Processes underpinning fish species composition patterns in estuarine ecosystems worldwide

Aim: We present the first global biogeographical regionalization of estuaries, assessing how dispersal limitation and/or environmental filtering mechanisms drive the patterns of fish assemblage composition among and within biogeographical regions. Location: Estuaries worldwide. Methods: A bootstrapped hierarchical cluster analysis was applied to define biogeographical regions based on pairwise beta diversity (βsim) of fish assemblages among 393 estuaries worldwide. Variables representing dispersal limitation and environmental features were used to disentangle the possible effects of assembly processes acting among and within biogeographical regions, through hierarchical partitioning of variation. Finally, we assessed the contribution of species to beta diversity through a method newly developed in this article. Results: Estuaries were grouped into seven major biogeographical regions explained by geographical dispersal variables. Dissimilarity in species composition within each biogeographical region was mainly determined by ecosystemconnectivity dispersal variables, with a minor influence of environmental variables (chiefly sea surface temperature). Marine species contributed highly to beta diversity between estuaries but less than expected by chance, whilst freshwater species had low contributions yet also less than expected. Estuarine (i.e. brackish) and diadromous species had low contributions to beta diversity but higher than expected by chance. Main conclusions: Fish assemblage composition in estuaries seems to be driven by dispersal limitation processes both among and within biogeographical regions, with only a minor role of environmental filtering evident within biogeographical regions. These findings contrast with the acknowledged importance of environmental variables as drivers of species richness patterns. Irrespective of biogeographical region, the ecosystem affinity of a species determined its contribution to beta diversity between estuaries. The new method presented here offers new opportunities for assessing the effects of taxonomic and functional aspects on beta diversity (βsim) across taxa and ecosystems, and thus contributes to improving knowledge of assembly processes.