A design of experiment aided stochastic parameterization method for modeling aquifer NAPL contamination. (March 2018)

• Record Type: Journal Article
• Authors: Li, Zelin; Chen, Bing; Wu, Hongjing; Ye, Xudong; Zhang, Baiyu
• Is Part Of: Environmental modelling & software. Volume 101(2018)
• Page Start: 183
• Page End: 193
• Publication Date: 2018-03
• Subjects: Design of experiment -- Stochastic -- Parameterization -- Subsurface NAPL contamination -- Uncertainty; 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.2017.12.014
• Languages: English
• ISSNs: 1364-8152

Abstract

Abstract: Numerical models have been widely applied in simulating subsurface Non-aqueous Phase Liquid (NAPL) contamination processes. In order to examine modeling uncertainties and improve simulation performance, a new hybrid stochastic - design of experiment (DOE) aided parameterization method was developed by using a coupled experimental and modeling approach. In a case study, an existing commercial groundwater modeling tool BioF&T 3D was applied to conduct numerical simulations of subsurface contamination processes based on flow cell experiments. Parameterization results indicated that porosity, distribution coefficient, and Henry's constant were the most significant parameters. The result also revealed their interactions. The DOE predicted responses were found reasonably close to the actual ones from the models' simulations. Monte Carlo simulation was applied to conduct uncertainty analysis within the narrowed parameters ranges, which were generated by centralizing the DOE optimized values, and the combinations of parameters were further updated when better responses were found. After parameterization, R 2 valued 0.80, 0.91, 0.89, and 0.90 for benzene, toluene, ethylbenzene, and xylene (BTEX), respectively. A good consistency (R 2 = 0.76 to 0.90 for BTEX) was also achieved during the model verification, which confirmed that after the parameterization processes, the simulation model can potentially be used for predictions under similar conditions. Highlights: BioF&T 3D was applied to conduct simulations of subsurface contamination processes. Simulating BTEX contamination and natural attenuation processes using flow cells. A hybrid stochastic - DOE aided parameterization (HSDP) method was developed. The HSDP method can efficiently identify key parameters and their interactions. Monte Carlo simulations were used for uncertainty analysis of the modeling system.