CTBTO Monitoring: Intelligence Source Guide

The CTBTO operates a global, purpose-built network for detecting nuclear explosions using seismic, hydroacoustic, infrasound and radionuclide sensors. Its bulletins are the authoritative technical record of nuclear test events – and almost none of the underlying data is public.

At a glance

Source	CTBTO Monitoring
Category	Conflict, Crime & Human Security › Military, Weapons & CBRN
Homepage	https://www.ctbto.org/
Machine interface	`https://www.ctbto.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, Measurement & Signature Intel
Mission domains	WMD / Proliferation

Nuclear-test detection network. — as catalogued in the platform’s own source registry.

The Preparatory Commission for the Comprehensive Nuclear-Test-Ban Treaty Organization runs the International Monitoring System, a network designed for one detection problem: distinguishing a nuclear explosion from everything else that shakes the ground, the ocean and the atmosphere. When complete the IMS comprises 337 facilities – 321 monitoring stations and 16 radionuclide laboratories – across four technologies. Fifty primary and 120 auxiliary seismic stations detect the ground motion of an underground explosion. Eleven hydroacoustic stations, a mix of moored hydrophone triplets and coastal T-phase seismometers, exploit the SOFAR channel to hear underwater and coastal events across entire ocean basins. Sixty infrasound arrays detect the sub-audible pressure waves of an atmospheric explosion. Eighty radionuclide stations sample particulates from the air, of which a subset also measure radioactive noble gases, and 16 certified laboratories provide confirmatory analysis. All of it streams to the International Data Centre in Vienna over a dedicated global communications infrastructure, where automatic processing produces a sequence of event lists that human analysts then review to produce the Reviewed Event Bulletin. The system is the most sensitive purpose-built geophysical detection network in existence, and it was built by treaty to serve States Signatories, not the public.

The analytical job the IMS does that nothing else does is discrimination with an audit trail. Global seismic networks run by academic consortia and national surveys will detect the same events; what the CTBTO adds is a verification architecture designed from the start to answer a specific legal question – was this a nuclear explosion – with defined processing, defined review, and a defined product that States Signatories can act on. The radionuclide component is the part with no substitute. Seismic, hydroacoustic and infrasound sensors tell you that an energetic event occurred and constrain its location, depth and character; only the detection of fission or activation products, particularly the xenon isotopes released by an underground test, converts that into evidence of a nuclear explosion. For CBRNINT and MASINT work this is the reference standard against which all open-source claims about a suspected test must be checked. It is also, for the analyst, an exercise in working with a source whose products you mostly cannot see – and that constraint is the single most important thing to understand about it.

Who publishes it, and why that matters

The organisation is a Preparatory Commission, not a fully constituted treaty body, because the Comprehensive Nuclear-Test-Ban Treaty has been open for signature since 1996 and has never entered into force. Entry into force requires ratification by a specific list of states in Annex 2 of the Treaty, several of which have not signed or not ratified. The practical consequence is that the CTBTO has spent nearly three decades building and operating a verification system for a treaty that is not legally in effect, funded by assessed contributions from States Signatories. The on-site inspection regime – the part of the verification architecture that would put inspectors on the ground – cannot be invoked, because it depends on entry into force. The political environment around the organisation has deteriorated: in November 2023 the Russian Federation revoked its ratification of the Treaty while remaining a signatory, and disputes over data sharing and station operation surface periodically. None of this has stopped the network functioning, and the IMS has continued to detect and characterise events throughout. But an analyst should understand that this is a technically excellent system operating on political foundations that are not stable, and that data availability is a policy decision that can change.

