8172 - Laserfiche WebLink

Home Browse Search

aggregation of species that were abundant either at many sites or only at one or a few sites. Locally or widely abundant fish species in cluster 1 included longnose dace, brown trout, brook trout, white sucker, Rio Grande sucker, Rio Grande cutthroat trout, and flathead chub. Cluster 2 was indicative of moderately .abundant fish species. This ohrster included cutthroat trout hybrids, rainbow trout, fathead minnow, and Rio Grande chub. The third cluster contained species that were represented by few specimens at any site, i.e., green sunfish, white crappie, Carosromus hybrids, kokanee salmon, channel catfish, river carpsucker, common carp, mosquitofish, and red shiner. Species-Environment Relations Fsh Species-environment relations were examined using CCA and DCCA. These procedures are multivariate techniques that combine ordination and multiple regression in an iterative fashion. The results of CCA on ranked relative abundance data for fish is shown in the biplot in Figure 17 for the first two canonical axes. These two axes explained abort 47% of the species-environment variation. The biplot from CCA corroborates the finding in principal components analysis that elevation, and variables that vary inversely with elevation, are the most important environmental variables. Further, it reinforces the finding of significant association between in distances between sites based on environmental factor scores White sucker-Rio Grande sucker hybrids. and distances based on fish presence/absence data (Box 2). Because of the way that axes are scaled, only the most important axes of environmental variation are shown in the biplot. All species except Rio Grande cutthroat trout, cutthroat trout hybrids, and brook trout were ordinated near the origin of the biplot and most were not shown, or labeled, on the biplot. Detrended CCA (DCCA) provided a more interpretable picture of species-environment relations (Figure 18) even though the first two axes accounted for only 25 % of the species-environment variance. In DCCA, axes are resealed so that the variation is spread out in the biplot. Most of the variation in the DCCA biplot is in the X-axis direction, which can be construed: to be most closely related to elevation. Fish species are arrayed in the biplot, from left to right, as high elevation species to low elevation species. Dotted lines are superimposed on the biplot to show three presumptive associations of fish species that are consistent approximately with the associations obtained with cluster analysis of presence/absence data (Figure 15). The left-most getup of fish species is the high altitude assemblage, the middle group is the trout stream assemblage, and the right-most group is a combination of the remaining three associations identified with cluster analysis, i.e., reservoir fishes, large river fishes, and warmer water fishes An interesting observation is suggested by the DCCA biplot. First, white sucker ("Cat com" in Figure 18) and Catostomus hybrids' ("Cat hyb") are at the lower left of the "trout stream assemblage" (middle species group in Figure 18) while Rio 17