An uncertainty and probabilistic sensitivity study of a hypothetical underground high-level waste (HLW) repository intersected by a vertical fracture or fault and under saturated conditions is presented. Several recently developed probabilistic methods, including the advanced mean value method and the adaptive importance sampling method, are applied to a previously developed one-dimensional analytical model. These probabilistic methods are based on a limit-state formulation and provide an effective means of computing performance probability distribution and probabilitybased random parameter sensitivities. A numerical example related to the transport of 237Np in a system of layered fractured rock is used to illustrate the application of these probabilistic methods for efficient uncertainty and probabilistic sensitivity analyses.