Seismic hazard increases with depth. We have never spent more on monitoring or understood the seismic risk management process better. Yet deep mines keep being temporarily shut down, and at times the consequences are far more serious. Set South Africa aside; since 2000, the record has not improved but is getting slightly worse, and still impacts the best-funded, best-instrumented tier-1 mines. 

The data is rarely the problem; we collect it well. The gap is elsewhere. It is in the controls that keep people safe underground: ground support, re-entry controls and mine design optimisation. They are implemented without knowing how much they reduce the risk by. Ask any onsite geotechnical engineer how much a control reduces risk, and the answer is silence. There is much room for improvement. 

Real risk, not just hazard, is seldom worked through in detail. So we keep being surprised by occurrences we could have anticipated and call them unforeseeable. They seldom are. They come from trusting controls whose effectiveness we seldom measure or review. 

The fix is to work towards SMART controls, not a new framework or more advanced sensors. In the acronym, we get ‘specific’ and ‘relevant’ often in place already; but skip the other 3 letters. ‘Measurable’: putting a number on how much a control reduces risk. ‘Achievable’: setting a risk target we can reach with current controls. ‘Time-bound’: setting a date to check whether the controls are effective and the risk is reduced, then act on the result and feed it back into the design. 

As mines go deeper and caving becomes more common, getting this right lets us continually improve on our mitigation, and turns seismic risk from a recurring surprise into something we manage proactively. If you can’t measure it, you can’t improve it.