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
`station_code`	string	IMS station identifier, conventionally a two-letter technology and role prefix plus a number and a station code – PS for primary seismic, AS for auxiliary seismic, IS for infrasound, HA and HT for hydroacoustic hydrophone and T-phase, RN for radionuclide. Encodes what the station can and cannot see.	Station location and technology, which determines detection geometry and therefore what a non-detection means.
`event_origin_time`	timestamp	Estimated time of the source event in UTC, derived from arrival times at multiple stations. It is a solution parameter with an uncertainty, not a measurement, and the uncertainty grows for poorly recorded events.	Correlation with reported test announcements, satellite observations and seismic bulletins from other networks.
`latitude_longitude`	string	Estimated epicentre or source location with an associated error ellipse. The ellipse is the analytically important part and is routinely dropped when the location is repeated in secondary reporting.	Geospatial correlation with known test sites, mining regions and facilities.
`depth`	int	Estimated source depth in kilometres, often poorly constrained and frequently fixed to a default value in automatic solutions. Shallow depth is a necessary but not sufficient condition for an explosion.	Discrimination workflow – depth is one of the first screening criteria applied.
`magnitude`	string	Body-wave magnitude mb from the seismic solution. Conversion from magnitude to explosive yield requires assumptions about coupling, geology and depth of burial that dominate the answer, so published yield figures are far less certain than the magnitudes behind them.	Comparison against the magnitude history of a known test site; nothing further without geological assumptions.
`event_type_screening`	enum	The outcome of the IDC's screening process, which classifies events as consistent or inconsistent with natural phenomena. Screening is a filter that reduces analyst workload; it is explicitly not a determination that a nuclear explosion occurred.	Prioritisation of events for analytical follow-up.
`radionuclide_category`	enum	Classification of a radionuclide sample by the significance of the nuclides detected, running from routine background through to samples containing fission or activation products of potential relevance. Category assignment is the trigger for laboratory confirmation.	Atmospheric transport modelling to identify possible source regions.
`noble_gas_isotope_ratio`	string	Relative activities of xenon isotopes such as Xe-131m, Xe-133, Xe-133m and Xe-135. Isotopic ratios are the discriminant between a nuclear explosion and civil sources such as medical isotope production or power reactor releases.	Source attribution when combined with backward atmospheric transport modelling.
`arrival_phase`	string	Identified seismic or infrasound phase at a station – P, Pn, Lg, Rg and so on for seismic. Phase identification drives location and is the substrate for discrimination ratios such as P to S amplitude comparisons.	Waveform analysis; the raw material for independent reanalysis by a national data centre.
`bulletin_product`	enum	Which IDC product a record belongs to – automatic Standard Event Lists produced in stages after data arrival, the analyst-Reviewed Event Bulletin, and the screened bulletin. Automatic and reviewed products differ materially and must never be quoted interchangeably.	Confidence assessment – a claim sourced from an automatic list carries far less weight than one from the reviewed bulletin.
`data_availability`	enum	Station uptime and data completeness state. The IMS publishes network performance information because a station being down is the reason an event may be missed, and it is the first thing to check before asserting a non-detection.	Detection capability modelling for a given region and time window.
`waveform_segment`	array	Continuous or segmented waveform data from a station channel, held in standard seismological formats. This is the underlying evidence and access to it is restricted to authorised users.	Independent reanalysis, cross-correlation against previous events at the same site – the most powerful discriminator available.

Coverage — and what is not in it

Global by design, and unusually close to it in practice – the IMS is one of the few sensor networks whose coverage was engineered against a uniform detection threshold rather than accumulated opportunistically. Seismic coverage is dense enough to locate events of modest magnitude anywhere on land, with sensitivity best in regions with good station geometry and degraded in ocean basins and parts of the southern hemisphere. Hydroacoustic coverage of the oceans is extraordinary for eleven stations because sound in the SOFAR channel propagates for thousands of kilometres, so a small number of well-sited hydrophones cover most of the world's water. Infrasound coverage is global but strongly modulated by atmospheric conditions – stratospheric wind direction changes seasonally and determines which azimuths a given array can hear. Radionuclide coverage is the weakest link and the most consequential one: particulate and noble gas transport depends on weather, and a detection may occur days after an event and thousands of kilometres away, or not at all. Time coverage begins with station certification, which has been a rolling process since the late 1990s, so the network's detection capability in 1999 is not the same object as its capability today and historical comparisons must account for that. Data flows continuously to Vienna; automatic products are generated within hours and the reviewed bulletin follows on a defined schedule of days.

Known blind spots

Absence of evidence here is not evidence of absence. These are the conditions under which CTBTO Monitoring will not show you something that is nevertheless real:

The network detects explosions, not programmes – subcritical experiments, hydrodynamic testing, computational stockpile stewardship and weapons development short of a nuclear yield produce no IMS signature at all, and the absence of detections says nothing about whether a weapons programme is active.
A fully contained underground test in favourable geology can release no detectable radionuclides, which means the confirmatory technology that distinguishes nuclear from conventional may simply have nothing to work with.
Very low yield tests, and tests decoupled by detonation in a large cavity, reduce seismic coupling by a large factor and can fall below the network's practical detection threshold in some regions.
Radionuclide detection is at the mercy of meteorology – wind direction, precipitation scavenging and station spacing mean a release can pass between stations or be washed out before sampling, and backward transport modelling produces a plausible source region rather than a location.
Infrasound sensitivity varies seasonally with stratospheric wind reversal, so the same event may be clearly recorded in one season and missed on the same azimuth six months later.
The system is deliberately not an attribution engine for anything except nuclear explosions; a chemical explosion, a mine collapse, an earthquake swarm or a bolide will be detected and located but the products are not designed to characterise them, and the screening process is calibrated for a different question.
Station outages, communications failures and political disruption to national station operation create real gaps, and detection capability statements that assume full network availability will overstate what was actually possible on a given day.
Yield estimates are not measurements. They are model outputs whose uncertainty is dominated by assumptions about the geology and emplacement conditions at the source, which are exactly the things an analyst outside the country does not know.
Almost nothing is public. The bulletins, station data and network performance information go to States Signatories through their National Data Centres, so an analyst without that access is working from press statements and from independent networks, not from the IMS.

