A comparison of various uncertainty models: An example of subsurface contaminant transport

Groundwater resources are under increasing threat of contamination and wasteful use in many parts of the world. Groundwater flow and integrated contaminant transport models are commonly used to predict the fate of contaminants in the subsurface environment. However, the lack of reliable data and complexity of the natural environmental systems, the predictions are subjected to large uncertainties. For reliable decision-making, these contaminant transport models are required to explicitly consider associated uncertainties in their parameters. This paper aims to compare the results of four common uncertainty models using an example of contaminant transport in groundwater. The research employed an advection–dispersion equation (ADE) to describe the transport of a contaminant in groundwater. For simplicity, two parameters – dispersion coefficient and velocity – were considered in the uncertainty analysis. Fuzzy set theory, one- and two-dimensional (1-D and 2-D) Monte Carlo simulations, and Probability Box (P-Box) methods were investigated. The cumulative distribution functions generated from these analyses were compared to evaluate the capabilities of these methods. The comparison showed that P-Box method provides a more comprehensive analysis with lesser assumptions as compared to other methods, and also found to be more pragmatic way to describe and propagate uncertainties in complex environmental systems. Furthermore, execution time required to perform uncertainty analysis using P-Box method is comparatively much less than 2-D Monte Carlo simulations.

► We study uncertainty analysis in subsurface flows.
► We compare the results of fuzzy set theory, one- and two-dimensional Monte Carlo simulations and probability box.
► We consider dispersion coefficient and velocity as uncertain parameters.
► The probability box method provides a more comprehensive analysis.
► The probability box is computationally more efficient.