11:45 AM - 12:15 PM
[SCG62-05] A weak fault model that does not require high pore pressure and their seismological background
★Invited Papers
Keywords:fault strength, stress, visco-elastic, pore pressure, dissolution-precipitation creep
1. Findings indicating that the fault is weak
Since Zoback (1987) pointed out that the shear stress acting on the San Andreas Fault is very small, a number of observations have shown that the strength of not only faults at plate boundaries but also inland faults is small. In the review by Copley (2018), it is stated that indirect inferences based on various geological and geophysical observations suggest that faults fail in earthquakes at shear stresses < about 50 MPa and effective friction coefficients below 0.3, possibly equivalent to about 0.05. Recently, a lot of convincing evidence based on high-precision seismic observation data has begun to emerge (e.g., Yoshida et al., 2014, Iio et al., 2022).
The reason why the strength estimated from the observed data is much smaller than that estimated from the friction coefficient of the fault measured in rock friction experiments (Byerlee, 1978) is often explained by the decrease in the normal stress acting on the fault plane due to high pore pressure (e.g. Rice, 1992). However, unlike plate boundaries, it is unlikely that fluids with high pore pressure will be continuously supplied to inland faults over a long period of time. Although there are cases where fluid activity that is thought to have risen from deep in the earth after a major earthquake is observed, the time scale for this is thought to be at most a few months (Sibson, 1988).
Analysis of fault rocks obtained from drilling of the San Andreas Fault has reported that strength is reduced by smectite and talc (Lockner et al., 2011, Moore & Rymer, 2007). These fault materials are stable in the shallow crust, where temperatures are relatively low, but it has also been reported that the strength of faults can decrease dramatically due to the plastic deformation of phyllosilicates even at greater depths (e.g., Wintsch et al., 1995; Bos and Spiers, 2002; Niemeijer and Spiers, 2005). However, since these fault materials generally exhibit velocity-strengthening frictional characteristics, these ideas can be applied to creeping faults, but it is thought that they are difficult to apply to seismogenic faults.
2. Finite element model of weak faults
Therefore, we constructed a new weak fault model that does not require high pore pressure and is capable of causing earthquake slips. Yamamoto et al. (2001, 2002) showed that a fault becomes weak when there is a part that can support normal stress but not shear stress. High-pressure pore water is exactly such a thing, but isolated viscoelastic regions are thought to relax over time and become similar.
We investigated in detail how the strength of a fault would change over a long period of time, when viscoelastic regions exist in patches along the fault using the finite element method. Specifically, we examined the spatiotemporal changes in stress and slip over a long period of time by periodically arranging rectangular regions (called gouge regions) with a cross-sectional shape perpendicular to the fault, along the fault. First, we set the normal stress to a constant value, then added shear deformation to the fault and investigated at what point the fault would begin to slip. We set up several different types of gouge region, from thin rectangles to squares, and in all cases we were able to reproduce the fact that macroscopic strength is lower than in the case of a homogeneous model.
This is a fault model, but it would be more accurate to describe it as a model of the host rock near the fault plane. When investigating the faulting, the properties of the fault plane are often the focus of attention, but it is also thought that three-dimensional treatment that includes the surrounding rock is important.
Since Zoback (1987) pointed out that the shear stress acting on the San Andreas Fault is very small, a number of observations have shown that the strength of not only faults at plate boundaries but also inland faults is small. In the review by Copley (2018), it is stated that indirect inferences based on various geological and geophysical observations suggest that faults fail in earthquakes at shear stresses < about 50 MPa and effective friction coefficients below 0.3, possibly equivalent to about 0.05. Recently, a lot of convincing evidence based on high-precision seismic observation data has begun to emerge (e.g., Yoshida et al., 2014, Iio et al., 2022).
The reason why the strength estimated from the observed data is much smaller than that estimated from the friction coefficient of the fault measured in rock friction experiments (Byerlee, 1978) is often explained by the decrease in the normal stress acting on the fault plane due to high pore pressure (e.g. Rice, 1992). However, unlike plate boundaries, it is unlikely that fluids with high pore pressure will be continuously supplied to inland faults over a long period of time. Although there are cases where fluid activity that is thought to have risen from deep in the earth after a major earthquake is observed, the time scale for this is thought to be at most a few months (Sibson, 1988).
Analysis of fault rocks obtained from drilling of the San Andreas Fault has reported that strength is reduced by smectite and talc (Lockner et al., 2011, Moore & Rymer, 2007). These fault materials are stable in the shallow crust, where temperatures are relatively low, but it has also been reported that the strength of faults can decrease dramatically due to the plastic deformation of phyllosilicates even at greater depths (e.g., Wintsch et al., 1995; Bos and Spiers, 2002; Niemeijer and Spiers, 2005). However, since these fault materials generally exhibit velocity-strengthening frictional characteristics, these ideas can be applied to creeping faults, but it is thought that they are difficult to apply to seismogenic faults.
2. Finite element model of weak faults
Therefore, we constructed a new weak fault model that does not require high pore pressure and is capable of causing earthquake slips. Yamamoto et al. (2001, 2002) showed that a fault becomes weak when there is a part that can support normal stress but not shear stress. High-pressure pore water is exactly such a thing, but isolated viscoelastic regions are thought to relax over time and become similar.
We investigated in detail how the strength of a fault would change over a long period of time, when viscoelastic regions exist in patches along the fault using the finite element method. Specifically, we examined the spatiotemporal changes in stress and slip over a long period of time by periodically arranging rectangular regions (called gouge regions) with a cross-sectional shape perpendicular to the fault, along the fault. First, we set the normal stress to a constant value, then added shear deformation to the fault and investigated at what point the fault would begin to slip. We set up several different types of gouge region, from thin rectangles to squares, and in all cases we were able to reproduce the fact that macroscopic strength is lower than in the case of a homogeneous model.
This is a fault model, but it would be more accurate to describe it as a model of the host rock near the fault plane. When investigating the faulting, the properties of the fault plane are often the focus of attention, but it is also thought that three-dimensional treatment that includes the surrounding rock is important.