Write the blind spot into the product. A statement that something “was not observed in CTBTO Monitoring” 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

This is the part most analysts get wrong. The CTBTO's data and products are made available to States Signatories, which access them through their National Data Centres under the Treaty's data-sharing arrangements; they are not an open dataset and there is no public API. If you work for a government that is a State Signatory, the route to IMS data runs through your NDC and is a matter of national policy, not of registration on a website. For researchers, the CTBTO has operated a mechanism for granting access to IMS data for scientific purposes under agreement, and the organisation has released data publicly in specific circumstances, most notably in support of civil emergency response. What is genuinely open is the public web presence: the CTBTO publishes press releases, executive statements, technical explainers, station network descriptions and public announcements when significant events are detected – the DPRK tests being the recurring example. That is what an open-source collector can ingest, and it is a communications channel rather than a data feed. Plan accordingly, and use independent seismological networks for the underlying geophysics.

Licence

Treat all IMS data and IDC products as restricted unless you have a written basis saying otherwise. Data provided to a State Signatory through its National Data Centre comes with national handling rules and with treaty-derived confidentiality expectations; onward disclosure is a national decision and often a restricted one. Material obtained under a scientific access arrangement carries the conditions of that arrangement, which typically constrain redistribution and require acknowledgement. Public web content on ctbto.org is publishable material issued for communication purposes and can be cited normally, subject to standard copyright and to the organisation's own terms. The general rule for this source is the opposite of most open-source work: assume closed, and document the specific authority under which anything else is held. Confirm current arrangements directly rather than relying on a description of them, because access policy here is politically contingent and has been revisited.

Rate limits and fair use

Not a rate-limited API in any conventional sense. The constraint on the public web presence is simple politeness: it is a small institutional site, updates are infrequent, and polling it more than a few times a day accomplishes nothing except making you conspicuous. For authorised users, throughput to the IDC is governed by the terms of the national arrangement and by the design of the data delivery mechanisms, not by anything you can tune. If your interest is continuous seismic waveform monitoring, the correct answer is to build that on open seismological data services, which are designed for high-volume programmatic access, and to reserve the CTBTO channel for the products only it produces.

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 CTBTO Monitoring 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
Public news and press releases	HTML	Weekly polling is ample; the site is not a high-frequency publisher	The only broadly available channel. Significant event announcements, executive statements and technical explainers. Treat each as a dated institutional statement and record it as such.
Station network descriptions and technology pages	HTML	Quarterly refresh	Reference material on which technologies are deployed where. Useful for building the detection-geometry context you need to interpret any event claim, and stable enough to cache.
National Data Centre channel	bulk	Continuous, per national arrangement	For government users only. Bulletins and waveform data arrive through the NDC under national handling rules. Everything about this route is governed by policy rather than by technology.
Scientific access arrangement	bulk	Per agreement	Research access to IMS data has been provided under agreement for scientific purposes. Terms are specific and constrain redistribution; approach the organisation directly rather than assuming eligibility.
Independent seismic data services as a substitute	bulk	Continuous	For the geophysics rather than the verification products, open seismological archives provide waveform and bulletin data programmatically and are the practical route for most analysts.

