The Fundão Mariana disaster, as seen from the geochemist's point of view

Mining waste accidents affect the business with an unfortunate regularity. The biggest ones make the headlines and sometimes contribute to the evolution of regulations. Scientific arguments are used, and sometimes misused, when the debate warms up, or when financial liabilities are at stake.

The disaster at Mariana occurred in November 5th 2015. SAMARCO Mining (a VALE and BHP Billiton joint venture) is responsible for the mining activities. The mud and water released by the dam failure affected the population of Bento Rodrigues village and residents of Gualaxo and Doce rivers banks (link to  https://pt.wikipedia.org/wiki/Rompimento_de_barragem_em_Mariana#/media/File:Bento_Rodrigues,_Mariana,_Minas_Gerais_(22828956680).jpg) . The mud affected the Doce river through to the city of Governador Valadares. The water (rich in Fe oxides) reached the ocean at Linhares city (ES state).

Visitors to Mariana by the end of November could not reach the Fundão Dam for safety reasons at that time, but only the Santerém Dam, just below Fundão dam, and the Barra Longa village area. This did not allow clear access to see much of Fundão dam damage.


Geochemistry and dam safety

Tailings dams stability is a matter of geomechanics, dam construction and/or foundation. Geochemistry may contribute to safety analysis, when the chemical properties of waste or wastewater may affect the stability of engineering.

The failure of the Los Frailes Aznalcollar dam, April 1998, was suspected to have been triggered by the corrosion of the underlying marl layer into wet clay by the acidic wastewater, transforming tough rock grounding into a sliding surface[1]. Geomechanics experts successfully tested dry marl before dam construction. Marl behaviour in contact of acidic waste may have not been tested.

Other issues with dams stability may be caused by groundwater or by interactions between waste and dam material. Geochemists should bring their expertise in such matters, which are of paramount importance for population safety and health downstream dams, and for environmental protection.

At Fundão Mariana where the mining site is, the geology is composed mainly of metassedimentary rocks (quartzites, metaconglomerate, slate, schists and metatuffs) and BIF.

The tailings were not known to be acidic and their composition was not suspected to have played a role in the accident. Data on their geochemical monitoring were not publicly available, to the best of our knowledge.

A daily hydrologic monitoring program was operated by CPRM/Geological Survey of Brazil until March 30th, 2015. CPRM sampled Mariana region (stream sediments, pan concentrates and soil) in 2009. CPRM also sampled stream sediments and water at Doce river in 2010. Summary results on the monitoring of river sediments (As, Cd, Cu, Hg, Pb, Sb, Zn) and river water (pH, EC, DO, As, Cd, Cu, Hg, Pb, Sb, Zn, F-, Cl-, Br-, NO2-, NO3-, SO4--) are available from CPRM (link to http://www.cprm.gov.br/publique/Hidrologia/Eventos-Criticos/Monitoramento-Especial-do-Rio-Doce-4057.html ).

Even if the mining signature is observable in the published data, it did not reach sensitive levels.

2009-2010 monitoring results show lower concentrations of heavy metals than the samples collected in November 2015. The mud from SAMARCO is product of BIF crushing/milling/floatation processes. The SAMARCO BIF is heavy metal poor (As, Cd, Cu, Pb, Zn etc.) comparing to the natural background from Mariana (gold bearing) rocks. Results of Fe and Mn are higher in the mud than those from the original stream sediments.

Since the disaster, CPRM/Geological Survey of Brazil and ANA/National Water Agency have been working and sampling water and sediments from the dam area, Doce river and some tributaries.

There is a sampling program (once a month) on 7 sites along the affected area, from Mariana, Minas Gerais (MG) State to Espirito Santo (ES) State (near the shore). Water, suspended sediments and stream sediments are the sampling media.

Other institutions are also working in the area (MG and ES states environmental agencies, universities).


Geochemistry and damage assessment

Understanding the damage and evaluating the risks

As soon as an accident happens, everyone investigates and provides their own conclusions. Life losses or property damage are easily identified, but the main causes and consequences require more attention.

The chemical nature of the released waste has to be properly understood: which kind of solids, sludge or water were released? What was their contents in potentially toxic substances, and their possible immediate or long-term effects? How far did the released waste go, where did it sit, and what possible effects its presence may cause? What are the possible effects on exposed populations downstream, which detailed investigations are needed to design an emergency plan, and what kind of prevention measures are needed?

