CONSTRAINTS ON NATURAL FAULT MECHANICS FROM OBSERVATIONS OF LOW-ANGLE NORMAL FAULTS COMPARED WITH LABORATORY ROCK-STRENGTH VALUES

Low-angle normal faults (LANFs; dip <30°) present perhaps the most difficult fault mechanical problem from an Andersonian perspective because they are at high angles to the subvertical maximum principal stress (sigma 1) inferred for extensional settings. Common explanations include rotation of sigma 1 away from vertical or rotation of the LANF from a steeper dip in order to reduce the angle between sigma 1 and the LANF to ~30°. However, field studies of the orientations of minor faults and fractures in rocks surrounding LANFs confirm that sigma 1makes a large angle with LANFs, refuting these models and supporting Andersonian stress orientations. Compounding the difficulty, Byerlee friction of 0.85 is more appropriate than 0.6 for the brittle upper crust.

 

The results from Mohr constructions that avoid common simplifying assumptions are compared to LANFs for which the depth and dip are well known. Common Mohr-Coulomb analyses do not explain slip on LANFs. These assume hydrostatic pore pressure, friction=0.6, and differential stress limited by either optimally oriented faults or by lithostatic load and tensile strength of zero. However, field relationships argue for ongoing mineralization and rock healing around many LANFs, indicating that a cohesionless frictional fault sandwiched between cohesive layers may be a better model. In such models, differential stress at shallow crustal levels is limited by tensile strength and (in the absence of elevated pore pressure) by cohesion at greater depth.  Such models explain slip on most LANFs at shallow depths (<3 km) but fail to explain slip on deeper LANFs. Elevated pore pressure also fails to explain slip on many LANFs. For hydrostatic pore pressure, reduced friction (0.4 to 0.6) explains several LANFs (dipping 20° to 30° at <7 km depth), but friction as low as 0.1 to 0.3 is required in extreme cases (dips of 3° to 15° at <10 km depth). However, most of these LANFs do not have weak materials along the fault plane, suggesting that chemical effects, slip weakening, or other processes are important and active in the crust but are not resolved at laboratory time and length scales. None of these models explain the initiation of brittle LANFs that did not follow some preexisting anisotropy, further suggesting that significant processes active in the crust have not been duplicated experimentally.