January 1998

If a nuclear test had been conducted, it would violate international law by defeating the "object and purpose" of the recently signed Comprehensive Test Ban Treaty (CTBT). It would also constitute a violation of the announcement requirements for the 1974 Threshold Test Ban Treaty, and would signal the end of a self-imposed moratorium on nuclear testing. Perhaps more critically, the allegations came at a time when the President was due to transmit the CTBT to the Senate for "advice and consent to ratification." The treaty is designed to significantly slow, if not stop, further development of nuclear weapons. Before the Senate can ratify such a treaty, members must be sure it will not jeopardize national security. We must be confident that other nations will not violate the treaty, and be able to verify that this is the case. If the Russians had conducted a nuclear test it would surely result in the scrapping of the CTBT.

On August 18, the Nuclear Test Intelligence Committee (NTIC) took a "fresh look at all the data" from the events of the previous few days. Their conclusions can be reconstructed using press reports and analysis, to ensure the inferred methods result in the reported outcomes. In addition to the satellite evidence of "dead ringer" activities at the test site, they concluded that "the location of the blast, while probably at sea, could have been at the test site," and that the waveforms displayed "explosive characteristics." They believed there was a serious possibility there had been a Russian nuclear test. The analysis in this paper suggests this was for the following reasons:

• The error ellipse for the location of the event was large and included the test site.
• In addition to this, the difference between the P- and S-wave arrival time at FINES was the same as for previous nuclear tests.
• ARCESS, the station usually used to study waveform characteristics, was not operational at the time of the event. When the distinguishing methods developed on ARCESS waveforms were applied to other stations they suggested the event was explosive in nature.

The allegations against the Russian government were leaked to the press resulting in headlines like, "Russia suspected of nuclear testing," and policy memos entitled, "Wake-up call from Novaya Zemlya: Zero-yield nuclear test ban is unverifiable, Russians will cheat, U.S. will suffer." At the same time, seismologists in private institutions collected data from open seismic stations. Using this data they were able to locate the event offshore, and determine its nature; the waveforms were characterized by S-waves, suggesting it was an earthquake. Despite publication of these findings, U.S. government officials still claimed, "the data is not conclusive; it lends itself to alternative interpretations. You cannot rule out that it was an explosion. You cannot rule out that it was an earthquake."

"In response to questions posed both inside and outside the government, Director of Central Intelligence, George Tenet ... convened a panel of experts to review all of the available intelligence and technical data on the August events." In their conclusions they stated that the seismic event was "almost certainly not associated with the activities at Novaya Zemlya," however, "experts cannot not say with certainty whether the ... event was an explosion or an earthquake." This ambiguous statement resulted in various interpretations by different newspapers. They ranged from the continued belief the Russians had probably conducted a clandestine nuclear test to the belief they had not. This will probably be the last official statement on the incident and will surely leave Senators concerned about the verifiability of the CTBT when voting on ratification.

The conclusions reached by the NTIC were reasonable, based on the data used. However, before drawing these conclusions, they threw out almost 80% of the available seismic data, using only two out of nine seismic stations with direct electronic access. They did not use data from stations if they were either:

• not part of their own seismic network, or
• they were not calibrated, meaning they had not recorded a previous nuclear test.

The adoption of these two constraints by the nuclear monitoring community reduces the number of seismic stations available to 120 International Monitoring System (IMS) stations, out of an estimated 10,000 currently deployed around the world. It also means extending the IMS is a waste of time; in the Novaya Zemlya region there have been no confirmed nuclear tests since 1990, any stations deployed since then are of no use.

Using data from just the five stations which are part of the IMS, I show that the seismic event can be confidently located in the Kara Sea, 60 km offshore, and 120 km from the test site. This location has been confirmed by many seismic analysts since the event. The error ellipse is small, 30 km in the longest direction, excluding the possibility of a location on land. This data provides no evidence of "explosive characteristics." Use of P/S ratios is inconclusive due to the distance of the event from the nearest operating station and its relatively small magnitude (mb = 3.25). The only reasonable conclusion that can be drawn from this is that the suspect seismic event was an earthquake. This is not surprising as there are two detectable earthquakes every week in the Novaya Zemlya region. Satellite observations alone are not enough to prove a nuclear test was conducted; subcritical tests, which are permitted, look the same as nuclear test. Without a seismic event to associate with the satellite observations, there is no evidence of an illegal nuclear test. The explanation offered by the Russian government, that a subcritical test was conducted, is completely consistent with the data available.

This is not the first time false allegations of nuclear testing have been made. It is not the first time it has happened in Novaya Zemlya either. The negotiations on the CTBT were almost derailed by a similar allegation leaked to the press on March 7, 1996. The same "dead ringer" activities were observed by satellite at the test site earlier in the year. The seismic equivalent of a smoking gun was later found in a search of the seismic records, a magnitude 2.5 event at 17:17 on January 13. After this event was ceased upon as evidence of a clandestine Russian nuclear test, careful locations placed it more than 200 km from the test site. As with the 1997 event, this fact never made it into the press.

• IMS usage. The best analyses of the August 16 event were produced when as much data as possible was used. Locations obtained using calibrated data are more accurate than those obtained from non-calibrated data, provided there is sufficient calibrated data. It is not the case if this criterion requires throwing out most of the data.
• IMS redundancy. There are many more seismic stations than the IMS network, some of them have records of past nuclear explosions. The use of these stations, when necessary, to supplement gaps in IMS data -- the use of KEV in the absence of ARCESS for example in the August 16 case -- can significantly improve the analysis.
• Use of the "too much to be a coincidence" philosophy. It seems this is currently being applied rather liberally by the U.S. governmental nuclear monitoring community perhaps due to a prevailing attitude that the Russians are always "up to something." More application of "Occam’s Razor," that the most likely explanation is the one with fewest assumptions, might be appropriate.

• Location of international monitoring equipment at nuclear test sites. Seismometers either on or near the test sites would confidently detect even small yield nuclear tests.
• Mechanism to resolve uncertainty. In cases where there is continued concern that a nuclear test has been conducted a mechanism is needed to resolve the situation. International inspections of this type will be permitted by the CTBT, once ratified.
• Transparency for subcritical tests. Many of the allegations that a nuclear test has been conducted are caused by subcritical tests. One way to prevent these damaging situations would be for nations to allow open reciprocal observation of these tests. The U.S. has the opportunity to set this precedent at its upcoming tests in 1998.

If the CTBT is to succeed, incidents such as the August event in Novaya Zemlya must be prevented. This study demonstrates that the August 16 suspect seismic event was detected by the IMS, and could be identified as an earthquake. Despite this, false allegations of a nuclear test were made by the U.S. During the course of the incident, government officials demonstrated their inability to correct mistakes publicly, demonstrating the need for publicly available data and independent analysis.