Report of the International Task Force for Assessing the Baia Mare Accident
This Report constitutes the response of the International Task Force for Assessing the Baia Mare Accident (the Baia Mare Task Force or 'BMTF') to the tasks given to us by Commissioner Margot Wallström, with the support of the Environment Ministers of Hungary and Romania, concerning the accidents which occurred at Baia Mare and Baia Borsa in Romania, in January and March 2000. These tasks were to consider:
- What happened, and why?
- What damage was caused?
- What other dangerous sites remain in the Tisa river basin system?
- What measures are recommended to minimise the risk of similar accidents?
We have tried to write the Report in a clear and succinct manner, in the hope that it will be easily accessible to those people affected by the accidents; from the outset the Commissioner asked the BMTF to ensure communication and dialogue with the people and the non-governmental organisations (NGOs).
The BMTF has travelled extensively throughout the affected areas in the Tisa river basin (the pollution had been largely diluted by the time that it reached the Danube River), and visited each of the accident sites.
We held intensive consultations with private citizens, local and regional officials, NGOs, and other professional associations along the course of the river through Romania, Hungary and the Federal Republic of Yugoslavia, from Baia Borsa and Baia Mare to Belgrade. Each of the three governments made available to us the results of their official investigations into the accidents and their impacts.
We were fortunate at an early stage to have access to the very valuable rapid assessments carried out by the UNEP/OCHA team in February and March 2000, led by Frits Schlingemann. We also had access to the results of studies undertaken by German and Dutch agencies and the WWF, and we commissioned a number of independent studies and investigations to assist us in addressing specific regulatory, engineering and environmental issues.
These were disastrous accidents which under different circumstances could have had more serious consequences. In the event, nobody died or became seriously ill despite the fact that 120 tonnes of cyanide and 20,000 tonnes of sediments containing heavy metals were released into the environment from the accidents. The impact might have been far more serious if the rivers had not been covered by ice for up to 200 km downstream of Baia Mare. Or, had the most severe floods for well over 100 years not occurred days after the second accident, with the result that heavy metals were washed away and dispersed, bringing heavy metal levels within the river system and on the flood plain back to levels which do not pose an immediate threat to health.
This must not, however, be allowed to obscure the fact that in many other locations in this region, and elsewhere in Europe, there are similar sites, both in active operation and abandoned, which pose the threat of similar accidents - the next time we may not be so lucky.
In our view, these accidents should act as the trigger for an urgent and thorough investigation of such sites. This should involve the preparation of an inventory, based on an agreed risk assessment methodology, so that all may be aware of the hazards and the need to minimise the risks posed. This is a task, the primary responsibility for which rests squarely on the shoulders of the countries themselves.
So far as the area of the Danube river basin is concerned, the International Commission for the Protection of the Danube River (ICPDR) should be the pro-active co-ordinating agency; but we strongly recommend that the European Commission takes the initiative to ensure adequate co-operation between all parties concerned. Such an inventory would serve as the basis for a programme of restoration and remediation. In addition, the relevant UN-ECE Conventions and Protocols should be quickly ratified by all UN-ECE member countries and the European Union.
We make a series of recommendations for the strengthening of existing regulations including emergency planning requirements, and for the addition of new regulations. In particular, we recommend that no new TMFs, where cyanide is used, should be based on the storage of water/slurry containing cyanide in tailings ponds open to the elements.
We feel that there is a clear need for a central Industry Guidance Document, which sets out and clarifies the many different regulatory requirements relating to the mining, extractive and ore-processing industries which are contained in a wide range of EU and national legislation - for EU countries and those countries seeking to become members of the EU, this task should be undertaken by the European Commission. Such a document would make it easier for both the industry and the public to be aware of the obligations of the industry regarding environmental protection, public safety and emergency planning. Our recommendations concentrate on what should, in our view, be done and not on how or at what level it should be done. In this context, the recent European Commission proposals are consistent with our recommendations.
