UNECE WORSHOP ON TDS, YEREVAN, November 2007 TAILING DAMS RISK ANALYSIS AND MANAGMENT Pavel Danihelka Eva Červeňanová

UNECE WORSHOP ON TDS, YEREVAN, November 2007 CONTENT: Examples of historical accidents Introduction to risk theory Risk analysis principles Basics of application of risk analysis to tailing dams safety Conclusion

UNECE WORSHOP ON TDS, YEREVAN, November 2007 EXAMPLES OF HISTORICAL ACCIDENTS At least 221 serious tailing dams accidents reported by UNEP*: * Mine name/ Location Incident Date Impact Baia Mare, Romania ,000 m3 cyanide contaminated water with some tailings released Baia Borsa, Romania ,000 t of tailings contaminated by heavy metals released Merriespruit, South Africa deaths, 500,000 m3 slurry flowed 2 km

UNECE WORSHOP ON TDS, YEREVAN, November 2007 Mine name/ Location Incident Date Impact Buffalo Creek, USA deaths, 500 homes destroyed Mufilira, Zambia deaths, 68,000 m3 into mine workings Omai, Guyana million m3 cyanide slurry released Placer, Philippines deaths, 50,000 m3 released Los Frailes, Spain released 4-5 million cubic meters of toxic tailings slurries Stava, Italy deaths, tailings flowed up to 8 km Aitik mine, Sweden million m3 water released Major tailing dams review – cont.

UNECE WORSHOP ON TDS, YEREVAN, November 2007 History of major tailing dams accidents Source: „ICOLD Bulletin 121“

UNECE WORSHOP ON TDS, YEREVAN, November 2007 BAIA MARE January 30, 2000 in Baia Mare (Romania) the biggest freshwater disaster in Central and Eastern Europe. Nearly 100,000 m3 of cyanide and heavy metal-contamined liquid spilled into the Lupus stream, reaching the Szamos, Tisza, and finally Danube rivers and killing hundreds of tones of fish and poisoning the drinking water of more than 2 million people in Hungary. Case study:

UNECE WORSHOP ON TDS, YEREVAN, November 2007 LOS FRAILES April 25, 1998 tailings dam failure of the Los Frailes lead-zinc mine at Aznalcóllar near Seville, Spain, released 4-5 million cubic meters of toxic tailings slurries and liquid into nearby Río Agrio, a tributary to Río Guadiamar. The slurry wave covered several thousand hectares of farmland, and it threatens the Doñana National Park, a UN World Heritage Area.

UNECE WORSHOP ON TDS, YEREVAN, November 2007 STAVA On July 19, 1985, a fluorite tailings dam of Prealpi Mineraia failed at Stava, Trento, Italy. 200,000 m3 of tailings flowed 4.2 km downstream at a speed of up to 90 km/h, killing 268 people and destroying 62 buildings. The total surface area affected was 43.5 hectares.

UNECE WORSHOP ON TDS, YEREVAN, November 2007 AITIK On September 8, 2000, the tailings dam of Boliden's Aitik copper mine near Gällivare in northern Sweden failed over a length of 120 meters. This resulted in the spill of 2.5 million cubic meters of liquid into an adjacent settling pond. Boliden subsequently released 1.5 million cubic meters of water from the settling pond into the environment to secure the stability of the settling pond.

UNECE WORSHOP ON TDS, YEREVAN, November 2007 VARIABILITY OF CAUSES OF ACCIDENT Inadequate management Lack of control of hydrological system Error in site selection and investigation Unsatisfactory foundation, lack of stability of downstream slope Seepage Overtoping Earthquake MAIN ROOT CAUSE: RISK ANALYSIS AND MANAGEMENT NEGLECTED

UNECE WORSHOP ON TDS, YEREVAN, November 2007 Distribution of causes of tailing dams accidents Source: ICOLD Bulletin 121

UNECE WORSHOP ON TDS, YEREVAN, November 2007 VARIABILITY OF CONSEQUENCES Flooding, wave of slurry Contamination of surface water, living organisms (biota), intoxication Drinking and irrigation water contamination (surface) Drinking and irrigation water (underground) contamination Soil contamination As consequence of 2),3),4)ad.5 : Food chain contamination » FREQUENTLY TRANSBOUNDARY EFFECT

UNECE WORSHOP ON TDS, YEREVAN, November 2007 Conclusion: Tailing dam is a risky installation able to cause major accident and so we have to treat it as major risk

UNECE WORSHOP ON TDS, YEREVAN, November INTRODUCTION TO RISK THEORY Definition of –Hazard –Risk Risk and its quantification (measurement) Principles of risk reduction/management

