Great Progress Made by Seismologists in Identifying Violations of Nuclear Test Ban Treaty

Advances in detection devices and methods of analysis have allowed seismologists to identify virtually all events that might be nuclear explosions of possible military significance under the Comprehensive Test Ban Treaty (CTBT), according to Prof. Lynn R. Sykes of Columbia University’s Lamont-Doherty Earth Observatory. Writing in the 29 October issue of Eos, published by the American Geophysical Union, Sykes analyzes 72 questionable events since 1960.

Verification was a major issue in the U.S. Senate debate in 1999, in which American ratification of the treaty was defeated. Since 1995, the International CTBT Organization has detected many small seismic events and has determined that many were earthquakes or otherwise identifiable. Under the treaty, however, the international organization is not charged with identifying all seismic events and is, in fact, not permitted to declare an event a nuclear explosion. Unresolved cases and possible nuclear events are referred to national CTBT agencies, whose work is usually classified.

Therefore, says Sykes, we do not always know how well the difficult cases are resolved. Only 72 events have been flagged in the literature or by the media as questionable or difficult to identify, a small fraction of all events recorded over the past 42 years, demonstrating the great progress made in verification, says Sykes. By studying technical characteristics of the seismic signals, nearly all of the 72 events have been identified as nuclear explosions, chemical explosions, earthquakes, or mine collapses, he writes.

Thirty years ago, problem events registered seismic magnitudes (mb) of 4.3 to 5.6, whereas today, most attention is focused on the mb 2.0 to 3.5 level. Since the magnitude scale is logarithmic, this represents an improvement factor of 300 in the size of signals that can be identified. It means that nuclear explosions 1,000 times smaller in their energy release can now be identified, says Sykes.

Since 1990, all of the problem events greater than mb 2.5 have received special study and have been identified. There is no evidence, says Sykes, that any countries have exploded nuclear devices since the CTBT was opened for signature in 1996, aside from India and Pakistan (which have not signed). Under the International Monitoring System, which began in 1995, seismic monitors have been placed close to, or in, countries that possess nuclear weapons, and the shorter the distance high frequency seismic signals have to travel, the better the identification of their source, Sykes writes.

Indian nuclear explosions on May 11, 1998, and Pakistani ones later that month were widely recorded and quickly identified. Sykes says the Indian claim of two tests with a combined yield of 0.6 kilotons on May 13, 1998, is surely exaggerated, as they created no detectable seismic signal. Similarly, media reports in 2001 of an Iraqi nuclear explosion in 1989 appear to be false, Sykes writes, as no seismic signal was produced.

Regarding suspected Russian tests on September 8 and 23, 1999, at Novaya Zemlya, Sykes concludes that, if they occurred at all, they were very tiny, that is, 0.001 to 0.02 kilotons. They may have been so-called “sub-critical” tests that release no nuclear energy, which are permitted by the treaty. Russia, China, and the United States have each conducted sub-critical tests since the treaty was signed in 1996. Sykes notes that to have military significance, tests must produce yields of at least five to 10 kilotons.

As for the future of verification, Sykes says that only on-site inspections can resolve any doubt in suspected events where the nuclear yield is zero, and this becomes possible once the CTBT enters into force. Short of that, occasional problem events may occur at the limits of detection. These will become fewer, as investigations of previous questionable events are published and the science of verification advances further, he concludes.

Contact: Harvey Leifert
(202) 777-7507
hleifert@agu.org