With time-lapse electrical resistivity tomography (ERT), transport processes in the subsurface can be imaged and monitored. The information content of obtained spatiotemporal data sets opens new ways to characterize subsurface spatial variability. Difficulties regarding a quantitative interpretation of the imaged transport may arise from the fact that data inversion used in ERT is generally underdetermined and that ERT data acquisition is often limited to two-dimensional (2-D) image planes. To address this problem, we conducted a numerical tracer experiment in a synthetic heterogeneous aquifer with preset variability and spatial correlation of the hydraulic conductivity and monitored the tracer breakthrough in a 2-D image plane perpendicular to the mean flow direction using time-lapse ERT. The tracer breakthrough patterns in the image plane were interpreted using equivalent transport models: an equivalent convection dispersion equation to characterize the spatially averaged breakthrough and a stream tube model to characterize local breakthrough curves. Equivalent parameters derived from simulated and from ERT inverted concentrations showed a good agreement, which demonstrates the potential of ERT to characterize subsurface transport. Using first-order approximate solutions of stochastic flow and transport equations, equivalent model parameters and their spatial variability were interpreted in terms of the local-scale dispersion and the spatial variability of the hydraulic conductivity. The spatial correlations of the stream tube velocity and of the hydraulic conductivity were found to be closely related. Consequently, the hydraulic conductivity spatial correlation may be inferred from stream tube velocity distributions, which can be observed with sufficiently high spatial resolution using ERT.