B. Malmqvist et Po. Hoffsten, Macroinvertebrate taxonomic richness, community structure and nestedness in Swedish streams, ARCH HYDROB, 150(1), 2000, pp. 29-54
Taxonomic richness and community composition were related to environmental
variables in 88 streams in an area of 60,000 km(2) in central Sweden. In al
l, 247 macroinvertebrate taxa were recorded and taxonomic richness observed
per site ranged from 21 to 77. Partial least squares regression analysis s
uggested that taxonomic richness was positively and most strongly correlate
d with channel width and catchment size. Correlations were also positive, b
ut weaker, with the percentage of lakes in the catchment, temperature and m
acrophyte abundance, whereas the degree of shading, moss coverage and dista
nce to upstream lakes showed negative associations with species richness. T
he number of 'rare' (found at <25 % of the sites) and 'very rare' (found at
<10 % of the sites) taxa, showed similar trends although the correlation w
ith distance below upstream lakes was positive rather than negative in the
'very rare' taxa category. The most taxa-rich sites were found in streams o
f intermediate size leading to a significant quadratic relationship between
catchment size and taxonomic richness. Canonical correspondence analysis i
ndicated that catchment area, distance from upstream lakes, substratum, tem
perature, altitude, amount of macrophytes, pH, colour, and Fontinalis spp.
were the variables explaining most variation in community composition. Ephe
meroptera, Plecoptera, Trichoptera and Diptera: Simuliidae showed nested di
stribution patterns, i.e. species-rich sites tended to be inhabited by spec
ies present also at species-poor sites. Some taxa, however, deviated from e
xpected distributions possibly because they are sensitive to biotic interac
tions, restricted to small streams, or specialised with respect to some hab
itat features. In comparisons with other studies of Swedish macroinvertebra
tes, we observed differences suggesting that the design of studies, includi
ng the selection of geographical regions and limited range of stream sizes,
can importantly influence the results.