A debate started in 1972 and lasted until 2009 about the identification of so-called problem or anomalous seismic events that some people believed to be difficult to identify. Occasional problem events such as these, which are likely to occur about once a year, require more detailed examination than most earthquakes do. The following examples illustrate that with special studies, all of them through at least 2009 can be identified as earthquakes. I pay particular attention in this chapter to very small seismic events near the Russian test site on Novaya Zemlya, several of which the U.S. government claimed were either nuclear explosions or possible explosions.
U.S. PAPER SUBMITTED TO THE UN IN 1972 ON ANOMALOUS SEISMIC EVENTS
In August 1972 the U.S. government presented a working paper, A Review of Current Progress and Problems in Seismic Verification, at the Conference of the Committee on Disarmament (CCD) of the United Nations. It listed twenty-five “anomalous” or “problem” seismic events that presented difficulties in identifying underground explosions as opposed to earthquakes using the difference between long- and short-period seismic waves by the Ms-mb technique, discussed in chapter 6. The U.S. working paper described them as “false alarms,” identifying them as likely earthquakes that lay either in or close to the nuclear explosion portion on an Ms-mb diagram.
Although no author is listed on the 1972 working paper, officials in the Department of Defense likely wrote the document with data processing being performed by a consulting firm. Like many other official U.S. reports, it emphasized many problems in seismic verification and indicated that the Ms-mb technique might well fail to identify future seismic events. Other than stating the need for more research, it did not take a problem-solving approach. These reports typically convey the sense that the “glass is half empty.” The report’s implication clearly was that seismic identification was not good enough to justify the United States’ entering into a full test ban treaty.
Although considerable information from other seismic stations was readily available, most of the original magnitudes in the 1972 report were determined using data from only one or two stations. Four years later, in 1976, Lamont graduate student Robert Tathum, research scientist Donald Forsyth (now at Brown University), and I made a thorough study of each of the twenty-five events included in the 1972 report, using available unclassified seismic data. We increased the number of good mb determinations from the original 83 in the 1972 report to 242. We published our results in a peer-reviewed scientific journal.
We showed that with additional data about half of those twenty-five events had Ms-mb values that put them clearly in the earthquake population—that is, those events ceased to be “anomalous.” Several of the remaining problem events occurred at depths of 15 to 30 miles (25 to 50 km). They could be identified easily because nuclear explosions have not been set off at depths greater than about 1.2 miles (2 km). The greatest depths of petroleum and other wells do not exceed 7 miles (11 km). The sole exception was a single super-deep hole drilled to a depth of about 10 miles (15 km) for scientific purposes by the Soviet Union in hard, ancient rock at a huge cost. It was unstable and shrank in size at those depths and would therefore have been unsuitable for a nuclear test.
This left one unresolved sequence of problem events in a small, very mountainous area of Tibet that occurred during a three-month period in 1968 and 1969, and some events in this swarm were larger than magnitude 5. Nevertheless, at several stations the first motions of their P waves clearly were downward, indicating they were earthquakes and not explosions. Other events in the swarm had waveforms almost identical to those of the larger events in the series, which led us to conclude that all of them were earthquakes. Even in 1974, we were able to identify all of the claimed twenty-five “anomalous” events as earthquakes. Today much better seismic data exist from stations within Asia, so that events like these can be readily identified.
Understandably, the United States and other countries do not want to miss even a single nuclear explosion, nor do they want to incorrectly identify an earthquake as an explosion. This is a big job, especially with small explosions, because the number of earthquakes increases by a factor of about eight to ten when seismic magnitude decreases by one unit. Additional methods besides the standard Ms-mb technique should be applied to distinguish earthquakes from explosions. For the few “problem” events, I recommend a thorough, unbiased, and problem-solving approach using a variety of seismic methods developed by experts with a long tradition of work in seismic verification. U.S. officials who reviewed the 1972 report prior to its submission to the UN should have caught the low number and incomplete nature of the magnitudes used.
