Global Tailings
Storage Facility
Failures
An analytical report on 110 years of mine tailings dam failures — patterns, causes, consequences, and the path to responsible operations.
A Crisis That Is Not Slowing Down
Tailings storage facilities (TSFs) hold some of the most hazardous material on Earth — billions of tonnes of finely-ground mineral waste, process chemicals, and contaminated water. When they fail, the consequences cascade: loss of life, destruction of downstream communities, contamination of water supplies serving millions, and ecological damage that persists for generations.
This report synthesizes data from 306+ documented TSF failures spanning 1914 to early 2026, drawing on the World Mine Tailings Failures (WMTF) database, the research of Lindsay Newland Bowker, the Earthworks Safety First initiative, and recent field reporting. The evidence is unambiguous: failure rates are increasing, consequences are worsening, and the structural conditions that produce failure are being replicated faster than the industry is learning from them.
"85% of catastrophic failures occur at active TSFs — those receiving tailings from current production. Failure is not a legacy problem. It is a present-day operational reality."
— Lindsay Newland Bowker, World Mine Tailings Failures, 2020The three dominant failure modes — slope instability, overtopping, and foundation failure — account for approximately 70% of all documented incidents. These are not exotic or unpredictable failure mechanisms. They are the same mechanisms identified in failures 50 years ago. The industry's failure to systematically eliminate known risks is the central lesson of this dataset.
Three critical failure events in early 2025 alone — in Zambia, Indonesia (two events), and Bolivia — demonstrate that the trajectory has not improved despite the adoption of the Global Industry Standard on Tailings Management (GISTM) in 2020.
110 Years of Failure Data
The documented record of TSF failures spans from 1914 to the present. While data completeness improves significantly from the 1960s onward, the trend is unmistakeable: failure frequency has grown alongside global mining production, and the shift toward lower-grade ore has dramatically increased the volume of tailings generated per tonne of metal produced.
The Supercycle Effect
The 1990–2010 commodity supercycle drove unprecedented production volume increases. As mines grew larger and processing facilities expanded to handle lower-grade ores, TSFs grew accordingly — but the engineering practices, staffing competence, and regulatory oversight did not keep pace.
Research by Bowker and Newland (2015) documents a direct statistical correlation between copper production volumes and tailings failure rates. As ore grades decline, more rock must be processed per tonne of metal, generating larger volumes of finer-grained tailings — which are inherently more susceptible to liquefaction.
Why Tailings Dams Fail
Analysis of 306+ documented failures reveals six dominant failure modes. Most are not singular causes — they interact. Overtopping is commonly triggered by inadequate freeboard combined with abnormal rainfall. Slope instability is accelerated by poor drainage that elevates pore pressures. Foundation failure can be masked by surface stability until catastrophic collapse occurs.
Upstream construction is associated with disproportionately high failure rates relative to downstream and centerline methods. The El Cobre failure (Chile, 1965) which killed 200 people, and the Brumadinho (Brazil, 2019) which killed 270, were both upstream-constructed dams.
The GISTM (2020) explicitly prohibits new upstream construction for facilities with high or extreme consequence classification — a direct response to this pattern.
Rainfall events appear as contributing factors in over 35% of all failures, underscoring the need to design TSFs for extreme precipitation — especially critical as climate change intensifies storm events.
Where Failures Concentrate
TSF failures are not uniformly distributed. They cluster in high-production mining nations, in regions where regulatory oversight is weakest, and along seismically active belts. The shift in the failure trajectory — from developed to developing countries — is a defining trend of the post-2000 period.
South America
Chile has had the most seismically-triggered TSF failures in the record — more than 20 facilities failed in the 1965 and 1985 earthquakes alone. Brazil's recent mega-failures (Mariana 2015, Brumadinho 2019) have shifted the global focus to iron ore tailings in Minas Gerais state, where dozens more facilities remain classified as high or extreme consequence.
Peru, Bolivia, and Colombia continue to report failures from both artisanal and large-scale operations, often with limited environmental monitoring and community notification.
