[S02-2] World overview of anthropogenic seismicity II

[S02-2-04] Very small repeating earthquakes on a geological fault at 1-km depth in a gold mine in South Africa

^○Makoto Naoi¹, Junya Yamaguchi¹, Masao Nakatani², Hirokazu Moriya³, Toshihiro Igarashi², Thabang Kgarume⁴, Osamu Murakami⁵, Thabang Masakale⁶, Yasuo Yabe³, Kenshiro Otsuki³, Hironori Kawakata⁷, Tsuyoshi Ishida¹, Luiz Ribeiro⁸, Anthony Ward⁸, Raymond Durrheim⁹, Hiroshi Ogasawara⁷ (1.Kyoto University, 2.University of Tokyo, 3.Tohoku University, 4.Council for Scientific and Industrial Research, 5.Tono Research Institute of Earthquake Science, 6.Open House Management Solutions, 7.Ritsumeikan University, 8.SeismoGen Cc, 9.University of the Witwatersrand)

We deployed an Acoustic Emission (AE) monitoring network consisting of 30 very sensitive AE sensors and 7 accelerometers at 1-km depth in the Cooke 4 gold mine in South Africa, where many earthquakes up to ~M 3 were induced by stress buildup due to mining. Naoi et al. (2015) analyzed data obtained by the network during a 2 month period, and found very small repeating earthquakes of –5.1 ≤Mw ≤–3.6 which occurred in a region of ~100-m extent on a geological fault. When applied to those repeater sequences, the Nadeau and Johnson (1998) empirical formula (NJ formula), which relates the amount of background creep and repeater activity and is well established for plate boundary faults, yielded an impossibly large estimate of the background creep. This implies that the repeaters were produced more efficiently, for a given amount of background creep, than predicted by the NJ formula.

In this study, we extended the analysis period to 14 months (from 7 April 2011 to 30 May 2012), and searched for repeaters among 3735 events that occurred on the fault, on the basis of waveform similarity and the proximity of their hypocenters relocated by the Double Difference method applied to relative travel-time accurately determined by cross correlation. Out of the 3735 events, 1572 events (42.1%) were identified as repeaters, which formed 360 groups. The largest group saw as many as 45 recurrences. Activities of some groups continued for the whole 14 months, whereas some groups exhibited a decrease in magnitude and disappeared eventually. We propose that the disappearance of the repeaters represents the dissipation of unstable patches of the fault, resulting from frictional wear of the contacts as fault creep progresses.