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 public channel honestly — Add the CTBTO in sources.php as an HTML publication source with a low cadence, and record in the source description that it is an institutional communications channel rather than a data feed. Mislabelling it as a monitoring feed will produce a coverage claim in coverage.php that the platform cannot support.
Capture announcements as dated events — Run collect.php against the news and statements pages and normalise each item into an event record with its publication timestamp preserved. The institutional statement, with its date, is the artefact of value – not a paraphrase of it.
Separate detection claims from institutional claims — During ingest.php processing, tag whether an item asserts a technical detection, describes network status, or is a policy statement. These three classes have different evidentiary weight and mixing them is the most common analytical error with this source.
Resolve geography carefully — Where an announcement names a location or a test site, resolve it to a coordinate through the platform's location handling and attach an explicit uncertainty. Never store a point without the uncertainty that came with it; a bare coordinate implies a precision the underlying solution does not have.
Cross-reference open seismic bulletins — Use correlate.php to align CTBTO announcements against independently published seismic event solutions for the same origin time and region. Agreement between an intergovernmental verification body and an academic network is meaningful corroboration; that is the check the platform should perform automatically.
Place on the country and theatre timelines — Push resolved events into timeline.php and country.php so a detection sits alongside diplomatic, sanctions and military reporting for the same state, which is where a nuclear test event acquires its analytical context.
Set alerting on the announcement channel — Define an alerts.php rule on new items from this source. Volume is low and precision is high, so this is one of the few sources where an unfiltered alert on any new publication is defensible.
Record the access constraint in the source metadata — Note explicitly in sources.php that the underlying IMS data is not held. Any report generated through reports.php that touches nuclear testing should carry that limitation, because a reader will otherwise assume the platform has access it does not have.

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.

As a sensor system, this is about as good as verification technology gets. The network was designed against an explicit detection threshold, the stations are certified against defined specifications, the processing chain is documented, and analyst review sits between automatic detection and the authoritative bulletin. The organisation has demonstrated its capability repeatedly and publicly against real events, and its performance during the DPRK test series is the most thoroughly examined case study in the field. Independent scientific scrutiny of IMS-derived results is extensive, which is the strongest quality signal available for any sensor network. The judgement to make is therefore not about whether the system works but about what you are actually holding. If you have IDC bulletins through an NDC, you have high-quality technical products with stated uncertainties. If you have press releases, you have accurate but heavily compressed institutional statements from which the uncertainties have been removed for a general audience, and you should not reason quantitatively from them. The gap between those two conditions is enormous and is invisible in a citation.

Characteristic false positives

Chemical explosions, mine blasts and mine collapses produce seismic signals that resemble explosions on the screening criteria, and industrial blasting is routine in many regions, so a shallow event with explosion-like characteristics is far more likely to be industry than a test.
Radioxenon detections are frequently civil in origin – medical isotope production facilities are the dominant global source and release quantities that swamp the signal from a small test, which is a known and much-studied confounder rather than an obscure edge case.
Automatic event lists contain events that human review later merges, relocates or discards, so any claim sourced from an automatic product before analyst review may not survive the review.
Yield figures quoted in the media are model outputs presented as measurements; the same magnitude will support a range of yield estimates spanning more than an order of magnitude depending on assumed coupling, and the choice of assumption is usually not stated.
Location error ellipses are dropped in retelling. A point coordinate propagates through secondary reporting as if it were surveyed, and analysts then match it to a specific facility that lies within the uncertainty but is not indicated by it.
Atmospheric transport modelling produces a possible source region with a probability field, and that field is routinely reported as if it named a country; the tails of these distributions are wide and cross borders.
Detection capability is treated as constant when it is not – a station outage, a seasonal wind reversal or an unfavourable transport pattern changes what the network could have seen that week, and non-detection claims made without checking network state are unsound.
Aftershocks, secondary collapse events and induced seismicity at a test site are separate events that can be miscounted as additional tests if origin times are read carelessly.

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

Event records do not age – a detection is a dated observation and remains valid indefinitely, and the IMS archive becomes more valuable over time because historical waveforms are the reference library against which new events are cross-correlated. What ages is everything around the record. Detection capability statements age quickly, because station commissioning, outages and network changes alter what the system could see; a capability figure quoted from a decade-old assessment does not describe today's network. Yield estimates age when better geological information about a site emerges, and published estimates for historical tests have been revised. Institutional and political statements age fastest of all, because the organisation's operating environment, membership behaviour and data-sharing arrangements change. A stale record here looks like a confident sensitivity claim – 'the network would have detected anything above X' – repeated years after the assessment that produced it, with no reference to which stations were actually reporting.

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

Serves these mission domains

WMD / Proliferation

Yields these data points

How each sector uses CTBTO Monitoring

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 nuclear operations, arms control support and CBRN defence planning, the IMS is the authoritative external check on whether a nuclear detonation occurred and where. Its practical military value is less about warning – national technical means are faster – and more about corroboration and about the international legitimacy of a finding, because an event characterised through a multilateral verification system is harder to dispute than one asserted unilaterally. The infrasound and hydroacoustic components have secondary value for detecting large conventional explosions, missile reentry and other energetic atmospheric events, which is a real capability worth knowing about even though it is not the network's purpose. The doctrinal caution is that this system detects explosions and says nothing about deployment, readiness or intent, so it belongs in an indications and warning picture as one narrow, high-confidence input rather than as a monitoring capability in its own right.

