Tailings dams: the path to zero failure

19 August 2015

Tailings dams are complex and unforgiving systems which evolve over a period of years. The recent failure of Mount Polley tailings facility in Canada has brought global attention to the fact that the reliability of tailings dams is dependent on consistent and thorough planning, design, monitoring and effective operation at every level. After an independent inquiry into what has been described as Canada’s largest industrial accident of its kind, the industry is being steered on a path towards zero failure. Suzanne Pritchard reports.

"There is no doubt in anyone's minds that this is the worst mining disaster to ever occur in this province. Our nations and all British Columbians have raised questions as to how such a disaster could occur," Chief Bev Sellars from the Soda Creek Indian Band remarked in the aftermath of the Mount Polley tailings dam failure in 2014.

Bill Bennett, BC Minister of Energy and Mines, added: "We have a responsibility, as the jurisdiction where this failure took place, to find out exactly why it happened, ensure it never happens again and take a leadership role internationally in learning from this serious incident."

Sudden breach

Early in the morning of 4 August 2014 the tailings dam at Mont Polley mine in the Canadian province of British Columbia was breached. Loss of containment was sudden. Estimates suggest that up to 17Mm3 of water and 8Mm3 of tailings were released into Polley Lake. As Hazletine Creek flows out of Polley Lake, the contaminated flow also continued into Quesnel Lake. The breach was finally stabilised on 5 August, and a water use ban which had been imposed by local authorities on nearby towns was lifted by 12 August.

Mine owner Imperial Metals Corporation said that monitoring instruments and onsite personnel had no indications of an impending breach. The company emphasised that the tailings were alkaline with an average PH of 8.5, and were not acid generating.

Described as the largest industrial accident of its kind in Canadian history, the breach release resulted in physical impacts to the downstream environment. This included the deposition of tailings, eroded earth, trees and woody debris into the lakes and creeks, plus the erosion and scour of embankments.

Imperial Metals Corporation issued a statement on 5 August 2014 which said: "Our first priority was, and continues to be, the health and safety of our employees and neighbours. We are grateful for no loss of life or injury, or personal property damage due to this event. We are deeply concerned and are working to mitigate the immediate effect and understand the cause of the breach."

By 18 August 2014 the Ministry of Energy and Mines had ordered an independent engineering investigation and inquiry into the breach. The independent panel was tasked to issue its final report by 31 January 2015. It was to determine why the tailings dam failed and provide recommendations which would be implemented by the government "as needed and where appropriate", to ensure that such an incident never happens again.

The panel members were appointed by the government with the support of the Soda Creek and Williams Lake Indian Bands. These experienced geotechnical experts with expertise in tailings management facilities were:

  • Norbert Morgenstern, advisor to consulting engineers.
  • Steven Vick, geotechnical engineer (Colorado).
  • Dirk Van Zyl, professor at University of British Columbia.

The Soda Creek and Williams Lake Indian Bands were consulted about the terms of reference for the panel and liaised with them. The Association of Professional Engineers and Geoscientists of British Columbia, and the Institute of Mining Engineering at the University of British Columbia also provided input.

In addition the Chief Inspector of Mines issued an order to all mining companies to conduct a dam safety inspection for every tailings storage facility at a permitted mine by 1 December 2014. The deadline for such annual inspections would have been 31 March 2015, and would not have required an independent third-party review, but the order accelerated the deadline and established the requirement for an independent review.

"The failure of the tailings facility at Mount Polley was a dark day for the mining industry not only here in British Columbia, but worldwide," Dirk Van Zyl commented. "It's extremely important for us to understand how this breach happened and why, so that we can move forward with the best possible practices in ongoing and future mining operations."

Foundation failure

Based upon its experience with both water and tailings dams the panel worked with a multiple hypotheses of failure. It determined that four classes of failure mechanisms required consideration:

  • Human intervention.
  • Overtopping.
  • Piping and cracking.
  • Foundation failure.

No evidence was found to support the first two theories. Notwithstanding a number of concerns, the panel did not find evidence that the breach was caused by piping and/or cracking which could have resulted in uncontrolled internal erosion.
Instead the panel concluded that evidence indicates the breach was the result of a failure in the foundation of the embankment, a failure that occurred in a glaciolacustrine (GLU) layer.