Asia and the Pacific
The Philippines has one of the highest per-country failure counts in the record, with copper and gold mining operations contributing numerous incidents since the 1960s. The Marcopper disaster (1996) contaminated 26 km of river and remains a landmark case in mining liability.
China's TSF portfolio — estimated at 12,000–15,000 structures — has undergone aggressive reform since 2015, including closure and de-risking programs. However, Indonesia's 2025 failures at the Morowali Industrial Park show that rapidly developing mining zones in the region remain highly vulnerable.
The trajectory of global TSF failures has shifted from developed to developing countries since 2000. Weaker regulatory frameworks, inadequate engineering capacity, and compressed production timelines create systemic vulnerability in mining-dependent economies.
— Global-scale impact analysis of mine tailings dam failures: 1915–2020, ScienceDirect 2021The Failures That Changed the Field
Certain TSF failures have had an outsized influence on policy, engineering standards, and public awareness. Each one revealed systemic gaps — and each one was followed by commitments that were, with varying success, translated into practice.
Brumadinho (Córrego do Feijão) — Vale, Brazil
What happened
- Upstream-constructed dam failed within 2 minutes of collapse onset
- 12 million m³ of iron ore waste released
- Toxic mud wave hit worker cafeteria and downstream communities
- Dam had received safety certification months before failure
- 270 killed — worst mining disaster in Brazilian history
Root causes and lessons
- Upstream construction method fundamentally unsafe for this scale
- Inadequate monitoring of internal pore pressure
- Third-party safety certification was not independent in practice
- Vale had been warned of dam instability as early as 2003
- Triggered global GISTM process and ban on upstream construction
In November 2025, the UK High Court ruled BHP strictly liable for the related Fundão/Mariana 2015 failure. A US$30 billion compensation agreement for both disasters was finalized in 2024 — the largest mining liability settlement in history.
Fundão (Mariana/Samarco) — BHP/Vale, Brazil
What happened
- 43.7 million m³ released — largest TSF failure in Brazilian history
- Toxic plume reached Atlantic Ocean 17 days after failure
- 668 km of watercourses contaminated
- Two villages completely destroyed
- As-built design was never fully implemented at failure time
Root causes and lessons
- Construction error in 2009 negated the drainage design concept
- Production pressure led to rushed construction without proper design
- Independent Tailings Review Board engaged too late
- Predicted failure extent vastly underestimated in EIA
- Design-as-built discrepancy was never caught in audits
Mount Polley Mine — Imperial Metals, Canada
What happened
- Foundation failure on a clay-rich glaciolacustrine layer
- 10 million m³ of water and 4.5 million m³ of tailings released
- 25 km of waterways affected including Hazeltine Creek
- Quesnel Lake — a drinking water source — contaminated
- No deaths but catastrophic environmental consequence
Root causes and lessons
- Glaciolacustrine foundation layer not adequately characterized
- Dam raised beyond design capacity without proper re-assessment
- Independent Expert Panel recommended ban on upstream construction
- Regulatory gap: no mandatory independent oversight mechanism
- Led directly to British Columbia TSF regulation overhaul
Ajka Alumina — MAL Zrt, Hungary
What happened
- Shear failure in a dyke 300m from the dam
- 1 million m³ of caustic red mud (pH 13) released
- Village of Kolontár overwhelmed, Devecser flooded
- Danube River contaminated downstream into Croatia
- 10 killed, 120 injured
Root causes and lessons
- Low-probability design: inadequate safety margins
- Water pressure buildup from excessive precipitation
- EU Seveso Directive did not classify TSFs as major hazard
- Forced the reclassification of bauxite residue as hazardous
- Prompted EU review of mining waste directive
Buffalo Creek — Pittston Coal, West Virginia, USA
What happened
- Upstream-constructed coal slurry dam failed after heavy rainfall
- 500,000 m³ of black waste released at 80 km/h
- 16 communities destroyed in minutes
- 4,000 people left homeless; 125 killed
- $65 million in property damage (1972 dollars)
Root causes and lessons
- Three coal slurry dams stacked in sequence — each held by next
- Pittston Coal called it an "act of God" — courts disagreed
- No engineering review, no regulatory oversight
- Led to US SMCRA (Surface Mining Control and Reclamation Act 1977)
- Pioneering case in corporate mine waste liability
The Crisis Continues
The period from June 2024 through March 2025 saw at least five major TSF failures on three continents, occurring against a backdrop of the global mining industry's stated commitment to the GISTM. These failures demonstrate that adoption of standards at the corporate level has not translated into safety outcomes at all operating sites.
