Spatial and Temporal Variability of Macroinvertebrate Assemblages in Boreal Streams: Implications for Conservation and Bioassessment

In this thesis, I studied spatial and temporal variability of macroinvertebrate assemblages of boreal streams. The main objectives were (i) to characterize macroinvertebrate assemblage types across large geographical extents, and to assess the utility of assemblage types and landscape and stream type classifications as the basis of stream bioassessment and conservation programs (ii, iii). I also examined the relative roles of large-scale spatial trends and local environmental conditions in structuring macroinvertebrate assemblages (iv). Finally, I assessed (v) if assemblage classifications produce temporally stable and predictable assemblage types. Stream macroinvertebrate assemblage structure exhibited continuous variation instead of distinct assemblage types. Although ecoregions clearly accounted for a considerable amount of variation of macroinvertebrate assemblage characteristics, a combination of regional stratification and prediction from environmental factors would probably yield the most comprehensive framework for the characterizations of macroinvertebrate assemblages of boreal headwater streams. Differences in macroinvertebrate assemblage structure, as well as a group of effective indicator species for different stream types, suggest that landscape classifications could be used as a preliminary scheme for the conservation planning of running waters The strength of the relationship between assemblage structure and local environmental variables increased with decreasing extent, whereas assemblage variation related to spatially variables showed the opposite pattern. At the largest scale, spatial variation was related to latitudinal gradients, while spatial autocorrelation among neighbouring streams was the likely mechanism creating spatial structure within drainage systems. These results suggest that stream bioassessment should give due attention to spatial structuring of stream assemblage composition, considering that important assemblage gradients may not only be related to local environmental factors, but also to biogeographical constraints and neighbourhood dispersal processes. The classification strengths of macroinvertebrate assemblages based on data of three years were overall rather weak, and more importantly, the compositions of the site groups varied considerably from year to year. Such wide and continuous variation was also mirrored by low and inconsistent predictability of classifications from environmental variables. The observed level of temporal variation in assemblage structure may not be a serious problem for predictive approaches frequently used in bioassessment of freshwater ecosystems. For conservation purposes, however, alternative approaches (e.g. physical surrogates of biodiversity) need to be considered.