Key facts
- A tailings dam is an embankment that holds the fine waste left after ore is crushed and processed, usually together with process water.
- The Stava failure (Italy, 1985) killed 268 people; Brumadinho (Brazil, 2019) is reported to have killed about 270.
- Church Rock (USA, 1979) is the best-documented failure at a uranium mill; EPA records give about 94 million gallons of liquid and about 1,100 tons of solids.
- Release volumes in published sources often differ by a factor of two or more, so ranges are shown rather than single numbers.
What a tailings dam is and why its failure is different
When ore is crushed, ground and chemically processed, the valuable mineral is separated and the rest leaves the plant as a slurry of fine particles in water. This residue is called tailings. Most operations store it behind an embankment, built in a valley or ring-shaped on flat ground, forming a tailings storage facility. Unlike a water-supply dam, a tailings dam is usually raised in stages while the mine operates, and the stored material is itself part of the structure's behaviour.
The consequences of failure are unusual for two reasons. First, a saturated tailings deposit can liquefy and travel as a fast mudflow, far beyond the distance water alone would reach. Second, the material can be acidic, alkaline, toxic or radioactive depending on the ore and the process. The diagram above shows the main embankment types; the way the raised wall is built relative to the stored tailings is central to the causes discussed on the page about tailings dam safety.
How to read a failure chronology
No single global register is complete. Academic reviews draw on the database compiled by the International Commission on Large Dams (ICOLD) and on other collections; one widely cited review of European and world incidents drew on an ICOLD synthesis of 221 incidents. Such lists differ in what they count as a failure, in the period covered and in how they treat volumes.
For that reason the table below is deliberately short. It lists events that are documented by an investigation panel, a government or UN agency, or a peer-reviewed study. Volume figures are the ranges in those sources. Casualty figures are only given where an official or institutional count exists. Entries marked with an asterisk concern a uranium facility; others are included because their lessons apply to every kind of tailings storage.
Chronology of major documented failures
| Year | Site | Country | Material | Consequences (as documented) |
|---|---|---|---|---|
| 1979 | Church Rock* | USA | Uranium mill tailings and acidic liquid | About 94 million gallons of liquid and about 1,100 tons of solids entered Pipeline Arroyo and the Rio Puerco (EPA); site later placed on the National Priorities List. |
| 1985 | Stava | Italy | Fluorite tailings | 268 deaths; about 180,000 m3 released; three hotels and dozens of houses destroyed. |
| 1994 | Merriespruit | South Africa | Mine tailings | 17 deaths after overtopping in heavy rain; flow reached the town about 2 km away. |
| 1998 | Aznalcóllar | Spain | Sulphide-ore tailings and acidic water | Several million cubic metres entered the Guadiamar River, whose floodplain borders the Doñana wetlands. |
| 2000 | Baia Mare | Romania | Cyanide-bearing process water | More than 100,000 m3 released after a breach on 30-31 January; cyanide reached the Somes, Tisza and Danube systems (UNEP/OCHA). |
| 2010 | Ajka (Kolontár) | Hungary | Alkaline alumina red mud | 10 deaths; roughly 600,000 to 1 million m3 depending on source. See the Ajka case study. |
| 2014 | Mount Polley | Canada | Copper-gold tailings and water | Foundation failure in a glaciolacustrine layer, found by the 2015 independent review panel; volume estimates differ widely between sources. |
| 2015 | Fundão (Mariana) | Brazil | Iron-ore tailings | 19 deaths; estimates of tailings released range from about 32 to about 44 million m3. |
| 2019 | Brumadinho (Córrego do Feijão) | Brazil | Iron-ore tailings | About 270 deaths; estimates of release range from about 9.7 to about 12 million m3. |
Patterns visible across the cases
The entries differ in ore, climate and decade, yet several patterns recur in the investigation reports.
- Rapid raising on weak ground. Stava combined an upstream-built bank, poor drainage and a steep outer slope; the Mount Polley panel traced the breach to a foundation layer that site investigations had not identified.
