Alaska: Seismic Signal Crossroads of Underground Nuclear Tests

Alaska bills itself as being the air crossroads of the world. Polar routes of many of the world's airlines converge on the state in passing from one continent to another. There is another sort of "traffic" that crosses the state of which few people are aware. These are the travel paths taken by seismic waves generated by underground nuclear tests in the Soviet Union and the U.S.

Although the news is hardly news any more, and it rarely makes the front page, both countries conduct nuclear tests on an approximately monthly basis. Seismic sensors in Alaska are particularly well situated to detect and compare the explosions from the different test areas. The Nevada test site near Las Vegas lies 2500 miles to the southeast. The Soviet Union's test site in Siberia lies 2500 miles to the northwest. Alaska lies in between.

Actually, the Soviets do not restrict their testing to one location, but utilize several. In addition to Siberia, they commonly use eastern Kazakh Province near the Himalayas, the Ural Mountains, and the desolate islands of Noveya Zemlya in the Barents Sea.

The tests--at least those of any significant size--are easy to detect and identify seismographically. Their "signatures" differ from those of an earthquake in the higher frequency ranges, and of practical necessity, they occur at more shallow depths than most earthquakes. In addition, earthquakes result in impulses both toward and away from the epicenter, depending on where the recording station is located, while explosions always produce an initial impulse away from the source. In fact, this characteristic of nuclear explosions provides a convenient check for seismologists to assure that their signals have not somehow become reversed in the recording process.

It is interesting, but probably not significant, that most Russian tests generate energy equivalent to about a magnitude 6 earthquake, while the majority of U.S. tests scale around magnitude 5 or lower. Also, the Americans have a penchant for setting off their explosions precisely at an even quarter-hour in the early morning, while the Russians detonate theirs at random times. This almost makes it appear as if they set them off first and then record the time at which they occur.

As easy as it is to identify nuclear tests, it is almost surprising that the exchange of seismic signals between East and West began only during this past October. Moscow delivered a signal from one of its seismographic stations to the U.S. via satellite with relays in Czechoslovakia, West Germany and England. In exchange, data collected from six stations in Canada and the U.S. were relayed to Moscow.

The November 9 issue of Science reported that the exchange was to be the first of many by seismologists from 34 countries who are participating in this novel three-month experiment organized by the Conference of Disarmament in Geneva. The purpose of the experiment is to rehearse a key mechanism for verifying compliance with a comprehensive ban on nuclear weapons tests.

It might be asked why such cooperation is needed if it is possible to detect explosions with monitoring equipment situated outside the testing nation. For one thing, there is a great deal of difference between detecting the detonation of, say, a 50 kiloton device and one of only two kilotons. The signal from the smaller event would obviously be much harder to detect and identify. Further, it is probably possible to mask even larger tests by conducting them in huge underground caverns, which would effectively "decouple" the blast from the surrounding earth. This possibility has long been a stumbling block in the way of test ban treaty verification.

Although we are presently undergoing a feeling-out period aimed toward agreement on a test ban treaty, the arrangement is not entirely satisfactory to DARPA, the Defense Advanced Research Projects Agency. Ralph Alewine, director of DARPA's Geophysical Services Branch, expresses concern that data from only one Soviet station are presently being sent. In addition, he notes that while two Soviet nuclear tests were conducted in late October, signals from the tests somehow did not appear in the Soviet station data provided to the U.S.