You are here:
Effects of Nitrogen Inputs on Freshwater Wetland Ecosystem Services–A Bayesian Network Analysis
Citation:
SPENCE, P. AND S. J. JORDAN. Effects of Nitrogen Inputs on Freshwater Wetland Ecosystem Services–A Bayesian Network Analysis. JOURNAL OF ENVIRONMENTAL MANAGEMENT. Elsevier Science Ltd, New York, NY, 124:91-99, (2013).
Impact/Purpose:
This manuscript explores the use of Bayesian Network analysis in determining how nitrogen loading affects ecosystem services from wetlands, and the factors that modify these effects.
Description:
Wetlands can provide a balance between regulating water quality and one aspect of mitigating climate change, by reducing the quantity of reactive nitrogen (Nr) reaching downstream receiving water bodies, while emitting negligible amounts of nitrous oxide (N2O) during incomplete denitrification. However, increased Nr inputs to freshwater wetlands potentially affect the interaction between N2O emissions and outflow water quality. The purpose of this study is to evaluate parameters that influence the effects of Nr inputs on Nr removal, as well as the interaction between N2O emissions and outflow water quality, using a Bayesian Belief Network (BBN). The BBN was developed by linking wetland classification, biogeochemical processes, and environmental factors. Empirical data for 34 freshwater wetlands were gathered from a comprehensive review of published peer-reviewed and gray literature. The BBN was implemented using 30 wetlands (88% of the freshwater wetland case file) and evaluated using a single test file containing 4 wetland cases (12% of the freshwater wetland case file). Sensitivity analysis suggests that N2O emissions, outflow water quality, and Nr removal have strong influence, soil Nr accumulation has moderate influence, and HGM classification, total denitrification, latitude, and soil pH have weak influence on the interaction between N2O emissions and outflow water quality. The BBN implies it is not average annual total Nr load entering the wetland, but the Nr removal efficiency that influences the interactions between N2O emissions and outflow water quality. Even though the network has a very low error rate (based on 4 wetland cases) indicating a high predictive accuracy, additional testing and larger training and testing datasets would increase confidence in the model’s ability to provide robust predictions and to reduce the uncertainty resulting from an incomplete dataset and knowledge gaps regarding the interactions between N2O emissions and outflow water quality.
URLs/Downloads:
The Effects of Nitrogen Inputs on Freshwater Wetland Ecosystem Services–A Bayesian Network Analysis