SPECIAL REPORT: Major Exercise Tests CTBT On-Site Inspections

Oliver Meier

SEMIPALATINSK, KAZAKHSTAN — It was a clear and sunny day when the earth shook in Arcania. Several seismic stations that are part of the International Monitoring System (IMS) that is monitoring compliance with the Comprehensive Test Ban Treaty (CTBT) picked up the event a short time later and transmitted the data in near real time to Vienna. There, the International Data Center of the Comprehensive Test Ban Treaty Organization (CTBTO) determined that the seismic event on Aug. 22 had a magnitude of 4.05 and placed it in the middle of the so-called Barrier Zone, where Arcania had conducted several nuclear test explosions in the 1970s and 1980s. Arcania claimed to have closed its nuclear test site in 1989 and stated that an earthquake triggered the seismic network. But Arcania’s neighboring state and regional competitor, Fiducia, remained apprehensive, not least because it had obtained its own information about suspicious activities prior to Aug. 22 in the Barrier Zone. Because both states are CTBT members, Fiducia requested an on-site inspection (OSI) to clarify what happened in Arcania on Aug. 22. That request was granted, and a team of international inspectors from 22 different countries began assembling immediately in Vienna.

This scenario provided part of the background information that was given to 40 international inspectors who arrived on Sept. 1 in Almaty, Kazakhstan, which doubled as the fictional Arcania’s capital city, Utopium. The team was taking part in the Integrated Field Exercise 2008 (IFE08), a large-scale simulation organized by the Provision Technical Secretariat (PTS) of the CTBTO to test crucial elements of the treaty’s OSI provisions. Once the CTBT has entered into force, such challenge inspections can be requested by any state-party to clarify whether a suspicious event was indeed a violation of the ban on nuclear tests.

Prior to the IFE08, the PTS had run several so-called directed exercises to test specific components of the OSI regime. The IFE08 was different. The month-long simulation was the largest and longest of its kind to date. Two hundred participants and observers traveled to the base of operations, at a remote location on the former Soviet nuclear test site Semipalatinsk. About 6,000 man-days were spent in the field, which represents a sizable portion of the annual monitoring activities of much larger organizations such as the International Atomic Energy Agency (IAEA) or the Organization for the Prohibition of Chemical Weapons (OPCW). Fifty tons of equipment had to be transported to the base (and back), and inspectors applied 10 different inspection technologies. All in all, the PTS spent about $6 million on the simulation, not including contributions of 672,000 euros (around $900,000) from the European Union and equipment from eight states.

Most importantly, the inspection was unprecedented because virtually all OSI components were tested in combination. “This was not a trivial exercise in terms of its parameters,” Tibor Tóth, the executive secretary of the PTS, concluded in an Oct. 9 Arms Control Today interview.

The inspectors’ assigned task was to find out what had really happened on Aug. 22. The exercise itself had a few more fundamental and important purposes, including a test of inspection equipment under realistic conditions, the training of future inspectors, and a trial of operational procedures. The IFE08 also was supposed to send a political signal that more than 10 years of OSI preparations have paid off and the organization will be ready to conduct an on-site inspection if and when the CTBT enters into force.

Assessing Logistics and Equipment

The IFE08 took place as significant progress has been made in the last few years in establishing the IMS, which has the primary task of identifying and locating suspicious events that might be indicative of a nuclear test explosion. When the IMS is complete there will be 321 monitoring stations and 16 radionuclide laboratories located in 89 countries. Because the CTBT prohibits nuclear test explosions in all environments—underground, in the atmosphere, or in the oceans—the system applies four different monitoring technologies and has comprehensive global coverage. A total of 170 seismic stations will monitor for the distinctive tremors produced by a nuclear test explosion, 60 infrasound stations will be able to detect sound waves typical for atmospheric tests, and 11 hydroacoustic stations will monitor the oceans. There will be 80 radionuclide stations capable of detecting telltale signatures of fission products produced by a nuclear explosion. As of Sept. 12, 232 IMS stations were certified, 31 were in a testing mode, and 36 were under construction. The PTS estimates that 85 to 90 percent of all IMS stations will be completed by 2009. Progress toward completion of the network of stations will depend on the degree of cooperation of the handful of countries that have not yet signed or ratified the treaty. Although not yet fully completed, the IMS is already operating and delivering data. For instance, numerous IMS stations detected North Korea’s Oct. 9, 2006, nuclear test, which was a relatively small yield event of only 0.6 kilotons TNT equivalent. (See ACT, October 2008.)

