Numerical and analytical model for dipole tracer tests in heterogeneous aquifers

Heterogeneity of hydraulic conductivity in aquifers is a topic touched upon by many researchers, but very few analyses were reported on the characteristic parameters of solute transport using stochastic methods. We study he effects of aquifer heterogeneity by running numerical simulations of dipole tracer tests in the software MODFLOW 6, instigated with Python scripts using the FloPy package. Random hydraulic conductivity fields are drawn by the Python package GStools from a log-normal distribution, characterized by a mean, variance and correlation length. A series of ensembles was run with varying variances and correlation lengths. Normalised mean breakthrough curves of the ensembles were compared for homogeneous, stratified and heterogeneous aquifers. Outcomes show that an increase in variance leads to an earlier arrival of the tracer, due to the occurrence of preferential flow paths, but will also lead to more tailing. An increase in correlation length enhances the effects of an increase in variance. For a short correlation length (i.e. much smaller than the distance between the wells) the heterogeneous solution tends towards the solution for a homogeneous medium, while for a longer correlation length, the solution tends towards the solution for a stratified medium. Based on the ensemble results we conclude that a fully heterogeneous aquifer shows a different breakthrough curve than stratified or a homogeneous structured aquifers. Therefore it cannot be treated the same way.