UNECE WORSHOP ON TDS, YEREVAN, November 2007 DEFINITION OF TERMS SOURCE OF DANGER = POTENTIAL TO CAUSE DAMAGE

UNECE WORSHOP ON TDS, YEREVAN, November 2007 RISK = PROBABILITY x GRAVITY OF ACCIDENT (EVENT)

UNECE WORSHOP ON TDS, YEREVAN, November 2007 RISK DANGEROUSITY IDENTICAL RISK DIFFERENT DIFFERENCE: MANAGEMENT OF RISK

UNECE WORSHOP ON TDS, YEREVAN, November 2007 INITIAL EVENT SYSTEM 1 SYSTEM 2 SYSTEM 3 FLUX OF DANGER Source system Target system Initial event Other conditions Flux of danger DOMINO EFFECT: CATASTROPHE Example: Stava accident

UNECE WORSHOP ON TDS, YEREVAN, November 2007 Flux of danger:  Movement of material  Flux of energy  Flux of information Targets system :  Population around tailings dam  Environment Surface water Underground water Soil Living organisms  Material and financial losses (direct)  Functioning of enterprise (including indirect losses)

UNECE WORSHOP ON TDS, YEREVAN, November 2007 Sources of danger : –Having potential (energy) to cause damage –Having potential to weaken structure, resistance, resilience of our system (tailing dam and its environment) Direct to dam stability Indirect including human error To consequences

UNECE WORSHOP ON TDS, YEREVAN, November 2007 QUANTIFICATION OF RISK RISK MATRIX A – banal case B – frequent accident with low consequences (minor injury, small contamination,...) C – disaster with high probability (walking in minefield) D – disaster with low probability (nuclear power plant, major incident) BC AD PROBABILITY GRAVITY

UNECE WORSHOP ON TDS, YEREVAN, November 2007 Acceptability of risk PROBABILITY GRAVITY NON ACCEPTABLE CONDITIONALLY ACCEPTABLE ACCEPTABLE ACTION NECESSARY ACTION VOLUNTARY RISK MITIGATION

UNECE WORSHOP ON TDS, YEREVAN, November 2007 ACCEPTABILITY OF RISK Decision is socio-politic, not scientific Decision should include all stakeholders All types of risk should be evaluation together

UNECE WORSHOP ON TDS, YEREVAN, November 2007 How to decrease risk?

UNECE WORSHOP ON TDS, YEREVAN, November 2007 RISK ANALYSIS PROCESS Scenarios proposal ETA FTA AMDEC FMEA HAZOP WHAT-IF Etc. Selection of sources of danger Risk assessment Goals setting Barriers of prevention Feedback and control Risk management Residual risk IMPACT PROBALITY TECHNICAL BARRIERS ORGANISATION BARIERS

UNECE WORSHOP ON TDS, YEREVAN, November 2007 SOURCES OF DANGER  Direct to dam stability: Active environment (rain, snow, freeze…) Earthquake Geological conditions Domino effect  Indirect to dam (including human error): Wrong conception Construction failure Material failure Bad maintenance Lack of control  To consequence: Water and sludge movement Mechanical contamination by solid particles Chemical toxicity / ecotoxicity Radioactivity

UNECE WORSHOP ON TDS, YEREVAN, November 2007 SCENARIO PROPOSAL All plausible scenario should be involved in preliminary conspiration All stages of life-time must be considered Those having minor impact omitted Similar combined to groups Described as combination of events in time Finally, we are able to compare limited number of scenarios only

UNECE WORSHOP ON TDS, YEREVAN, November 2007 TOOLS HELPING TO DEFINE SCENARIO Examples of past accidents Near-misses and accidents on site Control list WHAT-IF ETA FTA AMDEC FMEA HAZOP Etc.

UNECE WORSHOP ON TDS, YEREVAN, November 2007 Past accidents analysis In site – during all life of it In similar places you operate, including near-misses. Mind the necessity of reporting. In mine industry generally –TAILINGS DAMS, RISK OF DANGEROUS OCCURRENCES, Lessons learnt from practical experiences, ICOLD- UNEP 2001, Bulletin 121, ISSN –APELL for Mining: Guidance for the Mining Industry in Raising Awareness and Preparedness for Emergencies at Local Level, Technical report No. 41, UN Publications 2001, ISBN

UNECE WORSHOP ON TDS, YEREVAN, November 2007 SCENARIO DESCRIPTION (DAM DESTRUCTION) „TOP“ EVENT CAUSESCONSEQUENCES SCENARIO 1 SCENARIO 2 EACH SCENARIO NUMBERED

