Unravelling direct and indirect effects of hierarchical factors driving microbial stream communities

Abstract

AbstractAimThe diversity and distributions of organisms are jointly influenced by local environment and large‐scale variables, such as land cover patterns, dispersal processes and climate. However, the drivers of microbial diversity are complex and the pathways of these drivers’ effects are to date largely unresolved, especially in freshwaters. We disentangled the causal direct and indirect effects of climate, land cover and water chemistry on stream diatom communities using hierarchical settings.LocationFinland, a geographical gradient of c. 1200 km (60°–70°N).MethodsWe used structural equation modelling (SEM) to analyse patterns in diatom richness, composition and the uniqueness of species composition in 143 streams. The community composition was defined as the value of the first axis in non‐metric multidimensional scaling and the uniqueness of species composition of each site as its local contribution to beta diversity.ResultsSpecies richness was mainly affected directly by energy and resource availability, increasing with nutrients but decreasing with growing degree days (GDD). The community composition was strongly influenced directly by conductivity and mainly indirectly by anthropogenic land use but also to a lesser degree by GDD (directly and indirectly), indirectly by precipitation and directly by the amount of boreal wetlands. The uniqueness of species composition increased with conductivity but decreased with nutrient concentrations with high number of unique species in southernmost and northernmost sites.Main conclusionsOur SEM analyses revealed some of the important links between climate, land cover and water chemistry, all of which influenced the microbial diversity. Energy availability has varying indirect and direct effects on diatom communities regardless of local stream conditions, whereas land cover patterns affect aquatic communities most likely indirectly through water chemistry. Collectively, these results suggest that it is important to consider environmental variables simultaneously at different hierarchically ordered scales in macroecological and biogeographical research.