SMALL SEISMIC EVENTS NEAR THE RUSSIAN ARCTIC TEST SITE
During the 1980s and 1990s and into the following decade, some analysts in the United States claimed that several seismic events were either small nuclear explosions or possible explosions. A number of these accusations were released to the press, particularly to the Washington Times. Some came at critical times, either during or soon after the negotiations for the Comprehensive Test Ban Treaty (CTBT) in 1995 and 1996 or the U.S. Senate’s debate on the treaty in 1999. It is important to note that all of these “problem” events were identified as earthquakes by either British, Norwegian, Livermore, or Lamont seismologists.
In 1987 Ryall and others at DARPA’s Center for Seismic Studies were not able to identify a small seismic event on or near Novaya Zemlya on August 1, 1986 (Figure 12.1), as being either an explosion or an earthquake. The large uncertainty in the computed initial location included a small area of land to the southeast of the Soviet arctic test site, but not the test site itself. Their results were widely distributed, and several people indicated that similar problem events were an obstacle to achieving a full test ban treaty.
FIGURE 12.1
Locations (triangles) and their uncertainties (open circles) of five small earthquakes on and near Novaya Zemlya from 1984 to 1997. The seismic magnitude mb is shown in parenthesis following the date of each event.
Source: Sykes, 1997.
Novaya Zemlya and its surrounding Kara and eastern Barents seas are not very active for moderate to small earthquakes (figures 12.1 and 12.2). Novaya Zemlya is about 625 miles (1000 km) long, similar in size to California. A 200 by 200 mile (300 by 300 km) region centered on the main test site is less active than a similar area centered on New York City. Both are intraplate regions; that is, they are located inside a tectonic plate rather than along an active plate boundary.
FIGURE 12.2
Seismic events and stations in the vicinity of the main Russian arctic test site at Novaya Zemlya. Stars show the locations of nuclear tests since 1977. Small circles indicate seismic events from 1999 to 2009 of magnitude greater than 2.0. The primary stations of the International Monitoring Service (triangles), the auxiliary station in Sweden (square), and three other publicly available stations (pentagons) are shown in white. The diamond locates the Kursk submarine disaster of 2000. Many of the events in Scandinavia, Finland, and mainland Russia were small mine blasts.
Source: National Academies Report, 2012.
The main Novaya Zemlya test site is located well north of the Arctic Circle, at 73 degrees north latitude, farther north than the northernmost point in Alaska. It is a frigid, largely mountainous region where polar bears are much more numerous than humans (figure 12.3). A glacier covers much of the northern island of the two parts of Novaya Zemlya.
FIGURE 12.3
Polar bear in front of the officers’ club at Novaya Zemlya.
Photo courtesy of Paul Richards, 2004.
In 1989 Marshall and two colleagues from the British verification group made an extensive study of the 1986 seismic event and published it in a peer-reviewed journal. They examined more data and performed more identification tests than Ryall and others had done. They concluded the event was an earthquake based on the ratio Ms/mb and a focal mechanism solution. Their uncertainty in location, which is shown in figure 12.1, placed it entirely beneath the Kara Sea. Their identification of the seismic waves P, pP, and sP on records from several stations at large distances allowed them to determine its depth as about 15 miles (24 km), another indication it was an earthquake. The paths those waves travel through the Earth are illustrated in figure 3.1. The thorough work of Marshall and colleagues convinced the nuclear verification community that the event clearly was an earthquake even though its magnitude mb was small, only 4.3. If it had been a standard, well-coupled nuclear explosion, its yield would have been about one kiloton.
Detection, location, and identification of seismic events in and near the Soviet arctic test site improved after the early 1990s with the installation and operation of powerful seismic arrays (a system of linked seismometers arranged in a regular geometric pattern) in northern and southern Norway, southern Finland, and to the north of Norway in Spitzbergen (figure 12.4). Each array provides an estimate not only of the distance to a seismic event but also of its azimuth (or direction). An array is like either a spotlight that shines brightly only in a narrow beam or a directional microphone that is sensitive to sound from one azimuth. A single station, on the other hand, is akin to a light bulb that shines equally in all directions.
FIGURE 12.4
Locations of four seismic arrays that continuously monitor the Russian test site on Novaya Zemlya (NZ).
Source: Kværna, unpublished figure 2016.
The four seismic arrays continuously monitor the Novaya Zemlya test site with an excellent detection capability down to magnitude 2.2 to 2.5 (5 to 15 tons = 0.005 to 0.015 kilotons). These capabilities today are made possible by the successful operation of those arrays and decades of work by Norwegian seismologists. The Russian test site at Novaya Zemlya is one of the world’s best-monitored places of high concern to the United States.
