Mount Polley — a Buried Glacial Clay Layer the Designers Never Found

Summary

Shortly after midnight on 4 August 2014, the Perimeter Embankment of the Mount Polley Tailings Storage Facility — a roughly 40-metre-high earth-and-rockfill dam impounding the wastes of a copper-gold mine near Likely, in the interior of British Columbia — broke open and discharged its contents into the headwaters of the Fraser River system. Some 7.3 million cubic metres of tailings, 10.6 million cubic metres of supernatant water and a further 6.5 million cubic metres of interstitial water — about 25 million cubic metres in total — surged out, scoured Hazeltine Creek from a metre-wide stream into a gouged channel tens of metres across, and emptied into Polley Lake and Quesnel Lake. No one was killed. What failed was not the dam but the ground beneath it: a continuous layer of weak glacial-lake clay, never identified by the foundation investigation, that lost its strength under load and slid the embankment off its base.

The facility had been raised almost annually since 1997 to keep pace with the mine. By 2014 the Perimeter Embankment stood about 40 metres high, with its downstream face built at a slope of 1.3 horizontal to 1.0 vertical — markedly steeper than the gentler profile a soft foundation demands. Beneath it, deposited in a lake between glaciations, lay a thin, continuous bed of glaciolacustrine silt and clay: the Glaciolacustrine Unit, or GLU. This clay was strong enough when its pore water could drain slowly, but when loaded faster than it could shed that water it behaved like a lubricated plane. The site characterisation had logged glacial sediments in places but never recognised the GLU as a single continuous, low-strength surface running under the breach location.

The trigger was load and geometry acting together. Each annual raise added weight; the steep downstream rockfill drove the load forward over the buried clay rather than spreading it back. As the embankment climbed past the GLU, the clay reached the point where it could no longer carry the stress without draining — and it could not drain. It failed in undrained shear, a sudden loss of strength, and a wedge of the embankment translated outward along the clay layer. The dam did not erode or overtop; it slid.

The Independent Expert Engineering Investigation and Review Panel, chaired by Norbert Morgenstern, reported on 30 January 2015 with a single unambiguous verdict: the dominant contribution to the failure resided in the design. The foundation investigation had failed to find the continuous GLU and to recognise its susceptibility to undrained failure, and the steepened downstream slope had removed the margin that might have saved it. Morgenstern's metaphor became the case's epitaph: building on the weak clay loaded the gun; building the steep slope pulled the trigger.

---

Timeline

1990s

Site characterisation

Foundation investigations for the tailings facility log glacial deposits in the Perimeter Embankment foundation but do not identify a continuous, low-strength glaciolacustrine clay layer beneath the breach area. Knight Piésold Ltd. is the original engineer of record.

1997

Embankment commissioned

The TSF begins operation. The dam is designed to be raised in stages over the mine's life, growing upward and outward as tailings accumulate behind it.

1997–2014

Near-annual raises

The embankment is raised almost every year to keep pace with production, steadily increasing the load imposed on the foundation soils, including the unrecognised GLU.

2005–2006

Stability analyses

Engineers run stability analyses on the rising embankment. The governing weakness — the continuous glaciolacustrine clay and its undrained behaviour — is not captured because the layer is not recognised as continuous.

2006

Downstream slope steepened

Because of a delay in rockfill delivery, a portion of the downstream slope is placed steeper than planned, at about 1.3H:1.0V. The interim steepening is never corrected.

~2011

Engineer of record changes

AMEC takes over as engineer of record from Knight Piésold. A recommendation to drill from the 2011 embankment crest to better characterise the foundation soils is not made; key field observations go unrecorded.

Early 2014

Continued raising

The embankment, now roughly 40 m high, continues to be raised. Pond water sits close to the crest, and the load on the buried clay approaches its undrained capacity.

Late July 2014

Pre-failure conditions

The reservoir level is high relative to the embankment. The foundation clay is now loaded near the threshold at which it can no longer carry stress without draining pore water it cannot shed.

4 August 2014, ~00:00

Breach

A section of the Perimeter Embankment fails by sliding outward along the glaciolacustrine clay layer. About 25 million m³ of tailings and water are released in hours.

4 August 2014, hours after

Downstream destruction

The flow scours Hazeltine Creek into a channel tens of metres wide, drains the upper portion of Polley Lake, and discharges into Quesnel Lake. No fatalities occur.

18 August 2014

Panel appointed

British Columbia, with the Williams Lake and Soda Creek Indian Bands, appoints the Independent Expert Panel: Norbert Morgenstern (chair), Steven Vick and Dirk Van Zyl.

