Scale dependency in fish beta diversity–hydrology linkages in lowland rivers

Aim: A key aspect of biodiversity research is to determine the environmental drivers affecting the degree to which ecological assemblages vary in space (beta diversity). The strength and significance of environmental drivers of beta diversity is, however, influenced by the spatial extent over which beta diversity is assessed. Beta diversity of riverine biota is affected by flow variability. We examined hydrology–beta diversity relationships at two spatial extents (reaches vs. entire systems) within rivers to determine if relationships with hydrological variables are scale dependent or generalisable across spatial scales.

Location: Murray–Darling Basin, Australia.

Taxon: Freshwater fish.

Methods: Fish assemblage data were sourced from two monitoring programmes that sampled sites annually during 2014–2019 and spanned either reaches (<100km) or the broader 'system' extent (>1000km) of five tributaries. Beta diversity was examined by calculating pairwise (incidence and abundance) dissimilarities for each year to compare temporal trends in beta diversity. Multi-site dissimilarities were modelled against hydrological variables using beta regression.

Results: Inter-annual change in assemblage composition was detected only at the extent of river reaches but not at the extent of river systems. Temporal variation in within-river beta diversity showed inconsistent patterns when compared between the two spatial extents. Within-river beta diversity relationships with hydrological gradients were inconsistent among rivers. Overall, statistical models explained much more variation in within-river beta diversity when assessed at the reach extent when compared to broader river system extents.

Main Conclusions: Our findings highlight that changes in within-river beta diversity are likely to depend on the spatial extent of sampling. Furthermore, inconsistent beta diversity–hydrology relationships among rivers suggest that both empirical evidence and theoretical predictions adopted in ecohydrology may not be transferable among river systems.