Regulation explainerMetals & Mining

Tailings dam monitoring under GISTM: instruments, roles and trigger action response plans

A tailings facility is monitored to confirm it is behaving as designed and to act early when it is not. The Global Industry Standard on Tailings Management sets expectations for accountability, surveillance design and how results reach the Engineer of Record. This page explains those expectations, the instruments involved, how alert levels drive a trigger action response plan, and where analytics help without replacing engineering judgement.

Reviewed 8 min read

On this page
  1. Why tailings facilities need continuous surveillance
  2. Standards and laws that shape tailings surveillance
  3. Accountable roles for a tailings facility under GISTM
  4. Instruments in a tailings surveillance programme
  5. How one alert moves through a trigger action response plan
  6. Where analytics help, and where they defer to the Engineer of Record
  7. No model certifies a tailings dam as safe
  8. Data architecture for remote tailings surveillance
  9. A hypothetical piezometer trend crossing an alert level
  10. Questions and answers
  11. Sources

Why tailings facilities need continuous surveillance

Tailings facilities hold fine-grained processing waste, often with water, behind embankments that may be raised many times over a mine's life. Failure modes such as internal erosion, slope instability, overtopping and liquefaction can develop slowly and then accelerate, so surveillance is continuous: it compares what the facility is doing with what its design says it should do, while there is still time to act.

Scrutiny scales with the consequence of failure. GISTM classifies facilities from low to extreme consequence, based on potential harm to people, the environment and infrastructure downstream, and the class sets how demanding design and review must be1.

Standards and laws that shape tailings surveillance

GISTM is the global reference point, but national law, permits and dam safety guidance set the binding requirements in most places.

Global Industry Standard on Tailings Management (GISTM)

Global industry standard

Applies whenAn operator commits to it, as ICMM members did when it was published in August 2020, or a lender, insurer or customer requires it1.

  • Six topic areas, fifteen principles and seventy-seven auditable requirements covering affected people, site knowledge, design and monitoring, governance, emergency preparedness and disclosure1.
  • A monitoring principle requiring surveillance systems that manage risk at every phase of the facility lifecycle, including closure.
  • Named accountabilities: an Accountable Executive, a Responsible Tailings Facility Engineer, an Engineer of Record and independent review.

ICMM members' conformance commitment[^1]

ICMM member companies

Applies whenThe operator is a member of the International Council on Mining and Metals.

  • Conformance for extreme and very high consequence facilities by 5 August 2023, and for all other facilities not in a state of safe closure by 5 August 20251.
  • Public disclosure of progress; ICMM's 2025 report acknowledges that full conformance is taking longer than planned2.

Guidelines on Tailings Dams, Revision 1 (ANCOLD)[^3]

Australia, professional guidance

Applies whenAustralian practice is the design and operating reference, alongside state dam safety and mining law.

  • Guidance on consequence categories, design criteria, surveillance and closure of tailings dams3.

Dam Safety Guidelines and mining dams guidance (Canadian Dam Association)[^4]

Canada, professional guidance

Applies whenOwners or provincial regulators use CDA guidance as the basis for dam classification, safety reviews and surveillance.

  • Dam classification, periodic dam safety reviews, surveillance and emergency preparedness, with guidance on applying them to mining dams4.

National Dam Safety Policy (Lei 12.334/2010, as amended by Lei 14.066/2020)[^5]

Brazil

Applies whenAny mining dam in Brazil.

  • Prohibits building or raising mining dams by the upstream method5.
  • Requires an Emergency Action Plan (Plano de Ação de Emergência) for dams that store mining tailings5.

Accountable roles for a tailings facility under GISTM

Accountable Executive
A senior executive, answerable to the chief executive, accountable for the safety of tailings facilities and for limiting the consequences of a failure.
Responsible Tailings Facility Engineer (RTFE)
The operator's engineer responsible day to day for the facility's integrity, who runs surveillance and works directly with the Engineer of Record.
Engineer of Record (EOR)
The qualified engineer or firm that confirms the facility is designed, built and performing in line with its design intent, and that sets or reviews thresholds.
Independent Tailings Review Board (ITRB)
Senior independent experts who review the facility across its lifecycle and report to the Accountable Executive; for some facilities a senior independent technical reviewer fills the role.
Trigger Action Response Plan (TARP)
A pre-agreed table linking each monitored parameter's alert levels to specific actions, owners and response times.

Instruments in a tailings surveillance programme

InstrumentWhat it measuresWhat it can revealLimits to plan for
Vibrating-wire piezometersPore-water pressure in the embankment and foundationA rising phreatic surface, seepage paths, undrained loadingPoint readings; sensors drift or fail over the facility's life
Inclinometers and shape arraysLateral movement below the surface, by depthDeveloping shear surfaces in the embankment or foundationManual reads are infrequent; casings can shear off
Survey prisms and GNSS monumentsSurface displacement and settlementCrest deformation, bulging, differential settlementLine of sight and weather affect robotic total stations
Seepage weirs and flow metersVolume and turbidity of seepageInternal erosion when flow rises or carries finesNeeds rainfall and pond level as context
Satellite InSARMillimetre-scale ground movement over wide areasSlow deformation across the whole facility and its abutmentsRevisits take days; vegetation and steep slopes reduce coverage
Drones and visual inspectionImagery, photogrammetry and thermal conditionCracks, sinkholes, wet areas, beach width and freeboardWeather and flight approvals; needs trained review
Pond level and rain gaugesWater balance inputs and freeboardOvertopping risk, and context for every other readingA single gauge can fail silently without redundancy

Instrument choice follows the facility's credible failure modes and the performance objectives agreed with the Engineer of Record.