As the negotiations for the Nonproliferation and Comprehensive Test Ban treaties commenced in the 1990s, any small seismic events that were located on or near the two islands of Novaya Zemlya were leaked to the media as possible or likely Russian nuclear explosions. Four small seismic events from 1986 to 1997 were cited by the U.S. Defense Department, including one by the secretary of defense, as being of either suspicious or uncertain origin.
The event shown in figure 12.1 of magnitude 2.7 on December 31, 1992, took place near the main Russian test site. Since it occurred on New Year’s Eve, some Americans and Russians have said most people at the test site would have been too drunk to conduct a nuclear explosion and Soviet authorities would not have permitted it to be conducted on New Year’s Eve. It is important to note that of the forty-two underground nuclear explosions on Novaya Zemlya, none occurred between December 4 and May 7, times of harsh arctic winter conditions and no daylight. Hence, a test on New Year’s Eve seems highly unlikely. Nevertheless, skeptics drew attention to it as a problem event.
Fortunately, the ratio of high-frequency P to S waves, like those shown in figure 12.5, also indicates the 1992 event was a very small earthquake. If it had been a well-coupled explosion in hard rock at that test site, which it was not, its yield would have been very small, about 0.025 kilotons (25 tons).
FIGURE 12.5
Ratios of high-frequency P to S waves discriminate (identify) Novaya Zemlya nuclear tests from earthquakes. Left-hand side compares 6–8 Hz (cycles per second) seismic waves at Kevo, Finland, for the 1997 Kara Sea earthquake (above) with a nuclear test in 1990 (below). Right-hand side shows P/S values for five nuclear tests, nonnuclear explosions related to the sinking of the Kursk submarine, and earthquakes on and near Novaya Zemlya. The 1997 and more recent earthquakes in 2007 and 2009 are labeled.
Source: National Academies Report, 2012.
The yield of another small seismic event, on January 13, 1996, near the coast of the northern island of Novaya Zemlya (figure 12.1), would have been about 0.013 kilotons (13 tons) if it were an explosion. Even if it or the earthquake of New Year’s Eve 1992 had been decoupled (muffled) explosions detonated in a cavity in hard rock, their yields likely would have been smaller than one kiloton. Leaking stories to the press about the smallest detected event, such as that of 1996, clearly is making a mountain out of a molehill—especially one such as this located 125 miles (200 km) from the Russian test site. Salt, which is the easiest common rock in which to form large cavities, is not present on Novaya Zemlya in any appreciable thickness.
U.S. CLAIMS THAT THE 1997 EARTHQUAKE WAS A RUSSIAN NUCLEAR EXPLOSION
A great governmental furor and media attention in the United States focused on a small seismic event in the Kara Sea on August 16, 1997. Officials in Moscow maintained that the event was a small earthquake in the Kara Sea to the east of Novaya Zemlya. Nevertheless, on August 28, 1997, the Washington Times carried a lead story by B. Gertz titled “Russia Suspected of Nuclear Testing.” The Washington Post and the New York Times published similar stories the next day, the latter with the headline “U.S. Suspects Russia Set Off Nuclear Test.” Remarks quoted in the press, such as “This one certainly had characteristics that at least would lead some to believe that there had been an explosion that caused the event,” emphasized the likelihood of a clandestine nuclear explosion. Buried in the body of the text, however, was a statement that whether it was a nuclear explosion or a small earthquake was still in doubt.
This was the first of several allegations by U.S. officials of cheating by Russia after the UN General Assembly overwhelmingly adopted the Comprehensive Nuclear Test Ban Treaty (CTBT) in September 1996. The widely publicized allegation about the August 16 event and a formal protest by the United States to Russia came at a very sensitive moment, just as President Clinton, who appeared at the UN on September 22, 1997, reported that his administration would submit the treaty, which he had signed, to the U.S. Senate for its advice and consent.
Russia reported that a chemical explosion had been conducted two days earlier, on August 14, 1997, to test the compression of plutonium. They stated it was a so-called hydrodynamic test that released no nuclear energy. The United States had conducted and announced hydrodynamic tests in previous months at the Nevada Test Site. Hydrodynamic tests are permitted under the CTBT. Apparently, an earthquake two days after August 14 led officials in the United States to jump the gun, believing that the seismic event of August 16 was a small nuclear test.
