Recolonisation patterns of benthic invertebrates: a field investigation of restored former sewage channels

The response of benthic invertebrate assemblages to stream restoration is the subject of numerous recent studies, but the process of recolonisation of restored river sections is still poorly understood. Field studies are often flawed by uncertainties about whether the observed species colonised the restored sections after restoration or whether they were present before restoration. We investigated restored streams in the Boye catchment in an urbanised area in Germany. The streams were formerly open sewers and thus were not inhabitable by benthic invertebrates, except for Oligochaeta. Restoration included the construction of underground sewers for wastewater and morphological remodelling of the stream bed. In the spring of 2012, we sampled seven restored former channels connected with near‐natural upstream sections, which were never used as sewage channels and are in good status morphologically, and six unconnected restored former channels. Restoration measures had been conducted between one and 19 years before sampling. Additionally, we sampled 21 near‐natural sites within the catchment and eleven near‐natural sites in neighbouring catchments. Near‐natural sites were considered to be potential source sites from which benthic invertebrates might colonise the restored sites. We recorded 128 taxa and categorised them into five dispersal classes that reflect dispersal capabilities and degree of ecological specialisation. Assemblages at restored sites were characterised by lower numbers of taxa and/or high abundance of hololimnic taxa and poorly dispersing winged species and by higher species numbers and abundance of strongly dispersing generalists. A distinct recolonisation sequence was derived from the observed patterns, in which winged, strongly dispersing generalists colonised most rapidly and were followed by hololimnic species, weakly dispersing generalists and habitat specialists. Restored sites connected with near‐natural upstream sections were colonised more rapidly than unconnected restored sites, particularly by habitat specialists. Almost 90% of the recolonisation events originated from sources within a distance of 5 km. We observed a succession from pioneer assemblages to more mature communities, which resembles that of the surrounding near‐natural sites. In conclusion, assemblages in connected, restored sites needed nine to 19 years to reach maturation, while the settlement of assemblages in unconnected sites is expected to require more time.