We analyse the probabilistic nature of the time of travel of conservative solutes to a pumping well operating within a heterogeneous aquifer. The latter is modelled as a three-dimensional, doubly stochastic composite medium, where distributions of geological materials and hydraulic properties are uncertain. The problem is tackled within a numerical Monte Carlo framework. We study the importance of uncertain facies geometry and uncertain hydraulic conductivity and porosity on predictions of solute time of travel to the pumping well by focusing on two special cases in which: (1) the facies distribution is random, but the hydraulic properties of each material are fixed, and (2) both facies geometry and material properties vary