IAEA Incident & Trafficking DB: Intelligence Source Guide

The IAEA Incident and Trafficking Database is the authoritative record of nuclear and radioactive material out of regulatory control – thefts, losses, discoveries and seizures reported by states since 1993. The detailed records are closed; the annual factsheet is what you get.

At a glance

Source	IAEA Incident & Trafficking DB
Category	Conflict, Crime & Human Security › Military, Weapons & CBRN
Homepage	https://www.iaea.org/resources/databases/itdb
Machine interface	`https://www.iaea.org/`
Format	`HTML`
Access	Restricted — eligibility-gated (member states, institutions or vetted users) The platform catalogue records this as open. That is wrong, and the correction is explained under Access, licensing and what you may do with it below.
Disciplines	CBRN Intelligence
Mission domains	WMD / Proliferation

Nuclear/radioactive material trafficking incidents. — as catalogued in the platform’s own source registry.

The ITDB is the IAEA's information system on incidents involving nuclear material and other radioactive material that has fallen outside regulatory control. It was established in 1995 and holds records reaching back to 1993. Participation is voluntary: states that join designate an official point of contact who reports incidents to the Agency, and the Agency confirms each report with the reporting state before it enters the database. Records are classified into three groups. Group I covers incidents that are connected, or likely connected, with trafficking or malicious use. Group II covers incidents where the intent could not be determined. Group III covers incidents with no connection to trafficking or malicious use – the theft of a radiography source from a vehicle, the loss of a gauge in transport, the discovery of an orphan source in a scrap yard, unauthorised disposal. Group III is by a wide margin the largest of the three, and understanding that is the beginning of using this source properly. The Agency publishes an annual factsheet summarising the cumulative and yearly totals with a breakdown by group and by material type, and uses the database internally to support member states, guide nuclear security assistance and inform the Nuclear Security Series. The detailed incident records are not public.

For CBRNINT work the ITDB does one thing no other source does: it converts scattered national reporting on nuclear and radiological security failures into a single, definitionally consistent, longitudinal series that a state's regulator has actually confirmed. Individual incidents surface in press reporting, in national regulator bulletins and in customs announcements, but those are inconsistent in definition, incomparable across borders and heavily biased by what makes news. The ITDB is the corrective. Its analytical value is in the shape of the aggregate rather than in any individual entry: the persistent dominance of Group III incidents tells you that the dominant nuclear security problem worldwide is regulatory and custodial rather than adversarial; the small and relatively stable count of incidents involving weapons-usable material tells you something important about the actual state of the illicit market; and the composition of material types tells you which industrial sectors are leaking. That is the empirical foundation for radiological threat assessment, and analysts who reason about dirty bomb risk without it are reasoning from anecdote.

Who publishes it, and why that matters

The IAEA is an autonomous intergovernmental organisation reporting to the UN General Assembly and Security Council, funded by member state contributions. The ITDB sits within its nuclear security function, which is legally distinct from its safeguards function – safeguards verify that declared nuclear material is not diverted to weapons programmes under legally binding agreements, while nuclear security is an assistance and coordination role with no verification authority. That distinction determines everything about the database. The IAEA cannot compel a state to report an incident, cannot investigate one, and cannot publish an incident a state has not confirmed. The Agency's incentive is to maintain the participation of as many states as possible, which means confidentiality is not an administrative preference but the condition on which the database exists at all. A state that believed its reports would be published in identifiable form would stop reporting. The consequence for the analyst is a source of high internal integrity and deliberately low public resolution, and no amount of pressure will change that. Participation has grown steadily over three decades, which is the main reason apparent incident counts rise over time.

Provenance is the first question to ask of any dataset and the one most often skipped. Who collects it, what their incentive is, whether they publish a methodology, and whether they correct the record when they get something wrong all bear directly on how much weight a finding drawn from it can carry.

What a record actually contains

The fields you will be working with, what each one means, and whether it is something you can pivot on. Read the meanings carefully — more analysis is wrecked by misreading a field than by failing to find one, and a field that looks like an observation is often an inference.