At Fundão Mariana, the damage itself was caused by a massive flow (30 to 60 milion m3) of iron-rich mud, covering villages, fields and the valley. However, the great impact of iron hydroxide (https://www.instagram.com/p/_y10s1udBc/ ) in water should not be confused with the mud flow. The Risoleta Neves dam, pictured here, continued operations. This means that at this distance (about 100 km) from the accident, the spill is constituted by solute metals and suspended matter in water. This form of contamination went down to the Atlantic Ocean (a further 670 km downstream), not the mud flow itself.

How toxic was this mud? Iron oxide and silica are not contaminants per se, though they may choke river life by their excess. A paper by Luisa Massarani (http://www.rsc.org/chemistryworld/2015/11/brazil-mine-disaster-dam-collapse) reported substances like mercury, arsenic, chromium and manganese at levels exceeding drinking water limits. Testing of the tailings muds by SGS was reported by Samarco (http://www.samarco.com/wp-content/uploads/2015/11/26.11.15-Test-reports-show-that-tailings-from-Samarcos-Fund--o-dam-do-not-present-any-risk-to-people.pdf ) and they were found to be non hazardous, but no figures were provided. Contributions by our readers should be sent to our webmaster (link).

One important geochemical aspect of the damage is the loss of dissolved oxygen in the river system, as pinpointed by the Guardian          (http://www.theguardian.com/business/2015/nov/22/anger-rises-as-brazilian-mine-disaster-threatens-river-and-sea-with-toxic-mud ). This is likely to have at least as much impact as toxic element concentrations.


Damage remediation strategies

After the initial emergency phase, plans have to be drawn for site remediation. At the dam site, this is simple: rebuild the dam, or another one, but a better and safer one. In the damaged area, issues like solid waste and sludge removal and disposal, land reclamation and water protection become first priority and all require geochemical expertise. Expedited disposal of waste in emergency time may lead to further damage. Flushing downstream contaminated water or sludge may remediate the affected area and damage and spread the damage further.

At Fundão Mariana, downstream protection was not really an option, like it was in Aznalcollar for the Doñana National Park, as the flow reached quickly the sea. Remediation by downstream flushing is the quickest option, but it leaves a major part of the impact on the marine shore, as reported by Bruce Douglas in the Guardian (http://www.theguardian.com/business/2015/nov/22/anger-rises-as-brazilian-mine-disaster-threatens-river-and-sea-with-toxic-mud ). Tailings removal and site clean-up are not yet widely documented. Disposal is likely planned to be placed in a new tailings dam.


Monitoring damage and recovery

An essential task is to identify how far visual damage (flooding, mud) is accompanied by durable contamination of soil, water and sediment. These investigations have to be led in a human health risk perspective, but also in terms of broader environmental contamination, in order to manage long term effects. A sensitive aspect of contamination assessment is community involvement. Restricted communication may be desirable in some instances, but it may lead to community distrust and to further concerns, sound or not. Ultimately, it may shatter the social acceptability of the mining activities in the region.  

For the same reason, monitoring the effects of remediation and site recovery is not only needed but desirable, once again with public involvement and dissemination of results. Geochemical monitoring cannot be performed by public health experts alone: a thorough knowledge of geology, soil and water is required too. This is the mission of geochemists.

At Fundão Mariana, immediate as well as future ecological damage to the river milieu, sediments and living species is reported by researchers (http://www.rsc.org/chemistryworld/2015/11/brazil-mine-disaster-dam-collapse ). The short term effects are choking, then mud induration and permanent sterilisation. Long term effects include bioaccumulation of contaminants.

Monitoring river water quality, including dissolved oxygen, suspended matter, iron, pH, turbidity and conductivity, has to be performed on a regular basis at various places downstream the dam, in order to control milieu restoration (link to http://www.cprm.gov.br/publique/media/20151130_Deslocamento_Agua_com_Turbidez.png).

The occurrence of contaminants (heavy metals, process reagents) in river water, sediments and neighbouring soil, even at non-critical levels, implies the necessity of a geochemical monitoring program. Interpretation of contaminant data has to be more in depth than compliance to regulatory levels, to be useful for remediation management and for community risk assessment.



The implication of expert and dedicated geochemists is critical at all the stages of disaster management. It would also help to prevent future disasters if they are implied in facility design, operational management and monitoring.

[1] See Alonso, E. E., and A. Gens (2006), Geotechnique, 56(3), 165 – 183; Olias, M. et al. (2012), Environ Monit Assess 184:3629–3641; this hypothesis has yet to be further discussed.


-Joao Larrizzati & Bruno Lemiere
ukash haber seo seo seo google sıra bulucu google sıra bulucu kanun kanun kanun script encode decode script encode decode ukash