Finally, we draw attention to the recent adoption of the EU Water Framework Directive, which calls for day-to-day protection of the Europe Union's water resources to be carried out by river basin authorities. The monitoring of water quality and water quantity for water resources management will require that river basin authorities develop the necessary skills, resources and flexibility to undertake these tasks. In the case of the Danube river basin, the ICPDR will be expected to play a vital role in facilitating the co-ordination of monitoring and early warning arrangements, and will need greater funding from its member countries, and an enhanced capacity to take swift action where necessary.
A great number of people and organisations have assisted us with our inquiries. The European Commission provided both the initial impetus to establish the BMTF and also provided the resources without which we would have been unable to undertake our inquiries. The governments of Romania, Hungary, Slovakia, Ukraine and the Federal Republic of Yugoslavia provided us with their full support. Crucially, we received invaluable assistance from many local and municipal authorities and from non-governmental organisations. Without their input, our findings would be far less relevant. At an international level, the Dutch and Danish governments both provided the financial support for public communications initiatives proposed by the BMTF and we have also been greatly assisted by informal advice from the Irish and United States environmental protection agencies. On a technical level, we have received advice on regulatory and tailings pond management from a number of leading experts from Romania, Hungary and Germany. To all of these people, and the many others who have assisted us, we give our sincere thanks for helping us to complete our tasks.
The Baia Mare Task Force
Tom Garvey, Chairman
1. WHAT HAPPENED AND WHY
On the evening of 30 January 2000, a tailings pond burst at a facility near the city of Baia Mare, Romania which was reprocessing old mining tailings and re-depositing the waste sludge into a new tailings pond. This led to approximately 100,000 m 3 of waste water containing up to 120 tonnes of cyanide and heavy metals being released into the Lapus River, then travelling downstream into the Somes and Tisa rivers into Hungary before entering the Danube.
On 10 March 2000, another tailings dam burst in Baia Borsa in the same region close to the Ukrainian border. While some of this material was retained within the dam complex, 20,000 tonnes of sediments were then released into the Novat River, a tributary of the Viseu and Tisa rivers.
This section attempts to provide a brief account of what happened, and why, at the two sites concerned. While the BMTF was initially established to investigate the Baia Mare accident, its remit was subsequently enlarged to cover the Baia Borsa accident as well.
The Danube river basin and the accident sites
1.2 BAIA MARE - WHAT HAPPENED?
The first accident occurred near the city of Baia Mare at a tailings pond operated by AURUL, a joint-venture between Esmeralda Exploration of Australia and REMIN the Romanian state-owned mining company. AURUL had been established to re-process the tailings of an old abandoned tailings pond, extracting gold and silver and other metals and, at the same time, removing the tailings from their existing site close to a residential area to a more remote site 8km from the city of Baia Mare.
The facility had been designed as a 'zero discharge' process, with no emissions of process water to local rivers. The dust blown from old tailings ponds contains traces of heavy metals which can cause adverse health effects, so removal of the material provided an environmental health benefit to the town, while also providing an economic benefit to the town in terms of employment by AURUL. Consequently, the facility appeared to offer the prospect of a 'win-win' situation, with both environmental and economic benefits. In line with the design of the process, the tailings pond received process water containing high levels of cyanide, which is used in the gold extraction process. The process water was then recycled to the plant, thus reducing the amount of new cyanide that had to be added into the process and reducing the operating costs of the facility.
Schematic diagram of the AURUL process
The project began in 1992, and after a lengthy permitting process, AURUL started operations in May 1999. After seven months of operation, a dramatic failure of the retaining embankment wall led to the release of tailings water into local rivers.
The sequence of events leading to the accident was as follows:
- Throughout 1999, the tailings dam was developing as intended, with the hydrocyclones1 building the embankment walls higher as re-mined tailings and waste process water were pumped into the tailings pond.
- Over the course of the winter of 1999, a significant amount of rain and snow fell on the tailings pond, and the pond became covered in a thick layer of snow and ice.
- Heavy (but not exceptional) rain and snowfall in December 1999 and January 2000, combined with rapid snowmelt from 27 January 2000 as the temperature rose suddenly from below freezing to 9.5ºC, and nearly 40mm of rainfall on 30 January 2000 caused water levels to reach critical levels. The embankment walls became saturated and unstable as the snow melted directly on their surfaces.
