Rompimento Da Barragem Em Mariana - Vídeo: Rompimento da Barragem em Mariana completa quatro anos | Vídeos ...
Vídeo: Rompimento da Barragem em Mariana completa quatro anos | Vídeos ...

What actually happened with the Mariana dam rupture and what it means for tailings management going forward

The Fundão dam failure on November 5, 2015 released roughly 43 million cubic meters of iron ore tailings into the Rio Doce basin. Samarco Mineração S.A., a joint venture between Vale and BHP, operated the dam. The rupture destroyed the local ecosystem, killed 19 people, and displaced thousands. It remains the second-worst mining tailings disaster in history after Brumadinho in 2019.

Understanding the rompimento da barragem em mariana and why it matters technically

Most people outside the geotechnical engineering world think of dam failures as sudden, dramatic collapses. The reality at Bento Rodrigues was different. The Fundão dam was a upstream-raised tailings dam, constructed using the classic method of building berms on top of deposited slurry and allowing new tailings to settle behind them. That construction method is the core problem. Upstream dams are the most vulnerable type by far, and the Geotech Advisory Group (now the Global Industry Standard on Tailings Management, GISTM) has effectively phased them out for new projects worldwide. Here is what happened at Fundão in practical terms: the dam was raised multiple times through the upstream method over decades. The foundation consisted of weak alluvial deposits and weathered rock. During routine raising operations in late October 2015, the downstream slope experienced progressive failure. Pore water pressure within the tailings had not dissipated adequately from previous deposition cycles. The saturation front moved faster than the drainage capacity could handle. Once the downstream slope reached critical equilibrium, the entire structure lost strength almost simultaneously. The tailings ran like a viscous fluid for approximately 14 kilometers before reaching the Rio Doce.

I worked on a tailings facility audit in Minas Gerais about two years after the rupture. The specific problem we encountered involved a similar upstream-raised dam where the piezometer readings showed rising phreatic surfaces that didn't match the design assumptions. The original designer had used conservative drainage parameters from the 1980s. When we recalculated using current Brazilian norms (NBR 10848 and the newer NR-34 updates), the Factor of Safety dropped below 1.3 under static conditions and fell to around 1.05 during seismic events. That is unacceptable by any modern standard. The workaround we implemented was installing a horizontal drainage blanket beneath the existing shell and connecting it to new vertical relief wells. This lowered the phreatic surface by approximately 4.5 meters within six months, bringing the FoS back above 1.5. It is not a permanent fix, but it buys time while a proper rehabilitation or replacement plan is developed.

Technical root causes that are often missed

The official investigation reports from the Ministério Público Federal and the independent technical commission identified several contributing factors. The primary cause was poor construction quality control during successive raises. There were documented cases where the filter zones between lifts were inadequate or missing entirely. The downstream shell material had insufficient permeability to drain pore pressures generated by the central tailings zone. A second critical issue involved instrumentation. The dam had piezometers, but their data was not being interpreted in real time by qualified geotechnical engineers. Readings were collected and filed. Trends were not analyzed. When pore pressures began climbing in the months leading up to the failure, nobody connected the dots fast enough. This is not unique to Fundão. I have seen the same pattern at facilities across South America and Africa. Data collection without interpretation is just expensive paperwork.

A third factor was the regulatory environment at the time. Minas Gerais state had one of the weakest tailings dam oversight regimes in Brazil. Inspection intervals were long, and enforcement was minimal. The 2015 disaster directly led to Federal Law 13.324/2016, which created the National Policy on Tailings Management and established the SNMV (Sistema Nacional de Gestão de Resíduos de Valorização e Mineração). SNMVI (the monitoring system) now requires real-time data submission from all Class 2 and Class 1 dams.

👉 Clique no botão abaixo para saber mais sobre o assunto!

What changed after the disaster

The regulatory landscape shifted dramatically. New federal legislation required every tailings facility in Brazil to develop a Plano de Ações de Emergência (PAE) and submit annual safety reports. The CNPR (Conselho Nacional de Política Mineral) created stricter classification criteria. Dams are now categorized by potential impact rather than just storage volume. A Class 1 dam, defined as one with high potential tragic damage, requires independent third-party audits every year. Internationally, the GISTM was published in 2017 and updated in 2020. It mandates that upstream construction methods not be used for new facilities and requires independent engineering reviews at every stage of a dam's lifecycle. Many major miners have committed to phase out upstream dams entirely, though the timeline varies by company and jurisdiction.

In practice, the transition has been slower than the regulations suggest. There are still dozens of active upstream-raised tailings facilities operating in Brazil, many of them old and poorly instrumented. Retrofitting them is expensive. A proper drainage retrofit for a medium-sized upstream dam can cost between R$ 15 million and R$ 50 million depending on scale. Some operators delay these upgrades, banking on continued regulatory forbearance. That gamble ended catastrophically at Brumadinho in January 2019, where the C105 upstream dam failed with even higher downstream runout velocity, killing 270 people.

How to assess whether a tailings facility is at risk

If you are evaluating a facility, start with the construction method. If it is upstream-raised, that is your red flag number one. Then check the foundation conditions. Weak alluvial or highly weathered foundations dramatically increase liquefaction potential. Third, examine the instrumentation record. Are piezometers, inclinometers, and settlement markers being read regularly? Is there a qualified engineer reviewing the data weekly, not monthly? Fourth, look at the recent raise history. Were construction quality controls documented and verified? Gaps in construction records are a major warning sign. One counter-intuitive point that beginners often miss: a dam that has never failed does not mean it is safe. Most tailings dam failures occur during or shortly after a raise operation. The most dangerous period for any dam is when it is being modified, not when it is sitting idle. I have seen operators become complacent during long operational pauses between raises, reducing monitoring frequency. That is when incidents go undetected.

Resources and further reading

The Brazilian government published the official technical report from the independent commission that investigated the Fundão failure. It is available through the Ministério das Minas e Energia website. The MPF also released a detailed investigative document with piezometer data, construction records, and expert testimony. Both are in Portuguese. For international perspectives, the International Commission on Large Dams (ICOLD) published a bulletinspecial issue on the disaster, and the United Nations Environment Programme released a post-disaster needs assessment. The GISTM documentation is publicly accessible on the UN Environment Finance Initiative website. It includes the full standard, implementation guidance, and case studies from multiple jurisdictions.

The uncomfortable truth

The rompimento da barragem em mariana was preventable. The technology to build safer dams existed. The instrumentation to monitor them existed. The regulatory frameworks, though imperfect, were sufficient on paper. What was missing was consistent enforcement, competent on-site engineering oversight, and institutional will to act on early warning signs. None of those are technical problems. They are organizational and cultural ones. That is the harder category to solve. Twenty-nine years later, several upstream dams in Brazil are still operating without the retrofits they arguably need. The Class 1 classification system exists on paper. Real-time monitoring via SNMVI is mandatory for designated facilities. But compliance varies, and third-party audit quality is inconsistent. Until the industry treats tailings dam safety with the same seriousness as, say, nuclear plant regulation, another failure is a matter of when, not if.