4. Discussion
With this study we aimed to identify the independent effects of soil and climate on plant trait variation within model grassland communities. Our results point to three key findings. The first is that while individual biomass responses to climate and soil are species-specific, species dominance hierarchies (i.e., relative species contributions to community biomass) are stable. The second finding is that soil differences drive highly stochastic trait variation across both site and species. Lastly, climate differences lead to consistent trait responses in all three species, with species responses irrespective of site.
While these results exemplify the utility of plant-model communities as an ecological tool, caution is necessary when interpreting results. Firstly, we lacked statistical power to explore interactions between specific climate and soil effects. We therefore limit interpretations to descriptions of climate-specific differences and do not treat climates as representations of any single abiotic gradient. To address separating climate effects from soil effects in situ , we standardized the local soil to the reference soil at each climate, however these results must be interpreted in the context of the local climate. Further, our experimental set-up did not preclude micro-organism contamination of the reference soil and we did not characterize the microbial composition of local soils. Local soils likely contained rich microbial communities including Arbuscular Mycorrhizal Fungi (AMF). AMF are known symbionts of the three species in this study and affect interspecific competition and growth (Kyriazopoulos et al. 2014; Scheublin, Logtestijn, and Heijden 2007). Lastly, it was observed that reference soil treatments had larger root biomass, while concurrently having higher rates of mortality, potentially due to physical restriction belowground. With larger root biomass, reference soil communities likely experienced faster dry-down following precipitation, leaving individuals more susceptible to drought stress (Turner 2019). While these points of concern are common in ecological studies, they nonetheless should be considered when interpreting experimental results.