Diversity and stability of serpentine plant communities: Interacting effects of climate, soil type, and species functional traits

Climatic variation, perhaps the most important force structuring the world's ecological communities, now receives growing attention from ecologists because of the incontrovertible evidence for directional climate change, including increases in variability. One of the less understood aspects of climatic impacts is how they may be mediated by intrinsic differences among natural communities, in factors such as soil fertility, functional composition, functional diversity, invasion history, and human land use. My dissertation examines how soil mediates the response of species and communities to climate at local and regional spatial scales using experimental and observational approaches. In addition, I use species functional traits as a tool to better understand the mechanisms driving community response to both natural and experimental variation in climate.;First, I compared grasslands on infertile serpentine and 'normal' sedimentary soils with respect to temporal variability in species richness and composition using a 10 year data set. I found that variability in species richness and composition tracked precipitation on both soils, but variability was lower in grasslands on serpentine. Communities on serpentine had species with more "stress-tolerant" traits than non-serpentine communities. Within and between soils, variability in richness and temporal turnover were lower in communities scoring lower on a multivariate index of these stress-tolerant traits. In addition, I found that within 41 species found commonly on both soils, variability was lower on serpentine and was positively correlated with community biomass. Thus, I concluded that infertile soils reduce variability indirectly by selecting for stress-tolerant traits and directly by limiting productivity, suggesting that communities on infertile soils may respond more conservatively to predicted long-term directional changes in climate than communities on soils of normal fertility.;Second, in a study across a 1200-km rainfall/productivity gradient in California, I asked if dissimilarity between 96 pairs of plant communities on serpentine and adjacent non-serpentine soils is higher in wetter and more productive regions. I found that communities on serpentine soil had more "stress-tolerant" traits than non-serpentine communities regardless of the climate. Local diversity, species composition, and functional composition were all less strongly related to climate on serpentine than non-serpentine soils. As predicted, the dissimilarity between serpentine and non-serpentine communities was highest in wet and productive climates. Soil organic matter and Ca:Mg increased with increasing climatic productivity, but only on non-serpentine soils generating a difference between soils that was positively correlated with dissimilarity. These results lend support to hypothesis that increasing niche specialization along productivity gradients may contribute to the positive productivity-beta diversity relationship.;Finally, in a three year study, I experimentally tested the response of serpentine and non-serpentine communities to increases (via irrigation tanks) or decreases (via rainout shelters) in spring rain. I also compared the responses of endemics and generalists to changes in spring rain in the absence or presence of competition. Peak season biomass was significantly greater in the rain addition plots in the non-serpentine grassland compared to control plots, but depended on the year. There was no effect of either rain treatment on biomass production in the serpentine grassland. The survival, biomass production, growth rates, and seed production of soil endemics and generalists were all significantly reduced by competition, but were unaffected by changes in spring rain. Overall, endemics tended to perform better in serpentine soil and generalists tended to perform better in non-serpentine soil, suggesting that soil is an important factor for the establishment and survival of endemics and generalists. The effect of competition was similar on both soil types, suggesting that species interactions may be important even in low resource habitats. In conclusion, these results suggest that special soil or low fertility plant communities may be slow to respond to changes in climate compared to communities on more fertile soil and that community response may depend on the dominant species.