On Sunday August 7 1994, an explosion at the Moura No 2 underground coal mine in Queensland led to the deaths of 11 miners. This tragedy was the catalyst for a major shakeup in the approach to safety in all kinds of mines around Australia over the late 1990s and early 2000s.
Since that time, we have seen major improvements in safety performance. In 2003, there were 12.4 fatalities per 100,000 workers; a decade later the figure was down to 3.4.
However, since then progress has slowed if not stalled. Despite the industry's adoption of risk management systems, competency training, and a shift away from prescriptive regulation in the years following Moura, the rate of deaths and serious injuries has barely changed over the past decade.
Given the huge size and variety of Australia's mining industry, and the inherent dangers of the work, we may never reach a time when there are no deaths. But zero fatalities must still be the goal.
A rise in 'one-off' incidents
In the past, most deaths were due to what are called "principal hazards". These are major incidents such as fires, explosions and mine flooding that can kill or injure many people.
Most safety work has, for good reason, focused on these hazards, and by my count they are today involved in fewer than 20% of deaths. What this means is that today's tragedy landscape is more diffuse, with fatalities scattered across a range of different scenarios.
Now, most deaths are the result of "one-off" events such as being struck by objects, caught in machinery, falling from heights, or vehicle collisions. Addressing all these possibilities is more complex.
Mental health, fatigue, staff turnover
Human factors also loom large. Despite a huge increase in mine automation and remote operation technologies that reduce workers' exposure to hazards, there are indications of worsening mental health, rising fatigue and high staff turnover, which can erode corporate knowledge.
Psychological and social problems such as these affect an estimated 20% of the modern mining workforce. Although there are fewer workers on site, they are often under huge production pressures and the rosters can be very tough on family life.
Poor mental health can compromise decision-making and reduce vigilance, leading to safety problems.
Slow, steady improvement
There are some promising developments. The "critical control management" approach already adopted by Rio Tinto and Newmont, among others, has been highly effective. This is a method that identifies a relatively small number of vital controls that can prevent serious incidents, and directs resources towards rigorously designing, implementing and maintaining them.
We are also likely to see future safety gains from better equipment design, further advances in automation and remote operation, and mental health initiatives, such as Western Australia's Mental Awareness, Respect and Safety program.
But in an industry that has still averaged eight fatalities per year over the past decade, more safety reform is overdue. While new technologies and initiatives may be helpful, none will be a "silver bullet".
Queensland alone has staged three "safety resets" in the past five years, with little result. Real safety improvement will be slow and steady, and will come from diligently and consistently applying proven safety management techniques.