Twenty years of experiments in a grassland community do not support paradigmatic views of plant strategies, assembly, and trait-function relationships

Background/Question/Methods

Competitive exclusion is central to foundational models in community ecology, which often conclude coexistence occurs when other processes offset competitive differences. Furthermore, if plant responses to multiple processes are correlated, species may cluster into “strategies” such as stress tolerance based upon underlying plant traits. Despite the prevalence of these ideas, there is little data exploring species and community responses to large numbers of potential assembly drivers. Here, we take a novel approach using trait and community data collected within a single grassland at the Northern edge of the Great Plains. Since 1999, my research group has conducted 17 manipulations of biotic and abiotic factors (e.g. resources, competition, defoliation, etc.) within a 250ha area. This approach allows us to avoid confounding effects such as shifts in species pools and regional variation in climate. Each experiment had the same core design with control and treatment plots, treatments applied at least three years, and population sizes estimated as percent cover. We use these data, combined with trait information collected on-site, to test whether there is evidence for (1) competitive dynamics being a driver of assembly, (2) consistent plant strategies across treatments, and (3) whether ecological function can be inferred from plant traits.

Results/Conclusions

Most treatments caused shifts in community structure in wildly idiosyncratic directions. The least impactful treatments directly modified competition, challenging the idea that shifts in competitive dominance drive assembly. There was also no clustering of species across treatments indicative of consistent plants strategies (e.g. “stress-tolerant”). Focusing on individual drivers has likely given a misleading understanding of the hyper-dimensionality of a plant’s niche, at least when described but functional responses. In contrast to species responses, the impactful treatments had very similar effects on trait structure. However, species-level trait information was unable to predict species-level responses. A parsimonious explanation may be that measuring a trait does not inform function, as similar traits can be used differently by different species. Though competition does occur in this system, there is no evidence it plays a significant role in community assembly. We suggest that the maintenance of biodiversity at local scales in natural systems may be simple. If species vary in response to nearly everything, natural variation in anything will generate non-equilibrial realities, which should prevent species loss. We encourage ecologists to question the continued dependence of elegant models, which may not be particularly informative in describing the dynamics of natural systems.