RAPID ROTATION OF NORMAL FAULTS DUE TO FLEXURAL STRESSES: AN EXPLANATION FOR THE GLOBAL DISTRIBUTION OF NORMAL FAULT DIPS

We present a mechanical model to explain why most seismically active normal faults have dips much lower (30–60°) than expected from Anderson-Byerlee theory (60–65°). Our model builds on classic finite extension theory, but incorporates the possibility that the active fault plane may rotate as a response to the build-up of elastic stresses with increasing extension. We postulate that fault plane rotation acts to minimize the amount of extensional work required to sustain slip on the fault. In an elastic layer, this assumption results in rapid rotation of the active fault plane from ~60° down to 30–40° before fault heave has reached 40% of the faulted layer thickness. This rotation mechanism is different from the flexural rotation of the exhumed, inactive fault scarp due to isostatic readjustment and occurs over much smaller amounts of total extension.

In our model, fault rotation rates scale as the inverse of the faulted layer thickness, which is in quantitative agreement with 2D geodynamic simulations that include an elasto-plastic description of the lithosphere. We show that fault rotation promotes longer-lived fault extension compared to continued slip on a high angle normal fault, and therefore holds a strong control on faulting styles (i.e., multiple short-offset vs. dominant large-offset faults). We further argue that the globally observed range of normal fault dips can only be produced in brittle upper-crustal layers thinner than 15-20 km, and that rift zones characterized by thicker brittle layers should feature steeper normal faults. Finally, we predict that oceanic detachment faults should root into shallow-dipping slip surfaces at the axis of mid-ocean ridges. These predictions are testable with careful regional studies of normal fault seismicity coupled to independent estimates of the local rheological behavior of the lithosphere.