Abstract Green roofs are increasingly promoted for urban biodiversity conservation, but the value of these novel habitats is uncertain. We aimed to test two hypotheses: (i) green roofs can support comparable invertebrate family and order richness, composition and abundances to ground-level habitats and (ii) green roofs planted with native species from local habitats will support a richer invertebrate community at family and order level than other green roofs. We sampled the invertebrate community on green roofs dominated by native grassland or introduced succulent species in Melbourne, Australia, and compared these to the invertebrate community in ground-level sites close by, and sites with similar vegetation types. The only significant differences between the invertebrate communities sampled on green roofs and ground-level habitats were total abundance and fly family richness, which were higher in ground-level habitats. Second hypothesis was not supported as invertebrate communities on green roofs supporting a local vegetation community and those planted with introduced Sedum and other succulents were not detectably different at family level. The per cent cover of green space surrounding each site was consistently important in predicting the richness and abundance of the invertebrate families we focussed on, while roof height, site age and size were influential for some taxa. Our results suggest that invertebrate communities of green roofs in Melbourne are driven largely by their surrounding environment and consequently the effectiveness of green roofs as invertebrate habitat is highly dependent on location and their horizontal and vertical connection to other habitats.
Read full abstract