Fundão (Mariana) — the Flow-Liquefaction Breach That Poisoned 668 km of River

Summary

On the afternoon of 5 November 2015, the Fundão tailings dam — an upstream-raised iron-ore tailings embankment operated by Samarco, the joint venture of Vale and BHP Billiton, in the Germano mining complex near Mariana in Minas Gerais, Brazil — breached and released roughly 43.7 million cubic metres of saturated mine tailings down the Gualaxo do Norte and Rio Doce valleys. The mudflow reached the village of Bento Rodrigues within about thirty minutes, killed 19 people, and spread contamination along some 668 kilometres of watercourse to the Atlantic seventeen days later. It was, by volume and reach, the largest tailings-dam disaster on record at the time. The embankment did not overtop and was not undermined by piping. It liquefied — the saturated tailings sand momentarily lost its strength and flowed like a fluid.

The Fundão dam was raised by the upstream method, the cheapest and most common way to grow a tailings dam, in which each new lift is built partly on top of previously deposited tailings rather than on competent ground. The method is sound only if the tailings beneath remain drained, dense, and capable of carrying load. At Fundão they did not. A chain of construction and design decisions — an under-built starter drain that was never properly repaired, a forced realignment of the left abutment back over a deposit of weak fine "slimes," and a rising water table — left a wedge of loose, fully saturated sand sitting on soft clayey slimes near the left abutment. That sand was a loaded gun: dense enough to stand under static conditions, loose enough to collapse and liquefy if disturbed.

The disturbance came in stages over 2015. As the embankment was raised above the slimes, the soft layer extruded sideways under the load, stretching and loosening the sand above it and nudging the whole left flank toward instability. The process was already far advanced when, at roughly 14:30 on 5 November, three small seismic tremors registered in the region. Computer modelling later showed those minor shocks added a final increment of horizontal movement to the slimes and the overlying sand — enough to tip a system on the edge into collapse. Around 16:00 the left abutment slid, the sand liquefied, and the dam burst.

The Independent Panel commissioned by the owners — chaired by Norbert Morgenstern and reporting in August 2016 — concluded that the failure was a flow liquefaction caused by a specific, traceable sequence of construction defects, not an unforeseeable act of nature. The case, paired with the Brumadinho disaster that followed in 2019, drove a worldwide reckoning with upstream tailings dams and produced the first global tailings-management standard. Fundão is the canonical demonstration that a tailings dam is a soil-mechanics problem in which saturation and density, not the height of the wall, decide whether the structure lives or dies.

---

Timeline

2008

Fundão dam commissioned

Samarco begins depositing tailings behind a starter dam in the Fundão valley, planned as an upstream-raised structure to store the increasing waste from expanded iron-ore production.

2009

First defect in the base drain

A blanket drain meant to keep the downstream sand dry is built improperly; saturation begins to rise in the area that will later fail. The drain is never fully restored.

2011

Cracking and seepage on the left abutment

Construction problems and seepage prompt a redesign. To avoid a damaged zone, the embankment alignment on the left abutment is set back upstream — over ground where soft fine slimes had been deposited.

2012

Raising continues over slimes

The upstream-method raises advance, now built above a wedge of weak, compressible slimes on the left abutment, with loose saturated sand placed above them.

2013

Design changes increase saturation

Modifications to the dam's design and deposition raise the phreatic (water) surface within the sand, leaving more of the tailings fully saturated and prone to liquefaction.

2014

Lateral extrusion underway

Loading from the rising embankment squeezes the soft slimes sideways. The extrusion stretches and loosens the overlying sand, progressively reducing its strength near the left abutment.

Mid–late 2015

Warning signs accumulate

Cracking, settlement and instability are observed on the left abutment. Remedial measures and monitoring are undertaken, but the underlying liquefiable condition is not arrested.

5 Nov 2015, ~14:30

Minor seismic tremors

Three small earthquakes (around magnitude 2–2.6) are recorded in the region, roughly 90 minutes before failure, adding a final increment of movement to an already-failing slope.

5 Nov 2015, ~16:00

Flow liquefaction and breach

The left abutment slides; the saturated sand liquefies and flows. The dam breaches, releasing about 43.7 million m³ of tailings into the valley below.

5 Nov 2015, ~16:30

Bento Rodrigues destroyed

The tailings wave reaches Bento Rodrigues, some 8 km downstream, within about half an hour, burying much of the village. Of the 19 dead, most are mine workers; several are villagers.

22 Nov 2015

Contamination reaches the sea

The plume travels down the Gualaxo do Norte, Carmo and Doce rivers, reaching the Atlantic 17 days after failure, having affected roughly 668 km of watercourse.

