To understand the role biota play in resilience or vulnerability to environmental change, we investigated soil, plant, and microbial responses to a widespread environmental change, increased nitrogen (N). Our aim was to test the plant–soil threshold hypothesis: that changed biotic structure influences resilience to accumulated changes in N. For six years, we removed one of two codominant species, Geum rossii and Deschampsia caespitosa, in moist-meadow alpine tundra in Colorado, USA. We also manipulated nutrient availability by adding carbon (C) or N, separately and in combination with the species removals.
Consistent with our hypothesis, Geum was associated with soil feedbacks that slowed rates of N cycling and Deschampsia with feedbacks that increased rates of N cycling. After a four-year initial resilience period, Geum dramatically declined (by almost 70%) due to increasing N availability. In contrast, Deschampsia abundance did not respond to changes in N supply; it only responded to the removal of Geum. Forbs and graminoids responded more positively to Deschampsia removal than to Geum removal, indicating stronger competitive effects by Deschampsia. The changed biotic interactions appear to have community-level consequences: after six years of Geum (but not Deschampsia) removal, evenness of the community declined by over 35%.
Increased N affected the soil–microbial feedbacks, particularly in association with Geum. Microbial biomass N declined at higher N, as did the activities of two C-acquiring and one N-acquiring extracellular microbial enzymes. In the presence of Geum, N fertilization slowed the activity of phenol oxidase, a tannin-degrading enzyme, suggesting that microbes shift from degrading Geum-derived compounds. In the absence of Geum, acid phosphatase activity increased, suggesting increased phosphorus limitation in association with Deschampsia.
With continued N deposition forecast for this system, these results suggest that initial resilience of Geum to increased N will be overwhelmed through elimination of microbial feedbacks. Once Geum declines, the loss will indirectly facilitate Deschampsia via competitive release. Because Deschampsia exerts strong competitive effects on subordinate species, increased Deschampsia abundance may be accompanied by a community-wide drop in diversity. We conclude that plant–soil feedbacks through the microbial community can influence vulnerability to exogenous changes in N and contribute to threshold dynamics.