Annual changes in summer patterns of mesohabitat distribution and associated macroinvertebrate assemblages

Abstract

The summer distribution of distinct mesohabitats (gravel, sand, silt, Ranunculus beds and emergent vegetation patches) at four 100-m sections along the length of a lowland river was studied over 3 years, together with their associated faunal assemblages. Habitat proportions in the sections showed little change between 1995 and 1997 at the uppermost site, in contrast to major changes downstream, which included an increase in percentage cover of Ranunculus from 10% in 1995 to 50% in 1996. There did not appear to be a significant relationship between summer or winter discharge and the proportions of mesohabitat, but there were indications that stream macrophyte areas were high when the preceding winter flows were low. Multivariate analysis of the faunal/site matrix showed that the principal source of variation was the mesohabitat composition. Superimposed over this was a site component where the uppermost site was separated from the intermediate sites, which in turn were separated from the most downstream site. Annual changes in faunal composition appeared to be the least significant overall but this varied with site, with little difference at the uppermost site and clear differences between 1995 and the other two years at the other sites. Ranunculus beds showed the least change between years. The highest number of taxa was found in emergent or marginal vegetation at all sites, with Ranunculus and sand supporting the lowest number. Mean abundance was greatest in Ranunculus and gravel samples, intermediate in emergent vegetation and silt, and lowest in sand. In order to investigate the implications of habitat changes, mean abundance of taxa per mesohabitat per year was compared with total abundance in the 100-m section for which habitat information was available. The total mean abundance of macroinvertebrates per mesohabitat was greatest in 1995 but this is not reflected in the total abundance per section. Mesohabitat abundances were greatest in Ranunculus, silt and emergent vegetation in 1995 but these mesohabitats were poorly represented in the section. In 1996, Ranunculus increased in area and this was reflected in higher total abundance. In 1997, the mean number of animals per mesohabitat was less than in previous years and this resulted in low overall abundance for the reach. Thus, the changes in habitat proportions controlled annual changes in total reach abundance and composition. Such changes have implications for the interpretation of biomonitoring and river habitat survey results whereby misleading conclusions may be drawn without considering annual and spatial fluctuations in habitat proportions. Copyright © 2000 John Wiley & Sons, Ltd.