| Event | Location | Date | Severity | Key Factor |
|---|---|---|---|---|
| Chinchorro TSF | Peñablanca, Chile | June 13, 2024 | Moderate — limited flow due to thickened tailings | Overtopping from 100mm rainfall event |
| Sino-Metals Leach Zambia | Chambishi, Zambia | Feb 18, 2025 | Catastrophic — 50M litres into Kafue River | Cascade failure between TSF cells |
| PT Huayue Nickel Cobalt (HNC) | IMIP, Sulawesi, Indonesia | Mar 16, 2025 | Major — wave of red tailings, 341 families affected | Heavy rainfall, questionable base construction |
| PT Qing Mei Bang (QMB) | IMIP, Sulawesi, Indonesia | Mar 21, 2025 | Fatal — 3 workers killed | Built over infilled pond; prior landslide signs ignored |
| Laguna Kenko | Llallagua, Bolivia | Mar 16, 2025 | Fatal — 2 killed, 47 homes destroyed | Legacy tin mine TSF being re-mined without re-engineering |
Zambia, 2025
The Sino-Metals failure contaminated the Kafue River — the most important waterway in Zambia and drinking water source for 60% of the country's 20 million citizens. 700,000 people in Kitwe lost water supply. 176 farmers filed an $80 billion lawsuit against Chinese-linked mining firms. As of August 2025, the US Embassy was still ordering personnel out of affected areas citing ongoing contamination.
A 2014 government audit had already identified systemic mismanagement at Copperbelt TSFs — including the Chambishi complex. Warnings were not acted upon.
Indonesia, 2025
Two failures in five days at the Indonesia Morowali Industrial Park (IMIP) — one of the world's largest nickel processing zones. Google Earth imagery from January 2025 (two months before the March failures) shows a prior landslide at the same facility, indicating pre-existing instability that was not remediated.
The IMIP sits on the Matano Fault — a highly active seismic zone. The construction of a TSF on top of an infilled pond (as with PT QMB) is a fundamental geotechnical error that should not have passed any competent review.
Bolivia, 2025
The Laguna Kenko failure at a legacy tin mine being re-processed illustrates a growing global risk category: closed or abandoned TSFs being recommissioned without adequate re-engineering. As commodity prices rise, operators reprocess legacy tailings for residual metals — but the original structures were not designed for re-use and may have deteriorated significantly.
The failure destroyed 47 houses and impacted 70% of the downstream community of Andavilque. Heavy metals from the tailings were already leaching into the environment before the catastrophic breach.
"In all three 2025 cases (Zambia, Indonesia, Bolivia), heavy rains appear to have factored into the failure — but rainfall is not a cause of dam failure. Tailings dams must be designed to withstand the rainfall patterns of their location, especially as climate change intensifies weather extremes."
— Earthworks, Safety First Initiative, April 2025The Same Mistakes, Over and Over
The most troubling finding from 110 years of failure data is not the scale or frequency of failures — it is how consistent the contributing factors are. The database reveals that the mining industry has not systematically applied the lessons from catastrophic failures. The same warnings appear in reports from 1972, 1985, 2000, 2014, and 2025.
Upstream construction persists despite known risk
Upstream-constructed TSFs are overrepresented in catastrophic failures. Despite the GISTM's 2020 ban on new upstream construction for high-consequence facilities, thousands of existing upstream-constructed dams remain in operation globally.
Production pressure overrides engineering judgment
In nearly every major failure with a detailed post-incident report, evidence emerges that warnings were known but production timelines took precedence. Fundão, Brumadinho, Mount Polley, and Buffalo Creek all share this pattern.
Third-party audits are not truly independent
The Brumadinho dam received a safety certificate months before collapse. The Independent Tailings Review Board at Fundão was engaged after construction was already underway. Independence must be structural, not nominal.
