Stress amplification in rigid blocks of lower-crustal shear zones is controlled by bulk strain rate

Seismic failure of dry lower-crustal rocks requires very high differential stress on the gigapascal-level. Among the mechanisms proposed to generate such high stresses is the so-called jostling block model, in which stress is amplified in rigid blocks within lower-crustal shear zone networks, leading to seismic failure. The model is based on field observations from the Musgrave ranges, Australia and the Nusfjord ridge, Lofoten, northern Norway, where pseudotachylytes (quenched frictional melts produced by coseismic slip) occur within the aforementioned structural setting.

Here we present numerical models to test if stress can be amplified in jostling blocks to the levels necessary to fracture dry, intact, lower-crustal rocks, and on which timescales such stress amplification can be achieved. Our models are based on the geometries and material properties determined in the Nusfjord locality. We systematically test the influence of strain rate, viscosity, loading conditions (pure vs. simple shear), and geometry (shear zone thickness, spacing, angle) and find that the bulk strain rate has the most significant impact on both the magnitude and rate of stress amplification. At high to moderate strain rates of 10^-10-10^-12 s^-1 stress amplification to the required level is achieved in years to hundreds of years, while lower strain rates are insufficient to reach the required stress levels. Average long-term strain rates in the in the crust are on the order of 10^-13-10^-15 s^-1, and transiently high strain rates are reported from both field localities mentioned above. Our numerical results are thus well-supported by the rock record. Furthermore, we find that a high viscosity contrast in our models is necessary to reproduce the geometries observed in the field. A third notable contributor to the magnitude of stress amplification that can be reached in the jostling-block geometry is the loading conditions. Specifically, we find that the impact of pure shear on stress amplification is greater compared to simple shear. Shear zone angle and spacing typically have a minor effect. In contrast, increased shear zone width leads to a reduction of stress in the blocks as strain is accommodated fully by the viscous shear zones, and elastic loading of the rigid blocks is no longer necessary to accommodate bulk strain.

Our results clearly demonstrate that, geometric and material properties contribute to stress amplification in different ways, but that strain rate is the controlling factor. In fact, our results indicate that at moderate to high strain rates, stress amplification to levels necessary for failure of intact lower-crustal rocks in shear zone networks is not only plausible, but inevitable.