Effect of clay content and distribution on hydraulic and geophysical properties of synthetic sand-clay mixtures

Osterman, Gordon and Sugand, Madhuri and Keating, Kristina and Binley, Andrew and Slater, Lee (2019) Effect of clay content and distribution on hydraulic and geophysical properties of synthetic sand-clay mixtures. Geophysics, 84 (4). E239-E253. ISSN 0016-8033

Text (Author copy of accepted manuscript)
GEO_2018_0387.R1_Submitted.pdf - Accepted Version
Available under License Creative Commons Attribution-NonCommercial.
Download (853kB)

Abstract

Geophysical methods show promise for detecting the spatial variability of subsurface clay content and its effect on subsurface hydraulic properties. We have developed a laboratory study that examines the influence of clay content and distribution on the relationships between hydraulic conductivity K and the physical and geophysical properties of the media. Two geophysical methods are investigated: spectral induced polarization (SIP) and nuclear magnetic resonance (NMR). We used synthetic sediment mixtures of sand and up to 10% kaolinite clay by mass; the clay was homogeneously mixed or was present as large (approximately 5 mm) clusters distributed through the sample. The K varies moderately well (normalized root-mean-square error [Nrms error] = 0.393) with the pore-volume normalized surface area S-por a proxy measure of clay content, in the homogeneous samples and poorly (Nrms error = 0.507) when the clustered samples are included in the fit. The SIP parameters indicate moderately good to excellent fits with S-por for homogeneous samples (Nrms error = 0.0783-0.139) and moderately good to good fits for clustered samples (Nrms error = 0.140-0.336), and the coefficients describing the polarizability of the samples depend on the clay distribution. NMR parameters vary moderately well with S-por in the homogeneous samples (Nrms error = 0.341-0.412) and poorly (Nrms error = 1.08-6.04) in the clustered samples. The SIP parameters vary moderately well with K (Nrms error = 0.301-0.466); however, the relationship between the SIP parameters and K is compromised by the nonnegligible polarization of the clay clusters. NMR parameters indicate good to excellent fits with K (Nrms error = 0.0789-0.116). For SIP and NMR, fitting homogeneous and clustered samples together does not compromise the fit quality. These results suggest that the geophysical measurements are better predictors of K in heterogeneous porous media than bulk measures of pore geometry such as S-por.