30 January 2015

Panel reports

The Panel concludes the dominant cause resides in the design — an unrecognised continuous GLU subject to undrained failure, combined with a steep downstream slope.

The Build: A Dam Raised Every Year on Unmapped Ground

The Mount Polley Tailings Storage Facility was not a single act of construction but a structure that grew for seventeen years. When the open-pit copper-gold mine opened in 1997, its tailings dam was designed to be raised in stages — climbing upward and pushing outward as the impounded waste rose behind it. This is ordinary practice for tailings dams, and it carries an ordinary hazard: the foundation is loaded progressively over decades, and any weakness in it is tested not at the moment of construction but slowly, as the embankment finally reaches the height at which the ground gives way.

The Perimeter Embankment was founded on glacial terrain. As the ice sheets of the last glaciation advanced and retreated across the interior plateau of British Columbia, they left behind a complicated stack of deposits — tills, sands, and, critically, fine sediments laid down in lakes that formed between or beneath the ice. One such bed was the Glaciolacustrine Unit: a thin, continuous layer of silt and clay deposited in still water, soft, low in strength, and prone to a particular and dangerous behaviour under rapid loading. The foundation investigations had encountered glacial sediments — they were logged in places — but the investigation never resolved the GLU into what it actually was: a single continuous, low-strength surface running beneath the embankment at the location that would eventually fail.

Two decisions compounded the blind spot. First, the embankment's downstream face was built steep. By 2006 a portion of it had been placed at roughly 1.3 horizontal to 1.0 vertical — steeper than planned, the result of a delay in rockfill delivery, and never afterward corrected. A steep downstream slope concentrates the embankment's weight forward and downward, driving shear stress into the foundation rather than spreading it over a wide base. On competent ground this is tolerable; over a buried clay layer it is the difference between standing and sliding. Second, the engineering analyses that should have caught the danger could not, because the governing element — the continuous clay surface and its undrained strength — was never in the model. The single most important feature of the site was the one the design did not know existed.

The Failure Sequence: A Clay Layer That Could Not Drain

The mechanism was undrained shear failure of saturated clay, and it is worth stating plainly because it is the heart of the case. Clay derives much of its strength from the friction between its particles, and that friction depends on the water pressure in the pores between them. Load a saturated clay slowly and the pore water drains away, the particles press together, and the clay carries the stress. Load it faster than it can drain and the water cannot escape; pore pressure spikes, the effective stress between particles collapses, and the clay loses strength almost completely. It stops behaving like soil and starts behaving like a slick plane. The GLU beneath Mount Polley was exactly such a clay, and the embankment loaded it faster than it could drain.

For seventeen years the dam was raised, almost every year, and every raise added weight to the foundation. The steepened downstream slope ensured that this weight bore forward, concentrating shear stress along the buried clay rather than distributing it. Through most of that history the GLU held, because it had not yet been loaded to the point where its undrained capacity was exceeded. By mid-2014 the embankment had reached about 40 metres and the pond stood high against the crest. The stress on the clay had climbed to the threshold at which it could no longer carry the load without shedding pore water it had no way to shed.

When that threshold was crossed, the failure was not gradual. The clay yielded in undrained shear, lost its strength, and a wedge of the embankment translated outward along the now-slick layer — a foundation slide, not an erosion event. The Perimeter Embankment opened. In the hours that followed, about 25 million cubic metres of tailings and water poured out: 7.3 million cubic metres of tailings, 10.6 million of supernatant water, 6.5 million of interstitial water. The release scoured the metre-wide trickle of Hazeltine Creek into a channel tens of metres across, drained the upper part of Polley Lake, and emptied into Quesnel Lake, a deep cold-water lake feeding the Fraser salmon fishery. It was, by volume, among the largest tailings failures in Canadian history. That it killed no one was a function of geography and timing, not of margin in the design.

The Reckoning: 'Failure Resides in the Design'

British Columbia, together with the Williams Lake and Soda Creek Indian Bands, appointed an Independent Expert Engineering Investigation and Review Panel on 18 August 2014 — Norbert Morgenstern as chair, with Steven Vick and Dirk Van Zyl, three of the most senior figures in geotechnical and tailings engineering. Their report, delivered on 30 January 2015, was direct. The dominant contribution to the failure resided in the design. The design had not taken into account the complexity of the sub-glacial and pre-glacial geology of the foundation; the investigation had failed to identify the continuous GLU and to recognise that it was susceptible to undrained failure under the stresses the embankment imposed. The steep downstream slope had removed the reserve that might have prevented the slide — had the slope been flattened to 2.0 horizontal to 1.0 vertical, the Panel found, the failure would have been avoided.

