 tree bases of the Bercy District of Paris were surveyed annually from 2009 to 2015. For each species and each street, we used SPOMSIM software to identify the best‐fit metapopulation model between four models with different colonization and extinction functions: propagule rain model (PRM) and Levins’ model with or without rescue effect. Results demonstrated that species more often conformed to the PRM in streets near green spaces, which suggested that green spaces could act as sources for the populations in those streets. Species with seeds with long‐term persistence more often conformed to the PRM, indicating that a soil seed bank helps species invade entire streets. Finally, a higher percentage of species with a short height conformed to models with a rescue effect, which indicated that those small species resisted the effects of weeding by the city technical services better than taller species. Synthesis and applications. This study showed how biological traits of species and geography of the district determine the dynamics of plants in the streets, and these results may provide important information for biodiversity management in cities.