HISTORICAL CONTINGENCY AND BIOTIC DETERMINISM IN COMMUNITY ASSEMBLY: A LONG-TERM EXPERIMENT OF GRASSLAND DYNAMICS

Abstract

Starting with studies about ecological succession over a century ago, ecologists have been interested in the extent to which community assembly is predictable and the degree to which assembly is influenced by colonization history and initial conditions. Today, ecologists agree that there are two general processes by which communities may assemble. The first is deterministic assembly, in which environmental conditions determine the outcomes of species interactions, and thus under common environmental conditions communities converge in composition over time. The second is historically contingent assembly, in which colonization history - regardless of environmental conditions - create unique biotic conditions and unique assembly trajectories that produce divergent communities, driven by stochastic community drift or by priority effects. It is recognized that these processes are not mutually exclusive but work in concert to produce observed, mature communities in nature. To tease apart the roles and influences of these two processes, a field experiment in experiment was established in eastern Kansas that experimentally manipulated grassland communities to a wide variety of initial conditions in terms of species richness, species composition, and the relative abundances of constituent species. The subsequent convergence or divergence of these communities over eight years of community development was assessed not only in terms of species composition but also in functional trait and phylogenetic composition in order to integrate information on the niche requirements of coexisting species, which influences the intensity of those interactions. Furthermore, analyses of community composition were conducted using abundance-weighted and presence/absence data in order to emphasize abundance dynamics and occupancy dynamics over the course of community assembly. These analyses revealed that assembly processes varied both among different levels of community organization (i.e. species, functional trait, or phylogenetic composition) and among weighting schemes within a level of community organization: species abundance is strongly deterministic while species occupancy is historically contingent; functional trait abundance is historically contingent while functional trait occupancy dynamics are strongly deterministic; phylogenetic dynamics are historically contingent regardless of weighting scheme. When functional trait analyses were broken down into individual traits, it was found that dynamics in among-plot dissimilarity varied widely, converging with respect to some and diverging with respect to others. This indicated that the broad trends in aggregate functional trait composition may belie the variable dynamics of different traits during assembly. Overall, these results provide a comprehensive and in-depth experimental study of plant community assembly processes by including and comparing both functional trait and phylogenetic composition in the assessment of community convergence or divergence, and also by considering abundance dynamics separately from occupancy dynamics. In doing so, the varied influences of deterministic and historically-contingent assembly processes are better understood. It is worth noting, however, that despite a relatively long-term experimental dataset, eight years is a fairly short timespan on an ecological scale. As such, these results only characterize dynamics at the very onset of community assembly, which may crucially impact later dynamics, but may still be species- and context-dependent and in flux, with a capacity to change over time. However, these results are still valuable to our overall understand of the long-term processes of community assembly, with the potential to improve restoration and conservation efforts in the face of habitat degradation and global climate change.