Field	Type	What it means	Pivot value
`incident_group`	enum	Group I for incidents connected or likely connected with trafficking or malicious use, Group II for undetermined intent, Group III for incidents unconnected with trafficking – theft, loss, discovery, unauthorised disposal. This is the single most important field and the one most often ignored in secondary citation.	Threat interpretation; Group III totals belong in a regulatory discussion and Group I totals in a security one.
`incident_date`	timestamp	When the incident occurred or was detected, which are not the same thing – an orphan source discovered in a scrap consignment may have been out of control for years, and the recorded date reflects detection.	Temporal trend analysis; the detection date bounds rather than fixes the loss.
`reporting_state`	string	The participating state that confirmed the incident to the Agency. This is where the incident was reported from, which for a seizure at a border is the detecting state and not necessarily the state of origin.	National regulatory context; the authority that would hold the underlying case file.
`material_type`	enum	Broad classification of what was involved – nuclear material such as uranium or plutonium, other radioactive material such as sealed sources, radioactively contaminated material, or a combination. Weapons-usable material is a small minority of records.	Isotope and application; which industrial or medical sector the material came from.
`isotope`	string	The specific radionuclide where recorded. In practice the recurring names are the industrial and medical workhorses – caesium-137, cobalt-60, iridium-192, americium-241, strontium-90 – because those are the sources in widest circulation and therefore the ones most often lost.	Application type, and by inference the sector and the likely custodial failure mode.
`activity_or_category`	string	The quantity or the IAEA source category, running from Category 1 for the most dangerous sealed sources down to Category 5. Category is a far more useful risk descriptor than activity alone because it is normalised against the harm a source can cause.	Radiological consequence assessment; regulatory control requirements applicable to that category.
`incident_circumstance`	enum	How the material came to be out of control – theft, loss, discovery, unauthorised disposal, unauthorised possession, or attempted sale. The circumstance distinguishes an adversarial event from an administrative one.	Investigative pathway; whether the matter is a criminal case or a regulatory one.
`recovery_status`	enum	Whether the material was recovered, and if so when. Non-recovery is the field with the most operational significance, because unrecovered Category 1 and 2 sources are a standing hazard rather than a closed case.	Ongoing risk assessment; the population of material still missing.
`detection_means`	string	How the incident came to light – portal monitoring at a border or scrap facility, regulatory inventory reconciliation, law enforcement action, or public reporting. This field describes the detection system, not the event.	Detection capability assessment; where a state's controls are working and where they are not.
`malicious_intent_indicator`	enum	Whether the reporting state assessed an intent to use the material maliciously or to traffic it. Rarely affirmative, and the rarity is itself one of the database's most important findings.	Escalation to security rather than regulatory handling.
`cross_border_element`	enum	Whether the incident involved movement across an international border. Relevant to whether the case engages international legal instruments and cooperation mechanisms.	Applicable convention; the states whose authorities would need to cooperate.
`confirmation_status`	enum	Whether the state has confirmed the incident to the Agency. Unconfirmed reports do not enter the published totals, which is why open-source incident counts and ITDB counts diverge.	Reconciliation between media-reported incidents and the official record.

Coverage — and what is not in it

The database covers incidents from 1993 onwards, reported by a large and growing set of participating states – well over a hundred, with the current figure given in the annual factsheet. Coverage is therefore best understood as a function of two things: whether a state participates, and whether that state's regulatory system detects and reports incidents at all. Both vary enormously. States with mature regulatory infrastructure, comprehensive source registries, portal monitoring at borders and scrap facilities, and a culture of regulatory reporting will generate many recorded incidents, most of them Group III and most of them minor. States without those systems generate few records regardless of what is happening on their territory. The cumulative confirmed incident count since 1993 is in the low thousands and the annual factsheet is the only authoritative statement of it; the annual rate has been broadly of the order of a hundred-plus incidents per year in recent factsheets, dominated by Group III. Material coverage is overwhelmingly sealed radioactive sources from industrial radiography, medical therapy, well logging and gauging applications. Incidents involving highly enriched uranium or plutonium are a small number of cases across the entire history of the database, and that scarcity is a substantive finding rather than a gap.

Known blind spots

Absence of evidence here is not evidence of absence. These are the conditions under which IAEA Incident & Trafficking DB will not show you something that is nevertheless real:

Non-participating states contribute nothing, so entire regions can be absent from the record while having significant unaddressed problems, and their absence looks like safety.
The database records what regulators detected, so a state with no source registry and no portal monitoring will report almost nothing regardless of how much material is out of control on its territory.
Material that is stolen and never missed does not appear, and the discovery of orphan sources decades after loss demonstrates that this category is real and unquantified.
Conflict zones and areas outside effective state control are systematically invisible, precisely where facilities holding sources may have been abandoned or looted and no regulator exists to notice.
Reporting is voluntary in practice even for participants, and a state may decline to report an incident it considers sensitive without any mechanism forcing disclosure.
The public factsheet is aggregate only, so an analyst cannot see which state reported what, what isotope and activity were involved in a specific case, or whether a specific loss was recovered.
Rising incident counts over time are confounded by increasing participation and improving detection, which means the trend line measures the observing system as much as the phenomenon.
Attempted sales and scam offers of purported nuclear material are a recurring category and the database's classification of intent depends on the reporting state's assessment, which varies in rigour.
There is no linkage in the public record between an incident and any subsequent prosecution or recovery, so the disposition of most cases is unknowable from this source.

Write the blind spot into the product. A statement that something “was not observed in IAEA Incident & Trafficking DB” is defensible; a statement that it “did not happen” is not, and the difference is what survives cross-examination.

