Beyond Brittle Deformation: Insights into Seismogenic Slip Processes from Natural and Experimental Faults

Abstract:

Shear displacements in upper crustal faults are typically localized within cm- to m-thick high strain fault cores composed of interlayered tabular domains of cataclasite and gouge. Evidence from exhumed/exposed seismic faults shows that the great majority of co-seismic slip is taken up along narrow (<10 cm) ultracataclasite slip zones, containing thin (<100μm) principal slip zones (PSZ) bounded by sharp, polished and striated principal slip surfaces (PSS). Even in unconsolidated materials deformed near to the surface, seismogenic slip is observed to localize within discrete, narrow PSZs.

Theoretical studies suggest that in all but the shallowest settings, the natural PSZs may be sufficiently thin to generate localised frictional heating that potentially promotes thermally-activated dynamic weakening mechanisms. We can recreate these processes in the laboratory using displacement-controlled friction experiments performed in a rotary shear apparatus on fault gouges of known composition deformed at seismic slip rates (v > 1ms^-1) and normal stresses of up to 20 MPa. A sequential sampling approach is used in which slip is arrested at different stages of the observed friction evolution (e.g. post-compaction, peak friction, steady state after weakening). This allows the evolution of gouge microstructures and deformation mechanisms in the experimental samples to be: a) related to the evolving temperature regimes in the PSZ and changing mechanical behavior; and b) compared to natural PSZ/PSSs.

Using this approach we have investigated the behavior and deformation mechanisms of gouges made of common, rock-forming minerals (calcite, clays, olivine, quartz) both in pure form and, in some cases, as mixed compositions deformed under a range of experimental conditions. We have studied the effects of varying confining pressure, fluid content (room humidity vs water saturated) and composition (de-ionized water vs brine) and slip rate (e.g. seismic vs. sub-seismic). Our findings – and those of others – reveal a startling diversity of ‘non-brittle’ micro- to nano-scale deformation processes (e.g. viscous GBS, particulate flow). This has implications for our understanding of the frictional strength of faults, the recognition of past seismogenic events in natural examples and the forecasting of future earthquakes.