Verifying the nuclear ban: Lessons from South Africa

In July, 122 UN member states agreed on a legally binding treaty banning nuclear weapons. It is likely to enter into force this year. But as ban advocates are well aware, the road toward nuclear disarmament will continue to be a rocky one. Structural, institutional, and technical challenges—all of them deeply intertwined with international politics—stand in the way of the treaty’s implementation.

Among the key implementation challenges will be to devise realistic, credible pathways toward nuclear disarmament—and to establish an effective verification regime that can specifically address nuclear reversal. Treaty signatories addressed these questions in broad terms during negotiations, but they left many details to be resolved after the treaty’s entry into force. And today, no multilateral regime exists for verifying dismantlement of nuclear weapons, either within the ban treaty apparatus or outside it.

Several key lessons about the design and implementation of a verification regime can be drawn from the example of South African disarmament. South Africa is the only nation ever to have developed nuclear weapons and subsequently eliminated them. When the nation admitted in 1993 that it had possessed—and already eliminated—a small nuclear stockpile, the global nonproliferation regime was called upon to perform a number of unique tasks. For example, inspectors from the International Atomic Energy Agency (IAEA), whose work usually entails confirming that peaceful nuclear programs are not illicitly put to military uses, were for the first time asked to verify the elimination of an existing military program that had successfully produced nuclear weapons—work that requires a quite different skill set. Likewise, agreements had to be reached about the information that South Africa would share with inspectors, and how the information would be shared. Even after South Africa’s denuclearization was verified, the agency faced a new challenge: implementing agency safeguards in a newly disarmed state. Examining South Africa’s experience in dismantlement and verification provides important insights into some of the key issues that will face the new treaty.

Disarmament pathways. The ban treaty in its final form (unlike in some earlier drafts) provides two distinct pathways for nuclear disarmament. The major difference between the two pathways is the sequence in which states join the treaty and eliminate their nuclear weapons.

One pathway foresees that nuclear-armed states may, prior to “coming clean” and joining the treaty, choose to eliminate their nuclear arsenals and dismantle their nuclear weapons facilities or convert them to peaceful uses. This pathway is directly inspired by the South African experience of dismantlement prior to joining the Nuclear Non-Proliferation Treaty (NPT) as a non-nuclear weapon state. The other pathway offers nuclear-armed states the opportunity to join the ban treaty first and dismantle their weapons and related facilities later. Following this path would entail a set of mandatory procedures and deadlines regarding elimination of warheads and the dismantlement or conversion of a weapons program’s various elements.

Whichever pathway a nation chooses, the treaty stipulates certain verification requirements that can be broadly characterized as belonging to two separate missions. The first mission is verification of the irreversible elimination of a nuclear weapons program (or its conversion to peaceful uses). The treaty assigns a key role in this mission to an as-yet undefined international body, referred to in the treaty as the “competent international authority.”

The second verification mission involves bringing newly disarmed states under the umbrella of the IAEA safeguards regime. The treaty does not specify the type of safeguards to be applied in that case, which leaves the door open to the possibility of a new instrument tailored to the situation. But the treaty does set out general principles for applying safeguards in denuclearizing countries: The denuclearizing nation must provide credible assurances of non-diversion of declared nuclear material and of the absence from the country of undeclared nuclear material or activities. The ban treaty, by taking this approach, ultimately envisions an extension of IAEA safeguards to all current nuclear weapon states when they decide to disarm and join the ban treaty. Although South Africa’s denuclearization history relates most specifically to the first pathway, it also reveals lessons relevant to the second pathway.

Case study: South Africa. In March 1993, South African President F.W. de Klerk announced that his country had previously possessed a small arsenal of nuclear weapons—but had dismantled them. South Africa’s decision to eliminate the weapons, and to either dismantle associated facilities or convert them to peaceful activities, had been made in early 1990. In July 1991, South Africa ratified the NPT as a non-nuclear weapon state, by which time the weapons, related components, and design data had all been destroyed. Nearly two years passed between South Africa’s accession to the treaty and its admissions regarding previous nuclear weapon activities.

