Relative importance of climate and spatial processes in shaping species composition, functional structure and beta diversity of phytoplankton in a large river.

Although metacommunity dynamics of lentic phytoplankton are relatively well-documented, studies on the role of environmental and spatial processes in shaping phytoplankton communities of large rivers are still scarce. Here, we examined six phytoplankton data sets, which were collected in 1978-2017 from large river-scale segments (mean spatial extent 1117 km) in the Danube River. Our aim was to elucidate role of climatic, spatial and temporal predictors in variation of phytoplankton beta diversity using variance partitioning for compositions of species and functional groups sensu Reynolds. We hypothesised that phytoplankton beta diversity (measured as average distance to group centroid) would be positively related to both climatic heterogeneity and spatial extent used as a proxy for dispersal limitation. Additionally, we tested alternative dispersal models to evaluate different spatial processes structuring phytoplankton community. Our results revealed that spatial variables were more important than climatic factors in controlling both species and functional group composition. Climatic heterogeneity showed significant positive relationship with beta diversity. In contrast, there was no significant relationship between beta diversity and spatial extent, suggesting that spatial effect on beta-diversity was attenuated by anthropogenic disturbance. The better performance of non-directional model compared to model of water directionality suggested that spatial dynamics of phytoplankton metacommunity was in large part regulated by differences in the regional species pools. Spatial and temporal variables outperformed environmental (including climatic) factors in explaining phytoplankton metacommunity structure, indicating that phytoplankton exhibited strong biogeographical patterns. Thus, dispersal limitation interfered with species-sorting processes in determining phytoplankton community structure. In conclusion, our findings revealed that the development of a more reliable bioassessment program of the Danube River should be based on separation into basin regions.