4:30 PM - 4:45 PM
[SCG45-42] Laboratory Frictional Healing Measurements under In-situ Slow Slip Conditions in the Nankai Trough, Japan
Keywords:Fault Zone, Nankai Trough, Megasplay, Frictional Healing, Slow Slip, Laboratory Experiments
Slow slip events (SSEs) are now recognized as an important mode of fault slip, due to their relation to ordinary earthquakes as well as potentially providing information on the faults on which they occur. Shallow SSEs are especially important in this regard, because they can be closely observed by geophysical monitoring, and their habitant can be directly sampled by scientific drilling. Both of these apply to shallow SSEs which occur in the Nankai Trough offshore Japan.
Here, we conducted laboratory friction experiments on two samples obtained from within the megasplay fault (IODP Site C0004) and the frontal thrust region (Site C0007) of the Nankai Trough accretionary prism. We measured the time-dependent strengthening, or frictional healing, of these samples in slide-hold-slide tests using a rotary shear device at 4 MPa effective normal stress, at room temperature, and under water saturated conditions. This approximates the in-situ conditions at 100’s of meters depth below seafloor, from which the samples were recovered. We employ hold times from 10 to 1 million seconds, a larger time range than typical slide-hold-slide studies, and background driving velocities of 0.01 and 10 µm/s, which covers the approximate range of slip velocities for SSEs observed in at the Nankai Trough.
Preliminary results show that significantly larger healing consistently occurs for a background driving rate of 10 µm/s compared to 0.01 µm/s. To characterize the frictional healing measurements, we compared both log-linear and power law fits to the data. We used the MATLAB fitting algorithm which fits the curves directly, without converting the data to log-log space before the fitting procedure. We find that the power law provides a superior fit to the data for the slow background sliding rate of 0.01 µm/s, whereas for 10 µm/s the power law provides only a minimally better fit than the log-linear model. The power law generally predicts larger healing than projections based on the log-linear fit. Considering a recurrence interval of 8 and 15 months, as documented for the Nankai SSEs, we estimate stress drops of a few 100’s of kPa which represent up to a roughly 50% stress drop, which are considered large for SSEs. Assuming an intact wall rock strength based on previous measurements, extrapolation of our data indicates that healing rates are not rapid enough to completely heal the fault zone within a 15 month recurrence interval. This suggests that faults maintain enough weakness to preferentially focus repeated SSEs to the same zone within the repeat time of SSEs, potentially allowing for identification of SSE zones in structural characterization of core samples.
Here, we conducted laboratory friction experiments on two samples obtained from within the megasplay fault (IODP Site C0004) and the frontal thrust region (Site C0007) of the Nankai Trough accretionary prism. We measured the time-dependent strengthening, or frictional healing, of these samples in slide-hold-slide tests using a rotary shear device at 4 MPa effective normal stress, at room temperature, and under water saturated conditions. This approximates the in-situ conditions at 100’s of meters depth below seafloor, from which the samples were recovered. We employ hold times from 10 to 1 million seconds, a larger time range than typical slide-hold-slide studies, and background driving velocities of 0.01 and 10 µm/s, which covers the approximate range of slip velocities for SSEs observed in at the Nankai Trough.
Preliminary results show that significantly larger healing consistently occurs for a background driving rate of 10 µm/s compared to 0.01 µm/s. To characterize the frictional healing measurements, we compared both log-linear and power law fits to the data. We used the MATLAB fitting algorithm which fits the curves directly, without converting the data to log-log space before the fitting procedure. We find that the power law provides a superior fit to the data for the slow background sliding rate of 0.01 µm/s, whereas for 10 µm/s the power law provides only a minimally better fit than the log-linear model. The power law generally predicts larger healing than projections based on the log-linear fit. Considering a recurrence interval of 8 and 15 months, as documented for the Nankai SSEs, we estimate stress drops of a few 100’s of kPa which represent up to a roughly 50% stress drop, which are considered large for SSEs. Assuming an intact wall rock strength based on previous measurements, extrapolation of our data indicates that healing rates are not rapid enough to completely heal the fault zone within a 15 month recurrence interval. This suggests that faults maintain enough weakness to preferentially focus repeated SSEs to the same zone within the repeat time of SSEs, potentially allowing for identification of SSE zones in structural characterization of core samples.