Research on the Verifiability of the Comprehensive Nuclear-Test-Ban Treaty: An Overview of Work Supported by the Non-Proliferation, Arms Control and Disarmament Division of DFAIT

International Security Research and Outreach Programme
Non-Proliferation, Arms Control and Disarmament Division
International Security Bureau

INTRODUCTION

The challenge of verifying the Comprehensive Nuclear-Test-Ban Treaty (CNTBT) has featured as an important theme within Congressional and Parliamentary consideration of ratification of this treaty. The Canadian Department of Foreign Affairs and International Trade has supported a considerable amount of research aimed at addressing the issue of verifiability. This paper begins with an overview of the verification technologies and processes contained within the text of the CNTBT. It then reviews the research carried out by the Department’s Non-Proliferation, Arms Control and Disarmament Division, and considers, in light of two recent studies, the effectiveness of the CNTBT’s verification capabilities. The paper concludes by suggesting that while the verification of the CNTBT is an extremely complex task, it is achievable given existing verification technologies and methodologies as provided for by the Treaty.

THE CNTBT’s VERIFICATION PROCESS – A BRIEF OVERVIEW¹

The CNTBT was opened for signature on 24 September 1996. However, while a provisional secretariat has been established, many of the operational procedures of the treaty still remain to be finalized.

The basic obligations articulated in the text of the CNTBT state that the States Parties are required not to undertake nuclear weapon test explosions and to prevent such explosions at any place under their jurisdiction. Furthermore, they are to refrain from encouraging or participating in such explosions². The States Parties are also required to:

• establish International Monitoring System (IMS) facilities necessary for the effective monitoring of certain physical phenomena for evidence of nuclear explosions;

• provide data gleaned from national monitoring stations which are a part of the IMS, and ensure that raw data from nationally operated IMS stations are transmitted to the International Data Centre (IDC);

• participate in consultation and clarification activities;

• permit the conduct of on-site inspections; and

• participate in confidence-building measures.

In turn, the CNTBT Technical Secretariat (located within the CNTBT Organization, or CNTBTO) is required, among other things, to:

• provide technical support to the CNTBTO;

• negotiate arrangements or agreements with States Parties and international organizations regarding verification activities;

• offer assistance to the States Parties with regard to any verification issues raised under the Treaty; and

• generally assist the States Parties and the Executive Council in the implementation of the Treaty.

The CNTBT relies upon a variety of observables as evidence of a nuclear explosion. The principal observable is that of a seismic pattern resulting from a possible nuclear blast³. Other accepted observables detectable through remote sensors are: evidence of agreed-upon radionuclides (including noble gases such as Xenon), hydroacoustic disturbances, and infrasound patterns. The CNTBT also allows for the inclusion of data gleaned from National Technical Means (NTM), although the nature of the observables permitted remains unspecified. Importantly, the Treaty seeks to strengthen the ‘baseline’ for the assessment of observables by requiring States Parties to cooperate in an international exchange of seismological, radiological, hydroacoustic and infrasound observables.

The CNTBT also relies upon the presence of observables which can be detected through on-site inspections and overhead activities. These include radioactive gases, as well as passive and active seismological signatures from aftershocks in the case of underground tests. Other observables include physical objects which can be imaged through infrared and multispectral imagers, still or motion cameras, field glasses and the human eye. Aerial overflights are also permitted.

The procedural flow of data is shared by two distinct levels of authority; the States Parties, through their respective National Authorities, and the CNTBTO, particularly its Technical Secretariat.

States Parties may employ national technical means to augment IMS data used in a manner consistent with the principles of international law. States Parties are prohibited from such information collection methods where the use of such means is legally valid⁴.

Once the Treaty has come into force, the procedural flow of data within the Organization will involve three bodies: the Conference of the States Parties ("the Conference"), the Executive Council and the Technical Secretariat. Strictly speaking, the Conference and the Executive Council do not handle data – they simply make decisions concerning data. The Technical Secretariat is responsible for operating the International Data Centre (IDC) and the International Monitoring System (IMS). The Conference is composed of all the State Parties and comprises the principal organ of the Organization. It is tasked with the broad operation of the Treaty⁵. The Executive Council, designed to serve as the executive organ of the Organization, is to be composed of 51 members proposed by Conference members from each of a variety of geographic regions. The allocation of seats is designed to respect both geographic distribution as well as political and security interests. Elected by the Conference, the members of the Executive Council will be empowered, among other things, to supervise the activities of the Technical Secretariat, cooperate with the various National Authorities, and make recommendations for further consultative, clarifying or other activities⁶.

