Human actions have led to the destruction or degradation of natural habitats in virtually all parts of the Earth. Ecosystem restoration is one method to mitigate the effects of habitat loss. But restoration ecology is a young discipline and there is much left to be learned about how to effectively restore ecosystem functioning. This dissertation examines how soil amendments and planted herbaceous species diversity affect the restoration of ecosystem functions in wetlands, while also testing basic ecological questions that help us understand ecosystem function. Using data from the greenhouse and from the biodiversity and ecosystem function field experiment in Duke Forest, in Durham, NC, I examine how plant trait diversity, average plant traits, and environmental conditions influence nitrogen (N) removal from restored wetlands. Field data collected from a restored wetland in Charlotte, NC, enables me to examine how soil organic amendments influence the development of soil properties, processes, and plant communities. Finally, combining field data from both sites, I compare how soil properties influence denitrification potential in both restored wetlands.
One unanswered question in the research relating biodiversity and ecosystem function is whether species diversity or species traits are more important drivers of ecosystem function. The first portion of my dissertation poses several hypotheses about how plant traits, plant trait diversity (calculated as a multivariate measure of plant trait diversity), and environmental conditions are likely to influence two ecosystem functions, biomass N and denitrification potential (DEA), and then examines these hypotheses in a restored wetland in the Piedmont of N.C. Using multiple linear regression, I demonstrate that functional diversity (FD), of traits important for plant growth had no effect on biomass N, but two plant traits, leaf area distribution ratio (LADR) and water use efficiency (WUE), had strong negative effects. Soil inorganic N also had a positive effect. For DEA, FD of traits related to denitrification also did not have a significant effect, but there was evidence of a weak positive effect. Two plant traits had positive effects on DEA, aboveground biomass and aboveground biomass C:N ratio; two traits, belowground biomass C:N ratio and root porosity, had negative effects. Soil inorganic N and soil organic matter also had positive effects on DEA. Results from a Principal Components Analysis (PCA) clustering plant species in trait-space, suggest that <em>Carex</em>, <em>Scirpus</em>, and <em>Juncus</em> species tend to be associated with traits that maximize biomass N, while there is no specific region of trait space or set of species that correspond to high DEA. Instead, there are multiple plant trait combinations that can lead to high DEA. These results suggest that, even though plant diversity (as measured by FD) does not significantly influence biomass N or denitrification, plant trait diversity is important to maintaining multiple ecosystem functions simultaneously.
Restored wetlands tend to have lower levels of soil organic matter than natural reference wetlands. Low soil organic matter can limit nutrient cycling as well as plant survival and growth in restored wetlands. In the second portion of my dissertation, I examine how soil compost amendments influence the development of soil properties and processes as well as plant communities at a restored wetland in Charlotte, NC. Using two-way analyses of variance, multiple comparisons of means, and regression, I determine that available N and phosphorus (P) increase with increasing soil organic matter in both the low and high marsh. Total microbial biomass (MB) and microbial activity (measured by denitrification potential (DEA)) also significantly increase with increasing organic matter in both marsh communities, as does soil moisture. Neither total plant biomass (in the low marsh), nor plant species richness (in the high or low marsh) demonstrate any consistent patterns with soil organic matter level in the first three years post-restoration. These results suggest that compost amendments can positively influence some soil properties (i.e. soil available N, P, microbial biomass, and soil moisture) and some ecosystem functions including nutrient cycling (such as denitrification potential), but may have limited early impacts on plant communities.
In restoration ecology there is a general assumption that restoring ecosystem structure will also restore ecosystem function. To test this fundamental assumption, I examine whether two restored wetlands demonstrate similar general relationships between soils variables (i.e. do the two systems have similar soil ecosystem structure), and whether the importance of each soil relationship is the same at both systems (i.e. do the two systems demonstrate the same soil function). I use structural equation modeling to both pose hypotheses about how systems function and to test them using field data. I determine that the same model structure of soil relationships is supported by data from these two distinct, yet typical urban restored wetland ecosystems (that is, the two systems have similar soil structure). At both systems higher soil organic matter is the most important predictor of higher DEA; however, most of the other relationships between soils variables are different at each system (that is, the two systems are not functioning in the same way). These results suggest that some fundamental relationships between soil properties and microbial functioning persist even when restored wetlands have very different land-use histories, plant communities, and soil conditions. However, restoring similar soil ecosystem structure does not necessarily lead to the restoration of similar soil function. Ultimately, I hope this research advances our understanding of how ecosystems function and improves future wetland restoration efforts. / Dissertation
| Identifer | oai:union.ndltd.org:DUKE/oai:dukespace.lib.duke.edu:10161/620 |
| Date | 22 April 2008 |
| Creators | Sutton-Grier, Ariana E. |
| Contributors | Richardson, Curtis |
| Source Sets | Duke University |
| Language | en_US |
| Detected Language | English |
| Type | Dissertation |
| Format | 4389506 bytes, application/pdf |
Page generated in 0.0029 seconds