On-site inspections are intended to provide an additional, crucial layer of security to the verification system. If an ambiguous event is detected, any state-party may request an on-site inspection to seek clarification. In addition to information collected through the IMS, a requesting state may put its own evidence on the table to support its case that a state may have conducted a nuclear test. A positive vote of at least 30 of the 51 states that are members of the CTBTO’s Executive Council is required for an inspection to proceed. Because evidence from a clandestine nuclear test may disappear quickly, time is of the essence. A decision on a proposed inspection will have to be made no more than 96 hours after an inspection request was filed, and the director-general will then launch the inspection as quickly as possible. No more than 72 hours after arrival at the point of entry in the inspected state, inspection activities must begin.

A real on-site inspection can only be requested after the CTBT has become legally binding. But the treaty’s entry into force still depends on ratification by nine specific states, including nuclear-weapon possessors China, Israel, and the United States, which have signed but not ratified the treaty, as well as India, North Korea, and Pakistan, which have not even signed the accord. Thus, for the time being such inspections remain a theoretical possibility. Nonetheless, the CTBT requires the verification regime to be “capable of meeting the verification requirements of this treaty” when the accord enters into force. Therefore, preparations for the inspection regime have been ongoing since 1997.

Inspection team leader Wang Jun, in an Oct. 14 interview with Arms Control Today, called the fact that the exercise was conducted “a tremendous success, especially for a multilateral verification regime that is still in its build-up phase.” Organizers said the exercise put most components of an OSI to a realistic test, and with a few exceptions, inspection equipment met or outperformed expectations. Wang pointed out that the IFE08 marks the first time that a passive seismic detection network with mini-arrays, as opposed to single sensor stations, was used in an OSI exercise, and he said it performed extremely well. The teams deployed 28 such mini-arrays to monitor for aftershocks that could be indicative of the collapse of cavities created by underground nuclear blasts.

“These sensors covered about 80 to 90 percent of the inspection area, and the covered areas were highly sensitized,” Wang said. The system demonstrated its capability when it picked up and located a number of small conventional explosions that were apparently set off by locals at the margins of the inspection area on Sept. 10-11, even though the equivalent yield of the blasts was only on the order of kilograms of dynamite. Radionuclide sampling and gamma radiation surveys conducted from the ground and the air confirmed the real radioactive background survey results conducted just before the exercise. In the continuation period, inspectors applied techniques, especially aerial magnetic field mapping, to identify anomalies in the underground geological structures, which over real test sites produced very good results. “We now have full confidence that these inspection techniques would work” in a real inspection, Wang concluded, adding, “Had there been OSI-relevant radionuclides present, we would have found them.”

Some of the equipment had been donated by member states just before the exercise, creating the additional challenge for inspectors of using unfamiliar instruments. For example, the Chinese government had donated a mobile detector for argon-37, a radionuclide that does not occur naturally but can be found in the aftermath of a nuclear test. This was the first time the detector was operated in an inspection environment, and inspectors praised its performance.

Naturally, not all aspects of an on-site inspection can be simulated in an exercise, even a large one. Some components may be too intrusive to be applied at a former nuclear weapons test site; others, too expensive to be used for testing purposes. Thus, the treaty foresees the possibility of drilling in order to confirm the presence of radionuclides underground, which would provide indisputable proof of an underground nuclear explosion. At a cost of several million dollars per hole, however, drilling was ruled out for the IFE08. Other technologies could not be applied for safety and security reasons. For example, the Kazakh inspection hosts, possibly with the backing of former test site operators in Russia, insisted that soil samples could not be taken. Similarly, air samples could be collected but were not to be analyzed. Although the procedures for taking and analyzing of such samples were successfully tested, the results of this analysis had to be fabricated by scenario controllers. Harmonizing real and artificial data created additional difficulties.

The success of a future on-site inspection will also depend on the reliability of basic gear such as tents, toilets, and generators. The test of such equipment turned out to be more demanding than many had expected. The Kazakh steppe is a harsh environment, and the base was a minimum four-hour drive away from the next town, the formerly closed nuclear city of Kurchatov. Strong winds, rain, and icy temperatures made setting up the base of operations an unexpected challenge and delayed the beginning of some inspection activities, such as the initial overflight over the inspection area, for a few days.

