Effect of microhabitats, mesohabitats and spatial position on macroinvertebrate communities of a braided river

Abstract Habitat modelling aims to predict changes in the structure of aquatic communities as a function of habitat availability. It is a primary tool to inform management actions and to search for the best compromise between biodiversity conservation and water supply. The construction of these models requires in-depth knowledge about the main hydrological and geomorphological drivers that affect local communities. However, these drivers act at different spatial scales and determining which investigation scale is the best trade-off between model accuracy and model transferability is crucial. The present work aims to evaluate the mesoscale for habitat modelling of aquatic macroinvertebrates, by testing the effect of microhabitat (flow velocity, water depth, substrate), mesohabitat (riffles, glides, backwaters, isolated ponds) and spatial position (coordinates and derived spatial variables) on macroinvertebrate community variability. Multivariate spatial analyses were used to analyse the macroinvertebrate data collected in a braided reach of the Trebbia River (N Italy). Mesohabitat was a good predictor of macroinvertebrate community composition, although both microhabitat and space also showed a significant effect. The outcome of this work highlights the transferability of the results across mesohabitats, which supports the use of mesoscale modelling for macroinvertebrate distribution in braided rivers.