By the time the story appeared in the Washington Times, twelve days after August 16 and at the start of the Labor Day holidays, a strong consensus had developed among seismological experts in the United States, the UK, Norway, and Canada that the event was, in fact, a small earthquake with a magnitude of 3.3 (figure 12.1). If it had been a well-coupled nuclear explosion, its yield would have been about 0.1 kiloton (100 tons). In fact, it was well located beneath the waters of the Kara Sea about 60 miles (100 km) southeast of the Russian test site and farther away from it than the 1986 earthquake.
Nonetheless, nearly two months after the 1997 event, several top U.S. officials were either unaware of this scientific consensus or chose to ignore it, maintaining that the nature of the seismic event was ambiguous. A quick initial location, incorrectly made by U.S. agencies, apparently was passed up chains of command in the government even though it became obvious just one day later that the event was located well at sea. Clearly, revised estimates need to be made and communicated to high officials in a timely manner as more critical data become available, particularly when an event is or may be of potential concern to national security.
High officials and some government agencies in the United States had difficulty admitting they were mistaken—that the seismic event was merely a small earthquake and initial leaks to the press were incorrect. When the correct information was finally released officially months later, the event, once a front-page claim, became a small back-page story.
Both the preliminary International Data Center (IDC) of the International Monitoring System (IMS), then situated in Arlington, Virginia, and the U.S. National Data Center played key roles in recording and analyzing seismic data on August 16, 1997. Under the terms of the Comprehensive Test Ban Treaty, the IDC and its successor in Vienna are not allowed to identify “problem” events as earthquakes, nuclear explosions, or chemical explosions. Those tasks as well as yield estimations are reserved to national CTBT authorities. The Air Force Technical Applications Center (AFTAC) operates the data center for the United States, but its results are classified.
Seismic stations of the International Monitoring Service and other stations in northern Europe (figures 12.2 and 12.4), except for the sensitive seismic array ARCESS in northern Norway, recorded the 1997 event. Although ARCESS was being repaired following a power surge, data from other seismic stations in the region were sufficient to locate and identify the event. The Finnish standard station KEVO near ARCESS, while not part of the IMS, recorded the earthquake as it had many previous nuclear explosions at Novaya Zemlya. KEVO’s data were readily and quickly available over the Internet from the U.S. consortium for seismology called IRIS.
This illustrates the need to be ready to analyze available data rapidly from many stations for a “problem” event. When data from KEVO and several other stations in Scandinavia became available the next day, it was clear that the event was in fact in the Kara Sea, not on land (figure 12.1).
Many U.S. seismologists like me learned about the 1997 seismic event from press reports twelve days after it occurred. Several of us at Lamont, including Paul Richards, Won-Young Kim, and I, immediately examined seismic data from the event. We found that the use of high-frequency P and S waves provided clear additional evidence that the event of August 16, 1997, was an earthquake. It falls in the earthquake population of Figure 12.5 and not in that of nuclear explosions at Novaya Zemlya. In addition, Norwegian seismologists identified a very small aftershock of that earthquake by cross-correlating the signals from the two events.
The 1986 and 1997 earthquakes occurred near sites of Soviet nuclear waste disposal that would not be auspicious sites for conducting offshore nuclear explosions. (The occurrence of the small earthquakes there is not related to the disposal of wastes.) In addition, rocks in the area are very old, and it is not a site of present or recent volcanic activity.
On September 13, 1997, John Diamond of the Associated Press, under a headline “Study Supports Russia Test Denial,” wrote, “New findings in a secret Air Force study indicate that a tremor thought to have been a Russian nuclear test occurred underwater, pointing to the likelihood it was an earthquake.”
Jeffrey Smith, a staff reporter on national security for the Washington Post, wrote an extensive front-page article on October 20 titled “U.S. Officials Acted Hastily in Nuclear Test Accusation: CIA Hesitates to Call Russian ‘Event’ a Quake.” Smith wrote, “A high-priority classified alert issued by the CIA on Aug. 18 quickly caught the eye of senior policy makers. The bulletin came from the government’s Nuclear Test Intelligence Committee, an interagency scientific group, and said that Russia probably had conducted a nuclear test two days earlier on an island near the Arctic Circle. Officials at the National Security Council swung into action, convening an interagency meeting two days later and ordering a full-court press to collect an explanation from Moscow.”
