Wednesday, November 7, 2007 - 11:00 AM
323-5
On Complex Problems in Wetland Biogeochemistry.
Ronald Corstanje, Rothamsted Research, Hertfordshire, England, John R. White, Oceanography and Coastal Sciences, Louisiana State University, Energy Coast and Environment Building Room #3239, Baton Rouge, LA 70803, and Andrew V. Ogram, Soil and Water Science, University of Florida, Gainesville, FL 32611.
Describing the biogeochemistry of a wetland is a complex and heterogeneous problem, due to the dynamic interactions of biotic and abiotic variables that vary both spatially and temporally. The use of simple statistical methods, in which this structure is determined as a continuous response (regression analysis) or as set of responses to discrete levels in factors (analysis of variance) has been used extensively in wetland research to relate elements in the biogeochemical cycles or determine the extent to which a particular factor may influence a process. However, as more information is obtained about processes in wetlands, and larger sets of measurements over more wetland properties are produced in studies, we obtain more complex datasets which these techniques cannot deal with. Yet the problem is still understanding the processes that occur in wetlands, identifying the constituent elements that take part in these processes and the actors involved. This requires an appreciation of the inherent multivariate structure present in these datasets. Here we explore complex data with common multivariate methods such as cluster analysis, principal component analysis and stepwise discriminant analysis and we present a series of models of association using confirmatory factor analysis or structural equation modelling. Placing our current level of understanding in a model forces us to synthesize the information available to us in an unambiguous, coherent manner. Scale is probably one of the most confounding factors when synthesizing and modelling wetland processes, as it directly relates to what exactly is being described. Understanding the effect of scale on many of the conceptual and computational models of processes and actors in wetlands will be a key challenge for the future.