🕵 National intelligence

For CBRNINT and MASINT this is the reference standard and the principal source of publicly acknowledgeable technical evidence on nuclear testing. The analytical craft lies in the discrimination problem: separating explosions from earthquakes, nuclear from chemical, and test-derived radioxenon from the civil isotope production background. Every one of those separations is probabilistic and every one has a literature. For collection management, the network's own performance data is a prerequisite for writing an honest confidence statement about non-detection, which is the judgement most often demanded and least often defensible. The strategic point about this source is its dependence on political arrangements: access runs through National Data Centres under national policy, and the treaty under which the whole system operates has never entered into force, so the continuity of the collection is a policy question that belongs in your risk register.

👮 Law enforcement

Direct law enforcement use is rare and specific. The relevant intersections are the investigation of unauthorised nuclear activity, where a detected event provides technical anchoring for a case, and support to prosecutions or inquiries where the timing and location of an explosion is in dispute. The realistic route is through national channels rather than direct approach: the IMS is not an evidence service, its products are provided to states under treaty arrangements, and their admissibility in any national proceeding is a question for the national authority holding them. Where an investigation touches trafficking of nuclear or radioactive material, the incident-based records held by the IAEA are the operationally relevant source and this one is context. Approach through your national data centre or nuclear regulator, and do not expect a response to an unsolicited request.

🔍 Private investigation and corporate security

Limited but not zero. Corporate and private investigative work touching nuclear supply chains, sanctions exposure, or insurance and continuity assessment for facilities near test sites may need to establish whether and when an event occurred, and the public CTBTO statements provide a citable, authoritative answer for the small number of events that matter. The realistic posture is to use the public communications and open seismological data rather than to attempt access to restricted products, which will not be granted to a commercial investigator. The professional failure to avoid is over-claiming: writing a client report that implies access to verification data you are actually inferring from press statements and open seismic bulletins is the sort of thing that destroys a firm's credibility when tested.

📰 Journalism and OSINT media

For reporting on nuclear testing this is the source that settles the factual question, and the CTBTO's public statements are unusually clear and technically careful compared with most institutional communications. The reporting failures are consistent and avoidable: converting magnitude to yield without saying that the conversion depends on unknown geology, printing a coordinate without its uncertainty, and treating an automatic detection as a confirmed finding. Two habits fix most of this – always ask whether a figure is a measurement or a model output, and always ask whether a radionuclide detection has been separated from the civil isotope production background. Independent seismologists will speak on the record about these questions and are the right second source, because they can characterise uncertainty in a way an institutional press office is not structured to.

🌍 NGO, humanitarian and human rights

Disarmament and non-proliferation organisations use the IMS record as the evidentiary backbone for advocacy on the test ban, and the network's continued operation is itself a policy argument – it demonstrates that verification is technically solved and that entry into force is a political rather than a technical problem. For humanitarian and environmental organisations, the radionuclide network's civil applications matter: the same sensors that look for test-derived isotopes measure atmospheric radioactivity generally, and the organisation has made data available in support of civil emergency response. The discipline for advocacy use is to be precise about what the system can and cannot detect, because overstating its reach – implying it would catch any weapons-related activity – hands an easy rebuttal to opponents and misleads the public about what a test ban actually constrains.

🎓 University and research

The IMS is a major scientific asset well beyond arms control. Its waveform and infrasound archives support research on earthquake source physics, ocean acoustics, atmospheric dynamics, bolide detection, iceberg calving, volcanic eruption monitoring and marine mammal acoustics, and the noble gas network underpins a substantial atmospheric transport modelling literature. Access for research has been provided under specific arrangements, and the terms constrain redistribution, so plan the data management side of a project before designing the analysis. The most productive research posture is comparative: run your method on open seismological data first, establish that it works, then seek IMS access for the specific stations and windows where the network's unique geometry or sensitivity is genuinely required. Publication of results derived from IMS data carries acknowledgement and review obligations under the access terms.

Playbook: working CTBTO Monitoring 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 access you actually have

Before anything else, determine whether you are working from IDC products, from a scientific access arrangement, or purely from public statements and independent networks. These are three different sources with three different confidence ceilings. Write the answer down, because every judgement you produce inherits it and a reader six months later will need to know which one you were in.