The heart of the data management efforts of the Organization, however, will lie within the Technical Secretariat. It will be responsible for, among other things, the supervision of the operation of the International Data Centre (IDC), the receipt, processing, analysis and reporting of International Monitoring System (IMS) data, the provision of technical assistance and support for monitoring stations through the IDC, the receipt for and processing of requests for on-site inspections, as well as their facilitation, and the arrangement for assistance to the States Parties through their respective National Authorities, where necessary⁷. The IDC will serve as the focal point for the collection and dissemination of data within the Technical Secretariat. The States Parties may communicate with the Technical Secretariat through their National Authorities and the Secretariat’s Director-General.

The designation of the IMS and the IDC as an ‘information clearing house’ for the States Parties has been done in part to address the problem of inequities in the relative abilities of States Parties to collect and compile data. The IDC is mandated to provide a variety of standard products to the States Parties, including integrated lists of signal detections, Event Lists and Bulletins, Standard Screened Event Bulletins, and executive summaries of archived data as well as IDC products. Services include data collection, authentication, quality control and storage, system monitoring, and product dissemination⁸. These products and services also neatly resolve the question of standardization, because all States Parties utilizing the data provided through the IDC will essentially be basing their conclusions upon the same data sets (excepting those with access to NTM).

The GSETT-3 experiment⁹ sought to develop a data collation system capable of receiving information from seismic as well as non-seismic sources, and suggested that any such system would have to possess a high degree of automation if it were to accomplish this task within both budgetary and temporal constraints. The experiment suggested that the system would also have to handle an incoming data flow of approximately 10 gigabytes per day, and that event lists could be made available to members within a few hours, and a bulletin issued within 48 hours¹⁰.

In some instances, States Parties may find that the data gathered through the IMS/IDC leaves certain questions unresolved. In such circumstances, they may request clarification or an on-site inspection. Where States Parties request an OSI, they must follow a predetermined timeline, as well as specified procedures for data collection.

While inspection teams can employ approved equipment, the inspected State Party does have the right to verify that all such equipment is indeed approved. The OSI team determines the means by which the inspection will proceed. This is done in a manner consistent with Treaty provisions, including those relating to managed access. In their effort to collect data concerning observables in order to verify compliance, inspectors may rely upon the following techniques:

• Position finding from the air and at the surface to confirm the boundaries of the inspection area and establish co-ordinates of locations therein, in support of the inspection activities;

• Visual observation, video and still photography and multi-spectral imaging, including infrared measurements, at and below the surface, and from the air, to search for anomalies or artifacts;

• Measurement of levels of radioactivity above, at and below the surface, using gamma radiation monitoring and energy resolution analysis from the air, and or under the surface, to search for and identify radiation anomalies;

• Environmental sampling an analysis of solids, liquids and gases from above, at and below the surface to detect anomalies;

• Passive seismological monitoring for aftershocks to localize the search area and facilitate determination of the nature of an event;

• Resonance seismometry and active seismic surveys to search for and locate underground anomalies, including cavities and rubble zones;

• Magnetic and gravitational field mapping, ground penetrating radar and electrical conductivity measurements at the surface and from the air, as appropriate, to detect anomalies or artifacts; and

• Drilling to obtain radioactive samples¹¹.

Inspection teams may also avail themselves of information gathered from overflights of slow fixed or rotary wing aircraft. The Treaty permits the use of field glasses, passive location- finding equipment, video cameras and hand-held still cameras from such aircraft: teams may also use installed sensors for multi-spectral and infrared imagery, gamma spectroscopy and magnetic field mapping.

All inspection teams are required to observe provisions of managed access for the purpose of protecting equipment and areas deemed confidential by the inspected State Party¹².

