Upscaling solute transport in naturally fractured porous media with the continuous time random walk method

Sebastian Geiger, A. Cortis, J. T. Birkholzer

    Research output: Contribution to journalArticlepeer-review

    74 Citations (Scopus)


    Solute transport in fractured porous media is typically "non-Fickian"; that is, it is characterized by early breakthrough and long tailing and by nonlinear growth of the Green function-centered second moment. This behavior is due to the effects of (1) multirate diffusion occurring between the highly permeable fracture network and the low-permeability rock matrix, (2) a wide range of advection rates in the fractures and, possibly, the matrix as well, and (3) a range of path lengths. As a consequence, prediction of solute transport processes at the macroscale represents a formidable challenge. Classical dual-porosity (or mobile-immobile) approaches in conjunction with an advection-dispersion equation and macroscopic dispersivity commonly fail to predict breakthrough of fractured porous media accurately. It was recently demonstrated that the continuous time random walk (CTRW) method can be used as a generalized upscaling approach. Here we extend this work and use results from high-resolution finite element-finite volume-based simulations of solute transport in an outcrop analogue of a naturally fractured reservoir to calibrate the CTRW method by extracting a distribution of retention times. This procedure allows us to predict breakthrough at other model locations accurately and to gain significant insight into the nature of the fracture-matrix interaction in naturally fractured porous reservoirs with geologically realistic fracture geometries. Copyright 2010 by the American Geophysical Union.

    Original languageEnglish
    Article numberW12530
    Pages (from-to)W12530
    JournalWater Resources Research
    Issue number12
    Publication statusPublished - 2010


    Dive into the research topics of 'Upscaling solute transport in naturally fractured porous media with the continuous time random walk method'. Together they form a unique fingerprint.

    Cite this