Jeffrey Smith reported that Harold Smith, assistant to the secretary of defense, said that other scientists at the Pentagon shared his belief that the initial CIA report was wrong. He quoted Harold Smith: “I personally think it was an earthquake. We now know that they would have been well advised to wait until they had more data and could reach an accurate conclusion.” The article quoted my colleague Richards: “Not only was there a mistake made, but there was no effort to retract it.” Smith also quoted Eugene Herrin, who for the previous fifteen years had chaired the military’s principal seismological advisory panel: “somebody jumped the gun. Based on what I know, it was not an ambiguous event…. It’s an earthquake.”
I wrote a twelve-page article on the earthquake for the November/December 1997 issue of the Federation of American Scientists’ F.A.S. Public Interest Report. I spoke to Jeffrey Smith about my forthcoming article, and he was able to obtain information from sources in the U.S. government to which I was not privy. Smith’s article appeared on the same day I participated in a news conference held in Washington, DC, related to the August 16 event that was sponsored by the Coalition to Reduce Nuclear Danger.
One of the important documents I furnished to Jeffrey Smith was a copy of a fax sent by Frode Ringdal of Norway to Alewine of the U.S. Defense Department. In it Ringdal clearly put the event in the Kara Sea more than 60 miles (100 km) from the Russian test site. I realized that the date of the receipt stamp and fax numbers were still on the top of my copy of the fax. Alewine’s office received it a week before Gertz quoted him in the Washington Times indicating that the event was a possible explosion. This vital information was a “smoking gun” that Smith used to get more material for his story. He did an excellent job obtaining information from many sources for his front-page story.
Jeffrey Smith went on to state on October 20, “But the nuclear intelligence committee, which the CIA chairs, did not formally begin backpedaling until two weeks after the event, causing one official to describe it as “the last to join the crowd.” Articles of October 21, 1997, in the Washington Post and the New York Times indicated that the CIA was involved; no names were mentioned, but several people have told me independently that Larry Turnbull, a known hawk, was a chief player. He had stated at a previous international meeting at Princeton that he worked for the CIA. Hence, he was not working undercover.
Robert Bell, director of defense programs at the U.S. National Security Council, also participated in the news conference on October 20. He stated that considerable activity was observed at the Russian test site on August 14, two days before the seismic event. He also said that no seismic waves were detected on August 14 from Novaya Zemlya and that a plane sent out found no indications of radioactive release. Interestingly, on October 20, two whole months later, Bell still seemed uncertain about the nature of the seismic event of August 16.
A CIA press release on November 4, 1997, stated that “a seismic event occurred on August 16, 1997, in the Kara Sea. That seismic event was almost certainly not associated with the activities at Novaya Zemlya and was not nuclear. However, from the seismic data, experts cannot say with certainty whether the Kara Sea event was an explosion or an earthquake.” Nonetheless, many experts in the United States and abroad had concluded weeks before that it was an earthquake. While the CIA press release goes some distance in saying the event was not nuclear, no mention was made publicly that U.S. officials had earlier made a rush to judgment. In my opinion, the experts cited in the 1997 CIA press release should have been fired.
The early press releases about the seismic event of August 16, 1997, led some conservative commentators and organizations to continue to dig in their heels about its being a nuclear explosion long after a scientific consensus emerged that it was an earthquake. For example, Jeffrey Smith’s article stated, “Frank Gafney, director of the Center for Security Policy, said evidence his group is gathering bolsters the case that the tremor resulted from a nuclear test.” Gafney, much like Richard Perle, frequently opposed various arms control treaties and proposals. Neither he nor Perle is a scientist, but each often has claimed to have inside information.