Regulatory warnings are systematically ignored
In Zambia (2025), a 2014 government audit had identified the exact systemic failures that led to the Chambishi collapse. Eleven years passed. In Brazil, Vale was warned of dam instability at Brumadinho as early as 2003.
Design-as-built divergence is common and undetected
Multiple failures involve discrepancies between approved design and actual constructed condition. Fundão's drainage design was fundamentally altered in 2009 due to construction error — but the ITRB was not engaged until after the dam was already operational.
Inactive and closed TSFs are underestimated risks
20% of catastrophic failures occur at inactive or closed facilities. Legacy tailings are increasingly being re-processed — often without adequate re-engineering. Climate change is also destabilizing structures that were designed for historical rainfall patterns.
EIA consequence underestimation is routine
At Fundão/Mariana, the EIA predicted waste would flow 3.5 km. It traveled 668 km. Consequence modelling must account for maximum credible failure scenarios, not minimum plausible ones. The difference is who dies.
Competence gaps are pervasive and growing
Research by Caldwell and Bowker suggests that even among the world's top 40 mining companies, there are insufficient qualified tailings engineers to manage global portfolios. The commodity supercycle hollowed out technical capacity precisely when it was most needed.
Climate change is a multiplier, not a standalone cause
All five 2025 failures involved heavy rainfall as a contributing factor. As extreme precipitation events intensify and become more frequent, the design standards built on historical rainfall data become dangerously inadequate.
What Responsible Practice Looks Like
The engineering and management knowledge to prevent most TSF failures exists today. The gap is not primarily technical — it is organizational, regulatory, and economic. The following framework draws on the GISTM (2020), the Earthworks Safety First Guidelines, the Mt. Polley Expert Panel recommendations, and the work of Bowker, Caldwell, and others.
8.1 Design and Construction Standards
8.2 Monitoring and Instrumentation
Modern TSF monitoring has advanced significantly. Real-time pore pressure monitoring, satellite-based InSAR deformation monitoring, and automated alert systems now allow operators to detect early warning signs of instability with days or weeks of lead time. These technologies are available — but their deployment is uneven.
8.3 Governance and Accountability
Structural independence of oversight
The Engineer of Record, the Independent Tailings Review Board, and the qualified person responsible for safety certification must have legally protected independence from mine operators and their consultants. Nominal independence is insufficient. The Brumadinho certification failure illustrates this precisely.
Consequence-based regulatory classification
Every TSF must be classified by its potential failure consequence — the number of people, the sensitivity of downstream environment, the volume of tailings. Regulatory requirements should escalate with consequence classification. High and extreme consequence TSFs require the highest design, monitoring, and oversight standards, regardless of operator size or country income level.
Publicly accessible global TSF registry
As of 2025, no comprehensive global registry of TSFs exists with standardized risk data. WMTF estimates only ~10% of the world portfolio has publicly available consequence data. A mandatory, publicly accessible registry — with annual disclosures on design, consequence classification, and inspection status — is the single highest-leverage policy intervention available.
Building code model for tailings regulation
Bowker (2017) proposes the "building code" regulatory model for tailings: a legally enacted code with professionally staffed enforcement agencies, mandatory third-party inspections, and clear competence requirements for practitioners. This model has successfully governed increasingly complex high-rise construction. It can do the same for tailings facilities.
GISTM 2020: Progress and Gaps
The Global Industry Standard on Tailings Management, released in August 2020 following the Brumadinho disaster, represents the most comprehensive voluntary standard ever produced for TSF management. It covers design, construction, operation, monitoring, closure, and stakeholder engagement. But it is voluntary — and its coverage of the global portfolio is limited.
The GISTM is a necessary but insufficient condition for a safe global TSF portfolio. Without government mandates requiring its adoption, verified third-party audits, and mandatory public disclosure, the standard will protect the reputations of ICMM members while the broader global portfolio remains unaddressed.