South Africa has plentiful uranium reserves. During World War II, it was discovered that uranium ore was a by-product of the country’s large gold-mining industry, and South Africa eventually became a uranium exporter. After the decision to build nuclear weapons was made in the early 1970s, a uranium enrichment plant known as the Y-Plant was built to produce highly enriched uranium (HEU). A second enrichment plant, the Z-Plant, was also built, though apparently it produced only low-enriched uranium. Facilities for weapon design, fabrication, and testing were also constructed. By the time South Africa decided to dismantle its nuclear weapons and the programs associated with them, its nuclear arsenal comprised six gun-type HEU weapons and a partially completed seventh.

After South Africa joined the NPT in July 1991, it signed a Comprehensive Safeguards Agreement with the IAEA and submitted an initial declaration of nuclear material and facilities. After de Klerk’s admission in March 1993 of the prior existence of a nuclear weapons program, South Africa offered full transparency about the former program and gave agency inspectors broad access to any site in the country. The agency expanded its inspector presence, which was already considerable, through the addition of nuclear weapons experts.

Skill sets and information sharing. Under ordinary circumstances, the main activity of the IAEA Safeguards Department is to verify states’ declarations regarding civil nuclear activities. Verifying the dismantlement of an extant nuclear weapons program is a considerably different undertaking. South Africa’s former nuclear weapons program was small enough that the agency could organize a group of in-house weapons experts and send them to South Africa to augment the “regular” inspection presence there. But nearly all of the current nuclear-armed states have weapons programs much larger than South Africa’s, so a much larger cadre of nuclear weapons experts would be required. And the “competent international authority” envisioned in the ban treaty would have to be staffed by experts with a different skill set than that which is typical among their IAEA colleagues.

How to handle sensitive information is another aspect of verification under the ban treaty that would have to be resolved. To state the obvious, nuclear-armed states maintain extreme secrecy regarding the details of nuclear weapons technology. Secrecy prevents not only adversaries but also potential proliferators, including terrorist organizations, from obtaining military secrets. But if the decommissioning of a nuclear weapons program is to be verified, some level of detail must be shared with the verification body. In South Africa’s case, thousands of technical documents were destroyed before accession to the NPT. This was not surprising, but it nonetheless made later verification efforts more challenging for both the IAEA and for South Africa.

A related issue involves the methods and technologies that the verification body will use. For some time, technical means have been under development that might allow inspectors to check what must be checked without seeing information that they should not see. But no procedure is foolproof. What do inspectors need to see and what should they not see? How much proof is enough? How much doubt is acceptable? Answering such questions will entail political decisions. Meanwhile, the verification body will require—in addition to a large cadre of inspectors and analysts with the appropriate expertise—systems and procedures up to the task of handling and protecting highly sensitive information.

Safeguards and the disarmed state. Once the disarmament verification body has completed its work and the IAEA begins implementation of safeguards, a new set of challenges will have to be faced. For example, it is important that formerly nuclear-armed states provide the agency with as much detail as possible regarding the design and production activities of nuclear facilities, whether operational or shuttered, to be brought under safeguards. In South Africa, a few nuclear facilities had been declared to the agency prior to 1991—but only a few. Because the agency had relatively little prior knowledge about the country’s facilities, establishing that South Africa’s initial safeguards declaration was correct and complete entailed a lengthy process. In particular, clarifying the operating histories of certain facilities—especially the Y-Plant—required a significant amount of effort by both the agency and the South African authorities.

In such a situation, the disarming state should gather and organize all the data that could possibly have a bearing on the verification of its declarations. The state should be as transparent as possible—as early as possible. But the South African experience shows that discrepancies and ambiguities are virtually inevitable, so any verification regime should have the means to address and resolve such issues. Then again, the correct approach may sometimes be for the verification body, as a political matter, to let an issue rest. Managing such situations takes time, skill, and patience.

