The study of faulting and rupture processes requires direct and on-the-field observations. In the Earth crust deep mines provide for this purpose an unique opportunity by giving access to the focal depth of earthquakes. To investigate the physics of earthquakes we take part in two projects in South Africa, providing continuous monitoring of microseismisity by using data acquisition systems and sensors installed in the depth of about 3,5 km. Both sites (Mponeng and Tautona mines) are located near Johannesburg. Many earthquakes having magnitude more than 3 occurred recently there were induced by earlier mining. Very often large earthquakes in South African deep gold mines occur on existing geological structures such as dyke contacts or natural faults. We selected Mponeng mine to investigate influence of dyke as a stress concentrator on the process of rock fracturing and Tautona mine - to investigate the process of earthquake nucleation in the vicinity of Pretorius Fault.
Fig.1:Local observation network in Mponeng gold mine
We have recently completed the instrumentation of the first site at a depth of 3540 m (Fig. 1). Gold Reef is 90 m above the site. The reef to the west of the dyke has been already mined out and a 150 m x 150 m area east to the dyke will be mined in 2008. Primary objective here is to monitor micro-seismicity around the dyke as the stress changes caused by nearby mining, we plan to keep observation for a few years. Motivated by the existence of characteristic ultrasonic Acoustic Emission (AE) activities known from laboratory rock failure experiments, we have equipped the site with a trigger-mode recorder with a 500 kHz sampling frequency, which can be controlled remotely via Internet. Eight AE sensors (squares in Fig. 1) and one hydrophone, both with frequency range up to 170 kHz, have been installed in the boreholes drilled from the tunnel going through the both contacts of the dyke dipping vertically. One tri-axial accelerometer with a frequency response flat up to 25 kHz is installed in the hole #36.
Fig.2:An example of AE signal
Figure 2 shows an example of an event captured both with the AE sensors and the accelerometer. The AE sensor signals are rich in high frequency components up to 150kHz, showing that such high frequency could be still captured after 60m traveling (determined from S-P time). From the accelerometer record, magnitude is estimated to be about -3. Some other events were captured only with AE sensors, suggesting that there are further smaller events with main power above the 25 kHz frequency range of the accelerometer. Looking into high frequencies we hope to give new insides in the physics of small ruptures in order to link observations from laboratory experiments and tectonic earthquakes. We plan to keep the observation going until the mining of the area is finished. We will compare the micro-seismic activity with the stress field of the area, which is monitored by two strainmeters installed in the site (hole #16 and 40). The stress field is expected to undergo notable changes owing to the mining process. Two pairs of electrodes have been installed to look for electric signals associated with rock failure. The performance of the AE network can be evaluated by ultrasonic transmission tests done along the whole length of hole #22.
Fig.3: Map of NELSAM underground laboratory and boreholes crossing the Pretorius main fault zone.
This side was selected for two long term international projects called DAFSAM (Drilling Active Faults Laboratory in South African Mines) and NELSAM (Natural Earthquake Laboratory in South African Mines). Both interlinked projects focus on building an earthquake laboratory in deep gold mines in South Africa equipped with variety of sensors for detailed observation. GFZ Potsdam is going to take part in these projects providing continuous microseismic monitoring.
Given to the expected proximity of 3-D dense instrument array to the hypocenters of future earthquakes, the observations to be made during the DAFSAM-NELSAM project are expected to contribute key data on:
- The scales and processes of nucleation, eventual size of the ensuing dynamic rupture, and additional preceding signatures (e.g., geochemical and electromagnetic anomalies, cascading events). The detailed properties, dynamics, and energetics of the rupture process (velocity, geometry, crack vs. pulse mode of rupture, dynamic versus geometric sources of heterogeneity, possible opening motion).
- The orientations, magnitude, and heterogeneity of the stress and strain fields in the vicinity of an active fault, and their variations during seismically active and calm periods.
- The Structural, mechanical, and geochemical features related to active fault zones and rupture zones, and their relations to the seismic observations.
By directly observing the processes related to initiation and rupture propagation of mining-induced earthquakes, this research will yield important information about natural earthquakes and will provide a link between seismology and laboratory earthquake investigations. This work is expected to result in improved understanding of a) how faults are loaded and how they fail, b) the role that material properties and physical conditions play in modulating the fault failure, and c) the way faults interact at the surface and in the depth. Strong interplay with laboratory and theoretical modeling efforts are expected to result from this study. This work will provide new insights into how earthquakes nucleate and how damaging seismic energy is released.