Scaling the effects of predation and disturbance in a patchy environment.

The effects of hydraulic disturbances on the impact of two predatory benthic invertebrates on their prey were examined in a stream at two distinct spatial scales. At the scale of small habitat patches (0.0625 m), hydraulic patch type was an important determinant of the microdistribution of prey and predators. Prey abundances were similar across all patch types at baseflow, but local densities were higher in patches identified as low-flow refugia after periods of high and fluctuating flow. The microdistribution pattern of predatory larvae of a caddisfly, Plectrocnemia conspersa, was similar to that of its prey, whereas predatory larvae of an alderfly, Sialis fuliginosa, did not shift their microdistribution significantly with discharge and were always most abundant in lowflow refugia. There was little evidence of an aggregative response of predators with prey, even though both predators and prey are mobile. Both predator species showed similar patch-specific patterns of per capita consumption rates: uniform consumption rates across hydraulic patch types at low and moderate flows, but highest in flow refugia during high flows. Species-specific patterns, however, were apparent in the magnitude and direction of differences between consumption rates during disturbance events, and in comparable patches at base flow: At high flow, consumption rates for P. conspersa were exaggerated (3.9 times higher) in flow refugia but "at par" in other patches; for S. fuliginosa they were "at par" in flow refugia but reduced in other patches (up to 3.3. times lower). These differences may be related to species-specific foraging behaviours (search vs ambush predators) and the influence of prey movements on feeding success. Using the patch-scale results only, it is difficult to predict the effects of physical disturbance on predation intensity at the larger scales of whole habitats, populations or communities. At the large scale (>200 m), net predator impacts were estimated over the stream reach, using a spatially explicit model that accounts, in an additive way, for habitat heterogeneity and patch-specific responses of predators and prey. The relationship between predator impact over the whole reach and hydraulic disturbance differed for the two predators. The predator impact of S. fuliginosa decreased with increasing hydraulic disturbance, as predicted by the harsh-benign hypothesis. There was no directional trend for P. conspersa, however, and maximum predator impact may occur at intermediate disturbance levels. For the prey community in this stream, predation pressure from S. fuliginosa appears to fluctuate directly with the discharge hydrograph, whereas predation from P. conspersa may be more persistent. Flow refugia may play a dual role in the sructure of stream communities by preventing catastrophic mortality of animals (predators and prey) from physical forces during disturbances, and by maintaining (or perhaps increasing) predation pressure. Summing the effects of species interactions in small habitat patches to the larger scale of a whole stream reach indicates that the scale of approach influences the observed patterns and their implied underlying process.