Hale A.N.,University of Pittsburgh |
Noble G.,University of Pittsburgh |
Piper K.,University of Pittsburgh |
Garmire K.,University of Pittsburgh |
And 2 more authors.
Hydrobiologia | Year: 2015
Understanding landscape influences on stream ecosystems is a challenging task due to the spatial complexity and connectedness of stream networks. Here, we control for longitudinal connectivity to provide a robust test of the relative importance of reach- and catchment-scale factors in determining macroinvertebrate community structure in southwestern Pennsylvania streams. We determined that sites separated by ≤510 m along the stream network had significantly correlated macroinvertebrate community scores. After controlling for this spatial autocorrelation, a partial least squares regression identified two factors that together accounted for 32% of the variation in community scores. In this model, two reach-scale factors—habitat assessment score and stream pH—were the most important factors for predicting a stream’s macroinvertebrate community score. However, landscape diversity was also important. Landscape diversity is a catchment-scale factor that was highly correlated with percent pasture/hay and measures of habitat fragmentation. Our results provide support for the idea that stream communities in undisturbed areas are heavily influenced by reach-scale characteristics. Furthermore, our results indicate that Pennsylvania natural resource managers should consider habitat score and stream pH after accounting for spatial autocorrelation when identifying restoration targets for impacted streams. © 2015 Springer International Publishing Switzerland