Plant-pollinator interactions on green roofs are mediated by substrate characteristics and plant community composition

Abstract Green roofs can support pollinator communities in cities. However, little is known about the influence of green roof characteristics such as substrate and vegetation type on the abundance and diversity of attracted pollinators. Here we aimed to assess how green roof design impacts their attractiveness to pollinators. Using mesocosms on a rooftop in Paris (France), we studied the impact of two substrate types, two substrate depths (10 and 30 cm) and either monocultures or mixtures of 5 plant species on plant pollinator interactions. In the case of mixtures, we also tested the effect of substrate type (natural soil vs. artificial substrate). We counted the number of floral units and recorded the visits by pollinators once a week from mid-June to mid-August. The pollinator assemblage visiting plant communities included 4 functional groups of pollinators: domesticated honey bees, bumble bees, solitary bees and syrphid flies. Effects of treatments on pollinator community composition were variable and plant species dependent. Deep monoculture treatments resulted in the highest number of floral units and visits. Although plants grown on natural soil had less floral units than on artificial substrate, both treatments resulted in a similar number of visits. This paper provides evidence that plant-pollinator interactions on green roofs are modulated by substrate type, substrate depth and plant community. We suggest that combining plant species with diverse flowering morphologies and phenologies can enhance pollinator diversity. When possible, increasing substrate depth can result in higher levels of attractiveness.