Disturbance and temporal variability in invertebrate assemblages in two South African rivers

Includes abstract.This thesis presents an examination of the relationship between floods as disturbances, the disturbance regime and the temporal dynamics of invertebrate assemblages, over the short term and at intra- and inter-annual time scales in the Molenaars and Berg Rivers in the Western Cape of South Africa. Invertebrate responses to individual floods were investigated by a field study that links the displacement of river-bed stones by a flood to change in invertebrate densities and community and population structure from before to after flood events. The magnitude of the hydraulic force acting on each marked stone during the peak of each flood was also calculated, providing a second measure of physical disturbance. Multivariate analyses of similarity, hierarchical clustering and multidimensional scaling were used for analysis of invertebrate patterns before and after floods. Size frequency data for 28 species or genera were analysed to explore changes in population structure over the flood season. Flood records were developed from the daily discharge hydrological record of both study rivers collected by the South African Department of Water Affairs and Forestry. Flood frequency analysis and descriptive statistics were combined with graphical methods to describe the flood regime and to test flood predictability in these rivers. Quantitative monthly samples of invertebrates from the Molenaars River collected over 17 months were used together with a further 2 ½ years of semi-quantitative monthly data, to identify intra- and inter-annual patterns in communities. Multivariate analysis of community patterns was combined with a range of indices that reflect community persistence and stability over periods longer than one generation. Population dynamics of the common species were also studied. Life history attributes, specifically seasonality of life cycle stages and generation time, were explored using size frequency data from the samples