Faulting mechanisms in high-porosity sandstones; New Red Sandstone, Arran, Scotland

Publisher Website
Google Scholar
Abstract
Summary: Faults in the ‘New Red’ aeolian sandstones of Arran are unusual, firstly for occurring as closely-spaced (less than 1 m) often conjugate sets affecting large volumes of rock, and secondly for forming upstanding fault zones with numerous anastomosing strands of granulated rock, each preserving a small increment of slip. Anisotropy, such as bedding and cross-bed sets, has no discernible effect on fault behaviour. In contrast to the underlying Carboniferous rocks, large displacements are rarely concentrated on a single fault plane within the high-porosity sandstones. The proposed cause is slip-hardening of each fault after a very small displacement (less than 10 mm) causing the next slip increment to be taken up through undeformed rock rather than on the original plane. The common factor in recent records of similar faults elsewhere is their occurrence in high-porosity sandstones. Because of the low grain-contact strength, these rocks are partly analogous to unconsolidated sediment. The high porosity promotes high grain-contact stresses which induce rapid cataclasis during initial slip. Grain fracture and spalling of iron oxide coatings and quartz overgrowths produce a seam with reduced grain size, poorer sorting, higher angularity and lower porosity than the unfaulted rock. These factors collectively strengthen the seam because the coefficient of friction is increased, even though cohesion is reduced. This results in a Mohr failure envelope that lies outside the envelope of the undeformed rock for most stress states. A transient pore pressure increase in the fault seam may be important during slip. Rocks deformed by this slip-hardened faulting preserve a record of each increment of strain. If the displacement on each individual fault seam is the same, the geometry of the total fault systems is directly related to the bulk strain. Quadrimodal systems observed by us in Arran, and by others elsewhere, are probably a response to triaxial strain and show that bimodal ‘Andersonian’ fault systems are only special plane strain cases. If the bulk strain is irrotational, both the orientation and relative magnitude of the principal strains might be estimated.