- On 30 January, the dam overflowed and washed away a stretch of embankment wall 25 metres long and 2.5 metres deep. Approximately, 100,000m 3 of tailings water containing cyanide began to flow into the nearby Lapus River. AURUL stopped processing operations and began activities to close the breach.
1.3 BAIA BORSA - WHAT HAPPENED?
The second accident occurred at the Novat tailings management facility (TMF) at Baia Borsa, belonging to the state-owned mining company REMIN SA. In addition to the primary dam, a second and third dam had been constructed downstream from the primary dam. The second dam was designed to support the primary dam structure as it reaches its final height. The third dam was built of concrete and was designed to collect water that leaked from the first two dams for re-pumping to the main pond.
The dam overflowed and burst on 10 March 2000. 100,000 m 3 of water and 20,000 tonnes of tailings sludge containing heavy metals flowed out of the dam. While some material was retained between the two lower dams, the rest flowed downstream of the dam and into the Novat and Vasar rivers, from where it entered the Viseu and finally the Tisa river. Due to the downstream area from the dam being a specially protected nature area, no emergency discharge of water is permitted from the dam. As a result, the dam system is a closed-cycle design with no discharge downstream of the third dam.
The process was designed as a partially closed-circuit system, with process waters discharged into the pond being recycled through the metal extraction facility. In addition, the pond also receives direct rainfall and indirect surface run-off from the surrounding hill-slopes. Evaporation and the recirculating of water back to the main mine complex was supposed to ensure that water levels in the pond were controlled, with the main stream in the valley bypassing the pond. The pond complex, however, contained no provisions for emergency discharges downstream of the pond. As a result, if water levels reached critical levels, there was no way that water levels could be reduced.
Over the winter of 1999-2000, heavy rainfall and rapid snowmelt caused the water level in the pond to reach critical levels and for the dam walls to become unstable. As the pumps designed to pump water out of the pond, back to the processing plant and into other ponds were not working, it was impossible to avoid the overflow and breaching of the dam wall.
The sequence of events leading to the accident was as follows:
- In early December, heavy rain fell into the reservoir and surrounding catchment;
- From mid-December 1999 to late January 2000, more than 120mm of precipitation as snow fell on the reservoir and surrounding catchment;
- Temperatures fell below 0°C from 21 December 1999 and were below -10°C from 22 January 2000;
- On 27 January 2000, the temperature began to increase rapidly, rising above 0°C on 30 January;
- On 8-10 March 2000, torrential rainfall and rapid snowmelt led to increasing water levels in the reservoir, and the pumping systems for the pond failed;
- On 10 March at 11-00am a breach 25m wide and 10m high occurred in the dam, leading to the spillage of 20,000 tonnes of tailings into the Novat river;
- The tailings overflowed the second and third dams to enter into the river system, flowing along the Ukrainian border, before flowing into Hungary.
1.4 WHY DID THE ACCIDENTS HAPPEN?
The reasons why the accidents happened are both clear, and a matter of general agreement and are discussed below.
1.4.1 The use of closed circuit systems
Both the AURUL and Novat tailings management facilities (TMFs) were based on closed circuit systems designed specifically to avoid the need to discharge effluents into local rivers and streams. All the process waters (including in the Baia Mare case waters with high concentrations of cyanide), were re-circulated back to the processing plant for re-use. While in principle this was a worthy objective, both in environmental and economic terms, the design in each case contained no provision for the emergency discharge of excess waters when overflow threatened. Without specific provision for avoiding overflows, such 'zero-discharge' systems are not suitable, in our view, for use in meteorological conditions of heavy and intense precipitation, such as those prevailing in this part of Romania, and should never have been condoned by the permitting authorities. The result was that, in each case, during severe weather conditions the additional volume of rain and melting snow could not be contained within the ponds, causing them to overflow.
Furthermore there was a problem in the case of Baia Mare with the stability of the embankment walls themselves. This arose because the Baia Mare facility used a recognised technique of embankment or dam wall construction (called 'construction by operation') which called for the gradual deposition of tailings of sufficiently coarse grade on the starter walls to ensure stable and continuous growth of the height of the embankment walls.
