UVM Theses and Dissertations
Format:
Print
Author:
Blodgett, Corrie A.
Dept./Program:
Rubenstein School of Environment and Natural Resources
Year:
2008
Degree:
M.S.
Abstract:
Contextual interactions at the ecosystem level arise from nonadditive relationships among model factors and are represented as statistical interaction in multifactor experimental designs. Because ecosystem experiments are rarely fully replicated, little is known about the importance of contextual interactions in ecosystem response and development. Thus, ecosystem-level models may contain substantial bias that is interpreted as random error. The component of variability in ecosystem response to soil, climate, and community associations was quantified using calcium (Ca), magnesium (Mg), and potassium (K) export, as well as volume of leachate as dependent variables in a factorial mesocosm experiment located in two Vermont locations over a 9-year period. The mesocosm treatments consisted of three factors that are known to influence nutrient cycling in forest ecosystems: location (climate), soil, and plant community.
The variance of total annual nutrient flux has been partitioned between the main effects, contextual interactions, and random effects using an analysis of variance and variance component analysis. Results show that the main factors consistently controlled a high percentage of variation in Ca export, with the soil factor explaining over 50% of the variation in export every year. Between 4% and 27% of variation in K export was explained by the interaction of multiple factors. Variation in Mg export was consistently explained by the location x soil contextual interaction, which notably explained 65% of variation in export in the 2001-2002 water year. The results of this experiment provide an example of how contextual interactions contribute to variability in ecosystem processes, making it difficult to accurately predict ecosystem response in a changing world. From these results we draw four general conclusions concerning ecosystem complexity; 1) context influences the response of an ecological process, 2) context influences the variability of the response, 3) context influences the interaction of factors, and 4) context influences the power of our statistical tests.
The variance of total annual nutrient flux has been partitioned between the main effects, contextual interactions, and random effects using an analysis of variance and variance component analysis. Results show that the main factors consistently controlled a high percentage of variation in Ca export, with the soil factor explaining over 50% of the variation in export every year. Between 4% and 27% of variation in K export was explained by the interaction of multiple factors. Variation in Mg export was consistently explained by the location x soil contextual interaction, which notably explained 65% of variation in export in the 2001-2002 water year. The results of this experiment provide an example of how contextual interactions contribute to variability in ecosystem processes, making it difficult to accurately predict ecosystem response in a changing world. From these results we draw four general conclusions concerning ecosystem complexity; 1) context influences the response of an ecological process, 2) context influences the variability of the response, 3) context influences the interaction of factors, and 4) context influences the power of our statistical tests.