13:45 〜 15:15
[SSS06-P06] Rock friction in and around aftershock cloud of the Orkney earthquake, South Africa, drilled by ICDP-DSeis project
キーワード：ICDP-DSeis project、2014 Orkney earthquake (M5.5), South Africa、Frictional property of a seismogenic fault、Evolutions of friction and AE activity with slip distance、Rotary shear test under dry/wet conditions
The 2014 Orkney earthquake (M5.5), South Africa, ruptured a sub-vertical left-lateral fault in 3.5-7 km depths from surface and ~10 km-long in N-S. ICDP-DSeis project established a drilling site on a tunnel at 2.8 km depth in a gold mine located just above the aftershock cloud of the Orkney earthquake. Three holes (Hole A; 817 m, Hole B; 700 m, Hole C; 96 m branched from Hole B) were drilled from the site. Among them, Holes B and C intersected the aftershock cloud and succeeded to recover samples from an active seismogenic fault zone for the first time in the world. Miyamoto et al. (2022) found that the fault zone coinciding with the aftershock cloud occurred in an altered lamprophyre dike in a metabasalt of Crown lava. The altered lamprophyre is characterized by high contents of talc, biotite, calcite, and amphibole.
It is known from a waveform inversion (Yasutomi, 2018) that a dynamic slip during the mainshock did not propagate to the fault zone around the intersection with the ICDP-DSeis holes. This suggests that the fault zone around the intersection is stable as a whole. On the other hand, the high aftershock activity implies instability of the fault zone, at least, in part. To understand these scale/time dependent friction of the fault zone intersected with Hole B, we measured evolution of frictional properties of the altered lamprophyre and Crown lava up to ~170 mm in cumulative sliding distance using a rotary shear apparatus. Rocks were ground to produce gouges and sifted with #140 mesh (106 μm openings) sieve. The gouge of ~1 g was sandwiched between steel forcing blocks with inner and outer diameters of 20 mm and 30 mm, respectively. To monitor AE (acoustic emission) activity during the frictional sliding, AE sensors (center frequency 0.8 MHz) are attached to top and bottom of the forcing blocks. Sliding rate was changed stepwise by every 1/3 order from 4.64 to 46.4 μm/s under a constant normal stress of 5 or 15 MPa. As humidity affects frictional properties, it was controlled to dry (circulating air of ~5 %RH in sample chamber) or wet (deionized water is added, but pore pressure is not controlled). The rate- and state-friction (RSF) law with two state variables was fitted to the transient response of friction to the stepwise change in sliding rate by the eye.
Friction coefficient of Crown lava is 0.6-0.7 on average irrespective of dry or wet. Under the wet condition, friction coefficient of the fault zone is 0.2-0.3, being much smaller than that of Crown lava. For a small sliding distance, friction coefficient of the dry fault zone is ~0.6 and smaller than that of Crown lava.
In case of Crown lava, rate dependence of friction (a-b value of RSF law) is negative regardless of dry or wet. The wet fault zone shows a-b>0 throughout the experiment. Friction property of the dry fault zone is a-b>0 at small sliding distances being consistent with Miyamoto et al. (2022), while a-b value evolves to neutral or slightly negative as the cumulative sliding distance increases.
Irrespective of dry or wet, AE activity (number of events per unit sliding distance) associated with shearing Crown lava gouges becomes lower as cumulative sliding distance increases. On the other hand, it begins active after some sliding in both cases of dry and wet fault zones.
These results are consistent with observations that the dynamic rupture of the Orkney earthquake fractured a fault in the altered lamprophyre dike, but that its propagation was terminated in the fault zone around the intersection with ICDP-DSeis holes. Further, in spite of its stability, aftershock activity can be high in the fault zone where the dynamic rupture propagation was arrested.
It is known from a waveform inversion (Yasutomi, 2018) that a dynamic slip during the mainshock did not propagate to the fault zone around the intersection with the ICDP-DSeis holes. This suggests that the fault zone around the intersection is stable as a whole. On the other hand, the high aftershock activity implies instability of the fault zone, at least, in part. To understand these scale/time dependent friction of the fault zone intersected with Hole B, we measured evolution of frictional properties of the altered lamprophyre and Crown lava up to ~170 mm in cumulative sliding distance using a rotary shear apparatus. Rocks were ground to produce gouges and sifted with #140 mesh (106 μm openings) sieve. The gouge of ~1 g was sandwiched between steel forcing blocks with inner and outer diameters of 20 mm and 30 mm, respectively. To monitor AE (acoustic emission) activity during the frictional sliding, AE sensors (center frequency 0.8 MHz) are attached to top and bottom of the forcing blocks. Sliding rate was changed stepwise by every 1/3 order from 4.64 to 46.4 μm/s under a constant normal stress of 5 or 15 MPa. As humidity affects frictional properties, it was controlled to dry (circulating air of ~5 %RH in sample chamber) or wet (deionized water is added, but pore pressure is not controlled). The rate- and state-friction (RSF) law with two state variables was fitted to the transient response of friction to the stepwise change in sliding rate by the eye.
Friction coefficient of Crown lava is 0.6-0.7 on average irrespective of dry or wet. Under the wet condition, friction coefficient of the fault zone is 0.2-0.3, being much smaller than that of Crown lava. For a small sliding distance, friction coefficient of the dry fault zone is ~0.6 and smaller than that of Crown lava.
In case of Crown lava, rate dependence of friction (a-b value of RSF law) is negative regardless of dry or wet. The wet fault zone shows a-b>0 throughout the experiment. Friction property of the dry fault zone is a-b>0 at small sliding distances being consistent with Miyamoto et al. (2022), while a-b value evolves to neutral or slightly negative as the cumulative sliding distance increases.
Irrespective of dry or wet, AE activity (number of events per unit sliding distance) associated with shearing Crown lava gouges becomes lower as cumulative sliding distance increases. On the other hand, it begins active after some sliding in both cases of dry and wet fault zones.
These results are consistent with observations that the dynamic rupture of the Orkney earthquake fractured a fault in the altered lamprophyre dike, but that its propagation was terminated in the fault zone around the intersection with ICDP-DSeis holes. Further, in spite of its stability, aftershock activity can be high in the fault zone where the dynamic rupture propagation was arrested.