Access, licensing and what you may do with it

Access model: Restricted — eligibility-gated: member states, accredited institutions or vetted users only

The detailed database is closed. Access is through designated ITDB points of contact in participating states, which means a national regulatory authority or the equivalent nuclear security body, and it is a government-to-government arrangement rather than a research resource. If you work for a participating state's competent authority, your route runs through that channel and is governed by national policy. What is genuinely public is the ITDB factsheet published annually by the IAEA, together with the Agency's nuclear security publications, press material and the Nuclear Security Series documents that draw on the database. That is a small amount of material, and building an analysis on this source means building it on aggregate statistics plus a careful reading of the methodology behind them. For incident-level detail, the practical open substitutes are national regulator disclosures, court records, and the open-source incident compilations maintained by non-proliferation research centres – all of which have different and generally worse definitional discipline than the ITDB, and none of which should be presented as ITDB data.

Licence

The published factsheet and the Agency's nuclear security publications are issued for public use and can be cited normally, subject to the IAEA's own terms and to standard attribution. Anything obtained through a national point of contact carries the handling rules of that channel, which will be restrictive and are set nationally rather than by the Agency. The important point is definitional honesty rather than copyright: because the detailed database is closed, any dataset described as ITDB data circulating outside official channels is not ITDB data, and reproducing it under that label misattributes it to the Agency. If you build a compilation of incidents from open sources, name it as your own compilation and state its selection criteria, because it will not be comparable to the official series and presenting it as though it were will mislead every downstream reader.

Rate limits and fair use

There is nothing to rate limit. The public artefacts are a small number of documents published on an annual cycle plus occasional press material. Collection means checking the ITDB page and the Agency's publications listing periodically – monthly is generous – and capturing each new factsheet as a dated document. Any collection design that involves frequent polling of iaea.org reflects a misunderstanding of what this source is. Spend the effort instead on the national regulator sources that publish incident-level information, which are numerous, heterogeneous and genuinely require engineering.

Licensing changes, and it changes without warning. A dataset that was free for research this year may not be free for commercial or evidential use next year. Confirm the current terms before you build a dependency on it, and record the terms you relied on alongside the data — the licence in force at the time of collection is part of the provenance.

Collecting it

How IAEA Incident & Trafficking DB is actually pulled, in the order you would set it up. Prefer the bulk or export interface over per-item lookups wherever one exists: it is kinder to the publisher, faster for you, and gives a reproducible snapshot rather than a series of point-in-time answers you cannot reconstruct later.

Method	Format	Cadence	Notes
Annual ITDB factsheet	HTML	Annual; check quarterly	The core public artefact. Capture the PDF with its publication date and extract the group breakdown and totals into a local time series. Retain every year's edition, because figures for prior years are occasionally revised as states confirm late reports.
IAEA nuclear security publications	HTML	Monthly check	Nuclear Security Series documents and Agency reports that draw on the database and explain the definitions. These carry the methodology you need to interpret the numbers and are more useful than the numbers themselves.
IAEA news and press material	HTML	Weekly	Occasional statements about specific incidents, assistance missions and detection cooperation. Low volume, and each item is a dated institutional statement worth preserving as such.
National regulator incident disclosures	HTML	Varies by regulator	Where the ITDB is closed, individual national authorities publish source loss and recovery information in their own reporting. This is the practical route to incident-level detail and it is entirely a per-country engineering problem.
Open-source incident compilations	CSV	Per publication cycle	Research centres maintain open compilations of reported nuclear and radiological incidents. Useful for case detail and useless for trend comparison against the official series, because the selection criteria are completely different.

Ingesting it into the platform

Every step below is idempotent and cursor-based: interrupt one and it resumes from where it stopped rather than duplicating rows or losing progress. Collection is recorded per source, so a feed that quietly stops publishing shows up as a stale timestamp instead of silently thinning your coverage.

Register the source with its access reality stated — Add the ITDB in sources.php as an annual publication source and record explicitly that incident-level records are not held. This prevents the platform's coverage.php view from implying a capability that does not exist, which matters because readers of a CBRN assessment will assume otherwise.
Capture the factsheet series as a time series — Ingest each annual factsheet through import.php and extract the group-level totals into structured records keyed by year and group. The series, not the latest number, is the analytical object, and prior-year revisions must be preserved rather than overwritten.
Preserve the definitions alongside the numbers — Store the group definitions and any methodological notes from each edition with the figures they describe. Definitions have been articulated differently across editions and a trend built across changing definitions is not a trend.
Separate the regulatory series from the security series — Tag Group III totals distinctly from Group I so that any downstream analytics.php view cannot silently aggregate a custodial failure count into a trafficking count. This single discipline prevents the most damaging misuse of the source.
Attach to country risk context — Where national regulator disclosures give country-level incident information, resolve them to country entities so that country-risk.php can reflect nuclear security regulatory capacity as a factor. Do not attribute ITDB aggregate figures to individual countries – the public data does not support it.
Build the isotope and application dimension — Maintain a local reference table mapping recurring isotopes to their industrial and medical applications and to their IAEA source category. This is what lets an analyst move from a bare isotope name to a statement about which sector is leaking and what the consequence of a loss would be.
Wire the referral pathway into the record — Store, in the source record and in any case template, the fact that suspected material out of regulatory control is reported to the national competent authority and, where appropriate, through the Agency's channels. An analyst who identifies something real needs the pathway in front of them, not in a policy document elsewhere.
Set a low-frequency alert — Create an alerts.php rule for new IAEA publications touching nuclear security. Volume is very low and precision is high, so an unfiltered alert is appropriate here in a way it would not be for a high-volume feed.

