The influence of environmental, biotic and spatial factors on diatom metacommunity structure in Swedish headwater streams.

Stream assemblages are structured by a combination of local (environmental filtering and biotic interactions) and regional factors (e.g., dispersal related processes). The relative importance of environmental and spatial (i.e., regional) factors structuring stream assemblages has been frequently assessed in previous large-scale studies, but biotic predictors (potentially reflecting local biotic interactions) have rarely been included. Diatoms may be useful for studying the effect of trophic interactions on community structure since: (1) a majority of experimental studies shows significant grazing effects on diatom species composition, and (2) assemblages can be divided into guilds that have different susceptibility to grazing. We used a dataset from boreal headwater streams in south-central Sweden (covering a spatial extent of ∼14000 km(2)), which included information about diatom taxonomic composition, abundance of invertebrate grazers (biotic factor), environmental (physicochemical) and spatial factors (obtained through spatial eigenfunction analyses). We assessed the relative importance of environmental, biotic, and spatial factors structuring diatom assemblages, and performed separate analyses on different diatom guilds. Our results showed that the diatom assemblages were mainly structured by environmental factors. However, unique spatial and biological gradients, specific to different guilds and unrelated to each other, were also evident. We conclude that biological predictors, in combination with environmental and spatial variables, can reveal a more complete picture of the local vs. regional control of species assemblages in lotic environments. Biotic factors should therefore not be overlooked in applied research since they can capture additional local control and therefore increase accuracy and performance of predictive models. The inclusion of biotic predictors did, however, not significantly influence the unique fraction explained by spatial factors, which suggests low bias in previous assessments of unique regional control of stream assemblages