Finally, inspection reports must contain the following information:

• A description of the activities conducted by the team;

• The factual findings of the inspection team relevant to the purpose of the inspection;

• An account of the cooperation granted during the on-site inspection;

• A factual description of the extent of the access granted, including the alternative means provided to the team, during the on-site inspection; and

• Any other details relevant to the purpose of the inspection¹³.

Unlike the procedural aspect of the CNTBTO, the analytical component of this Treaty lies predominantly with the States Parties. While the IDC "[applies] on a routine basis automatic processing methods and interactive human analysis to raw [IMS] data in order to produce and archive standard [IDC] products on behalf of all States Parties", the IDC is not permitted to offer any conclusions which might prejudice the final judgements of these members¹⁴. It should be noted that the Treaty makes efforts to ameliorate the problem of inequitable analytical capabilities among the States Parties by offering, at no cost, expert technical analysis of IMS data in order to help the members identify the source of an event. The IDC will, however, be placing a ceiling upon the proportion of resources devoted to such assistance efforts, with possibly no more than 10% of IDC resources being given to such efforts and with such activities prioritized by receipt of request, the number of requests received from a particular country , and the estimated level of effort¹⁵.

The Treaty allows for the use of NTM by individual States Parties. This goes beyond such remote sensors as satellites, seismic arrays, radionuclide stations and other hardware to also include communications interception technologies and human intelligence. It also involves national data assessment capabilities. Considerable tension existed during the negotiations for the CNTBT over the role of NTM with regard to data analysis/collection and the decision as to what constitutes evidence: the Treaty infers that such analysis will not be carried out in a manner prejudicial to state sovereignty. Moreover, there seems to exist a sentiment by some states that data gleaned through the process of the IMS was more credible in supporting a request by some states for an on-site inspection (OSI), and that NTM activities would be relegated to a secondary place in the hierarchy of data. The Treaty specifies that the IDC may call upon data from "cooperating national facilities"¹⁶. Little substantive literature exists concerning the manner by which NTM data given to the CNTBTO would be protected. The common presumption is that such information would be degraded prior to presentation to a level sufficient to defend the position of the State Party without compromising secrecy concerning NTM capabilities. No clear discussion exists on the manner in which data gleaned through NTM would be integrated into either the work of the IDC or a request for an OSI.

Where analysis suggests possible non-compliant behaviour on the part of other signatories, a State Party may proceed in a variety of ways. First, and without prejudice to its right to request an OSI, a State Party may request a clarification with regard to the event in question. Where such a request is made from one State Party to another, a clarification must be provided within 48 hours; if the request is forwarded through the Executive Council and the Director-General (who may be asked to assist in this matter, possibly through the provision of appropriate technical information from the Technical Secretariat), it must be forwarded no later than 24 hours upon its receipt and an clarification must again be given within 48 hours. Where a State Party considers the clarification unsatisfactory, it may request further information from the Executive Councilor a meeting of the Executive Council for the purpose of discussing the situation and possibly carrying out various protests or other measures¹⁷.

Alternatively, a State Party could request an on-site inspection (OSI). Such inspections, designed solely to clarify whether a nuclear explosion had taken place, would be based on "information collected by the International Monitoring System, on any relevant technical information obtained by national technical means of verification in a manner consistent with generally recognized principles of international law, or a combination thereof"¹⁸. The timelines of an OSI are quite rigorous¹⁹. All such requests for OSI must be presented to the Executive Council, which must in turn acknowledge receipt of the request within two hours and communicate the request to the State Party sought to be inspected within six hours. The CNTBTO Director General is responsible for ensuring that the request meets the requirements for a OSI request and then communicates that request to the Executive Council and all the other State Parties within 24 hours. Any clarifications of the State Party receiving the request must be received within 72 hours. The decision to proceed with the OSI rests with the Executive Council, following at least 30 affirmative votes, and must be made no later than 96 hours after receipt of the original request.