Phase 2 — Learn the network geometry for your region of interest

Identify which IMS stations of which technologies cover your area, and understand that seismic, infrasound, hydroacoustic and radionuclide coverage are four different maps. This determines the meaning of a non-detection more than anything else. Record the station set and note that certification and outages change it over time, so the geometry you note today is not the geometry that applied to a historical event.

Phase 3 — Separate the detection question from the characterisation question

Detection asks whether an energetic event occurred at a location and time. Characterisation asks what kind of event it was. Seismic, hydroacoustic and infrasound data answer the first strongly and the second only probabilistically; radionuclide data is what closes the second. Structuring your analysis around this split prevents the commonest error, which is treating a confident location as though it were a confident identification.

Phase 4 — Check the natural and industrial baseline first

For any candidate event, establish what the region's normal seismicity and industrial blasting look like. Mining regions produce a steady stream of shallow explosion-like events. If you cannot describe the baseline, you cannot say an event is anomalous, and the screening process the IDC applies exists precisely because most shallow events are not what an analyst hopes they are.

Phase 5 — Treat magnitude as data and yield as inference

Carry the magnitude with its uncertainty through your analysis and quote yield only with explicit assumptions stated – coupling, depth of burial, geology. Where you must give a yield, give a range with the assumption set that produced it. A single yield number without assumptions is not a finding, it is a summary of someone else's guesses about rock.

Phase 6 — Interrogate any radionuclide claim against the civil background

Medical isotope production is the dominant global source of atmospheric radioxenon and it is not subtle. Before treating a detection as significant, establish the known civil sources upwind and the isotopic signature that would distinguish them. Isotopic ratios, not the mere presence of xenon, carry the discriminating information, and any claim that omits the ratios has not done the work.

Phase 7 — Reconstruct the transport, do not accept the conclusion

Backward atmospheric transport modelling produces a probability field over possible source regions, sensitive to the meteorological fields used and to the release timing assumed. If a source region is asserted, ask what release window was assumed and how wide the field is. Reported source attributions frequently present the mode of a broad distribution as a location.

Phase 8 — Cross-correlate against the site's own history

For any location with a testing history, the most powerful available discriminant is waveform similarity to previous events at the same site, because a repeated source in the same geology produces a near-identical signal. This requires waveform access, which most analysts do not have – but it also means that published analyses which used it are far stronger than those which did not, and you should weight external work accordingly.

Phase 9 — Build the timeline from independent networks

Whatever your access, construct the event chronology from openly published seismic bulletins as well as from CTBTO statements, and note where they agree and diverge on origin time, location and magnitude. Divergence is informative and is usually about station geometry and phase picking rather than about disagreement on whether something happened.

Phase 10 — Place the technical finding in its political frame

A detected event acquires meaning from context – declaratory statements, site activity observed by other means, diplomatic behaviour, and the treaty status of the state involved. Push the event onto the country and theatre timelines alongside sanctions and diplomatic reporting. The technical fact is narrow; the analytical product is the fact plus what preceded and followed it.

Phase 11 — State the non-detection carefully or not at all

If you are asked whether a test occurred and the answer is that nothing was detected, the defensible formulation names the network state, the coverage for that region and technology, and the approximate threshold below which an event of that type would not have been seen. Anything shorter is an assertion you cannot support, and this is the single judgement where analysts most often overreach.

Phase 12 — Record the access limitation in the product

Close every assessment with an explicit statement of what you held and what you did not. If the analysis rests on press statements and open seismic data, say so in the product rather than in a footnote. The credibility of work in this area depends almost entirely on being precise about provenance, because the audience includes people who do have the underlying data.

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 Incident and Trafficking Database	extends	Covers nuclear and radioactive material out of regulatory control, which is the other half of the nuclear security picture. The CTBTO sees explosions; the ITDB sees material movement and loss.
USGS Earthquake Hazards Program	corroborates	Independent, openly published event locations and magnitudes with a programmatic interface. The practical first check on any claimed event, and available to everyone without an access arrangement.
International Seismological Centre	corroborates	The definitive long-term global seismological bulletin, compiled from many networks and revised over years. Slower than any operational product and more thorough than all of them.
EMSC	corroborates	Rapid European-Mediterranean event solutions with fast public dissemination, useful for early independent confirmation of origin time and location.
IRIS seismological data services	prerequisite	Open waveform archives and standard web services. If you want to do seismology rather than read about it, this is where the accessible data lives.
NORSAR	extends	A national research institution with deep expertise in seismic array processing and explosion monitoring, and a long publication record on discrimination methods relevant to this problem.
UN Office for Disarmament Affairs	prerequisite	Treaty status, ratification records and the formal disarmament context that determines what the CTBTO can and cannot do.
United Nations Treaty Collection	prerequisite	The authoritative record of signature and ratification status for the Treaty, including changes. Check here rather than relying on any secondary summary of who has ratified.

