Aim (1) To resolve theoretical debates regarding the role of environment versus dispersal limitation, the conservatism of niches across distances and the prevalence of environmental specialists in diatom communities. (2) To provide guidance on the use of diatom communities and other microbial analogues to analyse ecological response to environmental change.
Location Eight hundred and ninety-one lakes in five regional datasets from north-western Europe and four regional datasets from north-eastern North America.
Methods Lacustrine diatom communities were analysed at three scales: inter-continental, intra-continental and regional. Nested partial redundancy analyses (RDAs) were used to determine spatial versus environmental components of community variation. Weighted-averaging (WA) regression and calibration, as well as logistic and quadratic regressions, were used to detect niche conservatism and the prevalence of environmental specialists.
Results Community patterns indicate that dispersal limitation acts predominantly at the inter-continental scale, while at the regional (less than c. 1,000,000 km2) scale, a single environmental variable (pH) explains more than five times the community variation as spatial (dispersal-related) variables. In addition, pH niche components appear to be conserved at the inter-continental scale, and environmental specialization does not impose relative rarity, as specialists apparently readily disperse to suitable environments.
Main conclusions Analysis at multiple scales is clearly important in determining the influences of community variation. For diatom communities, dispersal limitation acts most strongly at the broadest scales, giving way to environment at the scales considered by most analyses. The availability of a wide variety of propagules with consistent niches across regions indicates that diatom communities reflect the succession of taxa according to local environmental conditions, rather than disequilibrium with the environment or adaptation of local populations. While multi-scale analyses must be undertaken for other groups to resolve debates over community drivers and determine appropriate scales for prediction, for diatoms (and probably other microbial communities), responses to environmental change can be inferred using analogue datasets from large geographic areas.