Metric Correspondence and Important Variables

The CCA analysis including all measured chemistry variables (log Conductivity, log SO₄, log TDS, log TSS, log TKN, log NH₃, log NO₂,₃, log PO₄, log NO₂, log TP) was significant at the p = .01 significance level (tested with 1000 random permutations). The eigenvalue of the first axis was .279. The best parsimonious analysis comprised of TP and TSS was significant, however only the first axis was significant (p = .002). The total inertia explained by the model was 7.04. The eigenvalue of the first CCA axis was .195. This model has poor explanatory power, however it is useful for exploring the relationship between the measured variables in the model and the BDM scores. The interpretation is that the diatom assemblages at the sites responded mainly to TP and TSS. The triplot of the reduced model (log TP and log TSS only) with BDM scores passively overlain shows scores increase (cleaner water) as log TP and log TSS decrease.