GNGTS 2019 - Atti del 38° Convegno Nazionale

56 GNGTS 2019 S essione 1.1 following power law equation: G (k) ∝ a . k –β where the exponent of the power law β is related with the fractal dimension, D, and the Hurst number (H): D = (7 – β / 2 ; H = (β – 1) / 2 Results anddiscussion. Four 3Dsurfacemodels from two localities (Forca di Presta andColli Alti e Bassi; Fig. 1) were built and analysed. Surface models contained several heterogeneities, such as fractures and striations (Fig. 1a and 1b). Faults are expected to show an anisotropic self- affine behaviour, that is, they scale differently in the slip and perpendicularly to slip directions. In agreement with Scholz (2019), this is likely to be the direct result of wear produced by friction resulting in striated surfaces (Fig. 1b). Consequently, the linear interpolation of the power spectra in a log/log graph should indicate this anisotropy in the form of a different Hurst number (Fig. 2c). Our results show that the analysed surfaces are only slightly anisotropic roughness ratio (ǀH perpendicular /H parallel ǀ < 1.1). However, previous works (Candela et al. , 2012) considering data from kilometre to micrometre scale indicated higher roughness ratio values (near 1.33) with Hurst values equal to 0.6 and 0.8, respectively along and perpendicular to slip. From a visual inspection of the surfaces, it is possible that the low value of roughness ratio is a consequence of weathering (un-related to faulting). Fig. 2 - a) and b) Surface models showing different heterogeneities such as fractures and slip marks indicators, c) Power Spectral Density analysis along and perpendicular to fault slip. A detailed evaluation of the fault roughness (avoiding fractures) indicates that roughness varies along dip. This variation could be quantified in terms of wavelength, Hurst values. A similar trend is followed by the pre-factor, a, and other indicators of the roughness based on the distance to the best fitting plane (indicating also asperity height changes). This also maybe be related to weathering of the surfaces. At the cm-scale, the roughness seems very isotropic. Therefore, the features creating anisotropy are likely to have a wide wavelength. These results