The contents of a draft inspection report must be available to the inspected State Party for comments, which may, within 48 hours, give explanations and identify information it feels is not relevant to the inspection. The Director-General may consider any proposals for change be made by the inspected State Party and will amend its comments. The Executive Council is given the authority to decide, in light of the OSI report, whether or not a State Party has acted in non-compliance with the terms of the Treaty or whether the right to request and OSI has been abused (considering always its requirement to consult with the States Parties involved)²⁰; the Conference, taking into account the recommendations of the Executive Council, may take appropriate measures in the event of evidence of non-compliance.

A SURVEY OF ISROP LITERATURE CONCERNING THE VERIFICATION OF THE CNTBT

During the negotiations for a CNTBT text, Canada’s position was informed by the following concerns:

1. That the C[N]TBT text properly address both its disarmament and non-proliferation dimensions;
2. That the C[N]TBT should be comprehensive (i.e., it should ban all tests in all environments for all time);
3. That the C[N]TBT be as universal as possible, providing adherence by NPT and non-NPT parties;
4. That the C[N]TBT should be credible and effective in scope (i.e., in addressing testing should it cover related actions such as site preparation and, if so, how?);
5. That the C[N]TBT must have an effective international verification mechanism; and
6. That the C[N]TBT must be a binding legal instrument with the requisite compliance provisions."²¹

The creation of an effective international verification mechanism would prove to be a central challenge: it soon became clear that any effort create just such a mechanism (such as the one described in the previous section) would have to grapple with two interrelated questions:

1. Can the technologies described above actually detect and provide data for the analysis of the aforementioned observables used to determine compliance?
2. Can the procedural structure described above facilitate the transmission of information in such a way as to allow for the accurate and timely dissemination of data?

Even before the initiation of negotiations once again for a CNTBT text in November of 1993, a great deal of preliminary research was carried out by the Group of Scientific Experts through their GSETT-1 and GSETT-2 experiments (held in 1984 and 1991 respectively)²². Their work tested the utility and effectiveness of an international seismic monitoring network in support of such a Treaty.

The Department of Foreign Affairs and International Trade, through the research wing of the Non-Proliferation, Arms Control and Disarmament Division, also supported research into the utility of seismology for verifying the CNTBT. In July of 1991, the division (then referred to as the Arms Control and Disarmament Division) published a study entitled Nuclear Test Ban Verification: Recent Canadian Research in Forensic Seismology²³. The study examined, in part, the physical basis for earthquake/explosion source discrimination and explosion yield determination, the technical problems pertaining to seismic monitoring of underground nuclear tests, and the basic problem solving strategy deployed by the forensic seismology research team at the University of Toronto. The study’s data source was based upon teleseismic verification using P wave recordings from both the old and the recently refurbished Yellowknife Seismic Array and regional (close-in) verification using high frequency L_g and P_n recordings from the Eastern Canada Telemetered Network (a group of stations installed to record earthquakes); its research focussed upon the propagation, attenuation, geological distortion of these two seismic waves.

The study’s analysis indicated that, given the large body of historical nuclear explosion data from the Yellowknife Seismic Array, it was possible to make accurate characterizations of P wave attenuation effects along the paths linking the Canadian listening post with seven active nuclear testing areas. Furthermore, it was demonstrated that the revised P wave attenuation results gleaned from this research led to more interpretable nuclear explosion source functions, including instances of successful unmasking of pP (an elusive depth phase of great importance in seismic source identification and explosion yield estimation. This investigation of the epicentral locations of more than 7,000 earthquakes also produced a clear picture of region-specific location correction terms for the Yellowknife Seismic Array. These correction terms, obtained using conventional frequency-wavenumber analysis of vertical-component data, would enable the array to improve the accuracy of its preliminary epicentral determinations for early detection of potentially anomalous seismic events. Finally, an important implication from this network calibration work was that the gap between seismic detection threshold (m_b 2.0-2.5) and identification threshold (m_b 3.5) in low-attenuation regions such as the USSR, is closing, and that a lowering of the identification threshold by less than 0.5 m_b unit would appear to be sufficient to permit monitoring of well-coupled explosions with yields down to a subkiloton level, and decoupled explosions with yields down to a level below 5 kilotons.

