Streamline-based simulation of advective–dispersive solute transport

Abstract

We extend the streamline method to model diffusion and dispersion in solute transport problems using an operator splitting technique. Fluid is moved along streamlines ignoring dispersion, and then component concentrations are mapped onto the underlying grid. Dispersion is included by solving the dispersive portion of the conservation equation on the grid. We quantify the effects of numerical dispersion resulting from remapping solutions from streamlines to the grid. We verify the method by comparison with one-dimensional analytical solutions and by predicting the results of a tracer test in a heterogeneous sand pack. We show the ability of our formulation to handle finely resolved geological models by running a 1.122 million cell field-scale problem. At both the laboratory and field scales we find long tails in the breakthrough curves. The average behavior is consistent with anomalous or non-Gaussian transport where wide variations in permeability result in a large spread of flow velocities. Local mixing caused by dispersion or diffusion can impact the overall recovery of tracer but does not significantly affect the late-time behavior of the plumes.