Under such extreme circumstances, it quickly became clear which parts of the equipment worked well and which did not. The tents, donated by the European Union, were praised all around. Indeed, one of the Kazakh hosts inquired whether the team would not like to leave several of them behind. Fancy electrical toilets, built to incinerate waste and thus meet the most stringent environmental standards (euphemistically named Cinderella), were unreliable but at least turned out to be an inexhaustible source for jokes among exercise participants. While rugged Kazakh generators hummed along happily under any conditions, modern computer-controlled generators brought in for the exercise choked repeatedly on impurities in Kazakh diesel fuel, bringing life in the camp to an effective halt because of the sudden interruption in the electricity supply.

Training the Inspectors

Because on-site inspections are likely to be rare events, treaty negotiators decided that, unlike the IAEA and OPCW, the CTBTO will not have a ready, standing inspectorate. Instead, the future organization’s director-general would be able to draw from a pool of trained inspectors to form a team of up to 40 inspectors once the request for an OSI has been granted. One essential element of the exercise was therefore to train potential recruits under realistic conditions.

There is hardly a better training site for CTBTO inspectors than a former nuclear test site. The Soviet Union conducted 456 test explosions between 1949 and 1989 at Semipalatinsk, including 340 underground tests. Most of the boreholes used for underground nuclear tests are accessible, and about 30 were located inside the mock inspection area, providing inspectors with a unique opportunity to check how instruments would pick up the characteristics of underground tests, such as cavities created by nuclear detonations. The real possibility of contamination also contributed to a realistic test of health and safety precautions.

Semipalatinsk was also an ideal location to practice inspections because of its size and relative inaccessibility. The inspection area for the exercise measured 25 kilometers by 40 kilometers, thus covering the maximum space permitted under the CTBT of 1,000 square kilometers. The immense size of the area was compounded by the virtual absence of roads, making travel difficult and taking valuable time out of the maximum six hours that Kazakhstan for safety reasons allowed inspectors to spend in the field every day.

“We were pressed for time and extended for space,” inspection team leader Wang complained. He referred to the fact that, under the treaty, the inspection period can last up to 60 days after Executive Council approval of the OSI plus a possible extension of another 70 days but in the exercise that period had been artificially shortened to just 23 days, out of which field activities were conducted on 17 days.

The purpose of the initial inspection period, which under the treaty can last up to 25 days from the day that the Executive Council has authorized the inspection, is to narrow down the search area and identify sites of particular interest. The treaty foresees that techniques to cover large areas, such as visual observation, should be used initially. Thus, inspectors conducted overflights in an MI-8 helicopter and a radiation survey from the air. Based on the findings of these wide-area searches, inspectors were sent off to specific locations on the ground to conduct such tasks as taking environmental samples.

Narrowing down the inspection area turned out to be more difficult than expected. The team was not able to detect radionuclides indicative of a possible nuclear explosion, but an underground test does not necessarily vent such materials. Moreover, the progress inspection report (PIR), transmitted to the CTBTO’s director-general Sept. 12, listed several possible indicators of prohibited activities. For example, inspectors observed trucks and trailers near a construction site and a borehole. What were these doing at a site with no construction activities? At a different location, the team saw a scraped area. Was this an attempt by Arcania to hide evidence of prohibited activities? Inspectors also observed an orange trailer and tent, which had no obvious purpose at a remote location. Also, it found a new-looking concrete platform of about four meters in diameter covering a borehole.

In fact, these artifacts were several of a dozen anomalies within the inspection area that could be indicative of activities that the inspectors were supposed to identify during the exercise. The identification of these hints in the PIR was in itself a success. The inspectors, however, were frustrated because they were unable to see whether these “artifacts” were simply anomalies or represented a consistent pattern. Because of this and because the seismic survey did not enable them to reach any firm conclusions about the likely location of the triggering event, the inspectors refused to exclude certain areas for inspection in the PIR. This report was then accepted as the basis for an extension of the inspection into the so-called continuation period, a result that generated much frustration on the part of Arcanian officials, even if inspectors did prioritize three areas for further inspection.

Advancing OSI Procedures

Many CTBT states-parties hope that lessons learned during the exercise will invigorate discussions on an operational manual intended to guide on-site inspections that has been under development in Vienna in the Preparatory Commission’s (PrepCom) Working Group B, which is overseeing the implementation of the treaty’s verification system.