The Report on CTBT Technical Issues by the U.S. National Academies in 2012 “drew three lessons from the handling of the 1997 event: (1) use all available data for accurate estimates of location and event characterization, (2) avoid the use of only a narrow range of azimuths such as those to southern Norway and Finland, and (3) provide a mechanism for new information to be updated to policy makers as it becomes available for occasional ‘problem’ events of this type.” Hopefully, these lessons have been learned, as small seismic events on Novaya Zemlya in 2007 and 2009 did not lead to leaks of false assessments to the media. Their high-frequency P to S wave ratios identified them as earthquakes (figure 12.5). Norwegian seismologists located the magnitude 2.8 earthquake of June 26, 2007, within 30 miles (50 km) of the main site where Russia had previously tested.
A visiting high-level Russian scientist told me later that Boris Yeltsin, president of the Russian Federation, was on vacation when the U.S. press reported the event of 1997 as a likely nuclear test. Yeltsin thought the head of the Russian Ministry of Atomic Energy had deceived him and that the agency had, in fact, conducted a nuclear test. He bawled out the minister over the telephone. Knowledgeable Russian geophysicists quickly conveyed to Yeltsin that the event was a small earthquake in the Kara Sea. My source told me this would have been a much more serious incident between the United States and the USSR if it had happened during the height of the Cold War.
While press reports about small seismic events on and near Novaya Zemlya largely ceased after 1997, some stories continued to be published in the Washington Times about work at the Russian test site and about Chinese earthquakes being nuclear explosions at or near their Lop Nor test site. Bill Gertz, who had previously leaked incorrect information about Russian seismic events in the 1990s, wrote a story on September 24, 1998, headlined “Blast in Arctic, Satellite Shows.” He wrote, “Vehicle activity photographed recently by a U.S. spy satellite indicates Russia is preparing to set off an underground blast at a remote Arctic nuclear-testing site, The Washington Times has learned.”
Gertz coauthored a similar story on June 18, 1999, under the headline “Inside the Ring: Small Nuke Test?” which said, “China set off a small nuclear-related blast over the weekend—days before U.S. Undersecretary of State Thomas Pickering arrived in Beijing to deliver the latest American apology, according to Pentagon intelligence sources.” Daryl Kimball of the Arms Control Association emailed me, “My sources tell me that: the report is based on satellite imagery only; that there is no seismic signature for this event from the official U.S. seismic stations (which are capable of detection to very low levels around Lop Nor); and that the AFTAC radionuclide surveillance plane has been sent and has found nothing related to the event.” China may have conducted a subcritical experiment (hydrodynamic test), as the United States and Russia had done already several times since the CTBT was signed in 1996.
IMPROVED DETECTION AND IDENTIFICATION CAPABILITIES
Detection and identification improved greatly after the Kara Sea earthquakes near Novaya Zemlya of 1986 and 1997 and especially after the earthquakes in Tibet and Central Asia in the late 1960s. The seismic events in Eurasia that the United States claimed as problems or anomalies in 1972 are shown as a bar at the upper right of figure 12.6. The bar at the upper left of figure 12.6 shows the improved capability in 2009 to monitor seismic events on and near Novaya Zemlya.
FIGURE 12.6
Problem seismic events are indicated at the top for various dates. Their yield in kilotons (kt), if they had been nuclear explosions, is shown on the horizontal axis at bottom. The bar indicates detection capability in 2009 for monitoring nuclear explosions at Novaya Zemlya. The lower part of figure, as modified from Kidder (1985), shows the frequency of U.S. underground nuclear explosions in Nevada from 1980 through 1984. TTBT indicates the 150-kiloton limit set by the Threshold Test Ban Treaty.
In 1985 Ray Kidder published the relative frequency of U.S. underground nuclear explosions in Nevada from 1980 to 1984, shown in the main body of Figure 12.6. He argued that peaks in his diagram near 7 to 20 kilotons and near 150 kilotons indicate tests with yields of high importance to the United States, whereas those with yields less than one kiloton were not. My work indicated the yields of past Soviet tests had a pronounced peak near 20 kilotons. The capability to monitor well-coupled explosions at Novaya Zemlya in 2009 is better than that of Kidder’s smallest, less frequent explosions.
Conversely, nuclear explosions with seismic magnitudes of 4.5 to 5.5, like those of the so-called anomalous events in the CCD report of 1972, are near the peak of testing in figure 12.6. If they could not have been identified, the ability to monitor nuclear tests of military importance would have been poor in 1972. Even then, however, they were identified as earthquakes in our special study of 1976.