Registered sources and their last-collected state are listed in sources.php, and the scheduled chain that keeps them current is in automation.php.

How it is wrong, and how to tell

Every dataset is wrong in characteristic ways. Knowing which ways is the difference between using a source and being used by one, and it is the part of source evaluation most often skipped because it is the part that takes work.

Within its scope, the ITDB is the most rigorous dataset in its field, and the reason is the confirmation requirement. Every record has been confirmed by the state that reported it, which eliminates the rumour, duplication and misidentification that dominate open compilations of nuclear incidents. The classification scheme is stable and applied consistently, and the Agency is careful in its public communication about what the numbers do and do not mean. The methodological weakness is entirely on the collection side rather than the recording side: participation is voluntary, detection capability varies by an order of magnitude between states, and reporting depends on a national decision at every step. The result is a database that is highly accurate about what it contains and structurally unable to say anything about what it does not. Judge it accordingly – individual figures are trustworthy, cross-country comparison is invalid, and time trends require you to hold participation and detection capability constant, which you cannot do from the public data. The Agency's own framing acknowledges these limits more candidly than most secondary citation of it does.

Characteristic false positives

The dominant error is treating the total incident count as a trafficking count when Group III – losses, thefts of industrial sources, discoveries, unauthorised disposal – is the large majority and describes regulatory failure rather than adversarial activity.
Rising annual counts are read as a worsening threat when growing participation and improving detection are sufficient to explain the trend, and the database cannot separate the two.
Country attribution is inverted at borders: a seizure is recorded by the detecting state, so a state with effective portal monitoring appears in the record repeatedly while the origin state may never appear at all.
Detection date is treated as loss date, compressing the timeline of an incident where a source may have been out of control for years before a scrap yard monitor found it.
Purported offers of weapons-usable material are frequently fraudulent – the material offered is often not what is claimed, or does not exist – and treating an attempted sale as evidence of available material overstates the market substantially.
Open-source incident compilations are cited as though they were ITDB figures; they use different definitions, include unconfirmed reports, and are not comparable to the official series in either direction.
Activity and category are conflated, so a low-activity source in a large-sounding unit is reported as a serious radiological hazard when the category system exists precisely to prevent that error.
Non-recovery is under-weighted because it is undramatic; the standing population of unrecovered Category 1 and 2 sources is a more consequential fact than any single seizure and it is the fact least often reported.

None of these make the source unusable. They make it a source that requires corroboration before an assertion built on it goes into a product, which is true of every source and admitted by few.

Ageing

Individual incident records are historical facts and do not expire, but their status can change – material recorded as unrecovered may be recovered years later, and a state may confirm an incident long after it occurred, which is why annual figures are revised. The aggregate statistics age in a subtler and more dangerous way. A factsheet is a snapshot of a reporting system as well as of a phenomenon, so a five-year-old total describes both a different year and a different set of participating states with different detection capabilities. Citing an old cumulative figure as current is a common and avoidable error. Assessments of national regulatory capacity age fastest of all: a state that had no source registry a decade ago may have built one under an Agency assistance programme, and the security posture around a specific facility can change completely following a change of government or a conflict. Treat any capability judgement older than a few years as a hypothesis to be re-established rather than a fact to be repeated.

What this source feeds

A source is only worth what it lets you conclude. These are the disciplines that collect through it, the mission domains it serves and the data points it yields — every one is a tag, so you can follow any thread from here into the rest of the library.

Collected by these intelligence disciplines

CBRN Intelligence

Serves these mission domains

WMD / Proliferation

Yields these data points

How each sector uses IAEA Incident & Trafficking DB

The same dataset is worked very differently depending on who you are, what authority you hold, and what you are ultimately producing. A military analyst is supporting a commander’s decision; a journalist is meeting a publication standard; an NGO caseworker is protecting a person. The records are shared — the constraints, thresholds and outputs are not.

🎖 Military and defence

For CBRN defence and force protection, the ITDB's aggregate picture is the empirical corrective to threat assessments built on scenario planning. The material most often out of regulatory control is industrial and medical sealed sources, not weapons-usable material, which means the realistic radiological hazard in most theatres is a contamination and denial problem rather than a fissile material one. That has direct implications for detection equipment selection, for reachback requirements and for the training emphasis given to responders. In stabilisation and post-conflict operations the operationally significant fact is the standing population of unrecovered and abandoned sources at hospitals, industrial sites and research facilities, which becomes a force health protection issue and a civil-military coordination task. Handle any specific finding through CBRN specialists and national authorities rather than acting on it independently, and remember that the Agency's assistance role gives states a route to support that a military actor should facilitate rather than substitute for.

