Classification schemes for phytoplankton: a local validation of a functional approach to the analysis of species temporal replacement

During ecosystem succession, phytoplankton species composition is hard to predict and although it is tempting to use taxonomic groups for predictive purposes, the conditions favouring their development are often cross-phyletic and, frequently, overlap. Another alternative is to consider functional groups. Reynolds (1997) proposed phytoplankton associations according to functional criteria, based upon identified coherent morphological and ecological properties. Here we apply data from the phytoplankton community of Lake Rodó in Montevideo, Uruguay (small, shallow, polymictic and hypertrophic lake under restoration) to test and quantify the effectiveness of the approach. The phytoplankton species were sorted into their main taxonomic groups and into the associations proposed by Reynolds. A canonical variate analysis was use to test the non-random occurrence of these classification schemes and to determine their discriminatory power. Both classification schemes, taxonomic and functional, showed a significant result, but classification into functional associations had a higher discriminatory power. The eigenvalue for the canonical correspondence analysis first axis for the functional associations was 0.708 and the cumulative explained variance for the species–environmental relationship was 78.6%. The environmental factors showed similar patterns between associations and individual species. Our data indicate that the scheme using functional associations does indeed capture much of the ecology of the phytoplankton.