Two-dimensional sensitivity analysis of contaminant transport in the unsaturated zone

A two-dimensional probabilistic transport model is developed by coupling a reliability algorithm to a two-dimensional unsaturated flow and transport model, The goal is to determine the significance of the uncertainty of each variable to the probability outcome, Model input includes the mean, standard deviation, and probability distribution of each uncertain variable. The uncertain variables are assumed to be uncorrelated to each other. The uncertain variables are saturated water content, residual mater content, van Genuchten model parameters alpha (a) and n, saturated hydraulic conductivity, bulk density, coefficient of distribution, molecular diffusion coefficient, and dispersivity, Coefficient of variation is used as the measure of uncertainty for each uncertain variable. Three soil samples are considered for model simulations. One half of the flow surface area is the source of a generic contaminant spill. The reliability analysis determines the probability of exceeding a specific target concentration value and also provides a sensitivity value that measures the relative significance of each uncertain transport variable to the probability outcome, Generally, saturated water content tends to be the most significant uncertain variable for transport in the unsaturated zone. Bulk density and coefficient of distribution also are significant for reactive transport. Molecular diffusion and van Genuchten model parameter n are significant in fine-grained soils and also at locations far from the contaminant source area. The uncertainty of saturated hydraulic conductivity and dispersivity have no significant influence on the probability outcome and can be considered as deterministic input variables.