Analysis of macroinvertebrate assemblages in relation to environmental gradients in Rocky Mountain streams

Using redundancy analysis (RDA) and canonical correspondence analysis (CCA) , we assessed relationships among chemical and physical characteristics and macroinvertebrate assemblages at stream sites sampled by the Regional Envi ronmental Monitoring and Assessment Program (R-EMAP) in the mineral belt of the Southern Rockies Ecoregion in Colorado. We contrasted results of analy ses where community structure was summarized as community metrics and analy ses based on genera abundances. Our objective was to identify metrics or ta xa diagnostic of major environmental stressors in these streams. When RDA w as used to analyze the community metrics data, three axes were significant, accounting for 96% of the metric-environment relation. The first RDA axis was correlated with dissolved cadmium, sediment zinc, and total suspended s olids, variables that indicate it was related to mining effects. The second and third RDA axes were correlated with water temperature, mean substrate embeddedness, mean canopy density at the banks, and a riparian human distur bance index for agriculture, variables associated with riparian and substra te alterations associated with grazing by livestock. When CCA was used to a nalyze the genera abundance data, four axes were significant, accounting fo r 45% of the species-environment relation. The CCA axes were correlated wit h total and dissolved iron, water temperature, dissolved and total organic carbon, mean bank height, and mean water surface gradient, variables associ ated with riparian disturbance from livestock grazing, but not with mining effects. Because CCA measures variation in community structure in terms of changes in the absolute abundances of different genera relative to one anot her, that analysis of genera abundances was sensitive to the effects of rip arian disturbance and stream size, but not to the general toxicological eff ects associated with mining that reduced the abundances of all genera. Comm unity metrics measure various aspects of community structure, including tax a richness, taxa relative abundances, and taxa relative dominance, and thes e metrics were sensitive to the effects of mining, riparian disturbance, an d stream size. Some community metrics, such as the percentage abundance of the most dominant taxon, the total number of individuals, the total number of taxa, and the number of chironomid taxa, may be used to diagnose the env ironmental stressors in these streams, while the results of the CCA for gen era abundances may be used to design new metrics for this purpose.