“The overall experience of the exercise should help discussions on the manual to move beyond the very theoretical way of thinking that we have unavoidably had in Working Group B,” Malcolm Coxhead told Arms Control Today in an Oct. 15 e-mail. Coxhead is currently leading efforts to draft the manual, which by 2005 had grown to an unruly 766-page draft with numerous unresolved issues. Coxhead has tried to develop so-called draft model texts to speed up that process, but agreement on a consensus document will require more time to achieve.

For the purpose of the IFE08, Working Group B agreed on a 174-page test manual to be used as a guide for inspectors and the inspected state-party. That document contains only a handful of brackets, indicating contentious issues, but several states-parties made clear that agreement on the test manual does not indicate that they have given up their reservations on the real text. In preparation for the IFE08, discussions on the manual paused in 2008 but are expected to resume in February 2009.

The exercise highlighted open issues between the treaty itself and the operational manuals as participants were trying to apply the two texts to the reality of an inspection. For example, there was a heated debate over whether Arcanian officials have the right to comment on the PIR before it is transferred to Vienna. Also, what is supposed to happen if inspectors accidentally gather information that is outside their mandate? Does all the data associated with that mission have to be deleted or just the additional information? Many such procedural issues, which could complicate a real inspection or could even provide a pretext for its termination, were identified during the exercise. An evaluation team comprised of 10 outside experts and PTS staff who monitored the IFE08 and drew independent conclusions will also deliver its findings to Working Group B.

John Walker, a British expert who played the role of team leader of the inspected state-party in the IFE08, told Arms Control Today in an Oct. 15 e-mail that “the challenge now is to make sure that identified lessons are valid, learned, and embedded.”

One of the lessons may be that some restrictions placed by states-parties on equipment to protect national and commercial secrets are superfluous. For example, some states insist that inspectors not be allowed to use satellite-based navigation instruments like the Global Positioning System (GPS) because it would enable them to pinpoint the location of sensitive sites or facilities. Such an approach would appear to be outdated given that today’s GPS functions are integrated into many mobile phones and the Internet provides access to high-quality satellite imagery. During the IFE08, the GPS was used for navigation throughout the exercise and without a problem.

The field exercise also demonstrated that the relationship between inspected state-party and inspected state is likely to fluctuate between cooperation and conflict. Group loyalty was strong. As in a real inspection, interaction between the inspected state-party and the inspection team was limited. The base of operations was divided into different zones for inspectors and host state. The two groups even took their meals at different times. Arcanian officials went so far as to compose a national anthem that inspectors reluctantly sang when the inspected state-party hosted a joint barbeque.

Many conflicts evolved around managed access, which is the process of the repeated negotiation of access rights between the inspected party and inspectors that is at the heart of modern verification regimes. The core problem is finding an adequate balance between the rights and obligations between the two sides. How far can inspectors go? How are they supposed to behave? How much information do they have to share with the inspected state-party? How much assistance does the inspected state have to give inspectors? Again and again, these issues were played out in the scenario.

For example, since the beginning of the exercise Sept. 1, the inspection team had wanted to gain access to an area that Arcania had declared off-limits. Arcania maintained that the area was unsafe because of ongoing operations to clear unexploded ordnance from a closed firing range. Even more frustrating for the inspection team, the area was located within a no-fly zone.

But the beginning of the continuation period on Sept. 14 provided a perfect opportunity to test the CTBT’s managed access provisions. On Sept. 15, Arcania for the first time granted inspectors access to the no-entry area, saying that ordnance clearance operations had finished. At the same time, however, and in accordance with treaty procedures, it did declare a so-called restricted access site (RAS) of 2 kilometers by 2 kilometers. “This is because of national security reasons, unrelated to the purpose of the inspection,” John Walker, playing the role of director of the Disarmament Division in the Arcanian Foreign Ministry, explained dryly during the daily 5:30 pm meeting between Arcanian officials and the inspection team leadership. Additionally, Walker limited access by the inspectors to the northern part of the no-entry area and up to 500 meters outside the eastern, northern, and western boundaries of the RAS. Plus, only certain routes were to be used to approach the RAS.

For Wang, that was not good enough. Although the inspection team leader recognized Arcania’s right to declare a RAS, he argued that further limitations on access were not permissible. Wang maintained that it is not for the inspected state to decide where inspectors could go outside the RAS and accused Arcania of violating treaty stipulations and inspection procedures.

Walker remained adamant. “Take it or leave it!” he exclaimed, citing strict orders to him from Arcania’s Department of Defense. In reaction, the inspection team leader decided to leave. “This behavior is not in accordance with the treaty,” he exclaimed and announced that inspectors would pack up the next day.

