Implications of uncertain bioreactive parameters on a complex reaction network of atrazine biodegradation in soil

We study propagation of uncertainty related to bioreactive parameters through the biodegradation reaction network of atrazine (ATZ) and its metabolites in soils. The work is motivated by the recognition that detailed analyses of these aspects in models of complex biogeochemical systems is especially critical when the feedback amongst hydraulic, chemical and microbial processes is strong. With an emphasis on microbial processes, the ATZ reaction network is exemplary because it includes 19 biological reactions mediated by 3 microbial functional groups that compete in 6 metabolic and co-metabolic degradation pathways under aerobic and anaerobic conditions. Propagation of model parametric uncertainty yields the probability distributions of ATZ, its metabolites, and the microbial biomass. It also enables us to identify key parameters that drive the long-term system dynamics. We base our analyses on an initial screening of model parameters and a subsequent moment-based global sensitivity analysis, and we show that the probability density of target outputs are multimodal as a result of nonlinearities in the degradation rate formulations. The employed global sensitivity metrics reveal that joint effects of the uncertainty of multiple kinetic parameters significantly contribute to output uncertainties, thus suggesting the presence of a high level of coupling between processes included in the network. Specific biomass affinity to a given substrate is found to be a lumped indicator that can facilitate the analysis of microbial growth when multiple and competing microbial functional groups coexist.