Exploring stochastic and niche-based community assembly dynamics in a long-term grassland field experiment

Background/Question/Methods

Community assembly theory explores how deterministic and stochastic processes interact to regulate the formation of ecological communities from regional species pools. Niche-deterministic or trait-based assembly models are based upon assumptions of interspecific limiting similarity and trade-offs in ecological strategy. These models predict restrictions on community membership through deterministic species interactions, creating non-random assortments of coexisting species with respect to ecological traits. In contrast, neutral-stochastic models assume that local communities arise randomly with respect to species traits. Here, community assembly is expected to strongly reflect the influence of stochastic and historical phenomena. Niche and neutral-stochastic processes are not mutually exclusive; both can operate in nature. Because these contrasting models make different predictions regarding species responses to environmental change, it is important to clarify the relative contributions of each to community assembly. We present initial results from a long-term field experiment designed to evaluate grassland assembly dynamics through the manipulation of initial community states on a common environmental template. The experiment was established in 2007 on abandoned agricultural land in eastern Kansas on which 270, 2 x 2 m experimental plots were planted to a wide variety of initial community states with respect to: plant species diversity; composition; and initial abundances.

Results/Conclusions

From 2008 (year after planting) to 2011, plant community composition in plots shifted significantly over time as the community assembled from initial states. Temporal shifts in composition were associated with dramatic local species losses across all plots (decreased alpha richness). Species losses were greater in plots planted to higher initial diversity, both in terms of absolute numbers and proportions of species lost. Consistent with stochastic extinction, incidences of species loss from plots were partially explained by initial abundance of species at planting, independent of taxonomic or functional guild identities. In addition, community composition diverged over time (beta diversity increased), reflecting the influence of different planting histories on subsequent trajectories of community assembly. However, a few key species were rarely lost from plots and rose to dominance over time even in plots where they were initially planted rare, indicative of niche-based dynamics. Our initial findings sketch a view of community assembly shaped by the interplay of niche-deterministic phenomena, stochastic processes and historical contingency. We briefly discuss the roles of key plant traits and phylogenetic-relatedness of species in modulating community assembly dynamics of this system.