Phylogenetic species-area curves

Background/Question/Methods

Species richness invariably increases with sampling area, but the processes underlying species-area curves differ. For example, curves derived from small-scale plot sampling of plants tend to reflect ecological processes, while curves derived from large-scale sampling of island radiations tend to reflect macroevolutionary processes. The two main ecological processes are species sorting: more habitats and more species adapted to those habitats are found as more area is sampled; and random placement: more individuals and more species are found as more area is sampled when individuals are randomly distributed. The two main macroevolutionary processes are the colonization/extinction of species among regions and in situ speciation of species within regions. While macroevolutionary processes can clearly cause phylogenetic patterns in species distributions, ecological processes can also since species characteristics often reflect phylogeny, termed phylogenetic signal. Here, we explore how these ecological and macroevolutionary processes produce phylogenetic patterns in species-area curves. We refer to phylogenetic species-area curves as the relationship between phylogenetic diversity and sampling area. We develop theoretical expectations for the shapes of phylogenetic species-area curves under each process, and then compare these expectations to forest plot plant data from the U.S. Pacific Norwest and Anolis lizard island data from the Caribbean.

Results/Conclusions

Ecological simulations that did not have phylogenetic signal in either species distributions along an environmental gradient (species sorting) or in species relative abundances (random placement) produced flat phylogenetic-species area curves. When there was signal, phylogenetic diversity accumulated slower under the species sorting process when plots were aggregated along an environmental gradient than if plots were aggregated irrespective of the gradient. Random placement produced no such difference. While the empirical forest area curves matched our expectations for species sorting, the curves were negative curvilinear with small spatial scales having more phylogenetic diversity than large scales.  Phylogenetic diversity did not vary with area under any macroevolutionary simulation. Instead, simulated island phylogenetic diversity was determined by the degree to which biota were assembled from ancestral colonizations. Similarly, there was no relationship between Caribbean island area and Anolis phylogenetic diversity. Phylogenetic diversity was determined by ancestral colonizations where assemblages derived from few colonizations and subsequent in situ speciation had lower phylogenetic diversity than those derived from multiple colonizations. Thus, even though species richness invariably increases with area, phylogenetic diversity does not and the shapes of phylogenetic species-area curves are greatly determined by the underlying processes.