The Arms Control and Disarmament Division has also supported research into non-seismic means of verifying a CNTBT. A May 1993 study entitled Non-Seismic Technologies in Support of a Nuclear Test Ban²⁴ examined a variety of non-seismic techniques, including overhead surveillance, chemical detection techniques, and radionuclide sampling activities.

Overhead surveillance comprises a variety of technologies. Spaceborne imagery (imagery acquired by satellite sensors) could rely on the capabilities of a number of commercial satellite systems, including the American Landsat, the French SPOT (Systeme Pour l’Observation de la Terre)or Russian satellite sources. Such images could be used to examine ground-level construction and human activity in large swaths over areas of interest (in either an anticipatory or reactive capacity), using the following variety of technologies: visible or near infrared imaging (that portion of the electromagnetic spectrum, ranging from 0.4 to 2 microns, which can record using reflected energy in either a single band or many bands to create a "multispectral" image); far infrared (also referred to as thermal infrared – this technology measures small variations in radiated energy, in the order of 0.2 degrees Celcius, and is mostly conducted at night to avoid the ‘thermal blanket’ created by normal daytime solar activity); and radar (using both real aperture and synthetic aperture technologies to illuminate an area with microwave pulses and then measure the small variations in reflected energy). Aerial imagery (gleaned from similar sensors attached to fixed or rotary-wing aircraft) could provide similar data; its chief advantage over spaceborne imagery is the greater flexibility of aircraft in overflying selected sites; however, it is more intrusive than satellite imaging efforts²⁵.

The following table indicates some of the uses of overhead imagery and an assessment of their utility to the CNTBT. Early warning refers to that information which could provide information relating to the preparation of a text. Historical data refers to that information which can provide a long-term historical record of known nuclear testing sites.

Taken from Non-Seismic Technologies in Support of a Nuclear Test Ban, p. 12.

Chemical detection techniques would seek to determine whether geochemical signatures are retained in the surface soils above a known or suspected nuclear test site. An April 1996 study entitled Nuclear Test Ban Verification: Geochemical Signatures in Soils to Detect Underground Nuclear Events²⁷ confirmed that significant geochemical anomalies are created by an underground nuclear test, with partial leaches targeting the amorphous Fe oxyhydroxide (including Mn oxide) phase and the Mn oxide (alone) phase used. Furthermore, samples from all test sites showed distinct geochemical anomalies in a wide variety of elements²⁸. However, it was subsequently decided by the negotiators that the CNTBT would not avail itself of this technology.

The utility of radionuclide sampling was considered in the April 1996 study using a tracer modelling methodology. This technique is well established in meteorological circles as a consequence of efforts to simulate the atmospheric transport of radioactive materials following the 1996 Chernobyl accident. This study concluded that efforts to alter this model, so that it would be integrated backward in time to reconstruct the location, time and strength of the source using initial conditions provided by measurements of radioactivity would remain an extremely challenging prospect. Instead, it was recommended that the surveillance of vented radionuclides from nuclear tests into the atmosphere could be accomplished through the use of sampling stations positioned around the globe.

Research also indicated that there was a strong role for on-site inspection activities within a CNTBT. A paper²⁹ suggested that such techniques had a role to play in the CNTBT, one which would extrapolate from the TTBT and PNET experiences. The author noted that

"For the TTBT, the verifying (inspecting party has the right to carry out:

• Direct, on-site hydrodynamic yield measurements (CORRTEX) in a ‘satellite’ hole for all nuclear tests declared in advance to have a planned yield over 50 kilotons;

• Collection of regional seismic data from three Designated Seismic Stations (DSS) on the territory of the testing side for all tests declared to have a planned yield over 50 kt. (The U.S. established its DSS near existing seismic facilities in the Russian Federation at Obninsk, Arti, and Novosibirsk. The Soviet, now Russian, DSS were set up in Newport, Washington, in the Black Hills of South Dakota, and in Tulsa County, Oklahoma);

• On-site confirmation of the geology near an emplaced underground nuclear device and the emplacement conditions, and observation of the actual emplacement process for all tests declared to have a planned yield greater than 35 kt;

• Hydrodynamic yield measurements of ‘reference tests’ conducted in so-called ‘non-standard’ conditions;

• Hydrodynamic yield measurements on a minimum number of nuclear tests, regardless of the declared planned yield, at existing test sites and any future test sites; and

• Hydrodynamic yield measurements on each detonation in a test series if the planned aggregate yield exceeds 50 kt.