Particular difficulties might arise when a facility that previously was part of a nuclear weapons program is to become part of a safeguarded, civilian nuclear program. In such a case, the IAEA would need to gain detailed knowledge not only of past production records but also plant design. That is, if the agency is to verify in the future that a facility has not been reconfigured to support its previous nuclear-weapons mission, it must know what the previous configuration was. This could be a sensitive issue. More broadly—and again, as was the case in South Africa—once a safeguards agreement is in place, it is important for a formerly nuclear-armed state to give the agency unfettered access to all sites associated with the defunct nuclear weapons program. This could include shuttered or abandoned sites, or even sites about which the agency has information from other sources. Coordinating such access could require considerable effort, both for the agency and for the nation.

Resources will also become an important consideration if nuclear-armed nations accede to the ban treaty and their civilian nuclear facilities come under IAEA safeguards. As the agency’s 2015 Safeguards Implementation Report shows, the IAEA now carries out inspections at a limited number of nuclear facilities in each of the current NPT-recognized nuclear weapon states. In 2015, the agency inspected three facilities in China, two in France, and one in the United States. But the inspected facilities represent only a small part of the large nuclear infrastructures in these three nations. For example, China operates 37 nuclear power reactors and is building 20 more. France operates 58 power reactors and the United States 99. When one considers that the IAEA inspected a total of 250 power reactors around the world in 2015, it becomes clear that a requirement to safeguard the civilian nuclear infrastructures of any of these three states, or indeed of virtually any nuclear-armed state, would necessitate a dramatic increase in the agency’s budget and infrastructure.

The end state? Finally, in the event a nuclear-armed state decides to join the ban treaty, two particularly important terms will have to be defined. First, what does “eliminated” mean? The treaty requires, for example, that existing nuclear weapon stocks be eliminated. But what constitutes an eliminated nuclear warhead? Must its highly machined nuclear materials be verifiably melted down? What is to become of other components, such as high explosives, special electronics, or materials like tritium? Must a weapon’s delivery vehicle, if there is one, be verifiably destroyed, or can it be converted for use with conventional warheads—and who should verify the conversion?

Second, what does “irreversible” mean? The treaty stipulates either the elimination or the irreversible conversion of facilities related to nuclear weapons programs—which might not be technically possible in some cases. In South Africa, for example, shafts that had been drilled in the ground for testing weapons—they were never used for this purpose—were rendered useless in the presence of IAEA inspectors. But in the case of, for example, an operating centrifuge enrichment plant that once produced HEU for weapons and now produces low-enriched uranium to fuel power reactors, finding a suitable definition of “irreversible” might be tricky. The difference between the two outputs—very roughly stated—is in the configuration of a facility’s piping. The agency might confirm that piping has not been reconfigured to once again allow the production of HEU—but it would be difficult to make such a conversion irreversible without simply destroying the facility.

While the ban treaty offers a broad framework for the verification of nuclear disarmament, it leaves many specific details for subsequent elaboration. Key among the challenges involved in reaching a world without nuclear weapons is to ensure that nuclear arsenals and their associated infrastructure have been verifiably dismantled. Currently, no multilateral verification regime exists for verifying dismantlement of nuclear weapons. Creating an effective regime for that purpose will require the international community to address key design and implementation questions—and the South African experience provides key insights into some of the issues involved. A verification authority with the appropriate resources and skill set will have to be established. Decisions about what information should be shared with this authority, and how to share it, will present a number of difficulties. Once disarmament has been verified, the implementation of IAEA safeguards in a newly disarmed state will bring a new set of challenges; one of the most important of these will be ensuring that the IAEA has the resources it needs to carry out its mandate in the state. Finally, agreement will have to be reached regarding some key concepts that are referred to in the treaty, but not specifically defined.