Morgenstern's framing endured because it was exact: the weak clay had loaded the gun, and the steep slope had pulled the trigger. The Panel did not rest there. It urged the industry away from the conventional water-retaining tailings dam toward best available technology — dewatered, filtered tailings that do not require a wet impoundment behind a perpetually rising embankment — and toward stronger independent review and corporate accountability for dam safety.

The professional reckoning followed years later. Neither Imperial Metals, the mine owner, nor any individual was criminally charged or fined under environmental or mining law, an outcome that drew sustained criticism. The accountability that did arrive came through the profession itself. Engineers and Geoscientists British Columbia disciplined the engineers responsible: Todd Martin, the senior geotechnical engineer, admitted his work had not followed prudent practice — including a failure to recommend drilling from the 2011 crest to better characterise the foundation and a failure to record key field observations — and Stephen Rice, the review engineer at AMEC, was found to have allowed a junior engineer with limited embankment experience to serve as engineer of record. Each was fined and assessed costs. The case became the modern textbook demonstration that a tailings dam is a foundation problem first, and that the ground you do not investigate is the ground that will carry your dam away.

---

Contributing Factors

01

An undetected continuous weak layer in the foundation

A thin, continuous bed of glaciolacustrine silt and clay — the GLU — ran beneath the Perimeter Embankment at the breach location. The site investigations logged glacial sediments but never resolved the GLU as a single continuous, low-strength surface. The most important feature of the foundation was absent from the engineers' model of it.

02

Undrained shear failure not recognised in design

The GLU was a saturated clay that loses nearly all its strength when loaded faster than its pore water can drain. The design did not characterise the layer as susceptible to undrained failure, and so the analyses never tested the embankment against the very mechanism that destroyed it. A failure mode that was not modelled cannot be designed against.

03

A downstream slope built too steep

A portion of the downstream face was placed at about 1.3H:1.0V — steeper than planned, after a rockfill delivery delay — and never corrected. The steep slope concentrated the embankment's weight forward, driving shear stress into the buried clay. Flattening it to 2.0H:1.0V, the Panel concluded, would have prevented the slide.

04

Progressive loading of a latent defect

The dam was raised almost every year for seventeen years, each raise adding load to the foundation. The GLU held until the cumulative load finally exceeded its undrained capacity. A defect loaded slowly over decades fails suddenly when the threshold is reached, long after the decisions that created it.

05

Eroded engineering oversight and accountability

Recommendations to drill from the 2011 crest to characterise the foundation were not made; field observations went unrecorded; a junior engineer served as engineer of record under inadequate review. The chain of professional judgment that should have caught the GLU was broken at several links, and no individual or owner faced criminal or environmental sanction. ---

Aftermath

The Mount Polley breach released about 25 million cubic metres of tailings and water into Hazeltine Creek, Polley Lake and Quesnel Lake on 4 August 2014, with no loss of life but with severe damage to a salmon-bearing watershed and a permanent scar across the landscape. It became the catalyst for a tightening of tailings governance in British Columbia and beyond. The province adopted stricter dam-safety requirements, mandatory independent tailings review boards, and a stronger expectation that operators move toward best available technology — filtered, dewatered tailings rather than wet impoundments behind perpetually rising embankments — the central recommendation of the Morgenstern Panel. Internationally the failure fed directly into the development of the Global Industry Standard on Tailings Management. Within the profession, Engineers and Geoscientists BC disciplined the responsible engineers, even as no criminal charge or environmental fine was ever laid against the owner. Mount Polley is now the standing byword for a single proposition: a tailings dam is only as strong as the ground it stands on, and undrained failure of an unmapped clay layer is the foundation failure that will breach an embankment that never overtopped.

---

Lessons

Characterise the foundation as a continuous system, not a set of boreholes: map every weak layer to its full extent, because a thin clay bed that is logged in places but never traced as a continuous surface is the surface your dam will slide on.
Test the design against undrained failure wherever saturated fine soils underlie a structure: assume the clay cannot drain as fast as you load it, and design for the strength it has at that instant, not the strength it would have if it could drain.
Do not let an interim steepening become permanent: a downstream slope placed steeper than planned to solve a delivery problem must be flattened back, because the slope angle is the difference between spreading load and concentrating it onto a weak foundation.
Treat progressive, staged loading as a slow-motion test of a latent defect: a structure raised year after year is loading its foundation toward a threshold, and the failure will arrive suddenly, decades after the decision that doomed it.
Keep the chain of engineering judgment intact and independent: require the investigation that the site needs, record what is observed, and never place under-qualified engineers in roles of record without competent review — the defect missed is the defect that fails. ---