8172 - Laserfiche WebLink

Home Browse Search

chironomid faunal regions, benthic correlation divided by the number of macroinvertebrate faunal regions, and permutations. aquatic ecoregions defined with the environmental variables was examined. Here, the matrices of distances between EXplora'or' Analyses for IB1 Development sites were compared using a matrix Several exploratory analyses were made permutation test (Dietz, 1983; Cowley, regarding the possible utility of summary 1987; Cowley and Atchley, 1990), e.g., the statistics for an IBI. First, the number of distance matrix for sites based on fish unique (found nowhere else) chironomid presence/absence, data ..was-compared.to-.the.--species--at-a-site-was- regressed on'the total distance matrix for sites based on the number of chuwomid species at the site. multivariate environmental- factors. Six Second, fivquency distrNutions for number matrix permutation tests were performed, of chironomWs per site and for number of i. e. , fish versus environment, chironomids fish species per site were examined. Third, versus environment, benthic the potential utility of fish species macroinvertebrates versus environment, fish member-elevation relationship outside the versus chironomids, fish versus benthic upper Rio Grande drainage was examined. macroinvertebrates, and chironomids versus Here, data from a survey of the Black River benthic macroinvertebrates. in southeastern New Mexico (Cowley, 1979; Cowley and Sublette, 1987) were The matrix permutation test is a compared with data from the upper Rio nonparametric method of comparing two Grande drainage. Fourth, the utility of an similarity or dissimilarity matrices; standard additional environmental variable, distance parametric and nonparametric correlation from headwaters, was examined using the tests are inappropriate. The null hypothesis Black River data. in the matrix permutation test is that the magnitude of elements of one matrix are random (uncorrelated) with respect to the ReSWtS magnitude of elements in the second matrix. The alternative hypothesis is that the magnitude of elements in the two matrices is correlated. For this report matrix Aquatic Ecoreglons of the Upper Rio , permutation tests are performed by making Grande Drainage in New Mexico 10,000 permutations. In each permutation, the rows and columns of one of the matrices Two aquatic ecoregions within the upper (the second) are shuffled randomly and the Rio Grande drainage of New Mexico were correlation coefficient between elements of defined using cluster analysis on scores for the two matrices is obtained. The test multivariate environmental factors (Figure statistic for the permutation test is the 1). These ecoregions were given the number of correlation coefficients from the provisional names Aquatic Ecoregion 1 and 10,000 permutations that exceed the original Aquatic Ecoregion 2 (Figure 6). Before correlation coefficient obtained before describing these ecoregions further, it is permutation. The empirical probability interesting to consider the multivariate (P-value) of the test is the number of correlations exceeding the original 9