"The emergence of phylogenetic clusters in stochastic niche communities"De, SatavishaPhylogenetic information can help reveal the ecological and evolutionary processes driving community assembly. For example, if environmental filtering selects for species with similar values of a conserved trait (i.e., closely related species are more similar), coexisting species should be more phylogenetically related than species randomly drawn from a regional species pool (more phylogenetically clustered). In contrast, if competition is selecting for species that differ from one another, then species should be more distantly related than expected (more phylogenetic overdispersed). However, recent advances in community assembly have shown that multiple clusters of species emerge on a trait axis via competition. We propose that there may be similar patterns of multiple clusters in the phylogeny. This differs from prior metrics of phylogenetic clustering or overdispersion, which only measures the degree of relatedness in the community. We use a model of a ‘stochastic niche’ community assembly combining stochastic birth, death, and immigration along with niche differentiation processes. We apply this model to different regional pools of species whose trait values were generated from different modes of trait evolution (i.e., more conserved or less conserved trait evolution) on a phylogenetic tree. We then determine the patterns of phylogenetic clustering in these communities using a novel algorithm for cluster detection in networks by incorporating species abundance and phylogenetic information. We find that there are multiple phylogenetic clusters and that the pattern of phylogenetic clustering is strongest when the trait evolved in a conserved way on the phylogeny. Thus, communities selected by environmental filtering can have distantly related species coexisting in multiple clusters. However, we also find significant but weaker clustering under less conserved or random trait evolution. Using our algorithm, we can quantify the variation in the strength of phylogenetic clusters arising from different mechanisms of community assembly (environmental filtering, competition, and stochastic birth/death processes.) This emergence of multiple clusters from conserved traits contrasts with prior studies which would expect simply more phylogenetically related species in a community selected by environmental filtering. Thus, our study provides a revision of community phylogenetic pattern expectations, as well as a new tool for detecting these phylogenetic patterns that can be deployed in the future on real ecological data. |
« back