8172 - Laserfiche WebLink

Home Browse Search

Recommendations Results of the present study indicate that aquatic ecoregion classification can be obtained using a combination of ordination and cluster analysis. For consistency with the present analyses, it is recommended that principal components extracted from a correlation matrix for environmental variables be subjected to varimax rotation to improve their interpretability. It is recommended that no data transformation or elimination of variables be made. Cluster analyses should use Ward's algorithm (Ward, 1963). Finally, it is suggested that the utility of additional environmental variables such as distance from headwaters, stream width, and stream depth. Other environmental variables such as stream flow and current velocity are subject to periodic fluctuations and may be difficult to relate to species distribution and abundance. In defining faunal regions, presence/absence data is preferred to relative abundance data and there appears to be merit in using 7accard dissimilarity coefficients. Faunal regions can be identified using cluster analysis on percent disagreement data and the cluster tree should be constructed using Ward's method. It is recommended that the utility of alternative classifications be investigated, i.e., trophic and reproductive guilds and chironomid subfamilies. It is recommended that data collection be continued. As the number of sites for which chironomid data are available increases, more powerful and definitive statistical analyses will be possible. Also, as the number of sites increases that have all four types of data, more general results will be obtained. At this point, it is impossible to predict how many aquatic ecoregions will exist in New Mexico. From a statistical standpoint, it will be desirable to have about 30 sites per aquatic ecoregion. This number of sites per ecoregion would enhance multivariate analyses within ecoregions. It is recommended that additional effort be made to improve two of the categorical environmental variables, i.e., substrate -composition and in-stream vegetation. Analyses would be improved if both of these categorical variables were reordered to reflect an a priori expectation for changes that might occur with elevation. Literature Cited Bailey, R. G. 1994x. Ecoregions of the United States (map). revised edition. Washington, DC, Forest. Service, U. S. Department of Agriculture. Scale 1:7,500,000, colored. Bailey, R. G. 1994b. Description of the ecoregions of the United States. second edition. Forest Service, U. S. Department of Agriculture, Miscellaneous Publication No. 1391 (revised). Bailey, R. G., P. E. Avers, T. King, and W. H. McNabb (eds.). 1994. Ecoregions and subregions of the United States (map). Washington, DC, U. S. Geological Survey. Scale 1:7,500,000, colored. Accompanied by a supplementary table of map unit descriptions. Compiled and edited by W. McNabb and R. G. Bailey. Bauman, R. W., A. R. Gaufin, and R. F. Surdick. 1977. The stoneflies (Plecoptera) of the Rocky kf ,to