Legal, ethical and operational constraints

Nothing about reading the CTBTO's public website raises a legal question. Everything about the underlying data does. IMS data and IDC products are provided under treaty-derived arrangements to States Signatories, and holding, using or disseminating them without authority is a matter of national law and national security regulation in most jurisdictions – in some, seriously so. If you are not certain you have authority, you do not. For researchers working under a scientific access arrangement, the terms of that arrangement govern retention, redistribution and publication, and they are not negotiable after the fact. There is a separate ethical dimension specific to this domain: analysis of nuclear test detection touches directly on state secrecy and on arms control verification, and publishing detailed capability assessments – particularly quantitative statements about detection thresholds in specific regions – can be read as information useful to a state seeking to test below the threshold. Keep public analysis at the level of methodology and published results rather than producing operational guidance on evasion. In most jurisdictions, export control and technology transfer rules may also apply to some analytical techniques in this area, which is an unusual constraint that catches researchers unfamiliar with the field.

Operational security

Reading the public website is unremarkable and reveals nothing beyond an interest in nuclear matters. Requesting access is a different act entirely: an approach to a National Data Centre or to the organisation identifies you, your institution and your area of interest to a body that reports to states, and that information will be handled according to national rules you do not control. Assume any access request is visible to your own government and potentially to others. For analysts working on a specific state's nuclear activity, the aggregate pattern of your open-source collection – which technical literature you read, which stations you look up, which historical events you research – is itself informative to anyone observing it, and this is a field with active counterintelligence interest on multiple sides. Compartmentalise the collection, avoid encoding target identifiers into queries against any external service, and keep the analytical work product on systems whose classification matches its sensitivity rather than its source.

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 CTBTO Monitoring is contributing anything, and they are worth baselining now so the answer is available later.

Whether every nuclear-testing claim in your finished products is traceable to either an IDC product, an independently published seismic solution, or an attributed institutional statement – with the category named.
Count of non-detection statements issued without an accompanying network-state and threshold qualification, which should be zero.
Number of yield figures published without their assumption set, tracked as a quality defect rather than a stylistic one.
Freshness of your stored network geometry and station status information relative to the events you are assessing.
Agreement rate between CTBTO public statements and independent seismological bulletins on origin time, location and magnitude for the same events.
Time from a public announcement to its appearance on the relevant country timeline in the platform, which measures whether the low-volume channel is actually being watched.
Proportion of radionuclide-related assessments that explicitly address the civil isotope production background.

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:

Non-detection is a claim about the network, not about the world, and it is only defensible when you can state which stations were reporting and what the threshold was for that region and technology on that date.
Seismic magnitude and explosive yield are related by assumptions about rock, not by a formula, and the assumptions dominate the answer – two competent analysts can differ by an order of magnitude on the same event without either being wrong.
The presence of radioxenon means very little on its own; the isotopic ratios are the discriminant, and any analysis that reports detection without ratios has stopped one step short of the finding.
Automatic event lists and the analyst-reviewed bulletin are different products and citing one as the other is the technical equivalent of quoting a draft as a publication.
Waveform cross-correlation against previous events at the same site is the strongest discriminant in the field, which means published analyses that used it deserve substantially more weight than those that reasoned from bulletin parameters alone.
Infrasound coverage changes with the seasons because stratospheric winds reverse, so an array that hears a given azimuth in January may not in July – this is physics, not equipment failure.
The system was built for a treaty that is not in force, and the on-site inspection regime that would resolve ambiguous cases cannot be invoked, which caps what verification can conclude no matter how good the sensors are.
Detecting explosions is not monitoring weapons programmes; subcritical and computational work leaves no signature here, and conflating the two produces a badly wrong picture of what a test ban constrains.
The most common failure in this domain is not technical error but provenance inflation – work built on press releases described in language that implies access to verification data.

Questions analysts actually ask

Can I download IMS data?

Not as an open dataset. Data and products go to States Signatories through their National Data Centres under treaty arrangements, and research access has been provided under specific agreements. If you need waveform data for seismological work, open archives run by academic consortia are the correct and far easier route.

Does the CTBTO announce every nuclear test?

It issues public statements about significant events, and it has done so promptly and clearly for the events that mattered. But the routine flow of technical products goes to states, not to the public, so the public record is a selected and summarised subset rather than the bulletin.