However, the mix of tailings used did not have the ratio of coarse to fine grades stipulated in the design and, in addition, the hydrocyclones used to distribute the tailings within the pond could not operate in the very low temperatures experienced before the accident. As a result, the embankment wall construction was interrupted at a critical time, leading to a reduction in the 'freeboard', and consequently to wall breaching and overflow.
In the case of Baia Borsa, the TMF was designed for a capacity of 2 million m³ of slurry, whereas only 400,000 m³ of slurry was actually deposited. As a result, the material available for dam construction was less than anticipated, leading to a lower dam wall then specified.
The monitoring of the water level of both ponds was by simple visual monitoring, difficult during times when the ponds are covered by an ice and snow cap. For Baia Mare, this form of monitoring was in conformity with the risk classification allotted to the pond by the operator, and accepted by the permitting authorities. In the case of Baia Borsa, while the monitoring procedures gave advance notice of the impending danger, it was not possible to reduce the level of the pond, and avert the accident, due to defective pumping equipment. In these circumstances it should have been clear that a combination of melting snow and ice cap, due to rapidly rising temperatures, accompanied by heavy and continuous precipitation could have had only one result. And this is exactly what happened during the period leading up to the accident, when heavy rain was accompanied by a rapid rise in temperature from -13° C to +9°C, thus adding the snow/ice melt to the on-going rainfall. Our investigations have established that such conditions while severe, were by no means unprecedented, and therefore should have been foreseen.
1.4.4 Regulatory oversight
Apart from the design and operational factors, we have grave reservations over the adequacy of the permitting procedures existing when the projects were planned and constructed. It is apparent that no adequate appreciation of the likely variability in 'water balance' (the difference between the volume of water entering the system, and the volume of water exiting the system by evaporation) was brought to bear. The initial Environmental Impact Study carried out by ICIM2 on the AURUL project, stated that 'the danger of the dam overflowing the embankment...in the event of heavy rainfalls is out of the question'.
From then on, after the entry into force of the Romanian law on Environmental Protection (No.137/95), and during the six years it took to obtain all of the necessary permits, there was a failure to perceive the inherent dangers of approving such closed circuit systems, with no provision for emergency discharge. It was difficult for the BMTF, during its meetings with all of the authorities involved in the permitting process, to discern where exactly the responsibility for the overall safety of the facility lay. The regulatory process is extremely complex and diffused and, in the view of the BMTF, needs urgent revision. The Romanian authorities have informed us that significant progress has already been made in this revision.
As a result, it is the conclusion of the BMTF that the accidents were caused:
- Firstly, by the use of an inappropriate design of the TMF;
- Secondly, by the acceptance of that design by the permitting authorities; and
- Thirdly, by inadequate monitoring and dam construction, operation and maintenance.
These were causal factors, which were triggered by the severe weather conditions, conditions which could and should have been foreseen. In the case of the Novat pond we fail to understand how this facility could have been brought into operation at all, as the local Environmental Protection Agency did not accept the Environment Impact Assessment submitted by the operators REMIN, and refused to issue an operating permit.
In summary, therefore, the reasons for the two accidents are as follows:
|Baia Mare||Baia Borsa|
In effect, these were two accidents waiting to happen, waiting for the necessary trigger of adverse weather conditions which was bound to come sooner or later.
2. IMPACTS OF THE ACCIDENTS
The accident at Baia Mare led to the release of a toxic 'plume' of a cyanide/copper mixture into the Tisa river. This plume travelled down the length of the river, into the Danube river and then into the Black Sea by which time it had become significantly diluted. It devastated large number of plant and wildlife species in the river systems, but once the plume had passed, water quality and sediment quality started to return to normal. As the plume consisted of dissolved materials, there was little deposition of pollutants. As a result, once the plume had passed, effectively no cyanide remained in the river or river banks.
The course of the plume from Baia Mare
Source: Apele Romane
The accident at Baia Borsa was very different in character. A mass of mud and water containing heavy metals was released into the Novat stream. Most of the mud has remained close to the tailings pond, with the polluted water being washed downstream into the Viseu and Tisa rivers. The majority of the heavy metals within the water were then deposited in the upper reaches of the Tisa river. Almost immediately after the accidents, major flooding occurred in the Tisa river. This had the effect of re-mobilising and then dispersing the heavy metals.
