Determinism vs. Stochasticity in Community Assembly Processes: The Role of Species Phylogeny and Dominance

The deterministic or stochastic nature of the rules that regulate which species co-exist in a community has been long debated in ecology. This thesis aims to answer (1) if understanding the degree of determinism in the co-existence of species is more or less informative than other sources of uncertainty affecting regional biodiversity, (2) if biodiversity is important to understand ecosystem processes, and (3) if dominance and determinism are correlated. Using a landscape model I show that the degree of determinism or stochasticity (1) controls regional species richness, and (2) causes more uncertainty on species richness than land use change or climate change in the Tropical Andes. In herbaceous communities around the world, diversity is important to improve predictions of biomass productivity and litter accumulation mostly when the diversity and biomass descriptors of the communities were partitioned into legumes, forbs and graminoids. Graminoids, grasses and Carex, and one forb lineage were more likely to be dominant species than non-dominant, while more than a dozen forb lineages were more likely to be non-dominant species. The environment affected the dominant plants the most, increasing the role of habitat filtering, with a less common effect of limiting similarity; in contrast, non-dominant plants were mostly affected by limiting similarity among the non-dominants, but with no signal of habitat filtering. This pattern was observed in herbaceous systems using a local dominant removal experiment and a global observational dataset. The results show that dominant species are a more deterministic subset of species that converge towards a predictable optimum constrained by environmental conditions. In contrast, non-dominant species are better described as a diverging group of species, potentially with multiple optimums, and therefore with a less predictable response. Understanding the differences between dominant and non-dominants can improve models of ecosystem services that rely either on biomass accumulation or on diversity.Ph.D