Global angiosperm family richness revisited: linking ecology and evolution to climate

Aim The global richness gradient of angiosperm families is correlated with current climate, and it has been claimed that historical processes are not necessary to understand patterns of plant family richness. This claim has drawn criticism, and there have been doubts about the quality of the data used to quantify the pattern. We revisit this issue using the Angiosperm Phylogeny Group (APG) III classification and revised range maps, and we incorporate an evolutionary variable, family age, to explore covariation between evolution and ecology and their links to climate via the tropical conservatism hypothesis (TCH). Location Global. Methods The richness pattern for 408 families was derived from range maps, and family ages were derived from a dated angiosperm phylogeny. Patterns were generated for all families, 143 families composed of trees, and 149 families composed of herbs. We also examined family range size patterns to test the extent to which extratropical floras are nested subsets of tropical floras. Ordinary least squares (OLS) multiple and partial regressions were used to generate climate models for richness, mean range size and mean age for each plant dataset and to evaluate the covariation between contemporary climate and clade age as correlates of family richness. Results We confirmed the strong association between contemporary climate and family richness. Age patterns predicted by TCH were also found for families comprising trees. The richness of herbaceous families, in contrast, was correlated with climate but the age pattern was not as predicted by TCH. Floras in cold and dry areas are strongly nested within richer tropical floras. Main conclusions Phylogenetic niche conservatism at the family level offers a likely explanation for the global diversity gradient of trees, but not for non-desert herbs, probably because of the faster evolutionary rates for herbs and less constrained evolutionary responses to climate change. Thus, it appears that multiple processes account for the overall angiosperm family gradient. Our analysis also demonstrates that even very strong associations of taxon richness and climate do not preclude evolutionary processes, as has been widely argued, and that climatic and evolutionary hypotheses for richness gradients are not mutually exclusive.