Page:Minority of One September 1961.pdf/12

... ABOUT THE PRETEXT FOR EVADING A NUCLEAR TEST BAN

American-Soviet negotiations of a ban of nuclear bomb tests have been stalled because of the American negotiators' claim that such explosions could not be seismologically differentiated from earthquakes. A number of seismological studies conducted by qualified scientists suggest that this claim is politically motivated and scientifically unfounded. Jay Orear, Associate Professor of physics at Cornell University, suggested in the April, 1960 Issue of the Bulletin of the Atomic Scientists, that "The main cause of difficulties at this conference (the December 1959 American-Soviet technical talks in Geneva) was due to the political instructions given both sides, not to scientific disagreements. To accuse the Soviet scientific delegation of politically inspired behavior is to some extent a case of the pot calling the kettle black."

The President is sending Ambassador Arthur H. Dean to Geneva, on August 24th, to hold one more round of talks with Soviet representatives. The acknowledged technical preparations of the Atomic Energy Commission as well as the added military emphasis in the Administration's policy strongly suggest that Mr. Dean's true mission is not to ban nuclear testing, but to provide a pretext for its imminent resumption by the United States.

Against this background, the article here presented has important political implications. - Ed.

By Sheridan Dausler Speeth

It has been argued that the U.S. should resume the testing of nuclear weapons because the Soviet Union may be doing so clandestinely. The ethical and political implications of this argument have been widely discussed before the general public. The empirical aspects of this problem have, unfortunately, received far less publicity. I would therefore like to tell about some work that has recently been done on detecting underground nuclear explosions. My discussion will be confined to the problem of underground explosions because the U.S. and the U.S.S.R. have agreed on the techniques for detecting other kinds of tests.

The major problem for the scientist, engaged in the design of a monitor system for the enforcement of a test ban treaty, is that of discriminating underground explosions from natural earthquakes. Let us begin by considering some of the differences between nuclear explosions and earthquakes which might be used in discriminating between them from a distance.

An underground nuclear explosion comes from a chain reaction which releases all of its energy in less than one millionth of a second. This energy release produces a temperature greater than a million degrees Kelvin in the next few millionths of a second. If the bomb is in a small chamber, this stage is followed by the expansion of the cavity until the pressure inside is equal to the pressure from the weighty rock and earth above. This results in a huge sphere, lined with about four inches of molten rock. Since the crushing of rock as the chamber expands results in a large transfer of mechanical (e.g., vibrational) energy to the surrounding earth, Edward Teller (U.S. technical consultant at Geneva) and others suggested that one could cheat with less chance of being found out if one were to test bombs in chambers so large that no expansion occurred after the explosion. Subsequent U.S. experiments with chemical explosions have shown that under ideal conditions this might reduce the apparent size of an explosion by a factor of three hundred. Under such conditions a pressure wave is spread through the room's atmosphere (which is so hot that electrons are not attached to their corresponding nuclei) and elastically absorbed by the surrounding walls.

Earthquakes are the sudden release of geological pressures which stress large volumes of material. To picture the mechanism, hold a pencil in front of you parallel to your shoulders gripping each end with your two hands as you would grip the handlebars of a bicycle. Now try to break the pencil by pulling your right hand in and pushing your left hand out. The break relieves the strain on the pencil producing two kinds of motion. (Toward you on the right and away on the left.)

Both explosions and earthquakes produce vibrations of the earth which are detectable at considerable distances. These waves are of three types; pressure waves, shear waves, and surface waves. To understand these, try to picture the following bizarre scene:

A group of people are standing in line as if waiting to enter a theater. Each has his hands on the shoulders of the person directly in front of him. At the end of the line someone sways forward and backward and this makes the person ahead of him do so too, and so on. The disturbance thus propagated from the end of the line to the front is a pressure or P wave. Were the person at the end of the line to change his swaying direction and begin to sway from left to right, pulling the man in front from left to right and so on, the disturbance sent along the line would be a shear or S wave.