— Adapted from WMTF analysis, Easter Sunday 2022The March 2025 failures in Indonesia and Bolivia — both post-GISTM — demonstrate that the standard's current scope and voluntary nature are not creating the safety outcomes it promises. The GISTM was updated in February 2025 during the African Mining Indaba, just days before the Zambia disaster.
From Reactive to Preventive: The Decision Framework
Moving the industry from reactive crisis management to systematic failure prevention requires integrating engineering, governance, workforce competence, and community accountability into a coherent framework. The following structure draws on the most effective elements of existing standards and the lessons from 110 years of failures.
Design for the worst credible failure scenario
Use Probable Maximum Precipitation (PMP) as the design storm. Eliminate upstream construction for new high-consequence facilities. Characterize foundation fully before first lift. Require independent review of every design departure.
Verify construction quality continuously
Every raise must replicate or improve on the approved design. As-built documentation must be independently verified. Deviations trigger immediate engineering review before work continues. No production pressure exception.
Monitor with meaningful decision triggers
Install a real-time monitoring suite proportionate to consequence classification. Define action levels that trigger response — not just review. Regular field inspections by qualified engineers. Annual stability assessment reviewed by Independent TRB.
Maintain structural accountability
Engineer of Record with continuous authority. ITRB genuinely independent of operator and consultants. Board-level officer accountable for TSF performance. Community Emergency Response Committees with real authority.
Plan and fund closure from day one
Closure cost must be estimated from feasibility stage and funded progressively. No mine should be permitted without a credible, funded closure plan. The entity responsible for long-term stewardship must be identified and resourced before first production.
WMTF estimates the global cost to de-risk all identified high-risk TSFs at approximately $80M per facility — or $0.01/metric tonne of global mineral production per year over 5 years. Against an average failure cost of $US2.5 billion per catastrophe (including public liabilities and community harm), the economic case is overwhelming.
10.1 The Competence Imperative
Perhaps the most underappreciated systemic risk in global tailings management is the shortage of qualified engineers with the specific expertise — geotechnical, hydrological, and process engineering — required to design, operate, and review TSFs. Research by Jack Caldwell (reported by Bowker) suggests that even among the world's top 40 mining companies, there are too few qualified tailings engineers to cover each company's global portfolio — and many major companies rely entirely on consultants.
This creates a systemic vulnerability: when a company's institutional knowledge of a specific facility is held entirely by a consulting firm with competing engagements, continuity is fragile. Knowledge that should reside in the owner is instead scattered across contractual relationships.
The Path Forward Is Known. The Will Remains Uncertain.
The technical knowledge, monitoring technology, engineering methods, and governance frameworks needed to prevent most tailings dam failures exist today. They are codified in the GISTM, elaborated in the Safety First Guidelines, demonstrated in hundreds of well-designed facilities worldwide, and validated by decades of research. The persistence of catastrophic failures is therefore not primarily a technical failure — it is a failure of governance, accountability, and organizational culture.
For Operators
Build genuine in-house competence
Every company with more than one operating TSF needs qualified tailings engineers with full-time responsibility for those facilities — not just consultant relationships activated at intervals.
Commission consequence assessments for all TSFs in the portfolio
Understand what failure of each facility would actually mean — for workers, communities, water systems, ecosystems. This is the foundation of proportionate risk management.
Invest in real-time monitoring proportionate to consequence
Piezometers, InSAR, automated alerts, and seepage monitoring are not costs — they are insurance. Early warning systems have value only if they trigger mandatory response protocols with defined escalation.
Establish funded closure plans before first production
The engineering and financial obligation for a TSF does not end at mine closure — it extends for generations. That obligation must be quantified and secured from the outset.
For Regulators and Governments
Mandate a publicly accessible TSF registry
Every licensed mine should be required to annually disclose: facility inventory, consequence classification, design type, inspection dates, and engineer of record. Without visibility, risk cannot be managed at the portfolio or national level.
Require financial assurance proportionate to failure consequence
Environmental bonding should be set at the actual cost of the worst credible failure — not the expected cost of a minor incident. Financial assurance must be held by an entity independent of the operator.
