Frictional Strengths of Subduction Thrust Rocks in the Region of Shallow Slow Earthquakes

Abstract

Earthquakes are expected to nucleate within velocity‐weakening materials; however, at the updip limit of the subduction seismogenic zone, the principal lithologies exhibit velocity‐strengthening behavior. At an exhumed analogue for present‐day subduction at Nankai (the Mugi Mélange, Japan) two examples of paleoseismic features occur within altered basalts, suggesting that they may be velocity‐weakening. We shear altered basalt and shale matrix from the mélange in the triaxial saw cut configuration at in situ conditions of deformation (Pc = 120 MPa, T = 150 °C) for two pore fluid factors (λ = 0.36, 0.7). The shale matrix exhibits a coefficient of friction between 0.4 and 0.5, and velocity‐strengthening behavior. Altered basalt exhibits Byerlee friction and velocity‐weakening behavior. In most experiments deformation was partitioned into Riedel shears, which have a lower percentage of clasts, clast size distributions with higher fractal dimensions, and shape factors indicating rounder clasts compared to the matrix between Riedel shears. We hypothesize that earthquakes preferentially nucleate within altered basalt at the updip limit of the seismogenic zone. More complex forms of deformation (e.g., shallow very low frequency earthquakes) may occur by mixing velocity‐weakening altered basalt into the velocity‐strengthening shale matrix. In subduction zones where the matrix is composed of shale, this complex behavior is limited to shallow depths (T < ~250 °C), above the updip transition to velocity‐weakening behavior for the matrix. Altered basalt is a ubiquitous subducting material, and future studies on its behavior through a range of subduction zone conditions are required for a full understanding of the mechanical behavior of subduction zones.

Evidence of Localized Failure Along Altered Basaltic Blocks in Tectonic Mélange at the Updip Limit of the Seismogenic Zone: Implications for the Shallow Slow Earthquake Source

“…In their derivations, the critical nucleation length ( L ) is described by where μ ′ is the shear modulus, C is a numerical constant (with a value between 1 and 1.5), D c is the critical slip distance determined through experiments, b is the evolution effect determined in rate‐and‐state experiments, and σ n is the normal stress on the fault surface. We calculate the critical nucleation lengths for altered basaltic blocks using the experimental parameters determined in triaxial deformation experiments on altered basalt at the in situ conditions of the Mugi mélange (Phillips et al, 2020), which are summarized in Table 1. We consider the elasticity of the system (altered basaltic blocks within a shale matrix) to be controlled by the shear modulus of the shale, which has been determined in laboratory velocity measurements on samples collected from the Mugi mélange ( V s = ~2,000 m/s corresponding to μ′ = ~11.5 GPa at pressure) (Hashimoto et al, 2013).…”

“…(a) Diameter of minimum nucleation patch size ( L c ) for unstable slip after Uenishi and Rice (2003) for varying stress drops and critical slip distances. Red dots show results from experiments on altered basalt from Phillips et al (2020) at pore fluid factors ( λ ) of 0.36 (low) and 0.7 (high). Arrow shows the expected evolution of the critical nucleation length as pore fluid pressures increase to near‐lithostatic conditions ( λ = 0.99).…”

“…Calculations of critical nucleation lengths for basaltic blocks from Mugi are presented in Figure 9a. The frictional properties were determined experimentally at in situ conditions, and the stress drops associated with velocity steps ranges from 1.02 to 0.44 MPa for pore fluid factors of 0.357 and 0.7, respectively (Table 1) (Phillips et al, 2020). The critical nucleation length at these pore fluid pressures is on the order of 1-2 m, indicating that rupture along the surface of the larger blocks in our study area could have been dynamic and produced microseismic events (e.g., normal earthquakes in terms of frequency distribution of seismic signals) (Figure 9).…”

“…In summary, at lithostatic pore fluid pressure, all of the observed blocks would be below the critical size to nucleate an earthquake and might slip aseismically, while if pore fluid pressures dipped, then the largest individual blocks could rupture to produce microseismicity. Phillips et al (2020) Used to Calculate Critical Nucleation Lengths (Figure 4a) Note. T = temperature, P c = confining pressure, λ = pore fluid factor, b = evolution effect, σ n = effective normal stress, Δσ = stress drop, D c = critical slip distance.…”

“…In this contribution, we examine faults along altered basaltic blocks from the Mugi mélange, an exhumed shallow subduction interface. We discuss the potential of individual slip surfaces to be related to slow earthquake phenomena, use frictional characteristics of these rocks (Phillips et al, 2020) and the geologic dimensions to predict the style of slip, and compare our observations from the updip limit of the seismogenic zone to geologic observations from the downdip limit.…”