UNECE WORSHOP ON TDS, YEREVAN, November 2007 RISK ASSESMENT: FREQUENCY x CONSEQUENCES (IMPACT) FREQUENCY: From past accidents (high degree of uncertainty) From initial events frequency and FTA by boolean algebra Avoid omitting of low frequency events (not to limit only to 100-year water or earthquake) Human factor extremely important

UNECE WORSHOP ON TDS, YEREVAN, November 2007 Frequency of „100 year“ flooding

UNECE WORSHOP ON TDS, YEREVAN, November 2007 One mythus: „ We operate it long time without accident, so safety is prooved“

UNECE WORSHOP ON TDS, YEREVAN, November 2007 CONSEQUENCES: Consequences to human lives, health and well being. Evaluation of consequences with stakeholders necessary Direct costs (remediation, compensation,...) Social disturbance Consequence to environment – short time and long time impacts Economical consequences and operability Indirect costs

UNECE WORSHOP ON TDS, YEREVAN, November 2007 Costs of Failure Physical failure: recent large failures $30 to $100 million in direct costs Environmental failure: some recent clean-up liabilities to several $100’s of millions Closure liability: some recent examples in $ 500 milon to $ 4 billion range Industry/investor impacts: Shareholder value losses and industry imposed constraints and costs amounting to many billions of dollars

UNECE WORSHOP ON TDS, YEREVAN, November 2007 CONSEQUENCES II : The scales of consequences should be defined before analysis is done (4-6 grades) All possible targets should have the same scales of consequences (e.g. Grade X is comparable in all target systems) The most serious consequence is selected Internal values of society/enterprise become to be clarified

UNECE WORSHOP ON TDS, YEREVAN, November 2007 Severity of impact – an example (source: Robertson GeoConsultants Inc.)

UNECE WORSHOP ON TDS, YEREVAN, November 2007 RISK ASSESSMENT Following frequency and gravity, scenarios are put to the risk matrix PROBABILITY GRAVITY

UNECE WORSHOP ON TDS, YEREVAN, November 2007 GOALS SETTING: Non-axeptable (red zone) scenarios: immediate action Conditionally acceptable (yellow zone) scenatios: action envisaged PROBABILITY GRAVITY

UNECE WORSHOP ON TDS, YEREVAN, November 2007 BARIERS OF PREVENTION / PROTECTION Source system Target system Initial event Other conditions Flux of danger INITIAL EVENT SYSTEM 1 SYSTEM 2 SYSTEM 3 BARRIER CATASTROPHE BARRIER OF FLUX DOMINO EFFECT INITIAL EVENT SYSTEM 1 SYSTEM 2 SYSTEM 3 PROTECTION OF TARGET REMOTION OF SOURCE

UNECE WORSHOP ON TDS, YEREVAN, November 2007 SAFETY MANAGEMENT Prevention part (even three part of bow-tie diagram) Emergency preparedness

UNECE WORSHOP ON TDS, YEREVAN, November 2007 NEAR MISSES: „HUNTING FOR DEVIATIONS“ ELIMINATED CATASTROPHE BIG ACCIDENTS / LOSSES SMALL ACCIDENTS/ LOSSES DEVIATIONS

UNECE WORSHOP ON TDS, YEREVAN, November 2007 Emergency preparedness Preparedness to accident, even with low probability Training and not only desktop one Information of all potentially involved Crisis management including training Open and honest communication with municipalities, emergency response teams, government bodies (inspection…) Communication with media

UNECE WORSHOP ON TDS, YEREVAN, November 2007 RECOMMENDATIONS 1) Detailed site investigation by experienced geologists and geotechnical engineers to determine possible potential for failure, with in situ and laboratory testing to determine the properties of the foundation materials. 2) Application of state of the art procedures for design. 3) Expert construction supervision and inspection. 4) Laboratory testing for “as built” conditions. 5) Routine monitoring. 6) Safety evaluation for observed conditions including “as built” geometry, materials and shearing resistance. Observations and effects of piezometric conditions. 7) Dam break studies. 8) Contingency plans. 9) Periodic safety audits

UNECE WORSHOP ON TDS, YEREVAN, November 2007 And something for thinking…

UNECE WORSHOP ON TDS, YEREVAN, November 2007 DO WE REALLY NEED ACCIDENT PREVENTION? You've carefully thought out all the angles. You've done it a thousand times. It comes naturally to you. You know what you're doing, its what you've been trained to do your whole life. Nothing could possibly go wrong, right ?

UNECE WORSHOP ON TDS, YEREVAN, November 2007 THINK AGAIN!

UNECE WORSHOP ON TDS, YEREVAN, November 2007 THINK AGAIN!

UNECE WORSHOP ON TDS, YEREVAN, November 2007 Thank you for your attention !