For the PNET, the verifying party has the right to carry out:

• the same hydrodynamic yield measurements and on-site inspection activities as with TTBT monitoring for peaceful nuclear explosions (PNEs) with planned yields of 50 kt and 35 kt, respectively; and

• data collection from a temporarily installed local seismic network around the site of the PNE to confirm the number of individual explosions in a group PNE where the planned aggregate yield of the explosions exceeds 150 kt."³⁰

This paper suggested that there were a host of activities which could be carried out by on-site inspection: they included invitational inspections to demonstrate that a large planned inspection was not nuclear, as well as ‘after the fact’ inspections to look for such tell-tale observables as local radionuclides, unusual anomalies in groundwater displacement patterns, and atypical and localized seismic activity: the could also potentially carry out various drilling efforts to look for the physical detritus of a nuclear explosion.

An explicit assumption underpinning support for the research described above was the belief that the CNTBT would benefit from verification synergies – that is, the interaction of different verification technologies and methodologies which, when carried out together, could produce results which would be greater than the sum of their undertakings on an independent basis. Such efforts would comprise the general detection of an event through seismic means, and then the ‘drilling down’ by the use of the other non-seismic technologies to identify the geographical location of the suspect site.

Such a synergy would occur in the following manner: first, seismic detectors would locate an anomalous reading, and identify a location within an 100 km² location, within 24 hours. Next, overhead imagery could further narrow down that location to roughly 25 km², over a space of roughly one week. Finally, in the event of an actual test, on-site inspection methods could pin-point the location of the anomaly over the space of 2-5 weeks³¹.

All of the above discussion has focussed on the appropriate verification technologies for the CNTBT: however, considerable attention has also been given to those institutional procedures needed to give shape to and support these verification activities.

This issue was first addressed by the research unit of the then Arms Control Division within the September 1993 Verification Backgrounder; its suggested that any process flow would have to include the following steps:

"1. The event is detected through the international seismic monitoring network and data is passed to an analysis centre;

2. Data from overhead sensors and from other means including national means are incorporated;

3. The analysis centre undertakes the collection, collation and synthesis of the information; accepts input from other member states and from the alleged violator; and provides the decision makers with a recommendation;

4. The decision to initiate an OSI includes member state participation/observation and hosting by the alleged violator;

5. The verification decision and follow-on action, if any."³²

In particular, this study emphasized the need for flexibility in the amount of analysis to be accomplished centrally, and noted that while the verification decision remained a national one there should be international agreement to the degree possible.

Subsequent research refined this idea by identifying the following steps which any verification process would have to pass through:

Process

Step 1. Anomaly Detection. Through the gathering and processing of data by means of the [International Seismic Monitoring System, ISMS], anomalies indicative of a possible nuclear test are identified, thus triggering the verification process. The [International Data Centre, or IDC] provides data to the Authority’s Analysis Division.

Step 2. Supporting Methods Employed. The Authority collates data with additional information provided by other means (e.g., detection of radionuclides) and national inputs from member states, as well as from ancillary means, including commercially available overhead imagery.

Step 3. Initial Analysis. In most cases regarding test explosions, analysis would be accomplished in the Technical/Analysis Division (TAD) of the International Authority (IA). The level of analysis remains subject for resolution.

Step 4. Assessment Forwarded. Analysis developed by TAD is forwarded immediately to the IA Secretariat (IAS).

Step 5. Consultation and OSI Decision. The IAS initiates a consultation process which provides a second opportunity for national inputs from member states as well as from suspected violators. The IAS collates data, consults and decides on the need for an on-site inspection.

Step 6. Inspection Deployment. If a decision is taken that an OSI is required for further precision and confirmation, the IAS initiates an OSI with national support from member states (i.e., provision of specialists, supporting staff, and access to advanced technologies) and which is facilitated by the suspected violator.

Step 7. Final Evaluation and Report. The final evaluation will be developed by the IAS and TAD with OSI data. The report is forwarded to the Executive Council.