"There were no warning signs," the Minister of Energy and Mines said at the launch of the report on 30 January 2015. "No one could have known what was going to happen unless they knew the strength and location of the subsurface materials in that layer. Apparently nobody did. The panel took three months to really understand the cause and no surface investigations could have detected this issue."

According to the panel: "the dominant contribution to the failure resides in the design. The design did not take into account the complexity of the sub-glacial and pre-glacial geological environment associated with the perimeter embankment foundation. As a result, foundation investigations and associated site characterisation failed to identify a continuous GLU layer in the vicinity of the breach, and to recognise that it was susceptible to undrained failure when subject to the stresses associated with the embankment."

The report also indicated that the failure was triggered by construction of the downstream rockfill zone at a steep slope. Had the downstream slope been flattened failure would have been avoided. Ironically, the slope was in the process of being flattened to meet its ultimate design criteria at the time of the accident.

Omissions associated with the site characterisation were likened to creating a loaded gun. If constructed unknowingly on the upper GLU constituted loading the gun, the panel added that building with a 1.3H:1V angle of repose slope over this stratum pulled the trigger.

High water levels at the tailings dam were also described as the final link in the chain of failure events. According to the report: "Immediately before the failure, the water was about 2.3m below the dam core. The panel's excavation of the failure surface showed that the crest dropped at least 3.3 m, which allowed overflow to begin and breaching to initiate. Had the water level been even a metre lower and the tailings beach commensurately wider, this last link might have held until dawn the next morning, allowing timely intervention and potentially turning a fatal condition into something survivable.

"Finally, the quantity of water had a great deal to do with the quantity of tailings released after the breach developed. It was water erosion that transported the bulk of the tailings, and these fluvial processes ended when the supply of water was exhausted. Had there been less water to sustain them, the proportion of the tailings released from the facility would have been less than the one-third that was actually lost."

In agreement

In its response to the final report Imperial Metals Corporation said that the panel's determination of the cause of breach was in agreement with its own investigations. The company stated that the panel had indicated water accumulation within the tailings storage facility was not a cause of failure but contributed to the release of tailings. Imperial Metals Corporation says it had earlier recognised that water levels would increase and had taken steps beginning in 2006 to address this issue.

The company also made reference to the panel's conclusion that the failure was attributed to the GLU layer, approximately 8m below the base of the dam in the area of the breach, and was not as strong as had been assumed in the design of the facility. Imperial Metals Corporation stated: "The panel noted that the omissions associated with site characterisations remained undetected notwithstanding the large number of experienced geotechnical engineers associated with the facility over the years."

It went on to add that construction over the 18-year life of the tailings storage facility was, at all times, carried out in accordance with design criteria provided by the Engineers of Record and approved by the Ministry of Energy and Mines.

The owner of Mount Polley mine also noted that the Ministry of Energy and Mines' September 2013 geotechnical inspection report stated that: "based on our observations and information reviewed we consider that the [Mount Polley] tailings storage facility is being designed, constructed, and operated in general conformance with the requirements of the geotechnical components of ...permits and accepted engineering practices." The Ministry also observed that a quality control and quality assurance programme was in place, and considered Mount Polley's revised operations, maintenance and surveillance manual to be well thought out.

Indeed the inspection panel had stated in its report that inspections of the tailings facility would not have prevented the failure. It also noted that regulatory staff are well qualified to perform their responsibilities and performed as expected.

"In summary," Imperial Metals Corporation said, "the tailings storage facility was constructed in accordance with design criteria established by the Engineers of Record and accepted by the Ministry of Energy and Mines. Had the GLU beneath been as strong as assumed by the design criteria, this sudden and unanticipated failure would not have occurred."

No longer business as usual

Having established the cause of the Mount Polley failure, the panel asked where it goes from here in order to ensure a similar disaster does not happen again.

Currently in British Columbia there are 123 active tailings dams which contain surface water in their impoundments along with tailings. Statistically, based on historical records, there is a 1-in-600 chance of a tailings dam failure in any one given year.

"While these numbers may seem small," the report said, "their implications are not. If the inventory of active tailings dams in the province remains unchanged, and performance in the future reflects that in the past, then on average there will be two failures every 10 years and six every 30. In the face of these prospects, the panel firmly rejects any notion that business as usual can continue."

