Topography of natural and artificial fractures in granitic rocks: Implications for studies of rock friction and fluid migration

The topography of fractures and joints has a strong influence on their frictional strength and fracture permeability. Important aspects of the surfaces are the size, distribution and density of contact spots between the surfaces, which can be calculated using topographic data. For a variety of tensile fracture surfaces in granite, including both natural joints and man-made fractures, spectral and statistical properties are similar, suggesting that man-mane fractures in granites can be acceptable substitutes for natural fractures in some experimental situations. The topographic data indicate that fracture surfaces at scales from 0.1 to 200 mm are approximately fractal and statistically self-affine. A log-log plot of power spectral density us spatial frequency, calculated from the fracture surfaces indicates a spectral slope of -2.3 +/- 0.2, yielding a fractal dimension D between 2.25 and 2.45. Shear fracture surfaces are also fractal, but have anisotropic roughness, which develops during the fracture initiation process. The amount of anisotropy that develops on the shear fracture surfaces is small in comparison to that exhibited by natural fault surfaces, which include anisotropy that results from post failure wear and from surface evolution. For fractures like those measured in this study, simple simulations indicate that the aperture, size and number density of wall-to-wall contacts between the surfaces change non-linearly with relative shear displacement of the surfaces. It is possible to make preliminary estimates of how the permeability and frictional properties of fractures and joints depend on the size, distribution and character of the contact spots. Furthermore, given that artificial fractures can be generated numerically with statistical properties similar to those of natural or man-made fractures, it is reasonable for systematic studies of the effects of roughness on friction and fluid flow to use computer designed surfaces to provide a range of variability that is not easily accessible in the laboratory. (C) 1997 Elsevier Science Ltd.