| Abstract: | An important goal for community ecology is the characterization and prediction of changes in community patterns along environmental gradients. We aimed to identify the major environmental correlates of diatom distribution patterns in boreal running waters. We classified 197 stream sites based on their diatom flora. Direct ordination methods were then used to identify the key environmental determinants of this diatom-based stream typology. Finally, we tested whether a regional classification scheme based on terrestrial landscapes (ecoregions) provides a reasonable framework for a regional grouping of streams based on their diatom flora. Two-way indicator species analysis produced 13 site groups, which were primarily separated by chemical variables, mainly conductivity, total P and water colour. In partial CCA, the environmental and spatial factors accounted for 38% and 24%, respectively, of explained variation in community composition. A high proportion (almost 40%) of variation explained by the combined effect (spatially-structured environmental) indicated that diatom communities of boreal streams incorporate a strong spatial component. At the level of subecoregions, classification strength was almost equally strong for all sites as for near-pristine reference sites only. Procrustes analysis indicated that spatial factors and patterns in diatom community structure were strongly concordant. Our data support the argument that diatom communities are strongly spatially structured, with distinctly different communities in different parts of the country. Because of the strong spatial patterns of community composition, bioassessment programs utilising lotic diatoms would clearly benefit from regional stratification. A combination of regional stratification and the prediction of assemblage structure from local environmental features might provide the most robust framework for diatom-based assessment of the biological integrity of boreal streams. |
