Monte Carlo analysis of uncertainty attached to microbial pollutant degradation rates

Prediction of bioremediation performance relies on models of microbial activity that are typically fitted to few data, which can lead to large errors in parameter estimates and uncertain prediction of reaction rates and degradation times. This paper presents a Monte Carlo approach to propagate the uncertainty about model parameters and error component through the Michaelis-Menten equation, yielding a probability distribution for both pollutant degradation rate and time for cleanup to some prescribed level. The procedure is illustrated using data related to the degradation kinetics of halogenated hydrocarbons by Methylomicrobium album BG8. It is shown that the assumption of homoscedasticity of the error variance in the Michaelis-Menten model is usually inappropriate, and analytical expressions are derived to account for the dependence of the error variance on the concentration of substrate. Depending on the substrate, the addition of formate might have a significant impact on the expected degradation rates and times, and the proposed approach allows one to test statistically such an impact for various substrate concentrations.