In the evening, a teleconference took place between the (virtual) CTBTO director-general and the inspection team leader to discuss the stalemate. In a closed-door meeting, Wang then proposed to Walker that the inspection team would conduct an additional overflight over the RAS, using gamma radiation monitoring equipment. This turned out to be an acceptable compromise.

As the continuation period of the exercise began, inspectors began to deploy a number of additional technologies, such as magnetic field mapping, ground penetrating radar, and electromagnetic conductivity measurements, to investigate sites of interest and to detect the type of cavity typical of an underground nuclear test. Being able to test in a real former test site now proved particularly valuable because inspectors were able to show that their instruments can pick up anomalies from real underground tests, despite the fact that some of these were conducted more than 20 years ago. Although the inspectors did find a number of additional clues, they still did not detect any positive proof of a recent underground nuclear explosion.

The Endgame

As it turned out, Arcania had something to hide but not an underground nuclear test. On Sept. 21, the penultimate day of inspection activities, chief inspector Wang came tantalizingly close to the reason for Arcania’s secretiveness when Walker gave in to the inspectors’ pressure and invited him to visit the RAS. Wang specifically wanted to see the orange trailer and tent that inspectors had been observing from a distance for some time. Walker personally invited Wang to observe the site, but again kept him at a distance. Wang requested an environmental sample, but Walker refused to let Wang near the trailer. Instead, he collected the soil sample himself and handed it to the chief inspector.

The reason for Walker’s unwillingness to allow full access was that Arcania had been in the process of preparing a nuclear test when the earthquake that triggered the inspection took place Aug. 22. The trailer was supposed to house the instrumentation to steer the test. The tent covered the borehole, and between the two places, a trench had been dug for instrumentation cables.

Basing the field exercise on this scenario drew some criticism, particularly from the chief inspector. The CTBT’s verification system is not tailored to detect test preparations because these are not covered by the treaty. At the time when the treaty was negotiated, prohibiting test preparations was seen as too daunting a task, and states decided to leave that issue outside the treaty’s scope. Test preparations do not leave many signatures. Wang therefore called the scenario “very dubious” because inspectors had been told to find indicators of an underground nuclear test. He also pointed out that most nuclear-weapon states still maintain their nuclear test sites and that inspectors might come across artifacts similar to the one at the heart of the exercise even though these are not indicative of an imminent test.

On the other side, Walker argued that the scenario was valid given that, “in a real situation, there could very well be nothing to find, but the team has to go about conducting a thorough inspection of the inspection area to identify all potential features of interest, explore them, and see whether they were of relevance to the purpose of the inspection, following up in greater detail as necessary. In this respect ‘red herrings’ are a legitimate inject.”

Although there was some disagreement about the scenario, it was generally acknowledged that the inspectors performed well under difficult conditions, identifying most of the artifacts that the scenario designers had placed in the inspection area. Tóth called the game “probably the most challenging scenario, from different angles.”

Coxhead, who participated in the IFE08 but was not involved in setting up the scenario, agrees that people in the control team came to realize during the exercise that “presenting a credible scenario for an interesting game is harder than we may have thought.” Coxhead, however, argued that field exercises have natural limits: “It is not easy to simulate a nuclear explosion, and we can’t let one off for real!”

Once evaluation of the IFE08 starts, it will become clearer whether the exercise did send a clear message, particularly to verification skeptics, that the OSI regime would be ready when the treaty enters into force. Seven countries, including China, Iran, and Israel, which have all signed but not ratified the treaty, as well as the OPCW accepted an invitation by the PrepCom to send diplomats and experts as observers to the exercise. Notably absent were representatives of the U.S. government. Preparations for the OSI regime have been severely hampered by the decision of the Bush administration in 2001 to boycott discussions on OSI and withhold funding for all PrepCom activities related to OSI. (See ACT, September 2001.) Nonetheless, a number of independent U.S. experts were involved in Kazakhstan as observers and as part of the evaluation team.

Walker said that when the CTBTO holds another major exercise in about four years, there will be a chance for nonsignatory experts to participate and “hopefully the lessons from 2008 will be fully realized.” Tóth believes that the IFE08 demonstrated that, “in case additional tools are needed, they are at our disposal under the treaty, we can use those tools and use them in an efficient way” and hopes that this message will be heard particularly by countries that remain outside the CTBT. “I hope this process will lead the nine countries whose ratification is necessary for entry into force one by one to join the train that is already leaving the station,” he said.