Earthquakes in wet rocks: competition between thermal pressurization, damage, and frictional melting?

Coseismic frictional melting is inefficient when thermal pressurization is the dominant dynamic weakening mechanism. Yet, pseudotachylytes (quenched coseismic frictional melt) are found along some exhumed fluid-rich faults, implying that melting occurred despite the fault being wet. To solve this conundrum, we studied a pseudotachylyte-bearing thrust hosted in cherts from the Jurassic accretionary complex in central Japan. The pseudotachylyte has sharp margins, except embayments into thermally eroded wall-rock chlorite and brecciated wall-rock cherts, and contains thermally eroded quartz clasts, providing evidence for a melt origin. Microstructural observations of co- to postseismic goethite cements, however, provide evidence of pressurized aqueous fluids along the fault, that led to localised precipitation in extensional breccia domains along and immediately around the fault. We suggest that the extraction of fluids into damaged wall rocks led to a fluid pressure drop and consequent fault restrengthening, leading to more efficient frictional heating and pseudotachylyte production. Numerical simulations of coseismic stress evolution involving thermal pressurization support a model where wall-rock permeability has a dominant effect on dynamic weakening along slip zones of sub-cm thickness. We use the microstructural record to estimate that ∼ 2 MJ/m 2 and ∼ 5 MJ/m 2 of energy is consumed on-fault as heat during thermal pressurisation and frictional heating, respectively. This two-stage, serial dynamic weakening process implies that dilatant hardening may limit slip on faults in wet rocks, unless sufficient elastic strain energy is available to overcome hardening and activate melt lubrication as an additional weakening process after the initial thermal pressurisation.