Ecosystem responses to warming-induced plant species loss and increased nitrogen availability in a Rocky Mountain subalpine meadow

Climate change is predicted to be an important driver of future biodiversity changes, especially in mountainous environments. Climate warming-induced plant species loss is likely to be non-random and based on species-specific susceptibility to rising temperatures. Experimental warming results from a subalpine meadow in Colorado suggest that warming adversely affects shallow-rooted forb species in this ecosystem. To examine the ecological consequences of losing this warming-sensitive species group, I experimentally removed all shallow-rooted forb species from otherwise intact subalpine meadow plots. Since experimental warming also resulted in increased soil nitrogen availability, I crossed the removal treatment with a nitrogen addition treatment to determine whether the loss of shallow-rooted forbs altered the community's response to a perturbation in nitrogen availability.; After three years of experimental species removal, tap-rooted forbs and grasses were able to fully compensate for the loss of shallow-rooted forbs with increased biomass production. Moreover, the remaining plant community yielded a larger biomass response to nitrogen addition when shallow-rooted forbs were removed, possibly because removal led to increased soil moisture. The loss of shallow-rooted forbs and addition of nitrogen did not have strong effects on nitrogen cycling beyond increases in the amount of nitrate moving down through the soil profile. Uptake of nitrogen into plant tissue was also not affected by either the shallow-rooted forb removal or nitrogen addition treatments, suggesting that nitrogen may not have been the most limiting resource during the experiment. I found that spatial heterogeneity generally had a greater influence on soil microbial community composition than any of the experimental treatments.; I conclude that the warming-induced loss of shallow-rooted forbs did not affect biomass production, nitrogen cycling, or soil microbial community composition, but did increase the ecosystem's sensitivity to a perturbation in nitrogen availability. My results demonstrate that non-random scenarios of species loss do not necessarily follow the pattern of decreased productivity, decreased nitrogen uptake, increased nitrogen losses, and dampened response to nitrogen addition with species loss that is predicted by theory and experimental results from randomly-assembled plant communities. Therefore, caution should be used in applying results from randomly-assembled communities to natural communities undergoing non-random species loss.