Science: Nuclear nerves put to the test: Don't let this chance to halt the arms race go up in smoke, says Patricia Lewis

Nearly 50 years after scores of thousands of people in Hiroshima and Nagasaki were killed by the first wartime nuclear explosions, the world is on the verge of banning all nuclear weapons tests for ever. In Geneva tomorrow representatives of 39 nations will begin negotiations for a comprehensive test ban treaty (CTBT).

Not before time. The treaty banning nuclear tests in the atmosphere was negotiated in 1963. It was to have been comprehensive, but was cut back to a partial test ban, to the dismay of many states which did not possess nuclear weapons.

A comprehensive ban on all nuclear explosions, negotiated within the next year, would tell the world that the states with such weapons are serious about halting the nuclear arms race, and thus make it easier to persuade other countries not to join it.

And, for political as well as technical reasons, successful negotiation of a CTBT would greatly enhance the treaty on the non-proliferation of nuclear weapons (NPT), which is in dire need of a boost after the recent Iraq and North Korea crises. The international community has to decide the future of the NPT at a conference in New York in the spring of 1995.

A comprehensive ban on nuclear explosions would erect an effective barrier to the long-term modernisation of nuclear weapons by states already possessing them. It would also hinder efforts to develop H- bombs by states that possess, or are on the brink of possessing, fission weapons.

The talks are unlikely to go smoothly. The politics will be difficult enough, but technical issues will tax the negotiators, too. The first is how to define a nuclear test.

Development of a nuclear weapon takes several stages. The design is first tested by computer simulation of an explosion. Some people would like to ban the use of such computer software in weapons laboratories, but that is not practical.

Weapons scientists who once said they could not maintain nuclear stockpiles without full testing now say they can not only maintain them under a CTBT, but could also design reliable warheads using computer simulation alone. Such statements have fuelled the anger of many non-nuclear weapons states, which say the only reason the United States, Britain and Russia are now agreeing to negotiate a comprehensive ban is that they no longer need to test.

However, the weapons designers may have been economical with the truth. Yes, it is possible to design warheads by computer only, particularly if you possess results from past tests. But experts at the research establishment at Aldermaston, Berkshire, privately voice doubts over how confident they could be about the reliability of radical departures from proven designs. The motivation behind the optimistic statements most likely stems from a desire to stay in business.

A more difficult issue is the banning of laboratory tests that have near-zero explosive yields; tests of nuclear weapons components; and 'inertial confinement' fusion. The first two sets of tests are conducted solely to develop nuclear weapons and so, in principle, should be banned along with test explosions. However, verifying that such experiments never take place is impossible.

The problem with banning inertial confinement fusion or laser fusion is that while the technique is applicable to nuclear weapons, it could also form part of a genuine research programme into fusion energy. One solution might be to sidestep the definition of what is a nuclear test. Instead, the treaty could ban all nuclear explosions in all environments, leaving the diplomats to argue over the fine print of how the treaty is to be verified and designing the verification regime to catch those nuclear tests of primary concern. The treaty could allow for an upgrading of verification so that if, in the long-term, more subtle forms of nuclear explosions were a cause for concern, the regime could be extended to catch those, too.

Internationally, it has been clear for a long time that a CTBT can be well verified using a combination of seismology, remote sensing by satellite, detection of radioactive debris in the atmosphere, on-site inspections, radiochemical analysis and notification of large-scale drilling and large industrial explosions. In spite of this, verification will pose problems, not so much because of the technology but more because of the cost and organisation of the regime.

Governments pay lip service to the need for verifying international agreements, but are rarely prepared to put in the cash required. The Vienna-based International Atomic Energy Agency (IAEA) monitors compliance with the global non- proliferation treaty on an annual budget of only pounds 42m and found it nearly impossible to carry out some of its tasks in Iraq owing to lack of funds. The US is an honourable exception, for it supports research into the technology of verifying treaties and funds its on-site inspection agency properly.

For a CTBT, the number of seismic stations, the extent and speed of the data transmission and analysis, the commissioning of satellite images and the numbers of extensive on-site inspections, as well as the training required, are items that will be looked at with an eye to cutting costs. However, since the average cost of a single US nuclear weapons test is dollars 80m ( pounds 55m), it could be argued that the nuclear weapons states should finance the verification regime out of the savings from ending their testing programmes.

Another bone of contention is who should carry out the monitoring for the verification of a CTBT. Should individual countries do their own monitoring and carry out on-site inspection on a bilateral basis, when more than 100 nations may be signatories to the treaty? Or should there be an international organisation charged with the task? If the latter, should it be a body specifically set up to monitor a CTBT, or the underfunded IAEA?

Political rather than technical issues will make or break the treaty. There is concern over whether France and China will insist on carrying out more tests before signing, and whether one or more of the nuclear weapons states - possibly even the UK - will stall the talks.

Yet this remains the best chance we have had for more than a decade to achieve a total ban on nuclear weapons tests. If it can be done by spring next year, the cause of nuclear non-proliferation will be significantly boosted. If there is no substantial progress by then, and if it is clear that key states are prevaricating, the world will have lost a great opportunity. That would be unforgivable.

The author is director of the Verification Technology Information Centre, London.

(Photographs omitted)