Integrated Bayesian Multi-model approach to quantify input, parameter and conceptual model structure uncertainty in groundwater modeling. (April 2020)

• Record Type: Journal Article
• Authors: Mustafa, Syed Md Touhidul; Nossent, Jiri; Ghysels, Gert; Huysmans, Marijke
• Is Part Of: Environmental modelling & software. Volume 126(2020)
• Publication Date: 2020-04
• Subjects: Conceptual model structure uncertainty -- Bayesian approach -- Input uncertainty -- Bayesian model averaging -- Uncertainty quantification -- Groundwater flow model; Environmental monitoring -- Computer programs -- Periodicals; Ecology -- Computer simulation -- Periodicals; Digital computer simulation -- Periodicals; Computer software -- Periodicals; Environmental Monitoring -- Periodicals; Computer Simulation -- Periodicals; Environnement -- Surveillance -- Logiciels -- Périodiques; Écologie -- Simulation, Méthodes de -- Périodiques; Simulation par ordinateur -- Périodiques; Logiciels -- Périodiques; Computer software; Digital computer simulation; Ecology -- Computer simulation; Environmental monitoring -- Computer programs; Periodicals; Electronic journals; 363.70015118
• DOI: 10.1016/j.envsoft.2020.104654
• Languages: English
• ISSNs: 1364-8152

Abstract

Abstract: A flexible Integrated Bayesian Multi-model Uncertainty Estimation Framework (IBMUEF) is presented to simultaneously quantify conceptual model structure, input and parameter uncertainty of a groundwater flow model. In this fully Bayesian framework, the DiffeRential Evolution Adaptive Metropolis (DREAM) algorithm with a novel likelihood function is combined with Bayesian Model Averaging (BMA). Four alternative conceptual models, representing different geological representations of an overexploited aquifer, have been developed. The uncertainty of the input of the model is represented by multipliers. A novel likelihood function based on a new heteroscedastic error model is included to extend the applicability of the framework. The results of the study confirm that neglecting conceptual model structure uncertainty results in unreliable prediction. Consideration of both model structure and input uncertainty are important to obtain confident parameter sets and better model predictions. This study shows that the IBMUEF provides more reliable model predictions and accurate uncertainty bounds. Highlights: Full Bayesian multi-model approach to quantify uncertainty of MODFLOW model. Simultaneously quantifies model structure, input and parameter uncertainty. DREAM with a novel likelihood function is combined with BMA. Neglecting conceptual model uncertainty results in unreliable prediction. Results in more reliable model predictions and accurate uncertainty bounds.