The course of the plume from Baia Borsa
Source: Apele Romane
2.2 PUBLIC HEALTH IMPACTS
2.2.1 Short-term impacts
The most remarkable feature of the accidents was the fact that, as far as we are aware, no one was killed or became seriously ill as a result of the accidents.
Swift and concerted action by municipalities and water supply companies along the Tisa and Danube rivers ensured that none of the pollution entered into the public drinking water supply. High groundwater levels also helped to minimise the contamination of groundwaters by the accidents. Bottles and bags of drinking water were provided to local residents while the mains water supply was halted as the pollution plumes passed. There was local contamination of wells in Bozanta Mare, close to the accident site in Baia Mare, and these people were subsequently provided with bottled drinking water.
2.2.2 Long-term impacts
The long-term health impacts of the accidents are uncertain at this stage. They are unlikely to relate to cyanide, which has dispersed and does not bioaccumulate. Any impacts would relate to the heavy metals added into the sediments and soils of the Tisa ecosystem. All the evidence to date suggests that, as a result of the accidents, heavy metal levels have not been elevated significantly in the river system, except for immediately downstream of the Novat dam. This situation needs to be kept under constant surveillance. As a result, it is important that some means is found to control the heavy metal sediments deposited immediately downstream of the dam, to ensure that they do not contribute, in the future, to increased heavy metal levels further downstream.
2.3 ENVIRONMENTAL IMPACTS
The data available to us on the environmental impacts of the accidents represent a preliminary assessment, and are the result of investigations conducted by VITUKI and evidence collected by WWF from a variety of sources. Investigations are ongoing in Hungary in the areas where the most significant damage occurred. Many experts take the view that one annual reproductive cycle may be an insufficient basis for reaching final conclusions as to the long-term impacts. We welcome this on-going monitoring of the ecological development and recovery of the river system.
2.3.1 Short-term impacts
The spill at Baia Mare led to immediate, and very severe, effects on plants and wildlife in the Tisa river. Due to the non-persistent nature of cyanide and the relatively natural condition of the upper Tisa river system, once the pollution plume had passed, the river ecosystem was able to begin the process of recovery.
Plankton in the Somes and upper Tisa river in Hungary (closest to the accident) was completely killed, and in the middle and lower reaches of the Tisa between 30-60% of plankton were killed by the passage of the pollution plume. Within days of the spill materials passing, phytoplankton and zooplankton had begun to recover throughout the entire river system.
Current evidence indicates that the number and species composition for phytoplankton has substantially returned to normal over all the stretches of the river. A major factor which appears to have assisted the process of recovery, is the flooding which occurred in March, and presented ideal conditions for plankton growth.
Molluscs and benthic organisms
The current available data indicates that considerable mortality occurred among molluscs and other benthic organisms in the upper reaches of the Somes. It is probable, that some species may have already been reduced in numbers in this region, through pollution that has occurred over many years. In the middle and lower stretches of the Tisa river (Hungary and the Federal Republic of Yugoslavia) the benthic organisms appeared to survive the cyanide spill although there is evidence that many species populations were reduced as a result of the pollution.
Of great symbolic significance as an indicator of ecosystem conditions has been the survival of the 'Tisa Flower' (Palingenia longicauda), a mayfly which hatched in large numbers this past summer throughout much of the Tisa River system. The conditions for hatching were apparently ideal following the spring floods. The fact that the cyanide appears to have been concentrated in the main channel of the river meant that the larvae survived and were able to hatch. Monitoring over the next 2-3 years will be required to confirm the recovery of the Tisa Flower, to reflect its breeding cycle.
In general it appears that the populations of macroinvertebrates were damaged from spill but not completely eliminated as originally feared. Recovery of the diversity of species and numbers has progressed rapidly during the first growing season following the spill.
The visible death of fish was a clear indication of the immediate impact of the accident at Baia Mare. Hungarian authorities report a total of 1240 tonnes of fish that were killed as a result of the spill. Of this amount 33.8% were predatory fish, 13.5% Carp, 8.1% Sturgeon, and 44.6% herbivorous fish. The fish collected included nearly all the fish known to be present in the river. It is clear, however, that not all fish were affected equally. The herbivorous fish, the Silver carp (Hypophthalichthys molitrix), in particular seemed to be very vulnerable to the cyanide and made up a large percentage of the fish that died.
