The effect of species composition on community responses to toxicants : a comparison of species sensitivity distribution curves and mesocosm studies.

Different species are affected at different concentrations of a toxicant. As communities differ in their species composition, they are also likely to be affected at different exposure concentrations. The concentration at which a community will be affected may be extrapolated from single-species tests or determined from multispecies exposures. This work compares the ability of two methods to detect the effect of species composition on communities’ responses to toxicant exposure. The sensitivity of 15 indigenous freshwater macro invertebrate taxa to a surfactant were experimentally determined in single-species toxicity tests and supplemented with published data. The sensitivity of indigenous macroinverterate taxa to the surfactant varied by four orders of magnitude. Half of the indigenous taxa tested were more sensitive than the standard test species Daphrtia magna. The single-species toxicity data were then used to generate species sensitivity distribution (SSD) curves. SSD curves are used to estimate the concentration of a toxicant that is hazardous to 5% of species (HC5), based on the assumption that communities will be able to compensate for this small level of effect. SSD curves were generated for 60 naturally occurring macroinvertebrate communities, all from low order, circumneutral streams. The SSD curves were found to be sensitive to changes in species composition, with community HC5 values differing by more than an order of magnitude. An HC5 value was also calculated using the complete dataset. This value was lower than the HC5 values derived for 92% of natural communities. To test the validity of these HC5 values, the responses of structurally distinct communities to surfactant exposure were determined under identical test conditions. Prior to running this experiment it was necessary to establish whether structurally distinct communities would remain distinct over time in the stream mesocosms. Three structurally distinct macroinvertebrate communities were used to colonise a stream mesocosm, and changes in community composition and structure were monitored over ten weeks. At the start of the study, there was considerable overlap in the species composition of the communities, providing the potential for the three communities to become more similar over time, converging to a ‘generic mesocosm community’. In fact, the commuities remained remarkably distinct from one another for the duration of the study. Two structurally distinct communities were then exposed to the surfactant in the stream mesocosms for 28 days. The community dominated by sensitive taxa was found to be more sensitive, as predicted, with changes in species richness and community diversity occurring at different exposure concentrations in the two communities. The two communities exposed to linear alkylbenzene sulphonate (LAS) in the experimental stream study were structurally distinct, but their species composition was very similar. As these curves are generated from presence absence data they were unable to predict the differences in sensitivity of the two communities. However, the HC5 values determined from the SSD curves were found to be protective or overprotective of these communities, based on the responses observed in the mesocosm study.