Beta diversity and taxonomic sufficiency: Do higher-level taxa reflect heterogeneity in species composition?

Virtually all studies exploring the use of taxonomic surrogates in assessing patterns of diversity have focused on clear shifts in the location of samples in multivariate space. The potential use of coarser levels of taxonomic resolution to detect patterns of variability in multivariate space, corresponding to beta-diversity in the case of presence/absence data, remains unexplored. Here we considered five ecological data sets of highly diverse marine molluscan assemblages to test the hypothesis that patterns in compositional heterogeneity would be maintained at coarser levels of taxonomic resolution. Italy, Norway, New Zealand and the Arctic. We used multivariate dispersion based on the Jaccard resemblance measure of presence/absence data as a measure of beta-diversity to test the null hypothesis that patterns of heterogeneity in species composition for molluscs would be maintained at coarser levels of taxonomic resolution. Tests to compare beta-diversities among groups (based on distances to centroids and using 9999 permutations) were carried out separately for each of five data sets at the species level and then for each of genus, family, order and class levels. Differences in multivariate dispersion at the species level (heterogeneity in the identities of species) were maintained for genera and for families, but not at coarser levels of taxonomic resolution (order or class). These results were consistent across all data sets, despite differences in their spatial scale and extent, geographical location, environmental and habitat features (benthic soft sediments, rocky reefs or kelp holdfasts). These results suggest that either genera or families may be used as effective taxonomic surrogates to detect spatial differences in beta-diversity for molluscs. The use of surrogates can provide considerable sampling efficiencies for biodiversity assessments. We consider, however, that a degree of caution and more work is needed, as heterogeneity at the species level may not be reflected by taxonomic surrogates at smaller spatial scales.