This project focuses on the evolution of fault strength during large earthquakes. Accumulating constraints from theoretical, field and laboratory investigations highlight the importance of thermal processes to dynamic weakening. In mature fault zones that have accommodated many large earthquakes and are characterized by gouge layers that greatly exceed the thickness of the mm-scale "principal slip surfaces" in which shear is localized, the thermal pressurization of pore fluids is particularly important for reducing the fault strength and limiting the extent of shear heating. Nevertheless, for slip distances within the typical range for large earthquakes and reasonable estimates of hydraulic transport properties and other controlling variables, predicted temperature increases do often reach the onset of melting, especially at mid to lower seismogenic depths (e.g.10km). Calculated fracture energies from these models are consistent with seismic estimates. However, field evidence of melt products from exhumed mature fault zones suggests that macroscopic melting is actually quite rare. Those melt products that are recovered display a range of layer thicknesses and crystal contents, which indicate that significant shear heating continued long after melt onset. An examination of the transition to melting in a finite shear zone is required to better understand the dynamics and strength evolution of large earthquakes, and how the fusion of gouge solids affects the system behavior. There are two main conceptual challenges: 1. the energy input for frictional heating is generally assumed to be proportional to the effective stress, which vanishes when macroscopic melt layers are produced and thermodynamic considerations require that the melt pressure balance the normal stress; 2. the typical initial crystal content of a finite shear zone at melt onset almost certainly exceeds the critical solids fraction (~50%) that allows for slurry mobilization at a finite effective viscosity and provides the viscous heat source necessary for the melt fraction to increase subsequently. The former consideration motivates a closer examination of changes to the effective frictional behavior as melting first begins at highly stressed (um-scale) asperity contacts and changes in the real area of contact occur. The latter consideration suggests the likely roll of melt onset as a mechanism for extreme localization, requiring slip in a finite zone to be actually accommodated on a series of short-lived effective shear surfaces between adjacent melting gouge particles. This project encompasses a multi-scale modeling effort that combines: 1. focused studies aimed at resolving key micro-scale interactions such as the factors that determine the shear zone width, and reductions to the effective friction coefficient due to flash-melting; 2. continuum models that describe the evolution of fault strength and temperature - including the influence of potential additional energy sinks and water sources; and 3. targeted efforts that explore the dynamic implications of the melt transition on rupture characteristics. The project speaks directly to the wider scientific debate over the average strength of mature, plate-bounding faults. Moreover, the evolution of fault strength is a key factor in determining whether slip distributions are best approximated by self-healing pulses or crack-like behavior. Recurrence intervals and long-term fault dynamics are also strongly influenced by the evolution of stress during earthquakes themselves, as are the patterns of seismic damage. The results forthcoming from this project will provide valuable insight into these and other broader scientific questions.
