Soil heterogeneity and preferential paths for gas migration C.H. Delahaye a,b, * , E.E. Alonso a a Geotechnical Engineering Department, Technical University of Catalonia (DIT-UPC), Barcelona, Spain b Institute of Mining Research, National University of San Juan (IIM-UNSJ), San Juan, Argentina Received 24 July 2000; accepted 31 July 2001 Abstract Preferential paths are often reported in connection with gas breakthrough experiments. These preferential paths may be explained, in a deformable soil skeleton, by the action of pressurised gas along a natural path. A natural path is defined in a spatially heterogeneous material by a set of connected spots of higher gas permeability. Along this path the soil will experience a marked desaturation and gas permeability will increase fast. The paper describes a procedure to simulate these effects. Soil heterogeneity is described by means of random fields, which describe the key soil properties for gas migration. Since actual data concerning soil variability at a small (sample) scale is very limited, a number of computer models of increased complexity have been developed. They maintain, however, a close relationship with the known macroscopic data of the average (homogeneous) sample. The effect of some hypothesis concerning the stochastic structure of soil variability and the role of the hydromechanical coupling has also been explored. Once the heterogeneous soil model is defined, a general purpose code for Thermo-Hydro- Mechanical analysis is subsequently used to perform the analysis. The soil is initially saturated and a non-wetting phase (gas) is forced through the saturated mass either at constant pressure or at a constant flow rate. The results of the different analyses are compared in order to advance in the basic understanding of gas migration through clay barriers. Finally, a discussion of the relevance of soil heterogeneity, on the basis of the analysis performed, is presented. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Soil heterogeneity; Preferential paths; Gas migration 1. Introduction In a previous paper (Delahaye and Alonso, 1997), gas migration through the soil was analyzed by means of a continuum multiphase flow approach. A fully coupled Thermo-Hydro-Mechanical code (CODE- BRIGHT, Olivella et al., 1994, 1996) was used in an attempt to simulate some gas migration experiments performed within the framework of MEGAS project (Volckaert et al., 1995). The most significant soil properties of Boom Clay (water retention curve, water and gas relative permeabilities) were taken from actual measurements reported in Volckaert et al. (1995) and Ortiz et al. (1997). A sensitivity analysis was also performed by varying the water retention character- istics of the soil and, specially, its mechanical proper- ties. It was found that the profile of degree of saturation along the sample, which was determined once the experiments were finished, could be matched quite precisely by the model. Other computed results were, however, very far from actual measurements (time to breakthrough, gas flow rate, variation of gas flow rate 0013-7952/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII:S0013-7952(01)00104-1 * Corresponding author. Geotechnical Engineering Department, Technical University of Catalonia (DIT-UPC), 08034 Barcelona, Spain. www.elsevier.com/locate/enggeo Engineering Geology 64 (2002) 251 – 271