Correspondence between regional delineations and spatial patterns in macroinvertebrate assemblages of boreal headwater streams

Geographical stratification may provide a useful framework for stream management programs, yet most studies testing the utility of such stratifications have been conducted in temperate regions. We studied the correspondence between regional delineations (5 ecoregions, 11 subecoregions), environmental characteristics, and benthic macroinvertebrate assemblages in 156 boreal headwater streams in Finland, using a combination of principal components analysis, nonmetric multidimensional scaling, and discriminant function analysis (DFA). Both stream characteristics and macroinvertebrate assemblage structure showed a closer correspondence to ecoregions than to subecoregions, a pattern partly attributable to the smaller number of classes at the ecoregion level. Based on environmental variables, DFA classified correctly 84.0% of the streams by ecoregion and 74.4% by subecoregion. For macroinvertebrate assemblages, 64.7% of the streams were correctly classified at the ecoregion, and 51.9% at the subecoregion level. Arctic-alpine and north boreal ecoregions supported the most distinct benthic assemblages. There were significant differences among ecoregions in mean species number, with lowest species richness in the middle boreal and highest in the south boreal streams. However, no clear latitudinal gradients in local species richness were detected. Relative abundances of functional-feeding groups also exhibited regional differences. Shredders dominated the middle boreal brown-water streams, whereas scrapers attained their highest abundances in the arctic-alpine and north boreal clear-water streams. Despite some congruence among ecoregions and macroinvertebrate assemblages, regional delineations alone do not provide a suitable background for stream biomonitoring programs. A combination of regional delineations and local characteristics probably provides a better framework for the classification of boreal headwater streams, and for the prediction of their benthic fauna.