🕵 National intelligence

For CBRNINT this is the baseline against which every claim about a nuclear black market must be tested, and the test is usually failed. The database's most strategically important content is negative: the persistent scarcity of confirmed incidents involving weapons-usable material across three decades, in a period when detection has improved substantially, is strong evidence about the actual state of the illicit market. Analysts should use it to discipline assessments rather than to generate leads, because it is aggregate, lagging and closed. The useful collection targets it points toward are the sectors that leak – industrial radiography, well logging, medical therapy sources, legacy Soviet-era radioisotope thermoelectric generators – and the states whose regulatory capacity is weak enough that losses go unnoticed. Where an intelligence question involves a specific incident, the route is through the national competent authority and the Agency's channels, and unilateral collection against a nuclear security matter creates diplomatic exposure disproportionate to the yield.

👮 Law enforcement

For law enforcement the ITDB is context rather than casework, but the context is decisive. It tells you that the overwhelming majority of nuclear and radiological cases you will encounter involve industrial or medical sources, that the offenders are frequently opportunistic thieves and scrap dealers rather than proliferators, and that purported offers of weapons-usable material are usually fraud. That should shape investigative posture: treat the material as a radiological safety hazard first, secure the scene through qualified responders, and involve the national regulatory authority immediately because they hold the registry that identifies the source and the obligations that attach to it. Cross-border cases engage the international conventions on physical protection and on nuclear terrorism, and the cooperation route runs through national central authorities and international policing channels. Never handle suspected radioactive material outside a qualified response framework, and never treat a seizure as a normal contraband case.

🔍 Private investigation and corporate security

Private investigators and corporate security teams meet this domain through industrial clients – radiography contractors, well logging companies, medical facilities, scrap metal processors – where source loss is a real operational and regulatory risk. The ITDB's aggregate picture is the argument for controls: sources are lost from vehicles, from storage, in transport and in disposal, routinely and everywhere, and the resulting liability is regulatory as well as reputational. The appropriate work is preventive – inventory discipline, transport security, contractor vetting, portal monitoring at scrap intake – rather than investigative. Where a loss has occurred, the correct action is immediate notification to the national regulatory authority, which is a legal obligation in most jurisdictions and not a discretionary decision for the client to weigh. A firm that advises otherwise is exposing both the client and the public.

📰 Journalism and OSINT media

This is a field where reporting quality is poor and the source is the fix. Most journalism about nuclear smuggling reasons from a handful of dramatic seizure cases and produces an impression the data does not support. The ITDB gives you the denominator: how many incidents, of what type, and how few involve material of weapons relevance. Two habits produce far better reporting – always ask which group an incident belongs to, and always ask whether the material offered in an attempted sale was ever verified to be what was claimed, because frequently it was not. The genuinely underreported story in this data is not smuggling; it is the routine loss of dangerous industrial sources and the population of them never recovered. Approach the IAEA and national regulators directly for context; both engage with press on this subject.

🌍 NGO, humanitarian and human rights

Non-proliferation and nuclear security organisations use the ITDB as the empirical anchor for advocacy on source security, on universalisation of the Convention on the Physical Protection of Nuclear Material and its Amendment, and on implementation of the Code of Conduct on the safety and security of radioactive sources. The most policy-relevant reading is the persistence of Group III incidents at scale over three decades, which shows that the unsolved problem is custodial and regulatory rather than exotic. For humanitarian and environmental organisations, orphan sources are a public health issue with a documented history of serious accidents involving people who found abandoned sources and did not know what they were – a hazard communication problem as much as a security one. Advocacy that overstates the fissile material threat is counterproductive; the accurate account is alarming enough and is much harder to rebut.

🎓 University and research

The ITDB underpins research on illicit trafficking, nuclear security governance and the effectiveness of detection architectures, and it is one of the very few longitudinal datasets in the security studies field with a consistent definitional scheme applied by a single body over thirty years. The unavoidable methodological problem is that the observing system changes over the observation period, so participation growth and detection improvement are confounded with any trend, and no public data allows you to adjust for them. Serious quantitative work in this area therefore either models the reporting process explicitly or restricts itself to claims robust to it. Detailed access runs through national channels rather than research application, so most academic work uses the published aggregates alongside open compilations, and the honest ones are explicit that these are different objects. Comparative work across states using published data is not defensible and appears in the literature anyway.

Playbook: working IAEA Incident & Trafficking DB end to end

A repeatable sequence from first pull to finished product. Each phase states what you are trying to establish, not merely what to click — the objective is a defensible chain of reasoning, not a completed checklist.

Phase 1 — Establish what question the aggregate can answer

The public ITDB supports statements about the global pattern of material out of regulatory control over time. It does not support statements about a specific country, a specific incident, or the current whereabouts of anything. Fix that boundary before you start, because almost every misuse of this source is a question asked of it that it was never able to answer.