25 Aug 2016

Independent Panel reports

The owner-commissioned expert panel, chaired by Norbert Morgenstern, publishes its report on the immediate causes, attributing the failure to flow liquefaction set up by a chain of construction defects.

The Build: An Upstream Dam Resting on Its Own Saturated Waste

A tailings dam is not a water dam. It impounds the ground-up rock slurry left after the valuable mineral is extracted — at Fundão, the sand-and-silt residue of iron-ore processing. The cheapest way to enlarge such a dam over decades is the upstream method: deposit tailings behind a low starter wall, let the coarse sand settle near the wall, and build each successive raise partly on top of the previously deposited tailings, marching the crest upstream. The wall thus stands, in part, on its own waste. The method works only if that underlying tailings stays drained and dense enough to behave like soil rather than slurry. Loose, saturated tailings sand has a treacherous property: under static load it can stand, but if disturbed it can collapse its grain structure, transfer its load to the pore water, and flow. That is flow liquefaction, and the upstream method's entire safety case is the promise that it will not happen.

Fundão's design depended on keeping the downstream shell of sand unsaturated, drained by a blanket of coarse material at the base. That promise was broken early. In 2009 the base drain was built defectively, and seepage and saturation crept into the sand. Then, in 2011, cracking and seepage on the left abutment forced a redesign. Rather than rebuild on the original line, the engineers set the embankment alignment back — upstream — onto ground where soft, fine-grained slimes had been deposited. Slimes are the clay-sized fraction of the tailings: weak, compressible, slow to drain, and never intended to sit beneath the structural part of the dam. From that point the left abutment of Fundão was a loose, increasingly saturated sand placed above a layer of soft clay it should never have been built over.

The design then made the saturation worse. Modifications through 2012 and 2013 raised the water table within the sand, leaving an ever-larger volume fully saturated. By 2015 the left abutment held everything a flow liquefaction requires: loose contractive sand, full saturation, and a soft underlying layer ready to move. The structure had been quietly assembled into its own failure mode, raise by raise.

The Failure Sequence: Slimes Squeezed Out, Sand Stretched Loose, Then Shaken

The mechanism that destroyed Fundão was lateral extrusion. As the embankment was raised above the soft slimes, the weight of the new fill squeezed the slimes sideways, the way a soft layer of clay bulges out from under a load. That horizontal spreading did two things. It pulled the overlying sand apart, stretching it and reducing its density at exactly the wrong place; and it imposed shear on a body of sand already at full saturation. Loose saturated sand under increasing shear is primed for collapse — its skeleton on the verge of transferring its load to the trapped pore water. Through 2014 and 2015 this process advanced steadily, registering at the surface as cracking and settlement on the left abutment. Remedial work and monitoring followed, but none of it addressed the buried condition: a contractive, saturated sand stretched over moving slimes, holding far more load than its skeleton could keep carrying.

By early November 2015 the left abutment was, in the Panel's reconstruction, already failing — a slow, internal march toward collapse that needed only a final push. That push arrived as seismicity. At about 14:30 on 5 November, three small tremors of roughly magnitude 2 to 2.6 struck the region. They were minor — far too weak to threaten a sound dam. But Fundão was not sound. Modelling by the Panel showed that the shaking produced a small additional increment of horizontal movement in the slimes and the sand above them. On a slope sitting at the edge of stability, that increment was decisive: it accelerated a failure process that was already well advanced. About ninety minutes later, around 16:00, the left abutment slid. The disturbance collapsed the loose sand's grain structure; the sand liquefied, lost essentially all of its strength, and flowed. The breach propagated through the embankment and released roughly 43.7 million cubic metres of tailings into the valley. The wave reached Bento Rodrigues within half an hour, killing 19 and erasing the village. The trigger was minor and the distinction is decisive: the tremors did not cause the failure, the construction defects did. The shaking merely fired a structure already cocked to liquefy.

The Reckoning: A Defect Chain, Not an Act of God

The owners — Samarco, Vale and BHP Billiton — commissioned an independent expert panel through the firm Cleary Gottlieb. Chaired by Norbert Morgenstern, one of the foremost geotechnical engineers of the era, the panel issued its report on 25 August 2016 and reached an unusually precise verdict. The failure was a flow liquefaction of the saturated tailings sand on the left abutment. It was produced not by a single error but by a chain: the defective base drain that raised saturation; the 2011 setback that placed the embankment over weak slimes; the design changes that lifted the water table; the lateral extrusion of the slimes that loosened and stretched the overlying sand; and finally the small seismic shocks that nudged an already-failing slope over the edge. To prove the mechanism, the panel reproduced in the laboratory the very behaviour observed in the field and showed that what happened in the laboratory matched what happened at the left abutment.

