Foreshocks and Premonitory Slip during Large Scale Friction Experiments

Abstract:

Fukuyama et al. (2013, AGU Fall Meeting) reported that using large-scale biaxial apparatus at NIED with a loading rate of 0.1mm/s, many foreshocks were observed and pre-slip was not evident under 2.6MPa normal stress (LB04-003), but pre-slip was always recognized prior to each stick slip and foreshocks were not common under 6.7MPa normal stress (LB04-005). To further investigate these phenomena to understand what controls the foreshock activity, we conducted another series of experiments using similar rock specimens made of gabbro. We conducted a series of ten experiments whose normal stress was 6.7MPa in the first seven experiments and was lowered to 3.4MPa in the rest of three. We could not find any foreshocks in 9 experiments and the only experiments we found foreshocks was the fifth one (LB09-005), which was under 6.7MPa normal stress with loading rate of 0.1mm/s and 0.01mm/s (loading rate was changed at the middle). It seems that foreshock activity does not depend only on the normal stress but is also related to the detailed fault surface condition, which was created by previous slip history. In Fukuyama et al. (2013), initial roughness of the simulated fault surface was #600 while in the present experiment it was #2000. These differences in initial condition might affect the foreshock generation. In the present cases, a most serious experiment with 1mm/s for 400mm slip (LB09-004) was included, which was a previous experiment of LB09-005. We think LB09-004 provided a damage distribution and that could help generate the foreshocks. It should be noted that foreshock hypocenters, estimated from the PZT array, are located on the grooves. And amount of grooves generated during the experiment was the largest for LB09-005. These features suggest the change in surface condition prior to LB09-005, which might facilitate the generation of foreshocks.