A numerical study of passive transport through fault zones

We have numerically simulated contaminant transport through heterogeneous fault zones that consist of three components: a core, a damaged zone and a protolith, each with a different permeability. The interface between each of the components was assigned a fractal topography. This was done to quantify the scaling of a contaminant plume through a fault, as large-scale faults can dominate regional flow. The numerical method used was a modified lattice gas scheme for flow in porous media, coupled to a finite-difference solution to the advection-dispersion equation. The introduction of heterogeneity leads to complex behaviour of a chemical plume in several faults with the same statistical properties. A range of arrival times is observed along with bimodal break-through curves. This behaviour was a common feature in each heterogeneous fault model. The mean and the variance of the plume were calculated for every simulation and indicated that the transport is anomalous. The mean scaled through time from t(1.00) to t(0.54) for the different models. Thus, the mean velocity decreased with time in each of the fault models. The minimum and maximum scaling through time of the variance of the plume was from t(1.0) to t(1.87) indicating anomalous transport. The simulated break-through curves were analysed using the continuous time random walk method that quantifies the scaling of migrating contaminants using a single parameter beta. For beta > 2 the transport is Gaussian and for beta < 2 the transport is anomalous and the variance scales as t((3-beta)) for 1 < beta < 2. The average value of 8 for the different models ranged from 1.65 to 2.00. We have also shown that the scaling of the transport can vary significantly over short distances relative to the size of our models. (C) 2003 Elsevier B.V. All rights reserved.