Implications of taxonomic resolution and sample habitat for stream classification at a broad geographic scale

Macroinvertebrate biological monitoring in Victoria, Australia, has required substantial resources over the past decade, and often results have taken years to reach water managers. Ways of reducing the time and effort required in this program were examined. The influence of taxonomic resolution and sample habitat on the classification and ordination of 165 stream sites in 27 catchments across Victoria was examined by progressively reducing the level of detail in the original data. Macroinvertebrate samples were collected from 2 habitats (riffle and edge) at each site. A 3rd data set was generated by amalgamating the individual habitats. These 3 habitat treatments were analyzed at family, genus, and species level, and a 4th taxonomic treatment was generated at species level including Ephemeroptera, Plecoptera, and Trichoptera (EPT) taxa only. All 12 data sets were analyzed using presence/absence data, and each was used to classify sites across Victoria into groups that were characterized using environmental variables. The Mantel test was used to compare the 12 analyses, and showed that each of the analyses produced similar patterns. A number of possible ways to reduce time and effort in broad-scale macroinvertebrate studies were evident: 1) single habitat sampling was sufficient for biological monitoring, 2) studies that require species-level discrimination may be able to reduce costs by identifying EPT taxa only, and 3) genus-level identifications offered no substantial advantage over family-level identifications. Overall, species-level identification appeared to be unnecessary for broad-scale monitoring programs and, in future, taxonomic effort could be reduced by identifying to family level only.