Phase 2 — Build the group-separated time series first

Extract Group I, Group II and Group III totals by year from the factsheet series and plot them separately. The three lines behave differently and tell different stories. An analyst who has seen those three lines will never again quote a combined total, and one who has not will do it repeatedly.

Phase 3 — Model the observing system before reading the trend

Note participation growth and known changes in detection capability over your window – portal monitoring deployment, regulatory assistance programmes, registry establishment. Any increase in reported incidents must be assessed against these. If you cannot describe how the observing system changed, you cannot interpret the trend and should say so.

Phase 4 — Anchor the material picture on category, not on drama

Work out which source categories and applications dominate the record and what a loss in each would actually mean radiologically. Category 1 and 2 sources are the ones with serious harm potential; the bulk of incidents involve material far below that. This step converts an incident count into a consequence assessment, which is what a decision-maker actually needs.

Phase 5 — Test the fissile material question explicitly

Count what the record contains on weapons-usable material across the full history and state it plainly. This is the single most policy-relevant number in the field and it is small. Reporting it accurately is more valuable than any speculative assessment, and it is the finding that most often changes a reader's mind.

Phase 6 — Trace the sectoral origin of losses

Map the recurring isotopes to their applications – radiography, well logging, medical therapy, gauging, legacy power sources – and identify which industries are the persistent sources of loss. Regulatory effort is allocated by sector, so a sectoral finding is actionable in a way that a national one is not.

Phase 7 — Separate detection success from problem severity

A state appearing frequently in incident reporting may have a good detection system rather than a bad security situation, and the two are almost impossible to distinguish from published data. Where you must comment on a country, comment on its regulatory infrastructure using the Agency's assistance and legal instrument adherence records, not on its incident count.

Phase 8 — Assess the unrecovered population

Identify what the record says about material never recovered, because that is the standing hazard and it accumulates. This is the finding most likely to be operationally useful to a responder or a planner, and it is the one that receives the least attention because nothing happens on any particular day.

Phase 9 — Corroborate with national regulator disclosures

For incident-level detail, work the national regulators that publish source loss and recovery information. Their definitions differ from the Agency's and you must not merge the two series, but they provide the case texture the aggregate lacks and let you sanity-check the shape of the global picture.

Phase 10 — Treat purported sales with structural scepticism

Where an attempted sale of nuclear material is reported, the prior probability that the material was as advertised is low. Establish whether anything was analytically verified and by whom. A great deal of published assessment in this field rests on offers that were never substantiated, and identifying that is often the most useful contribution an analyst can make.

Phase 11 — Route anything real through the competent authority

If your work surfaces credible information about material out of regulatory control, the action is notification of the national competent authority – the nuclear regulator or equivalent – and, where cross-border, the international channels through the Agency and policing cooperation. This is a reporting obligation in most jurisdictions and not an analytical option. Record the pathway in the case file so that the next person does not have to find it.

Phase 12 — Publish the limits alongside the finding

Any product using this source should state that the data is aggregate, that participation is voluntary, that detection capability varies, and that trends are confounded by both. This is not hedging – it is the difference between an assessment that survives expert review and one that does not, and the expert reviewers in this field are unusually well informed.

The platform ships this as a step-checked workflow in playbooks.php, so progress is recorded against a case rather than held in someone’s head.

What to pair it with

No single source carries a finding. These are the datasets that corroborate, extend or contradict this one — and a source that contradicts is worth more than one that agrees, because it is the only thing that will tell you when you are wrong.

Source	Relationship	What it adds
IAEA Nuclear Security Series	prerequisite	The guidance corpus that defines source categorisation, physical protection expectations and detection architecture. Read the categorisation guidance before interpreting any incident record.
CTBTO monitoring	extends	Detects nuclear explosions rather than material movement. Together the two cover the nuclear security picture from opposite ends – one sees the material leaving control, the other sees the event.
INTERPOL	extends	International policing cooperation on radiological and nuclear matters, including the operational channels through which a cross-border case actually moves.
World Customs Organization	extends	Border detection and enforcement practice, including the radiation detection architecture at ports and crossings that generates a substantial share of seizure reporting.
UNODC	extends	Criminal justice responses to nuclear and radiological offences, including model legislation and the prosecution side that the ITDB record does not follow.
UNODA	prerequisite	The disarmament and non-proliferation framework, including Security Council Resolution 1540 obligations on states to prevent non-state actor acquisition of nuclear material.
Nuclear Threat Initiative	corroborates	Independent analysis of nuclear security including country-level materials security assessments, which supply the national dimension the ITDB aggregate deliberately withholds.
James Martin Center for Nonproliferation Studies	corroborates	Open-source research on trafficking incidents and non-proliferation, including compilations that provide case detail. Different definitions from the ITDB – never merge the series.