Decision-Making Scenario

Step 8. Executive Council Action. Without prejudice to the rights of individual members to initiate action (see Step 10 below), the Executive Council may either conclude its investigation ore pursue further action in accordance with treaty provisions. It will also provide a report to member states as indicated immediately below.

Step 9. Member States Informed. Since the CTBT is envisaged as an open process, member states will be fully informed regarding the International Authority’s activities either directly or through a CTBT collective mechanism.

Step 10. Member State Action. Should a violation of the treaty be judged by a member state (in its individual capacity) to have occurred, the matter will be forwarded by the member state to the UN Security Council for consideration;

Step 11. Collective Action. Collective action could include sanctions or other restrictive actions³³.

In summary, the paper concluded that this verification process would require the development of an agreed-upon baseline capacity, and use anomaly detection as its trigger mechanism. However, further research on the subject of decision-making and its applicability to the CNTBT process (using the rationalist and formal methodology of decision theory, coupled with non-cooperative game theory)³⁴ offer some words of caution as to the limits of this verification process. When faced with the option of complying with or violating the terms of a treaty, each state must weigh two calculi: the loss experienced by a detected violation versus the benefit accrued by an undetected violation. – something termed the state’s value ratio. Calculations of this value ration must include the likelihood of each possible consequence – in effect, a political gauging of the state’s willingness to violate a treaty:

"Increasing the state’s value ratio, either by increasing the penalty following detection or reducing the benefits of undetected illegal behaviour, decreases the threshold level of inspection effectiveness needed to deter violation. Conversely, if the state discovers it has less to fear from a detected violation, or more to gain by violating, then the level of inspection effectiveness sufficient for deterrence increases."³⁵

It is clear that a successful verification process must rely upon the successful relative balance of the value ratio (an essentially political parameter, determined by the state leadership’s political preferences and goals), and the effectiveness of the verification process and technologies (a technical parameter, determined by the allocation of resources and intellectual capacity) in favour of compliance. However, new political incentives exogenous to this balance can always threaten to tip the scales, something which any verification effort must remain fully cognizant of.

IS THE CTBT VERIFIABLE?

It is interesting to observe that much of what was called for in the research above (together with research submitted from other Canadian government departments) is now reflected in the language of the CNTBT. However, it is one thing to have developed what would appear to be a viable verification process and complementary technologies on paper; the real test must lie in its application. Unfortunately, the Treaty remains unratified by certain crucial states, and so has not yet entered into force. The track record, however, of even the partial IMS in detecting relevant seismic events is impressive.

By way of documenting this, the Non-Proliferation, Arms Control and Disarmament Division published in December 1998 a study entitled CTBT Verification Related Case Studies of Three Recent Seismic Events: Novaya Zemlya, India and Pakistan: it examined the Novaya Zemlya seismic event of August 16, 1997, the Indian nuclear explosions of May 11 and May 13 1998, and the Pakistani nuclear explosions of May 28 and May 30, 1998. Regarding the Novaya Zemlya event, the study noted that despite the limited data available (the IMS system was not fully assembled, and the nearest IMS station to the event was not operational), seismologists were still able to conclude that the original event had the waveform characteristics of other earthquakes which had occurred in the same area. Its magnitude was estimated at 3.5 (a detection result much better than the system’s advertised identification threshold of magnitude 4), with about 50 to 100 tonnes equivalent yield, and was identified as having occurred under the floor of the Kara Sea³⁶.

On May 11, 1998, the Indian government announced that it had detonated three simultaneous underground nuclear devices: a fission device (with a claimed magnitude of 12 kt), a thermonuclear device (with a claimed yield of 43 kt), and a low-yield device (with a claimed magnitude of 0.2 kt). On May 13 it then announced that it had set off two other simultaneous low-yield explosions (0.3 kt and 0.5 kt respectively) in sand dunes at the same test site. It should be noted that American NTM failed to notice the Indian preparations, although the Pakistani efforts were detected. It is suggested that this might be due to the fact that the Indian shafts had already been dug, and that the Indian government took special pains to hid their efforts from American NTM.