P and S waves are produced by "seismic" events such as bombs and quakes; they are propagated through the earth's interior and emerge again at a continuous succession of points from the source. With an appropriate instrument, the seismometer, one can pick them up, graph them, and study their features. These waves are very low in frequency, and the range usually studied goes from one cycle every one- hundred seconds to one cycle per tenth second. There is a little-understood continuous source of background noise (called microseisms) which is very strong at around one cycle per second and less bothersome at higher and lower frequencies. There are various paths which a wave might take; this produces multiple effects much like echoes. A pressure wave (P) may be followed by a reflection from some distant part of the earth's surface (pP) and later by a refraction through the earth's core (PKP) and so on. The time of arrival of the various echoes can be used to judge the depths of earthquakes. These occur at depths of 60 to 400 kilometers, depths far deeper than any conceivable site of bomb tests.

Since an explosion pushes out in all directions whereas an earthquake both pulls and pushes, the ability to see clearly the direction of first motion in a seismogram would enable us to make an inspection decision. (A pull on any side means it must have been an earthquake.) Unfortunately, this only works for small distances when using single, high frequency seismometers on the earth's surface. This is due to the fading and distortion of the first motion signal combined with the interference due to noise (microseisms).

Tremendous progress has been made in overcoming these difficulties. Richard Roberts, of the Carnegie Department of Terrestial Magnetism told, in his report to the Holifield committee, of his experience with seismometers lowered into oil well holes. The microseismic noise disappears as you go further under the earth's surface. Jack Oliver and others of the Lamont laboratory obtained similar results from ocean bottom seismometers. John Gerrard, Director of Earth Science Research at Texas Instruments, Inc., described, in the Berckner Report to the State Department, how further signal enhancement may be achieved through the use of arrays of many seismometers. By carefully combining the inputs from such an array one can get a cancellation of noise while simultaneously increasing the strength of the signal. The square root of the number of seismometers used gives an estimate of the factor by which the signal-to-noise ratio can, in this way, be improved. Jack Oliver has shown how distortion problems are overcome when only the very low frequencies are studied, and John Tukey, Professor of Statistics at Princeton, has developed a computational technique for treating seismic data which makes it possible to determine first motion in spite of distortion.

My work in this area began with the following notion: Distinguishing an explosion from an earthquake may in many ways be similar to (and as complicated as) deciding which one of two of your friends is speaking on the telephone. Let us press this analogy by considering the input and response of both the seismometer and the telephone microphone. If your friend is in a normal room with plaster walls (plaster reflects 80 per cent of the sound which hits it), then there will be multiple arrival times for each of his vocal pressure waves, a parallel to the seismologists' P, pP, PP, and PKP, and other waves. The band- limiting performed by the telephone trans- mission corresponds to the narrow bandpass of most seismometers. If the friend's room were to contain machinery or other sources of noise, you would have to perform a task not unlike distinguishing a seismic signal from the noise of microseisms. If the voice decision were then made on the basis of vowel pronunciation, you would have demonstrated the ear's ability to use the information contained in the temporal dynamics of the short-time audio spectrum. The analogy could be indefinitely extended, but by now the experimental question should be obvious: Would any benefits for the seismologist accrue from having his seismometer output presented as an auditory display?

By playing back a magnetic recording of a seismometer's output at three hundred times the speed with which it was recorded, the frequencies are translated into a range where it is possible to listen to them discriminately. A series of experiments conducted last summer indicates that one can train listeners to distinguish the seismographs of earthquakes from those of underground explosions simply by their sound. Listeners were successful in separating one class events from the other in over 90 per cent of the cases presented, although the explosions studied were smaller than the "nominal yield" bomb discussed at Geneva. The events were monitored from as far as 4,000 kilometers.

What about cheating by the big-hole method suggested by Teller? The problem is much simpler than often claimed. Well over ninety per cent of the earthquakes in the Soviet Union occur in regions in which there are no

Mr. Speech, a full-time consultant for the Bell Telephone Laboratories, has been working on nuclear test monitoring systems. He has been recently assigned a Defense Department contract for research on a system dealt with in the article here presented. He is a member of the Acoustical Society of America, the Sigma Xi and the Society for Social Responsibility in Science.