Old-field Community Response to Multiple Interacting Factors of Global Change

Communities are structured by many factors including abiotic factors such as resource availability, and biotic interactions. Climatic and atmospheric change will affect the composition of plant communities through multiple interacting biotic and abiotic factors. Literature on the effects of single factors on plant communities is abundant yet there have been few experiments examining the effects of multiple abiotic factors associated with climate change. Moreover, plant communities are not controlled solely by environmental conditions, but by biotic interactions such as competition and facilitation. In this thesis, I used a field experiment to examine the effects of elevated [CO₂], warming, and soil moisture on in-situ old-field plant communities (Chapter 2). In addition, I conducted a separate field experiment to examine the competitive relationships among the constituent species to test whether an experimentally derived competitive hierarchy can predict relative abundances of species within plant communities (Chapter 3).

I examined plant community responses to treatments of elevated [CO₂] (+300 ppm), warming (+3 °C), and soil moisture availability applied to experimental plots within open-top chambers at Oak Ridge National Laboratory. In 2002, we constructed plots with plant communities consisting of seven common old-field species, including grasses, forbs, and legumes. Beginning in 2003, we tracked foliar cover, density and recruitment, and reproductive phenology for each plant species, and determined changes in community diversity and evenness over the course of two growing seasons.

We observed few interactive effects of treatments on plant abundance. Most of the observed plant responses to treatments were responses to single factors. Species-specific foliar cover was most strongly influenced by warming: warming reduced foliar cover of Trifolium pratense and Dactylis glomerata, but increased foliar cover of Andropogon virginicus and Solidgao canadensis. Foliar cover of Dactylis glomerata was lower in dry plots than in wet plots. During the second full growing season, plant species diversity, evenness, and richness were at least 10% lower in wet plots, where total foliar cover and dominance were greater than in dry plots. Interactive effects of treatments appeared only toward the end of the second growing season. For example, late in the growing season of 2004, cover of Dactylis was four times greater within wet plots under ambient temperatures than in all other treatment combinations (temperature x water interaction; P < 0.02). Overall, temperature and soil moisture availability were the most important environmental variables to plant community composition, yet their effects on species varied. The response of Dactylis glomerata to the water treatment appears to drive the response of the whole community.

A plant species competitive hierarchy—a ranked order from competitive dominant species to competitive subordinate species—theoretically should predict the abundance and composition of species within intact communities. Knowledge about competitive interactions within a community may help separate the direct effects of environmental conditions on plants from indirect environmental effects mediated by biotic interactions. I tested whether a competitive hierarchy derived from pairwise competition under field conditions would predict plant community structure of an intact assemblage. We used seven species (those used in the global change experiment described above) to conduct a pot experiment in field conditions wherein we constructed a competitive hierarchy from all possible combinations of the species. Concurrently, at the same site, we constructed polycultures consisting of the same species used in the pot experiment I calculated an abundance hierarchy based on an importance value derived from species foliar cover and biomass. The competitive hierarchy was not concordant with the abundance hierarchy. This indicates that pairwise comparisons are not robust enough to predict the abundance of plant species within complex communities.

I conclude that in our system water availability and warming, singly and interactively, were the most influential environmental variables on plant species abundances and community composition. In addition to environmental variables, plant species clearly affect the community by affecting the growth and abundance of the species around them. Plant-plant interactions may mediate community level responses to environmental conditions, but, a competitive hierarchy derived from pairwise comparisons is not sufficient enough to predict abundances of species within a complex community and, therefore, cannot be used to predict future community composition.