Effects of different intensities of physical disturbance on benthic macroinvertebrates colonizing artificial (brick) substrata were examined in upland (Birthday Creek) and lowland (Gorge Creek) tropical rainforest streams that differed in community structure. The movement of substratum (disturbance) resulting from spates of different magnitudes was simulated by tumbling and scrubbing of bricks. Colonization of undisturbed (control) bricks was examined at both sites. At the upland site, temporal changes in the benthic rock fauna following the summer wet season (December-February) were monitored. Rock samples revealed temporal fluctuations closely correlated with wet and dry seasons, which may reflect food availability and life-history characteristics. Both density and richness of macroinvertebrates were lowest following the wet season (March), with a peak in July. Individual taxa typically reflected this trend, with the exception of Simulium sp., which was abundant in March, possibly as a result of reduced densities of other taxa on the scoured substrata, combined with elevated now levels. Macroinvertebrate densities and colonization rates on bricks were significantly higher at Gorge Creek than at Birthday Creek, with no asymptote reached at Gorge Creek after 70 d exposure. In contrast, faunal densities at Birthday Creek stabilized after 21 d exposure, at levels directly comparable to the natural rock samples. Direct taxonomic comparisons between bricks and rocks at day 34 and day 64 revealed that the fauna colonizing the bricks closely matched natural rock communities, particularly by day 64, when only one taxon (Nousia sp. 2) was more abundant on bricks. Species richness essentially remained constant over the colonization period at both sites. Overall response to disturbance was similar in both streams. Increases in the intensity of physical disturbance typically resulted in a significant reduction in total abundance and, to a lesser extent, reduced diversity on the bricks sampled 24 h later. Responses to disturbance were highly taxon-specific and generally varied with season. The most common response was a decline in density as disturbance intensity increased (e.g., Chironomidae, Helicopsyche, Oecetis), whereas Simulium and Baetis were least affected by the treatments. Recolonization onto disturbed bricks appears to be largely a function of resistance and resilience of individual tara. The effects of disturbance on community structure were more profound at Gorge Creek, which was numerically dominated by two species (Cheumatopsyche sp. and Ferrissia sp;) which were highly vulnerable to disturbance. This result suggests that a natural disturbance of sufficient magnitude to move the substratum may alter community structure at Gorge Creek, but this effect may be short-term due to the high resilience of the fauna (expressed as recolonization rate) compared with Birthday Creek. Increasing the recolonization distance by disturbing a large area (100 m(2)) surrounding the disturbed bricks had no effect on density, richness, or taxonomic composition of the colonizing fauna, suggesting the fauna is well able to recover from this spatial scale of disturbance. Disturbance appears to play an important role as a structuring force of macroinvertebrate faunas of tropical Australian rainforest streams, which may be a direct consequence of the highly active hydrological regime of this region.
