Dispersal, Boundary Processes, and Trophic‐Level Interactions in Streams Adjacent to Beaver Ponds

Abstract

I combined long—term (10 yr) descriptive and short—term experimental studies in a headwater stream in northern Minnesota to assess: (1) the effect of annual variation in stream discharge and spatial proximity of beaver (Castor canadensis) ponds on lotic fish abundance and (2) the subsequent influence of discharge and fish predation on lotic invertebrate colonization. Considerable annual variation in fish density occurred in the stream over the 10—yr period, particularly in pool habitats. Increased fish density was associated with increased stream discharge and creation of beaver ponds downstream from the study site. Wiegert traps used to monitor directional (upstream vs. downstream) fish movement during the last 4 yr of the study indicated annual changes in fish density were associated with the amount of fish dispersal occurring along the stream segment. Downstream fish movement, out of an upstream beaver pond occurred primarily during periods of elevated stream discharge. Upstream movement, out of a downstream beaver pond, occurred over a broader range of discharge conditions. A controlled, "slpit—stream," experiment examining the effect of very low vs. elevated discharge on upstream fish movement indicated, however, that upstream movement of fish out of beaver ponds was also reduced by very low discharge conditions. Movement data for individual fish species revealed considerable variation among the taxa in the tendency for downstream vs. upstream movement, due to variation in the morphology of upstream vs. downstream beaver ponds and its subsequent effects on the composition of fish dispersing from these source areas. Most fish movement occurred over relatively brief time periods, suggesting life history and developmental processes were critical in influencing the timing of dispersal. Size structure of fishes captured in the stream indicated predominantly older age classes (>age I) of fish where dispersing along the stream. However, based on the occurrence of age O individuals only 1 of 12 species, the creek chub Semotilus atromaculatus), routinely reproduced in the stream. Experiments conducted in an artificial stream located below one of the beaver ponds indicated discharge and fish predation have potentially strong and interactive effects on invertebrate colonization in stream ecosystems. Differences in colonization of riffles and pools under low vs. elevated discharge and fish vs. no—fish treatments suggested, however, that the interactive effect of these factors on invertebrate colonization was variable over even small spatial scales. Elevated discharge increased invertebrate colonization in riffles but decreased invertebrate colonization in pools. Contrary to intuitive expectations, fish predation reduced invertebrate colonization more under elevated than low discharge conditions, particularly in pool habitats. Taken together, these results: (1) beaver ponds act as reproductive "sources" for fish on the landscape, while adjacent stream environments act as potential reproductive "sinks," (2) large—scale spatial relationships between beaver ponds and streams, along with the influence of discharge on the permeability of the boundaries between these habitats, are critical in controlling fish dispersal between ponds and streams and the subsequent abundance and composition of fish in lotic ecosystems, and (3) fish predation and discharge have potentially cascading effects on invertebrate colonization in lotic ecosystems.