How small a test can the network detect?

There is no single number. Detection capability varies by region, technology, station availability, geology and emplacement conditions, and decoupling in a large cavity substantially reduces seismic coupling. Any answer that is a single figure without a region and a set of assumptions attached is wrong by construction.

If radioxenon is detected, does that prove a nuclear test?

No. Medical isotope production is the dominant global source of atmospheric radioxenon and produces far more of it than a small test would. Discrimination depends on isotopic ratios and on transport modelling that places the release in space and time, and both steps have real uncertainty.

Why does the treaty still not being in force matter to my analysis?

Because the on-site inspection regime, which is the mechanism for resolving an ambiguous detection on the ground, depends on entry into force and cannot be invoked. Verification therefore terminates at remote sensing, and a finding can be characterised but not confirmed by inspection.

Is the network useful for anything other than nuclear testing?

Substantially so. The same sensors detect large conventional explosions, bolides, volcanic eruptions, ocean events and atmospheric radioactivity generally, and the archives support a wide scientific literature. That is a genuine secondary value, not a marketing claim.

How do I check a claim that a country conducted a test?

Start with independently published seismic event solutions for the stated time and region, check the location uncertainty rather than the point, look for a corresponding institutional statement, and ask whether any radionuclide evidence has been reported and whether isotopic ratios were given. Most claims fail at one of those four steps.

Can I cite CTBTO press releases in a formal assessment?

Yes, as attributed institutional statements with their date. What you cannot do is describe them as verification data or reason quantitatively from figures that have been simplified for a general audience. Say what you have.

What changed when Russia revoked its ratification in 2023?

Russia remains a signatory and the practical effect on network operation has been limited, but it removed a ratification that mattered for entry into force and increased the political uncertainty around the regime. For an analyst it is a reason to treat continuity of data-sharing arrangements as a risk rather than an assumption.

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:

IMS station naming conventions, which encode technology and role in the station code and are the fastest way to know what a station can detect.
IDC product hierarchy – automatic Standard Event Lists, the analyst-Reviewed Event Bulletin, and screened bulletins – which defines the confidence attached to any event record.
Standard seismological data formats used across the field, including miniSEED for waveform data and the widely used bulletin exchange formats, which is what makes independent reanalysis possible at all.
FDSN station and network identifier conventions, the common vocabulary between the IMS and open seismological archives.
The Comprehensive Nuclear-Test-Ban Treaty itself, whose Protocol defines the verification regime including the IMS composition, the IDC's functions and the on-site inspection procedure.
Atmospheric transport modelling conventions used for source region estimation, where the meteorological input and release assumptions must accompany any published result.
STIX 2.1 location and report objects for export into the platform's common model, with the caveat that geophysical event data expressed in STIX loses the uncertainty structure that makes it meaningful.

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.

CTBTO Preparatory Commission — CTBTO. The organisation's site: network descriptions, technology explainers, public statements and the institutional context. The starting point and, for most users, the only accessible one.
United Nations Treaty Collection — United Nations. Authoritative signature and ratification status for the Treaty. Check it directly rather than relying on any summary, because the status has changed.
UN Office for Disarmament Affairs — United Nations. The wider disarmament framework in which the test ban sits, and the formal record of state positions.
USGS Earthquake Hazards Program — US Geological Survey. Open, programmatically accessible event solutions. The practical independent check on any claimed detection, available without any access arrangement.
International Seismological Centre — ISC. The definitive revised global bulletin. Slow, thorough, and the right reference for historical event parameters.
IRIS seismological data services — IRIS / EarthScope. Open waveform archives and standard web services – the accessible substrate for doing the geophysics yourself.
European-Mediterranean Seismological Centre — EMSC. Rapid independent event solutions with fast public dissemination, useful in the first hours after an event.
NORSAR — NORSAR. Deep published expertise in seismic array processing and explosion monitoring, including accessible explanations of the discrimination problem.
IAEA — International Atomic Energy Agency. The other half of the nuclear picture – safeguards, security and the incident record – and the body you will need alongside this one for any full assessment.
Nuclear Threat Initiative — NTI. Country-level nuclear programme profiles and analysis that provide the political and programmatic context around a detected event.

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 registers the CTBTO's public channel for what it actually is, records each detection announcement as a dated event with its stated uncertainty preserved, cross-checks it against independently published seismic solutions, and carries the access limitation through to every product that touches nuclear testing.. Browse the full source catalogue, or follow any tag above into the rest of the library.