Why Mt Polley Mine Failed

Jack Caldwell

The failure of the Mt Polley tailings facility vindicated Karl Terzaghi and Ralph B. Peck, the two great engineers who founded the science and art of geotechnical engineering.  They always warned that the undetected geological feature, stratum, or layer was the cause of failure.  This is exactly what happened at the part of the embankment that failed at the Mt Polley mine—the original site investigation failed to identify or characterize a layer of glaciofluvial clay.  Subsequent work failed to establish correctly the strength of the layer.  And that layer failed when the load from the embankment exceeded the strength of the clay.

But many other factors contributed to the magnitude and consequences of failure.  Think of the domino metaphor:  there is a line of dominoes all standing upright next to each other; push the first one over and  it topples the next one that knocks over the next, until they all fall down and flow to Hazeltine Creek.  If only somebody had thought to remove one of the dominoes, or thought to glue one to the base.  That would have stopped the line of dominoes falling down and flowing to Quesnal Lake.

The Domino Effect
The first domino at Mt Polley was the clay that was too weak to resist the loads on it.

The second domino at Mt Polley was the steep slope of the embankment—it had been designed to be inclined at two horizontal to one vertical—but because of a shortage of materials, had been built much steeper, namely at 1.3 horizontal to one vertical. 

The third domino was the presence of water upstream of the embankment. The official report notes that if the elevation of the water had been one meter lower, there would have been time & opportunity to fix the failing embankment. 

The fourth domino was the fluid tailings that were able to flow out and flow far once the embankment failed.  The official report, Independent Expert Engineering Investigation Review Board Report on Mt Polley Tailings Storage Facility Breach (January 30, 2015), notes that if the pool of supernatant water had been kept far away from the perimeter, the tailings might have dried & consolidated and not been susceptible to flow in the absence of embankment containment. 

In short, the dominos were perfectly lined up to fail en masse and catastrophically.  And they did.

Prior Knowledge
The consultants and regulators knew the system was weak.  This is why they knew this:  conventional engineering practice is to characterise the overall safety of an engineering structure by calculating its factor of safety.  This is simply the ratio of the overall strength of the system to the loads that could cause failure.  Long-standing convention for dams is that the factor of safety should be 1.3 during construction, before the water or tailings are impounded.  The long-standing convention is that the factor of safety be 1.5 once dam operation or tailings placement begins.

When the Mt Polley dam failed its factor of safety was 1.3, not 1.5, in spite of the huge volume of tailings and water that it impounded.  This is the result of what can only be called “creative” word play on the part of the engineers who argued that because the embankment was continuously being built, 1.3 applied.  They argued that 1.5 should apply only at closure of the dam.

The regulators knew this and had protested.  They had written in a September 19, 2012 communication from the Ministry of Energy and Mines to the Mt Polley Mining Company:  The factor of safety for the main embankment is only marginally above the short-term design criteria of 1.3… AMEC [Mining consultant company]has interpreted Table 6-2 from the 2007 Dam Safety Guidelines somewhat differently than I have seen in the past. This table recommends a minimum factor of safety of 1.3 at the end of construction and ‘before reservoir filling’ and a factor of safety of 1.5 at the ‘normal reservoir level.’ AMEC has interpreted the construction period as the entire pre-closure period, and this is open to debate. However, I consider that sufficient mitigation measures are in place (i.e., piezometer trigger thresholds) to support this more liberal interpretation in this instance.

The simple fact is that the probability of failure of a system with a factor of safety of 1.5 can be thousands of time less than the probability of failure of a system with a factor of safety of 1.3.   In the case of Mt Polley, we now know that the probability of failure of a system with a factor of safety of 1.3 can indeed be one. 

Had the embankment been built to the intended two horizontal to one, the factor of safety would have been 1.5 and the probability of failure thousands of times less—it probably would not have failed.

Put Best Practices to Use
What does all this mean to future tailings facility design, construction, operation, and closure?  Steve Vick, one of the authors of the report, is quoted as saying that we cannot continue to build tailings facilities using one hundred-year old technology.  The report itself calls for the use of Best Available Technology: filter-pressed tailings; polymer-amended tailings; no water in the facility; and the use of independent peer review panels. 

The report includes a picture of the Greens Creek filter-pressed tailings facility.  This has operated safely since the 1980s.  I drilled& sampled the underlying soils in 1983.  They are glacial lacustrine soils of very low strength and great sensitivity—push them a bit and they change from a solid to a fluid.  It was obvious that a conventional tailings facility would not work.  To the credit of the mine owners they protested, but eventually went with filter-pressed tailings. 

At the Suncor oil sand mine, they are amending the mature fine tailings with polymer.  I know from personal observation that the resulting tailings go solid and strong at a rate far in excess of the unamended tailings. 

The Best Practices have been practiced for decades.  Now all that remains is for the industry to adopt them en masse. 

***

Jack Caldwell is a semi-retired civil engineer who has worked most of his career on mine tailings facility design, construction, operation, and closure in more countries than he cares to revisit.

Related Posts

Watershed Sentinel Original Content

5 Issues/yr — $25 print; $15 digital