Of significance is the possibility that some native, protected and endangered species may have been finally eliminated by the spill. In this category are Huso huso (Danube Salmon) and Acipenser güldenstaaedti (a highly protected Sturgeon species) which before the spill may have survived in the Tisa, although only in very small numbers. In addition, three other species (Stizostedion lucoperca, Stizostedion volgense and Lota lota) were almost completely eliminated.
Investigations in the Federal Republic of Yugoslavia indicate that the general fish population does not seem to have been significantly adversely affected by the spill. Some dead and dying fish were recovered in the Federal Republic ofYugoslavia, but four species of fish examined in detail in a study prepared for WWF showed no major population alterations, with the exception of Zander.
There are no reports or evidence that fish in the Danube died as a result of the cyanide poisoning.
One of the important questions that emerged following the accident was why were not more dead and dying fish observed near Baia Mare and on the Somes river system in Romania. The explanation appears to lie in the fact that the ice cover prevented dead and dying fish from being seen and these were then washed downstream to Hungary where they were collected. In addition it seems that previous pollution of this section of the river had reduced the numbers and diversity of fish there. The ice cover also presented conditions ideal for compounding the effects of cyanide because in the ice covered stretches, the oxygen of the river would have been reduced. This deadly combination would have been most significant in the upper Tisa river section of Hungary.
Birds & Mammals
Immediate observations indicate that the effects of the pollution on birds and mammals were limited. The ability of mammals and birds to sense the presence of the cyanide and the ice cover over much of the river likely prevented large contact with the pollution.
There was direct evidences that two White Tailed Eagles were poisoned by the cyanide. The longer term effects on birds however would likely be more significant in relation to reduced breeding success resulting from loss of food supply. The initial evidence from Hungary and Romania, however, is that no detectable signs of population loss can be determined to date amongst species considered most sensitive.
Studies of species presumed to be sensitive to the loss of their food base through the spill including Cormorants, Storks, White-tailed Eagle, Bank Swallows, and Kingfishers in both Hungary and Yugoslavia found little evidence that major population loss has occurred. These species have found sufficient food from alternative sources - fish ponds not affected by the spill.
At the top of the food chain of a river system are otters. Evidence indicates that the population of otters was reduced significantly in the period of the spill either through death or moving to other areas and that in the months following the spill the situation has improved. Recently in some places under Tokaj the otters have already returned to their regular habitats. The process is hopefully going on toward the upper stretches of the rivers, but it is to early to summarise the effects on the population.
Studies of bats also found that despite the presumed reduction of numbers because of reduced food supply the population may have increased this past summer. One explanation seems to be that the reduced pressure of fish on the insect population may have made more food available for the bats.
2.3.2 Medium to long-term impacts
The spill at Baia Borsa posed the greatest risk of medium-term impacts. Heavy metals persist in the environment and 'bioaccumulate' in living organisms. As a result, even relatively low concentrations can pose a threat to ecosystems and human health over the medium to long-term. Immediately after the heavy metal plume had passed through the river systems, heavy metal concentrations were above limit values in the river sediments. This situation was generally improved several weeks later, when severe flooding re-mobilised and dispersing heavy metals, bringing concentrations to below limit values. Nevertheless, this still represents an incremental increase in heavy metal levels, no matter how difficult to quantify. If heavy metal levels are assumed to be increasing over time, then any additional input of heavy metals over the 'base load' must be seen as a serious issue.
It appears that the majority of the heavy metals contained within the sediments from Baia Borsa have remained within 6-10km downstream of the Novat dam. Over time, they can be expected to migrate slowly downstream during flood events and become dispersed throughout the river system. This residual threat of heavy metal contamination needs to be the subject of ongoing monitoring and research.
1 Hydrocyclones are devices used to separate fine sediments and water from coarse-grained sediments. The fines and water are deposited in the centre of the tailings pond while the coarser sediments are deposited on the dam wall, building up the embankment walls.
2 ICIM: National Research and Development Institute for Environmental Protection.