The role of fluids in governing rupture modes and seismic radiation on mature faults

Abstract:

Accumulating geophysical evidence, including observations of heat flow, steep angles between inferred principal stress directions and fault traces, and the geometry of thrust-belt wedges, suggests that the shear resistance acting on mature faults must be low (< 20 MPa). A key ingredient in many explanations for such apparent weakness of faults is the presence of fluids in the fault shearing layer. Two contrasting hypotheses used to explain such low-stress, low-heat operation of mature faults are that they are either chronically weak, such as from low effective confinement due to pervasive fluid overpressure, or are quasi-statically strong but undergo considerable enhanced dynamic weakening at seismic slip rates. One particular mechanism for such enhanced weakening is the thermal pressurization (TP) of pore fluids. Here, we explore seismologically inferable quantities in fault models with parameter regimes consistent with low-stress, low-heat operation on matures faults, using fully dynamic simulations of earthquake sequences on rate-and-state faults with varying degrees of efficiency in enhanced weakening due to TP.

We find that fault models including a combination of these two hypotheses, with some degree of chronic fluid overpressure along with moderate enhanced dynamic weakening, produce crack-like to mildly pulse-like ruptures and work best for reproducing an existing range of observations, including nearly magnitude-invariant static stress drops, increasing average breakdown energy with rupture size, apparent stress between 0.1 to 1 MPa, and radiation ratios between 0.1 and 1.0. In contrast, fault models with more efficient weakening that produce sharp self-healing pulses result in higher values of radiated energy per moment (and hence higher apparent stress) than inferred teleseismically for natural megathrust earthquakes. The larger radiated energy per moment of self-healing pulses is similar to limited regional inferences for crustal strike-slip faults. Our findings suggest that (1) persistently weak fault models with moderate enhanced dynamic weakening may be plausible representations of natural subduction zones, unless teleseismic estimates underestimate radiated energy, and (2) crustal mature faults may differ, with less chronic but more dynamic weakening.