Legal, ethical and operational constraints

The public material is publishable and citable without difficulty. The legal weight of this domain sits elsewhere. Handling, possessing or transporting radioactive material is regulated in every jurisdiction and unauthorised possession is a criminal offence in most; an investigator who encounters suspected material must not handle it and must notify the national competent authority, which is a legal obligation rather than a judgement call. Cross-border cases engage the Convention on the Physical Protection of Nuclear Material and its 2005 Amendment, the International Convention for the Suppression of Acts of Nuclear Terrorism, and Security Council Resolution 1540, and the applicable cooperation mechanisms run between designated national authorities rather than between investigators. Information about nuclear security arrangements at specific facilities is protected in most jurisdictions and collecting or publishing it can itself be an offence, so an analyst working on this topic should be careful that a legitimate assessment does not become a description of a facility's vulnerabilities. The corresponding ethical constraint is straightforward and absolute: nothing written in this area should be usable as guidance for acquiring, moving or concealing radioactive material, and analysis should be framed around detection, regulation and response.

Operational security

Reading IAEA publications is unremarkable. Everything beyond that carries weight. An approach to a national competent authority about ITDB material identifies you and your interest to a body that coordinates with the Agency and with security services, and that is appropriate but should be a deliberate decision. Research patterns in this area attract attention: sustained open-source collection on nuclear material trafficking, specific facilities or detection architectures is exactly the profile that counterproliferation programmes are designed to notice, and being noticed for legitimate research is inconvenient rather than dangerous but is best anticipated. Keep facility-specific work compartmentalised, avoid assembling anything that reads as a target package even inadvertently, and be careful about correspondence with sources in states where interest in nuclear matters is itself hazardous. If your organisation does this work regularly, an established liaison relationship with the national regulator is worth more than any collection technique.

Two rules that hold regardless of jurisdiction. Collection that is lawful is not automatically proportionate, and a dataset assembled for one purpose does not carry consent for another. Where the records concern identifiable people, the question is not only whether you may hold the data but whether holding it serves the purpose you are accountable for.

Is it earning its place?

Sources accumulate. Feeds get added during an incident and are never reviewed again, and a decade later the pipeline is carrying dead weight that nobody dares remove. These are the measures that show whether IAEA Incident & Trafficking DB is contributing anything, and they are worth baselining now so the answer is available later.

Whether your factsheet time series is complete and each year's edition retained, since prior-year revisions are the only way to see late confirmations.
Proportion of your products that state incident group explicitly rather than quoting combined totals, which should be all of them.
Number of country-level claims made from aggregate data, which should be zero and is the easiest quality defect to audit.
Currency of your isotope-to-application-to-category reference table, which is what converts a bare incident description into a consequence statement.
Coverage of national regulator disclosure sources for your priority countries, tracked as a collection gap rather than assumed.
Time from publication of a new factsheet to update of your series and of any dependent country risk view.
Whether every case template touching radioactive material carries the competent authority notification pathway visibly, which is a safety metric rather than an analytical one.

Beware of volume. Indicator counts rise easily and say almost nothing. Unique contribution — findings this source produced that no other source in your stack would have — is the measure that matters, and it is usually far lower than anyone expects.

Tradecraft notes

The distinctions that separate a competent analyst from a fast one:

Group III dominance is the central fact of this dataset, and an analyst who quotes total incidents without separating the groups has converted a regulatory statistic into a security claim.
Detection date is not loss date, and the gap can be years – orphan sources have surfaced in scrap streams long after the facility that lost them ceased to exist.
A country that reports many incidents probably has good detection; a country that reports none may have nothing to report or may have no capacity to notice, and the public data cannot distinguish these.
Purported offers of weapons-usable material are frequently fraudulent and often involve material that was never what it was claimed to be, so an attempted sale is evidence about criminal intent rather than about material availability.
Source category tells you about harm potential in a way that activity alone does not, and using category is the difference between a radiological assessment and a number with a unit after it.
The unrecovered population is the standing risk and it accumulates quietly; it is the least reported and most operationally significant thing the database describes.
Rising counts across three decades are substantially an artefact of a growing and improving observing system, and any trend claim that does not address this is not an analytical finding.
Open compilations of nuclear incidents are useful for case detail and incomparable to the official series; merging them produces a dataset that means nothing and looks authoritative.
The Agency's confidentiality is the condition of the database's existence rather than an obstacle to be worked around, and analysis that treats it as the latter misunderstands why the data is good.

Questions analysts actually ask

Can I get incident-level ITDB data?

Not unless you are working through a participating state's designated point of contact, which means a national regulatory or nuclear security authority. The database is closed by design because voluntary state reporting depends on it. Plan your work around the published aggregates and national regulator disclosures instead.

Does a rising incident count mean trafficking is increasing?

No, and this is the most common error made with the source. Participation has grown and detection has improved substantially over three decades, both of which increase reported incidents independently of any change in the underlying phenomenon. The public data does not allow you to separate them.

How much weapons-usable material has actually been seized?

Very little, across the entire history of the database. Incidents involving highly enriched uranium or plutonium are a small number of cases since 1993, and the great majority of records involve industrial and medical sealed sources. That scarcity is one of the most important findings in the field.

What is an orphan source and why does it matter more than smuggling?

