[1] Temperature is believed to have an important control on frictional properties of rocks, yet the amount of experimental observations of time-dependent rock friction at high temperatures is rather limited. In this study, we investigated frictional healing of Westerly granite in a series of slide-hold-slide experiments using a direct shear apparatus at ambient temperatures between 20 C and 550 C. We observed that at room temperature coefficient of friction increases in proportion to the logarithm of hold time at a rate consistent with findings of previous studies. For a given hold time, the coefficient of friction linearly increases with temperature, but temperature has little effect on the rate of change in static friction with hold time. We used a numerical model to investigate whether time-dependent increases in real contact area between rough surfaces could account for the observed frictional healing. The model incorporates fractal geometry and temperature-dependent viscoelasoplastic rheology. We explored several candidate rheologies that have been proposed for steady state creep of rocks at high stresses and temperatures. None of the tested laws could provide an agreement between the observed and modeled healing behavior given material properties reported in the bulk creep experiments. An acceptable fit to the experimental data could be achieved with modified parameters. In particular, for the power-law rheology to provide a reasonable fit to the data, the stress exponent needs to be greater than 40. Alternative mechanisms include time-dependent gouge compaction and increases in bond strength between contacting asperities.

The deep limit to seismicity in continental crust is believed to be controlled by a transition from velocity‐weakening to velocity‐strengthening friction based on experimental measurements of the rate dependence of friction at different temperatures. Available experimental data on granite suggest a transition to stable creep at about 350°C (∼15 km depth). Here we present results from unconfined experiments on Westerly granite at both dry and hydrated conditions that show increasingly unstable slip (velocity‐weakening behavior) at temperature up to 600°C. A comparison of previously published experimental results with those presented in this study suggests that the rate and state friction parameters strongly depend on normal stress and pore pressure at high (>400°C) temperature, which may help explain regional variations in the depth distribution of earthquakes in continental crust. Temperature dependence of the rate and state friction parameters may also contribute to strong dynamic weakening observed in high‐speed friction experiments on crystalline rocks such as granite and gabbro.

We conducted a series of experiments to explore the rate and state frictional properties of gabbro at conditions thought to be representative of slow slip events (SSEs) in subduction zones. The experiments were conducted using a heated direct shear apparatus. We tested both solid and simulated gouge samples at low effective normal stress (5-30 MPa) over a broad range of temperatures (20-6008C) under dry and hydrated conditions. In tests performed on dry solid samples, we observed stable sliding at low temperatures (20-1508C), stick slip at high temperatures (350-6008C), and a transitional ''episodic slow slip'' behavior at intermediate temperatures (200-3008C). In tests performed on dry gouge samples, we observed stable sliding at all temperatures. Under hydrated conditions, the gouge samples exhibited episodic slow slip and stick-slip behavior at temperatures between 300 and 5008C. Our results show a decrease in the rate parameter (a 2 b) with temperature for both solid and gouge samples; friction transitions from velocity strengthening to velocity weakening at temperature of about 1508C for both solid and gouge samples. We do not observe transition to velocity-strengthening friction at the high end of the tested temperature range. Our results suggest that the occurrence of slow slip events and the downdip limit of the seismogenic zone on subduction megathrusts cannot be solely explained by the temperature dependence of frictional properties of gabbro. Further experimental studies are needed to evaluate the effects of water fugacity and compositional heterogeneity (e.g., the presence of phyllosilicates) on frictional stability of subduction megathrusts.