Indices for detecting differences in species composition: some simplifications of RDA and CCA.

ABSTRACT

We provide algebraic simplifications for the redundancy analysis (RDA) eigenvalue and the canonical correspondence analysis (CCA) eigenvalue in the special case of permanent plots sampled twice. The indices for RDA and CCA are interrelated and are intuitively interpretable. These simplifications also apply to simple split-plot designs and to a balanced design with two independent samples.