COS 8-9
Catchment land use and dispersal barriers affect fish and insect assemblage composition in urban streams

Monday, August 10, 2015: 4:20 PM
322, Baltimore Convention Center
Robert F Smith, Massachusetts Cooperative Fish and Wildlife Research Unit, University of Massachusetts, Amherst, MA
Gillian C Gunderson, Massachusetts Cooperative Fish and Wildlife Research Unit, University of Massachusetts, Amherst, MA
Allison H Roy, U.S. Geological Survey, Massachusetts Cooperative Fish and Wildlife Research Unit, University of Massachusetts, Amherst, MA
Background/Question/Methods

The composition of communities is determined by local habitat characteristics and the movement of individuals across the landscape. The remaining patches of natural habitats are typically less suitable and more fragmented than in undeveloped landscapes. Managing human impacts on stream ecosystems usually focuses on the links between catchment land use to in-stream conditions, and ignores landscape characteristics affecting dispersal. We examined the ability of 1) catchment land use, 2) stream network structure, and 3) landscape characteristics that affect dispersal to predict the composition of fish and insect assemblages in streams along an urbanization gradient. We hypothesized that catchment land use has a greater effect on assemblages than network structure or dispersal barriers and that the effects of dispersal barriers differ between fish and insect assemblages based on their dispersal abilities. Using a model selection procedure, we compared fit of regressions of individual landscape predictors and measures of assemblage richness, diversity, and abundance. Catchments and broader landscapes were divided into overlapping and non-overlapping sections for analysis to minimize spatial autocorrelation, and a regression approach was used to empirically determine the appropriate spatial extent to measure landscape characteristics affecting dispersal.

Results/Conclusions

Results of the model selection procedure supported our hypotheses. Percent developed land use had the greatest effect on all three measures of fish and insect assemblages. The variables measuring network structure and dispersal barriers had a secondary, but significant effect on fish and insect assemblages. Stream density in a radius around and along pathways from sampling locations had the greatest effects on insects and fish respectively. Dispersal barriers in a radius around sites and found in or along stream channels had the greatest effects on insect and fish assemblages as well. The relationships of stream density and dispersal barriers to assemblage composition are presumably due to the fact that insects can disperse by flying throughout terrestrial landscape and fish are confined to stream channels. The results indicated that the focus on in-stream environments is important for managing fish and insect assemblages, but that landscape characteristics affecting dispersal also need to be considered. The results also indicated that the effects of fragmentation differed among assemblages based on their different methods for dispersal.