ABSTRACT
Complexity and heterogeneity of the subsurface, combined with ubiquitous scarcity of data characterizing its properties, render quantitative predictions of subsurface processes fundamentally uncertain. While predictive uncertainty in groundwater modeling has long been, and to a large extent continues to be, ignored, its quantication is a touchstone of modern predictive science [1]. is chapter provides a brief introduction to probabilistic methods for uncertainty quantication (UQ), although other frameworks (e.g., fuzzy logic) might be adapted as well. e UQ strategies described in this chapter are Monte Carlo methods (Section 22.3), moment dierential equations (MDEs) (Section 22.4), direct numerical simulations (Section 22.5), and the method of distributions (Section 22.6). Section 22.7 provides an assessment of relative strengths and limitations of these UQ approaches.
