Solute transport in groundwater is characterized by a high level of uncertainty since it is impossible to completely measure the spatial distribution of key soil properties, such as the hydraulic conductivity. Several studies have shown how an heterogeneous hydraulic conductivity field may lead to the formation of preferential channels (or high conductivity channels) in which the solute flows. Given a realization of the hydraulic conductivity field, it is possible to estimate the location of preferential channels using a graph-theory based method. The minimum hydraulic resistance and least resistance path are efficiently computed, so that these metrics can be effectively employed to assess the uncertainty of preferential flows using computationally intensive stochastic methods, such as Monte-Carlo simulations. For example, these metrics can be used for site characterization, choosing locations where to sample the hydraulic conductivity in order to reduce the uncertainty of high conductivity channels. As a consequence, our preliminary studies displayed more than 40% reduction on first arrival time uncertainty, when compared to a regular grid sampling protocol with the same number of sampling locations. The iterative sampling strategy can be reproduced using LazyMole, an open-source tool implementing the algorithms to compute the minimum hydraulic resistance and least resistance path for a given hydraulic conductivity field.