The panel of experts went on to add that it does not accept the concept of a tolerable failure rate for tailings dams: "To do so, no matter how small, would institutionalise failure. First Nations will not accept this, the public will not permit it, government will not allow it, and the mining industry will not survive it."

Improvements to current practice are seen as providing an essential starting point on the path to zero failures. However the panel cautioned that best practices can only go so far in improving the safety of tailings technology that has not fundamentally changed in the past hundred years. It says that improving technology to ensure against failures requires eliminating water both on and in the tailings. Only this can then provide the failsafe redundancy that prevents releases.

"Simply put," according to the report, "dam failures are reduced by reducing the number of dams that can fail. Thus the path to zero leads to best practices, then continues on to best technology."

Recommendations for future best available practices (BAP) require considerations that go beyond stability calculations. It is important that safety is enhanced by providing for robust outcomes in dam design, construction and operations. Examples of best available practices were given as improving corporate design responsibilities and adopting independent tailings dam review boards.

When discussing BAP, the panel said it was disconcerted to find that, while the Mount Polley tailings dam failed because of an undetected weakness in the foundation, it could have failed by overtopping, which it almost did in May 2014.Or it could have failed by internal erosion, for which some evidence was discovered. "Clearly, multiple failure modes were in progress, and they differed mainly in how far they had progressed down their respective failure pathways," the report said.

While best practices focus on the performance of the tailings dam, best available technology (BAT) concerns the tailings deposit itself. The goal of BAT for tailings management is to assure physical stability of the tailings deposit. This is achieved by preventing release of impoundment contents, independent of the integrity of any containment structure. In accomplishing this objective, the panel says that BAT has three components that derive from principles of soil mechanics:

  • Eliminate surface water from the impoundment.
  • Promote unsaturated conditions in the tailings with drainage provisions.
  • Achieve dilatant conditions throughout the tailings deposit by compaction.

In its report the panel made seven recommendations to improve practice and reduce the potential for future failures (see below).

Finally, in fulfilling its terms of reference, the independent review panel considered what actions could have been taken to prevent the failure. It says: "From a purely technical perspective, apart from rectifying the deficiencies reviewed here, there is one that stands out. The design for the next raise of the dam had been submitted only days before the failure. In it was a buttress that would have extended along the perimeter embankment, including the breach section. Although this buttress was still not designed using the appropriate stratigraphy or undrained strengths, the panel determined that had it been in place, the failure would have been averted. The solution would have been correct, even if for the wrong reasons."

Profoundly important

"I am relieved to know the cause of the failure at Mount Polley and want to thank the panel for their important work," Bill Bennett, Minister of Energy and Mines said after publication of the final report. He also announced that the government would act immediately on key recommendations of the independent panel.

The Chief Inspector of Mines requires all operating mines with Tailings Storage Facilities (TSF) dams to provide a letter by 30 June 2015, to confirm whether foundation materials similar to those at Mount Polley exist below any of their dams. If those materials are present, the letters must also confirm whether sufficient investigations and testing were completed to properly understand the strength and location of those materials and that the dams were designed to account for suchconditions.

The province will also move to implement a new requirement that all operating mines with TSFs in British Columbia establish Independent Tailings Dam Review Boards. These boards will support improved engineering practices by providing third-party advice on the design, construction, operation and closure of TSFs.

Lastly, the province will initiate a code review to determine how to best implement the panel's remaining recommendations, including the adoption of best available practices and technologies.

It was also announced that the inspections and third-party reviews of TSFs at other permitted mines in British Columbia, as ordered by the Chief Inspector of Mines immediately following the failure at Mount Polley, did not identify any immediate safety concerns. The province said it is also investing in improved systems so that this information can routinely be made public going forward.

"We now know the cause of the failure at Mount Polley but the work doesn't stop there. What we learn from this incident and how we respond to ensure it never happens again is profoundly important to British Columbia and to the mining industry here and around the world," Bennett remarked. "The panel identified the cause of the failure and also raised other issues around the design, construction and operation of the tailings storage facility. Both the cause of the failure and these other issues are addressed through the panel's recommendations. We will implement all of these recommendations to ensure our mining industry is safe and can continue to grow and create jobs for families."

