Applications of sensitivity analysis to uncertainty quantification in variably saturated flow

In this paper, we present results demonstrating the effectiveness of a sensitivity analysis approach to uncertainty quantification of a variably saturated flow model. The basis for our method is a software system which simultaneously solves for solutions of large-scale nonlinear systems of equations and the sensitivity of the solutions to selected parameters. We present test cases showing the effects on the relative uncertainty of pressure due to heterogeneity in the absolute permeability and to differences in parameterizing the Van Genuchten curve soil parameters, α and n.