In many cases, geostatistical properties of multiple sets of joints, shear bands and faults are dealt together particularly for faults formed by shearing of joints and joint zones. Attempts have been made by some investigators to distinguish between these different modes of fractures and their distributions (Myers, 1999; Davatzes and Aydin, 2003a; b; Flodin and Aydin, 2004; and de Joussineau and Aydin, 2007) provide scanline data in which fracture modes are identified and their lengths and spacing properties were analyzed separately. One of the most comprehensive treatment of the subject of multisets of fractures by Bonnet et al. (2001) who attempted to distinguish between data for joints and faults and their fractal dimensions and linear sizes (Figure 1). The authors highlighted several issues such as band-width of observation (sampling and truncation effects); nonfractal scaling; clustering; orientation (fracture anisotropy); dimension of observation; and fracture types (joints versus faults). The plots show common fractal dimensions between 1.5 to 2.0. The major difference is in the linear sizes of the joint and fault systems, the latter of which is of interest here. Obviously the fault systems have larger linear sizes. The authors highlighted several issues such as band-width of observation (sampling and truncation effects); nonfractal scaling; clustering; orientation (fracture anisotropy); dimension of observation; and fracture types (joints versus faults).

Figure 1. (a) Fractal dimension and linear size of joint and fault systems. As expected faul systems have much greater linear size. (b) The most common fractal dimensions are between 1.5-1.6, and 1.9-2.0. From Bonnet et al (2001).