|Title||Evaluation of environmental degradation kinetics using hierarchical bayesian modeling|
|Publication Type||Journal Article|
|Year of Publication||2015|
|Authors||Yu Y., Steinschneider S, Reckhow DA|
|Journal||Journal of Environmental Engineering|
This paper introduces the use of hierarchical Bayesian modeling as a method to estimate how kinetic properties of environmental contaminants vary across experimental factors. The hierarchical modeling framework utilizes a multilevel statistical structure, in which kinetic rate constants represent contaminant degradation behavior for individual, lower-level experiments, but also exhibit a higher-level structure across different experimental conditions. The main benefit of this modeling approach is the ability to pool information between experiments in order to reduce the influence of outliers and produce more robust predictions of degradation rate constants in an out-of-sample context. The Bayesian estimation method is also very flexible, with the ability to relax distributional assumptions on parameters, account for heteroscedasticity in model residuals, include prior information or expert knowledge when available, and propagate all uncertainties into model predictions, all with relative ease. The benefits of the hierarchical Bayesian approach are demonstrated in a case study examining the pH dependence of hydrolysis rates of haloacetamides (HAMs). A comparison is provided against a simpler least-squares method to highlight the differences and benefits of the proposed method.