Seismic monitoring, analysis and interpretation are key components of the hazard assessment and risk management system on rockburst-prone mines. The Mine Health and Safety Council commissioned a task as part of the project "Minimising the rockburst risk" (SIM050302) to review current practice on South African mines and to formulate standards, guidelines, and descriptions of best practice. Standards were set for some areas of mine seismology practice such as network design and emergency response. In other areas of practice, such as data processing, it was only possible to provide guidelines. In yet other areas of practice, such as seismic hazard assessment and risk management, most methods currently in use have not been rigorously evaluated and validated. Many methods rely, to a greater or lesser extent, on local experience and subjective judgement. Tacit knowledge still has to be translated into explicit guidelines and procedures. It is thus deemed premature to set standards, or even firm guidelines. Current practice is described, and methods to evaluate the procedures are proposed.

[1] The Research Group for Semi-controlled Earthquakegeneration Experiments in South African deep gold mines (SeeSA) has continuously monitored strain changes with a resolution of 24 bit 25 Hz at the Bambanani mine near Welkom. An Ishii borehole strainmeter was installed at a depth of 2.4 km near the potential M $ 3 earthquake source area. Instantaneous strain steps of $10 À4 strains associated with two M2 events were observed within a length of seismic fault. These steps were followed by significant postseismic creep-like drift, but not preceded by forerunners. Analysis of the continuous 25 Hz data reveals many smaller steps with much longer durations (100 ms $ 100 s) than seen in normal earthquakes (À1 < M < 2) with source durations of 1 ms$50 ms. Some of the especially slow steps were preceded by accelerations in strain, the maximum being as large as one-third of the step.

The authors document strain changes up to ~10 -4 associated with two M>2 events 2.4 km deep at distances less than ~100 m from a strainmeter. This corresponds to a ~7 MPa stress change recorded only within the hypocentral area. This change was recorded with a sensitive, wide-dynamic-range, Ishii strainmeter. A 15-m hole was drilled subparallel to the fault strike, in order to continuously monitor slip-driving shear and normal strains with a sampling frequency of 25 Hz. Two M>2 events took place around the fault. For both events, relaxation in the maximum principal stress at a rate of 10 -6 /week was observed for several days prior to the main shock. In one of the two events, foreshocks were concentrated in the last several tens of seconds, accompanied by strain steps and logarithmic post-seismic deformation. However, no acceleration in deformation was observed, even with a resolution of 1/10,000 of a coseismic step for each event.

No abstract

No abstract