Report on the Mount Polley Tailings Storage Facility Breach. Independent Expert Engineering Investigation and Review Panel. January 2015. Province of British Columbia
The full report can be downloaded from https://www.mountpolleyreviewpanel.ca/
More information about the failure can be found at www.gov.bc.ca, www.imperialmetals.com


A who's who of the panel

An expert review panel of renowned geotechnical experts led the investigation into the Mount Polley tailings dam failure on 4 August 2014.

Norbert Morgenstern (Chair) - A well-recognised leader in the field of geotechnical engineering, he has extensive experience in dam engineering and has worked on over 140 dam projects worldwide. An emeritus university professor of civil engineering at the University of Alberta, he has released over 330 publications in the field of engineering.

Dirk J. A. Van Zyl has more than 30 years' experience in research, teaching and consulting in tailings and mined earth structures. For the last ten years, much of his attention has been focused on mining and sustainable development. He has been involved internationally in many mining projects which covered the whole mining life cycle, from exploration to closure and post-closure, in a large range of climactic and geographic environments.

Steven G. Vick - A leader in the field of dam engineering, Vick is a geotechnical engineer and internationally recognised review consultant who has worked with various engineering firms, project proponents and governments. In addition to extensive experience working on projects located in BC, his experience includes chairing the investigation of the Omai tailings dam failure for the government of Guyana and participating in the investigation of the New Orleans levee failures that occurred during Hurricane Katrina. Vick has written two books, including Planning, Design, and Analysis of Tailings Dams, which is the only text of its kind that has remained in print for over 30 years and is familiar to most experts in the field. He also speaks on the subject frequently -- his most recent keynote lecture was entitled "The Consequences of Tailings Dam Failure".


Seven recommendations to improve practice and reduce failures

1) To implement BAT using a phased approach:
For existing tailings impoundments - Rely on best practices for the remaining active life.
For new tailings facilities - BAT should be actively encouraged for new tailings facilities at existing and proposed mines.
For closure - BAT principles should be applied to closure of active impoundments so that they are progressively removed from the inventory by attrition.

2) To improve corporate governance:
Corporations proposing to operate a tailings storage facility (TSF) should be required to be a member of the Mining Association of Canada (MAC) or be obliged to commit to an equivalent programme for tailings management, including the audit function.

3) To expand corporate design commitments:
Future permit applications for a new TSF should be based on a bankable feasibility that would have considered all technical, environmental, social and economic aspects of the project in sufficient detail to support an investment decision, which might have an accuracy of +/- 10-15%. More explicitly it should contain the following:
* A detailed evaluation of all potential failure modes and a management scheme for all residual risk.
* Detailed cost/benefit analyses of BAT tailings and closure options so that economic effects can be understood, recognising that the results of the cost/benefit analyses should not supersede BAT safety considerations.
* A detailed declaration of Quantitative Performance Objectives.

4) To enhance validation of safety and regulation of all phases of a TSF:
Increase utilisation of Independent Tailings Review Boards.

5) To strengthen current regulatory operations:
Utilise the recent inspections of TSFs in the Province to ascertain whether they may be at risk due to the following potential failure modes and take appropriate actions:
* Filter adequacy
* Water balance adequacy
* Undrained shear failure of silt and clay foundations
Utilise the concept of Quantitative Performance Objectives to improve regulator evaluation of ongoing facilities.

6) To improve professional practice:
Encourage the Association of Professional Engineers and Geoscientists of British Columbia to develop guidelines that would lead to improved site characterisation for tailings dams with respect to the geological, geomorphological, hydrogeological and possibly seismotectonic characteristics.

7) To improve dam safety guidelines:
Recognising the limitations of the current Canadian Dam Association guidelines incorporated as a statutory requirement, develop improved guidelines that are tailored to the conditions encountered with TSFs in British Columbia and that emphasise protecting public safety.


August 2014 Mount Polley tailings dam in British Columbia breached on 4 August 2014, sending contaminated water surging into nearby lakes. Image taken on 5 August 2014 and shows nearly all of the wastewater in the retention basin had drained, exposing the silty bottom.
July 2014 The mine and surrounding landscape before the dam failed are captured on 29 July 2014.

Privacy Policy
We have updated our privacy policy. In the latest update it explains what cookies are and how we use them on our site. To learn more about cookies and their benefits, please view our privacy policy. Please be aware that parts of this site will not function correctly if you disable cookies. By continuing to use this site, you consent to our use of cookies in accordance with our privacy policy unless you have disabled them.