- Water control. Merriespruit failed by overtopping after heavy rain; Baia Mare involved insufficient storage for the precipitation at the time; Stava involved drainage problems.
- Warnings and oversight. Reports on Stava describe an earlier assessment that called stability marginal and dams that went without serious stability checks for some twenty years. At Church Rock, cracks had been noted in the dam wall before the breach.
- Downstream exposure. The large death tolls (Stava, Fundão, Brumadinho) occurred where people lived or worked directly below the dam.
These recurring causes are the subject of the page on causes of failure and risk factors, and the instrumentation that can reveal them is covered under monitoring and surveillance.
Where uranium tailings fit in
Uranium mill tailings contain long-lived radium-226 and thorium-230 and, in many cases, acid and heavy metals, so a release carries radiological as well as chemical concerns. Church Rock is the principal documented dam failure at a uranium mill in the public record and has its own treatment on the page about uranium mill tailings dam failures. For the general nature of these wastes see uranium mill tailings: characteristics and management.
Other releases at uranium facilities, such as process accidents or legacy contamination, are not tailings dam failures and are catalogued separately on the page on incidents beyond dams.
Limits of this chronology
Several well-known events are absent because the figures in available sources conflicted or could not be traced to an institutional report. The omission is not a judgement about their severity. Where an entry says that estimates differ, the spread usually reflects different definitions: some sources count only solid tailings, others include the stored water, and others include material eroded along the flow path.
Two further cautions apply. First, the word failure covers very different events: a rapid collapse that liquefies the stored tailings and runs for kilometres (Stava, Fundão, Brumadinho) is not the same as an overflow of process water (Baia Mare) or a breach that releases mostly water and fine solids into a river (Aznalcóllar, Mount Polley). Second, the consequences depend as much on where the people and ecosystems are as on the volume released. A comparatively small release of cyanide-bearing water can be a transboundary pollution event, while a large release in an empty valley may cause little loss of life.
Readers who need a precise figure for a specific event should consult the original investigation report, the regulator's file or the relevant UN assessment, all listed below where available.
Frequently asked questions
What is a tailings dam?
A tailings dam is an embankment that retains the fine, water-laden waste left after ore processing. It is often raised in stages during the life of the mine.
Which tailings dam failure killed the most people?
Among the events in this chronology, Brumadinho (2019) is reported to have killed about 270 people and Stava (1985) killed 268. Sources differ slightly on the Brumadinho count.
Has a uranium tailings dam ever failed?
Yes. The best-documented case is Church Rock, New Mexico, in July 1979, when a dam at a uranium mill breached and released acidic liquid and tailings solids into the Rio Puerco drainage.
Why do published volumes for the same failure differ?
Sources count different things: solid tailings only, tailings plus stored water, or all material mobilised along the flow path. The spread can be large, so this page shows ranges.
Is there a single official global list of failures?
No. ICOLD and academic groups maintain databases, but each has its own criteria and gaps, so any chronology is a selection.
Sources and further reading
- Stava 1985 Foundation — The Stava tailings dam failure, 19 July 1985 (2020)
- tailings.info — Stava tailings dam failure (case study) (2019)
- US Environmental Protection Agency — Biological assessment documents on the 1979 Church Rock spill (2010)
- Mount Polley Review Panel — Report of the Mount Polley Independent Expert Engineering Investigation and Review Panel (2015)
- UNEP / OCHA — Cyanide spill at Baia Mare, Romania: UNEP/OCHA assessment mission (2000)
- Hungarian disaster management authorities / UNECE — Hungary red mud disaster: presentation to UNECE conference (2010)
- Association of State Dam Safety Officials — Warnings unheeded: the 1994 Merriespruit tailings dam disaster (2012)
- Stava 1985 Foundation — The Aznalcollar tailings dam failure, 1998 (2020)
- Journal of Hazardous Materials (Rico et al.) — Tailings dam failures: a review of the European incidents in the worldwide context (2008)
Related topics
