An assessment of macroinvertebrate functional feeding groups as water quality indicators in the Buffalo River, eastern Cape Province, South Africa

The aim of this paper was to investigate the potential for using functional feeding groups (FFGs) as indicators of water quality conditions in rivers, using the Buffalo River, South Africa, as a specific example. Multivariate classification and ordination techniques were used to investigate species and FFG distributions in relation to a number of physico-chemical variables at 16 sites from the headwaters to the estuary of the Buffalo River.Two-way indicator species analysis (TWINSPAN) of species composition ranked most of the sites sequentially down the river, irrespective of water quality conditions. Ordination of FFGs from a set of riffle samples collected in mid-late summer showed only weak relationships between FFG distribution and water quality changes, except where variables changed sequentially down the river (e.g. pH and temperature). Individual species responses to water quality gradients were examined for nine riffle-dwelling species representing diverse FFGs. Following correspondence analysis of a matrix of environmental variables and species frequencies, some species showed strong associations with defined ranges of some variables. In particular, Adenophlebia auriculata (Leptophlebiidae, Ephemeroptera) from the headwater sampling site, was associated with low pH and low temperature. Simulium damnosum occurred under conditions of high turbidity, while Afronurus harrisoni was found under high concentrations of potassium, ammonium and nitrite ions.We conclude that although there was a distinct headwaters fauna in the Buffalo River, and sequential downstream changes in species composition, most FFGs (apart from shredders) were represented down the whole length of the river. FFG classifications are therefore unlikely to provide useful indications of water quality conditions in the Buffalo River.Using a categorical approach to classifying water quality variables, and by applying correspondence analysis to the resulting matrix, we recognised nine species that could be used to define water quality. These indicator species can be used to define tolerance ranges of the fauna for water quality conditions in different parts of the Buffalo river.