How one alert moves through a trigger action response plan

Scheduled readingsThreshold or trend alertVerify and re-readConfirmed data, contextAssessment, actionsEscalate if requiredActivate response plan01Sensor andlogger02Monitoringplatform03RTFE04Engineer ofRecord05AccountableExecutive06Site emergencyteam
  1. Sensor and logger

    Takes readings on schedule and buffers them through communication outages.

  2. Monitoring platform

    Checks data quality, evaluates thresholds and trend rules, and logs every alert.

  3. RTFE

    Verifies the reading, inspects the facility and owns the first response.

  4. Engineer of Record

    Interprets the change against design intent and advises on actions.

  5. Accountable Executive

    Authorises higher-level responses and resources.

  6. Site emergency team

    Runs the emergency response plan if escalation requires it.

  1. Sensor and logger to Monitoring platformScheduled readings
  2. Monitoring platform to RTFEThreshold or trend alert
  3. RTFE to Sensor and loggerVerify and re-read
  4. RTFE to Engineer of RecordConfirmed data, context
  5. Engineer of Record to RTFEAssessment, actions
  6. RTFE to Accountable ExecutiveEscalate if required
  7. Accountable Executive to Site emergency teamActivate response plan
Conceptual escalation sequence for a single alert; each real TARP sets its own levels, owners and response times with the Engineer of Record.

Where analytics help, and where they defer to the Engineer of Record

Statistical and machine-learning methods earn their place on the data side of surveillance; what a trend means for dam safety stays an engineering judgement.

Bad readings trigger false alarms or mask real ones

Early signalFlat-lining, step changes or physically impossible values from one sensor.

MitigationAutomate range, rate-of-change and neighbour-comparison checks, and flag suspect data before thresholds are evaluated.

Slow trends slip between reviews

Early signalEvery reading sits below the alert level while the rate of change keeps climbing.

MitigationRun rate-of-change and trend alerts beside absolute thresholds, using rules agreed with the EOR.

InSAR and in-ground data reviewed in silos

Early signalSurface movement arrives as a periodic report while piezometers are read daily, with no shared view.

MitigationAlign both on one time axis per embankment section, so a movement anomaly is checked against pore pressure immediately.

A model score mistaken for a safety verdict

Early signalA dashboard shows a failure probability with no engineering basis behind it.

MitigationPresent analytics as anomalies for review, and leave the condition assessment and TARP level to the RTFE and EOR.

No model certifies a tailings dam as safe

Data architecture for remote tailings surveillance

0 of 6 checked

A hypothetical piezometer trend crossing an alert level

Questions and answers

Is GISTM mandatory for tailings facilities?

Not as law. GISTM is an industry standard: ICMM members committed to conform, and other operators adopt it voluntarily or because lenders, insurers, investors or customers ask for it. Binding obligations come from national and state law, permits and regulator-endorsed guidance, which differ by jurisdiction, so many operators map GISTM against those rules.

How often should tailings monitoring data be reviewed?

There is no single frequency. Automated instruments may report many times a day, while manual readings and inspections follow the surveillance plan. GISTM expects monitoring data to be analysed at the frequency the Engineer of Record recommends, with deviations from expected performance reported to the EOR promptly. Review frequency usually rises after heavy rain, earthquakes, embankment raises or a new trend.

Can AI predict a tailings dam failure?

Not reliably, and claims that it can should be treated with suspicion. Failures are rare and every facility is different, so there is no credible dataset for training a failure predictor. Analytics can detect data faults, anomalies and accelerating trends sooner, and can combine InSAR with in-ground instruments. Those signals feed engineering review under the TARP; the judgement about safety stays with the EOR and the operator.

Does satellite InSAR replace in-ground instruments?

No. InSAR measures surface movement across the whole facility and its abutments, which helps find areas nobody instrumented. It cannot measure pore pressure, and its revisit interval and coverage gaps mean it complements piezometers, inclinometers and survey prisms rather than replacing them.

Sources

  1. Global Industry Standard on Tailings Management: structure, roles and member commitments — International Council on Mining and Metals · checked 10 October 2026
  2. Tailings Progress Report 2025 — International Council on Mining and Metals · checked 10 October 2026
  3. Guidelines on Tailings Dams: Planning, Design, Construction, Operation and Closure, Revision 1 — Australian National Committee on Large Dams · checked 10 October 2026
  4. Dam safety guidelines and mining dams guidance — Canadian Dam Association · checked 10 October 2026
  5. Lei nº 14.066, de 30 de setembro de 2020 (amends the National Dam Safety Policy, Lei nº 12.334/2010) — Presidência da República, Brazil · checked 10 October 2026

More in Metals & Mining

Back to Metals & Mining

Next step

Share your tailings instrument register and alert levels

Send the instrument list, current alert levels and how readings reach your RTFE and EOR. We will reply with where data quality, trend detection or communications redundancy could be strengthened, without changing who makes the engineering calls.

Discuss tailings monitoring