Enact legally binding standards based on the GISTM
Voluntary standards protect ICMM members' reputations but cannot reach the broader global portfolio. The building code model — legally enacted standards with verified inspections — is the appropriate regulatory vehicle for facilities that can kill thousands.
Protect whistleblowers and affected community voices
In the majority of catastrophic failures, warning signs were known to workers and local communities before failure. Creating safe, effective channels for these voices to reach decision-makers is a critical loss-prevention investment.
Every tailings dam failure documented in this report was preceded by conditions and signals that, in retrospect, were legible. The question the industry must honestly answer is not "how do we detect failures?" — it is "why do we not act on the signals we already receive?"
— Synthesis from 110 years of documented TSF failure casesThe minerals that the world needs for clean energy, technology, and infrastructure will require more mining — and therefore more tailings. The challenge is not whether to produce those tailings, but whether we have the organizational and regulatory systems to manage them responsibly. NEXGROW Academy's Tailings Management program exists to build precisely that capability: the judgment, technical grounding, and decision-making frameworks that responsible operations require.
The database of failure is long. It does not have to grow.
Reference Data Table
Selected major documented TSF failures. Full dataset compiled from WMTF database, published research, and post-2015 event reporting.
| Year | Facility | Country | Commodity | Primary Cause | Volume (m³) | Fatalities |
|---|---|---|---|---|---|---|
| 1965 | El Cobre (Old Dam) | Chile | Copper | Earthquake liquefaction | 1,900,000 | 200 |
| 1966 | Mir Mine | Bulgaria | Lead/Zinc | Overfilling — extreme rain | 450,000 | 488 |
| 1966 | Aberfan | Wales, UK | Coal | Liquefaction — heavy rain | 162,000 | 144 |
| 1972 | Buffalo Creek | USA | Coal | Structural failure — rain | 500,000 | 125 |
| 1974 | Bafokeng | South Africa | Platinum | Seepage and pipe failure | 3,000,000 | 12 |
| 1982 | Sipalay No.3 Pond | Philippines | Copper | Foundation failure | 28,000,000 | 0 |
| 1985 | Stava | Italy | Fluorite | Slope instability | 200,000 | 269 |
| 1992 | Tubu | Philippines | Copper | Foundation collapse | 23,243,000 | 0 |
| 1994 | Merriespruit | South Africa | Gold | Overtopping — rain | 600,000 | 17 |
| 1995 | Omai | Guyana | Gold | Internal erosion | 4,200,000 | 0 |
| 1996 | Marcopper | Philippines | Copper | Structural failure | 1,600,000 | 0 |
| 1998 | Los Frailes (Aznalcóllar) | Spain | Lead/Zinc/Cu | Foundation failure | 6,800,000 | 0 |
| 2000 | Baia Mare | Romania | Gold | Overtopping — snowmelt | 100,000 | 0 |
| 2010 | Ajka Alumina | Hungary | Bauxite | Dyke shear failure | 1,000,000 | 10 |
| 2014 | Mount Polley | Canada | Cu/Au | Foundation failure | 24,400,000 | 0 |
| 2014 | Herculano | Brazil | Iron ore | Landslide — geology | 1,000,000+ | 3 |
| 2015 | Fundão (Mariana/Samarco) | Brazil | Iron ore | Foundation instability | 43,700,000 | 19 |
| 2019 | Brumadinho (Córrego do Feijão) | Brazil | Iron ore | Upstream dam liquefaction | 12,000,000 | 270 |
| 2022 | Jagersfontein | South Africa | Diamond | Structural failure | Unknown | 1 |
| 2024 | Chinchorro TSF | Chile | Mining | Overtopping — 100mm rain | Limited | 0 |
| 2025 | Sino-Metals/Chambishi | Zambia | Copper | Cascade cell failure | 50,000,000 L | 0 direct |
| 2025 | PT HNC — IMIP | Indonesia | Nickel/Cobalt | Heavy rain + poor base | Unknown | 0 |
| 2025 | PT QMB — IMIP | Indonesia | Nickel | Built over infilled pond | Unknown | 3 |
| 2025 | Laguna Kenko | Bolivia | Tin | Legacy facility, inadequate re-engineering | Unknown | 2 |