The signals from the May 11 explosions were readily detected by seismic stations around the world, with an indicated location close to the test area identified by the Indian government. The magnitude estimated by the prototype International Data Centre (pIDC) was 4.7: the US Geological Survey (USGS) estimated it to be 5.3. The yield for the May 11 explosions has been estimated to be roughly 12 kt: no evidence was found to suggest separate explosions. No evidence for the May 13 tests has been detected by seismologists outside of India, even though the Nilore station was operating at the time: if the Indian claims as to yield were accurate, then the explosion – even if it was well coupled – would have most certainly been detected.

On May 28, the Pakistani government responded with five nuclear explosions; it claimed the total yield to be 40-45 kt. The explosions were recorded by 65 stations from the pIDC, with a magnitude of 4.6 (indicating a yield of about 10 kt). The USGS reported an m_b of about 4.8, while geological Survey of Canada estimated the yields to be 10-15 kt. The May 30 event was recorded by 51 stations used by the pIDC with a magnitude of 4.3, indicating a yield of about 5 kt.

In these three cases, the pIDC – using both IMS and non-IMS seismic data – demonstrated an ability to detect events down to about m_b = 3 or below, an order of magnitude below its design threshold (a capability predicted in the aforementioned forensic seismology study). With its full complement of stations and networks operating, including those of non-seismic technologies, it can be reasonably be expected that the pIDC (and eventually the IDC) would be able to do an even better job, especially if non-IMS data continues to be used. However, these experiences also indicated that it would be unwise for the CNTBT to ignore other forms of data available to it, including remote sensing data. While the failure of the American NTM to detect the Indian activity prior to the test should be seen as simply just that – a failure of NTM, it does point to the limitations of relying solely upon overhead imagery or using such imagery independently of seismic triggers. Rather, such imagery seemed to operate more effectively when used to investigate anomalous seismic events – something which would seem to support the principle of verification synergy also discussed above.

The Novaya Zemlya experience also points to the salutary ability of the pIDC network to address a potentially volatile issue before it became politicized. In short, even in its incomplete state, the pIDC is currently making a meaningful contribution to international security. Finally, it should be noted that criticisms that the international community’s inability to forestall the Indian tests or dissuade the Pakistanis from responding in a like manner are unfounded: as the research on decision-making indicated, these are political issues, and stand apart from the verification challenges addressed by the Treaty. At best, they suggest that it is perhaps too much to ask of a verification regime to claim that its anticipatory use might inhibit proliferation alone. Rather, such mechanisms should simply be used as triggers, with subsequent political pressure serving as the retributive deterrent.

Looking beyond the vicissitudes of verification technologies themselves, it is also possible to make a few observations concerning the verification process established within the Treaty’s text (and summarized in the first section). In a December 1998 study which comparatively assessed the verification processes of a number of arms control regimes³⁷, it was suggested that all successful arms control regimes shared a number of structural similarities: the use of previous experimentation to work out ‘bugs’; the appropriate integration of automated monitoring technologies with working groups of personnel in a time- and data-efficient manner; the standardization of data between information sharers; an emphasis upon technological simplicity; and the use of consultative mechanisms to address technical questions before they become political ones. The majority of these issues were raised in the research paper describing what might comprise an effective CNTBT verification process; as can be seen from the first section, they are also all reflected in the verification process outlined within the text of the present Treaty – something which augers well for the future of the CNTBT.

CONCLUSION

As the above sections have discussed, the challenge of verifying the CNTBT comprises two interrelated tasks: incorporating the appropriate verification technologies, and ensconcing them within an effective structure or verification process. The research carried out by the research arms of the Department of Foreign Affairs and International Trade’s Non-Proliferation, Arms Control and Disarmament Division has identified a number of verification technologies and processes which could be used by a CNTBT; many of these have subsequently been incorporated into the text of the Treaty. Finally, as the last two studies demonstrate, there exists evidence to suggest that not only is the present incomplete pIDC functioning above expectations, but the verification process for the CNTBT itself mirrors the success stories of other complex verification systems. This in turn suggests that the CNTBT is an eminently verifiable Treaty, and should be brought into force so that it can make a substantive contribution to international security.