Spatial and spacing distribution of joints at (over-) saturation in the turbidite sandstones of the Marnoso-Arenacea Fm. (Northern Apennines, Italy)

Despite that many field studies on joint attribute statistics are available in the literature, there is still an ongoing debate on the most effective statistical treatment to properly describe the distribution of a fundamental parameter like joint spacing. This may also relate to the objective difficulty of collecting very large datasets from single linear scanlines directly in the field, because of limited exposures. To overcome such a limitation, we selected a spectacular outcrop of the Miocene Marnoso-Arenacea Formation turbidite sandstones located near the Coniale village, in the Northern Apennines (Italy), where a vertical cliff provides a 100% continuous outcrop about 90 m high and 280 m long, with well-accessible exposures at its base, along the Santerno River. This setting allows combining data acquisition on a photogrammetric DOM (digital outcrop model) covering a representative sector of the outcrop, with ground truthing by direct measurements in the field. More than 2000 joints were identified in 9 linear scanlines traced in the DOM along mechanical layers with different thickness. Comparison with field data indicates that both joint orientation and spacing data from the DOM are very reliable. The scanlines include between 549 and 37 joints, depending on layer thickness, allowing to perform robust statistical analysis. The Kolmogorov-Smirnov (K–S), Cramer-von Mises (CvM) and Anderson-Darling (A-D) nonparametric tests were used to identify the probability density function(s) that best fit spacing data among exponential, Gamma, log-normal and normal distributions. The Gamma distribution turned out as the most probable in most cases. In some specific sectors along all scanlines, joint density is significantly greater than the average for the presence of joint clusters. In some cases, joint clusters are vertically aligned with those in underlying and/or overlying scanlines, thus resulting in thoroughgoing joint corridors. Comparison of spacing statistics inside and outside joint corridors/clusters indicates a typical transition from Gamma distributions outside joint corridors/clusters, to log-normal distributions inside them. The presence of joint clusters and corridors, as well as FSR, mode/mean ratio, JSR, and FSI values, suggests that diffuse jointing in the sandstone layers of the turbiditic succession reached the saturation stage before a normal probability density function was produced, and then further strain was accommodated by localized deformation in clusters and corridors