You are here:
MULTIVARIATE ANALYSIS OF MACROINVERTEBRATE ASSEMBLAGES TO DETERMINE IMPACTS ON ROCKY MOUNTAIN STREAM ECOSYSTEMS
Citation:
Griffith, M. B., B H. Hill, A. T. Herlihy, AND P R. Kaufmann. MULTIVARIATE ANALYSIS OF MACROINVERTEBRATE ASSEMBLAGES TO DETERMINE IMPACTS ON ROCKY MOUNTAIN STREAM ECOSYSTEMS. Presented at North American Benthological Society, Duluth, MN, May 26-29, 1999.
Description:
Using reduncancy (RDA) and canonical correlation analysis (CCA) we assessed relationships between chemical and physical characteristics and periphyton at 105 stream sites sampled by REMAP in the mineral belt of the southern Rockies ecoregion in Colorado. We contrasted results obtained with community structure measured by species abundance, genera abundance and community metrics, identifying taxa or metrics diagnostic of stressors in these streams. For species abundance (CCA), 4 axes were significant and accounted for 32% of the species-environment relation. The first axis was correlated with presence of mixed canopy and mid-layer vegetation, % fast water habitat, % coarse substrate, and nonagricultural disturbance and inversely correlated with dissolved organic carbon, embeddedness, width x depth, and % pool habitat. The second axis was correlated with TOC (total organic carbon), P, embeddedness, alkalinity, suspended solids (TSS), and agricultural disturbance and inversely correlated with % coarse substrates, and presence of herbaceous groundcover. For genera abundance (CCA), the axes were not significant. For metrics (RDA), 3 axes were significant and accounted for 95% of the metric-environment relation. The first axis was correlated with P, embeddedness, TOC, and dissolved As and inversely correlated with presence of mixed canopy or mid-layer, % canopy density, % coarse substrate, and % fast water habitat. The second axis was correlated with DOC (dissolved organic carbon), % pool habitat, and alkalinity and inversely correlated with dissolved Cd, Cu, and Zn; TSS; and % fast water habitat. Species abundance is sensitive to the subtle physical effects of mining or agriculture, but this sensitivity is lost with genera abundance. The metrics are more sensitive to chemical effects.