The significance of that finding is that it removed the alibi. A tailings failure is often described as if it were a flood or an earthquake — an external force overwhelming a structure. Fundão was the opposite: a structure that had been built, over years and through specific human decisions, into a condition where it would liquefy under the smallest disturbance. The minor tremor was a trigger, not a cause; remove it and the slope would likely have failed within months regardless. The defects were the disaster. That conclusion put the upstream construction method itself on trial. When Brumadinho — another Vale-linked upstream dam — failed by static liquefaction in January 2019 and killed 270, the verdict became inescapable. The two cases together ended the era in which upstream tailings dams could be presumed safe if the wall stood up.

---

Contributing Factors

01

A defective base drain that raised saturation

The blanket drain meant to keep the downstream sand unsaturated was built improperly in 2009 and never fully repaired. Seepage and a rising phreatic surface left an expanding zone of tailings fully saturated. Saturation is the precondition for flow liquefaction; without water filling the pores, loose sand cannot collapse onto its pore pressure. The drain failure quietly created the conditions for everything that followed.

02

An upstream realignment over soft slimes

When cracking forced a redesign in 2011, the embankment was set back upstream onto ground where soft, fine-grained slimes had been deposited. Slimes are weak, compressible, and slow to drain, and were never intended to underlie the structural dam. Building the load-bearing raises above them put a strong embankment on a soft, mobile foundation — the geometry that made lateral extrusion possible.

03

Lateral extrusion of the slimes loosening the sand

As the dam rose, the embankment load squeezed the soft slimes sideways. That spreading stretched and loosened the overlying saturated sand, reducing its density and imposing shear at the worst location. A loose, contractive, saturated sand is the textbook ingredient for flow liquefaction; the extrusion mechanism manufactured exactly that material on the left abutment over several years.

04

Design changes that lifted the water table

Successive modifications to the dam's design and deposition raised the phreatic surface within the sand, enlarging the saturated, liquefiable volume rather than shrinking it. Each change made the structure more vulnerable, not less. The cumulative effect was a left abutment holding a large body of sand at full saturation, loose, and already stretching over the moving slimes beneath it.

05

A minor seismic trigger on an already-failing slope

Three small tremors (about magnitude 2–2.6) struck roughly 90 minutes before the breach. Far too weak to threaten a sound dam, they added a small increment of horizontal movement that accelerated a failure already well advanced. The lesson is the inverse of intuition: the trigger was trivial, the structure fragile. A dam this close to liquefaction will fail on almost any disturbance, or none. ---

Aftermath

Fundão killed 19 people, destroyed Bento Rodrigues and Paracatu de Baixo, released about 43.7 million cubic metres of tailings, and contaminated roughly 668 kilometres of watercourse from the Gualaxo do Norte down the Rio Doce to the Atlantic, making it the largest tailings disaster on record by volume and reach at the time. The economic and ecological cost ran to billions, with restitution and litigation continuing for years. Its technical legacy is institutional. The Morgenstern panel's defect-chain verdict, reinforced catastrophically by Brumadinho in 2019, drove the mining industry, the United Nations Environment Programme, the Principles for Responsible Investment and the International Council on Mining and Metals to co-create the Global Industry Standard on Tailings Management (GISTM), published in August 2020 — the first worldwide standard for these structures. Brazil moved to prohibit new upstream tailings dams and to require the decommissioning of existing ones. The case also entrenched the practice of independent tailings review boards, formal dam-safety governance, and explicit liquefaction assessment using critical-state soil mechanics. Fundão became the byword for a single proposition: a tailings dam is governed by the saturation and density of the soil it is built from, and an upstream dam standing tall is no proof that the saturated sand inside it will not flow.

---

Lessons

Assess saturated tailings against flow liquefaction, not just slope stability: a loose, saturated, contractive sand can stand under static load and then collapse and flow — design for the strength it loses, not the strength it shows.
Never build structural raises over soft slimes or weak foundations; if an alignment must move, prove the ground beneath it, because a strong wall on a mobile soft layer will extrude, stretch, and loosen the very material that holds it up.
Keep the tailings drained and the phreatic surface low — and treat the base drainage as a primary safety system, because a defective drain that raises saturation is not a maintenance issue but the first link in a liquefaction chain.
Read a small trigger on a marginal structure as proof of fragility, not bad luck: a dam that fails under a magnitude-2 tremor was already failing; the right response is to find why the slope sat on the edge, not to blame the shock.
Govern tailings dams with independent review boards and explicit liquefaction analysis; the upstream method's economy is real, but its safety depends entirely on conditions that must be measured and proven, never assumed. ---