4.2 Soil effects
Our study design allowed us to isolate soil effects from climatic effects on community and species-specific trait dynamics. Surprisingly, we found that despite each study location having a unique combination of soil and climate, all communities experienced similar changes in plant traits. This community-level finding contrasts with findings at the species level, which revealed stark variation in trait responses. Interactions between climate and species were prevalent across most response variables. A lack of distinguishable patterns in the magnitude or direction of soil effects demonstrates high trait stochasticity within our plant communities, which has previously been attributed largely to environmental variation (Davison et al. 2010; Riginos et al. 2018). Random forest analysis supported this finding, with no increases in predictive power being found when examining soil effects. Nutrient availability has also been shown to influence community assembly within grassland communities (Guo et al. 2014), with community dynamics shifting away from niche-based determination towards stochasticity and species asynchrony in the short term under high nutrient availability (Zhang et al. 2016; Conradi, Temperton, and Kollmann 2017). While our study only examined these responses following one year of treatment, our findings demonstrate the short-term effects of nutrient variation leading to high grassland community interspecific trait stochasticity.
We accept our hypothesis that soil effects would lead to high trait variation across both species and climate. No species consistently had the largest changes for any measured trait. For example, L. corniculatus had greater maximum height at Stubai (1850 m) relative to other species but grew less than others at Graswang (850 m). Further, soil effects resulted in unique species trait variation across climates. Surprisingly, species with high values for one trait that is traditionally linked to fitness did not concurrently increase in other fitness-linked traits. For example, if a species had high survival rates at a given climate, this did not necessitate high values for traits such as leaf dry weight, leaf area, and biomass. This offers another trade-off example between community-level survivorship and individual fitness.
The combination of local soil and climate yields stochastic trait responses in our study for all species. Past experiments have documented shifts in dominance hierarchies depending on interactions between nutrient and climate treatments (Alatalo et al. 2014; S. Niu and Wan 2008). For example, Klanderud and Totland (2005) found that climate change and nutrient addition in grassland ecosystems caused changes in dominance hierarchies, community structure, and diversity. While nutrient addition alone increased the competitiveness of graminoid and forb species, the climatic treatment did not have this effect. This aligns with our results, underscoring the role of changing species interactions resulting from variation in soil nutrient composition on the trait responses of individuals. It is difficult to simulate natural soil conditions. Naturally occurring soil minerals extend well beyond standard nitrogen, phosphorous, and potassium. Many other minerals are known to have interactions with shifts in climate change-relevant plant traits such as transpiration or root acquisition of soil minerals (Lynch and St.Clair 2004). For this reason, natural system experiments remain the most complete look into the future of plant grassland communities. Given the importance of soil composition on community trait dynamics, we suggest further work investigating grassland community responses to climate change incorporate natural soil systems.