Fig 1.
Hypothesized relationships between ecosystem function and species richness (a), functional diversity (b), and phylogenetic diversity (c).
We predict a stronger relationship with ecosystem function, and thus a higher R2, for functional diversity (b) and phylogenetic diversity (c) than for species richness (a) because the former two measures incorporate information about the traits, or the evolutionary similarity of the different species in the community. Panel (d) depicts the results of variation partitioning, indicating our hypothesis that functional and phylogenetic diversity will explain all of the variation explained by species richness, as well as additional variation, both overlapping and unique.
Table 1.
Results of the linear models for predicting zooplankton community biomass (ln), ranked in increasing order of AIC.
Fig 2.
Zooplankton community biomass in the 23 ponds as predicted by the best diversity indices in each category: taxonomic diversity—species richness (a), functional diversity—abundance weighted functional divergence (b), phylogenetic diversity—abundance weighted standard effect size mean pairwise distance (c), and the variation partitioning for these three models with their adjusted R2 (d).
Significant model trends are shown as black lines. The grey bands indicate the 95% confidence intervals for the predicted values (a) and the slope of the regression lines (b,c).
Table 2.
Results of the linear models for predicting chlorophyll a (ln), ranked in increasing order of AIC.
Fig 3.
Chlorophyll a in the 23 ponds as predicted by the best diversity indices in each category: taxonomic diversity—species richness (a), functional diversity—functional richness (b), phylogenetic diversity—presence absence standard effect size mean pairwise distance (c), and the variation partitioning for these three models with their adjusted R2 (d).
Significant model trends are shown as black lines. Insignificant model trends are shown as grey lines. The empty circles indicate the pond that is an outlier to the predicted trend. The dashed lines indicate the model trend when this outlier is removed. The grey bands indicate 95% confidence intervals for the slope of the regression lines.