A radioactive source that has fallen outside regulatory control without necessarily being stolen – abandoned, lost, or disposed of improperly. It matters because the documented history of serious radiological harm involves people who found such sources and did not know what they were, and because unrecovered sources accumulate as a standing public health hazard.

Which isotopes turn up most often?

The industrial and medical workhorses – caesium-137, cobalt-60, iridium-192, americium-241, strontium-90 – because those are the sources in widest circulation. The distribution of incidents follows the distribution of use, which is why the sectoral analysis is more informative than the geographic one.

What should I do if my investigation encounters suspected radioactive material?

Do not handle it, do not move it, and notify the national competent authority immediately – the nuclear regulator or its equivalent. In most jurisdictions this is a legal obligation. Qualified radiological responders secure the scene; an investigator's role is to report and to preserve the surrounding evidence, not to assess the material.

Can I compare countries using ITDB figures?

No. The public data is aggregate and does not attribute incidents to countries, and even if it did, the count would measure detection capability as much as security performance. If you need a national assessment, use regulatory infrastructure and legal instrument adherence as indicators instead.

Are the open-source incident databases a substitute?

They are a complement for case detail and not a substitute for the series. They use different inclusion criteria, contain unconfirmed reports, and are shaped by media attention. Use them for texture, cite them as themselves, and never merge their counts with the Agency's.

How does this relate to IAEA safeguards?

It does not, legally. Safeguards are a verification regime under binding agreements covering declared nuclear material in peaceful programmes. The ITDB sits within nuclear security, which is an assistance and coordination function with no verification authority. Conflating them leads to expectations of enforcement power the Agency does not have here.

Standards, formats and interoperability

What this source speaks natively, and what it has to be translated into before a partner can consume it. Work that arrives in a recognised format is easier to defend, easier to hand over and easier to automate against:

IAEA source categorisation, ranking sealed sources from Category 1 to Category 5 by potential to cause harm – the correct risk vocabulary for any incident discussion.
The Code of Conduct on the Safety and Security of Radioactive Sources, with its supplementary guidance on import and export, which is the international expectation framework for source control.
The Convention on the Physical Protection of Nuclear Material and its 2005 Amendment, the binding instrument for protection of material in use, storage and transport.
The International Convention for the Suppression of Acts of Nuclear Terrorism, which criminalises possession and use of radioactive material with malicious intent.
UN Security Council Resolution 1540, obliging states to prevent non-state actors acquiring nuclear, chemical and biological weapons and related materials.
IAEA Nuclear Security Series recommendations including the physical protection recommendations, which define what adequate control looks like and therefore what a failure is.
Radiation detection instrument standards used at borders and scrap facilities, which determine what portal monitoring can and cannot see and therefore shape the incident record.
Transport regulations for radioactive material, whose packaging and documentation requirements are the control point at which many recorded losses occur.

References

Primary documentation and authoritative references for this source. Publishers revise and retire material, so treat the retrieval date as part of the citation and re-check before relying on any of it in a formal product.

IAEA Incident and Trafficking Database — International Atomic Energy Agency. The database's own page, including the annual factsheet and the definitions of the incident groups. Read the definitions before the numbers.
International Atomic Energy Agency — IAEA. The operator, its nuclear security programme, and the assistance activities that explain why detection capability differs so much between states.
IAEA publications — IAEA. The Nuclear Security Series and related guidance, including source categorisation. The methodological grounding for everything in this domain.
INTERPOL — INTERPOL. International policing cooperation on radiological and nuclear crime, and the practical route for a cross-border case.
World Customs Organization — WCO. Border enforcement practice and the detection architecture at ports and crossings that produces much of the seizure reporting.
UNODC — United Nations Office on Drugs and Crime. Criminal justice responses to nuclear and radiological offences, including model legislative provisions.
UNODA — United Nations. The non-proliferation framework including Resolution 1540 obligations, which is the legal backdrop to state reporting.
Nuclear Threat Initiative — NTI. Independent country-level analysis of nuclear materials security – the national dimension the ITDB aggregate deliberately does not provide.
James Martin Center for Nonproliferation Studies — Middlebury Institute of International Studies. Open-source research and incident compilations that supply case detail, with their own inclusion criteria that must not be conflated with the Agency's.
CTBTO — CTBTO Preparatory Commission. The detection side of the nuclear picture. Useful alongside this source for any full assessment of a state's nuclear activity.

Link integrity: every reference above was verified with a live request when this page was generated. Where a publisher had moved or withdrawn a document, the link was repointed at a preserved copy in the Internet Archive and marked as archived. Anything with no reachable copy anywhere had its link removed rather than left to rot — the source is still credited, it simply cannot be linked.

Put it into practice

The Quantus Intel threat intelligence platform operationalises this source: it holds the factsheet series as a group-separated time series rather than a single number, keeps the isotope-to-category reference that turns an incident into a consequence assessment, and puts the competent authority notification pathway in front of any analyst whose case touches material out of regulatory control.. Browse the full source catalogue, or follow any tag above into the rest of the library.