Foreign Relations of the United States, 1958–1960, Volume III, National Security Policy; Arms Control and Disarmament, Microfiche Supplement

405. Report of the Panel on Seismic Improvement¹

Washington, January 7, 1959

1. The Panel on Seismic Improvement (PSI), appointed by the Chairman of the President’s Science Advisory Committee, met in Washington on January 6 and 7, 1959, to review measures whereby it “would be reasonably feasible within the present state of seismic technology to improve the capabilities of the system recommended by the Geneva Conference of Experts to detect and to identify seismic events as either earthquakes or explosions without increasing the number of manned control posts in the system.”

2. The capabilities of the Geneva system with regard to underground tests have recently been reevaluated by another Panel on the basis of new data from the underground tests at HARDTACK II. The PSI did not attempt to evaluate further the specific capability of the Geneva System. The proposals recommended herein would increase the estimated capabilities of the Geneva System. It is noted that the data on nuclear shots used in these estimates was from Rainier and HARDTACK II and thus has all the limitations of that small sample of nuclear test conditions. The PSI has not concerned itself with the possible seismic effects of nuclear tests under different conditions or the possibilities of concealment by decoupling or other techniques.

3. The Geneva System of seismic identification places principal reliance on the assessment of a single phenomenon, i.e. the direction of displacement of the first arrival of the P-wave in a specified network of seismic instruments. The PSI considered a variety of seismic phenomena and techniques which have been suggested to increase the capability of the Geneva System without adding manned control posts, including:

(a) Evaluation of the first motion of the P-wave with aid of approximate inverse transfer functions.

(b) Surface wave phenomena using long-period instruments.

(c) Unmanned, auxiliary seismic stations.

(d) Larger arrays of seismometers at manned control posts.

(e) Improvements from increased knowledge of the transmission properties of the earth by experience in operation of the system.

(f) After shocks as a diagnostic feature.

(g) Radiation asymmetry at the source.

(h) Use of computers in data analysis.

(i) Use of higher frequency seismic signals.

(j) Detectors on ocean bottom.

(k) Detectors in deep hole.

(l) Diagnostic possibilities of microseisms.

(m) Focal depth of disturbances.

5. [4.] On the basis of its review of the above techniques, the PSI believes that the following four promising approaches are within the present limits of technology and should be considered:

(a) Analysis of long-period surface waves. The capability of the Geneva System for the identification as earthquakes of seismic events equivalent to 5 kilotons or larger by the analysis of long-period surface waves is conservatively estimated at 25% and may be much larger.

(b) Network of unmanned auxiliary seismic stations. A triangular network of such unmanned stations is suggested, spaced 170 km apart between the stations of the 1000 km grid proposed in Geneva for the seismic regions of the world. If this network were installed and effectively operated, it would very greatly increase the capability of the Geneva system for identifying as earthquakes those seismic events occurring in interior areas and corresponding in energy to underground tests equivalent to one kiloton or larger.

(c) Larger arrays of seismometers at manned control posts. It is reasonably certain that the replacement of 10 distributed seismometers per control post as recommended by the Geneva Conference of Experts with arrays of approximately 100 distributed seismometers would increase the signal to noise ratio by a factor of from 1.5 to 2.5. This would substantially increase the capability of the system to identify small seismic events as natural earthquakes.

(d) Detectors in deep holes. A detection method which employs a seismometer in a hole at a depth of several thousand feet is being investigated at the present time. The method offers possibilities for improving the detectability of small signals by a factor of the order of ten, provided that the technological problems of operating instruments at the required depths can be solved. The results to date are incomplete but encouraging. If the factor of ten can in fact be achieved, it would drastically increase the capability of the system to identify small seismic events as natural earthquakes. More definitive results are expected by mid 1959.

The proposals (a), (b), and (c), above, are discussed more fully in the appendix attached.

6. The PSI invites attention to the fact that detection of aftershocks by specially and immediately implaced seismometers can be used by an inspection party as an aid in establishing that an unidentified seismic event was in fact a natural earthquake.

7. The PSI believes that seismic research has not in the past been supported as strongly as many other areas of science. Vigorous research in seismology is certain to produce many improvements that cannot now be predicted. It is certain that this would lead to an improvement in detection capabilities. The PSI will shortly report its detailed recommendations concerning research in seismology.

8. The PSI urges that sample detection networks be established without delay as representative systems to disclose operational and design problems and provide a firmer basis for the assessment of detection capabilities.

9. The Geneva Conference of Experts recommended supplying new instruments to the existing world network of seismic stations. The PSI believes that this recommendation should be implemented within the next year even if it must be done unilaterally by the U.S.

• Lloyd V. Berkner (Chairman)
• Victor H. Benioff
• Hans A. Bethe
• John Gerrard
• David T. Griggs
• Jack H. Hamilton
• Julius P. Molnar
• Jack E. Oliver
• Frank Press
• Carl F. Romney
• Kenneth Street, Jr.
• John W. Tukey

Appendix

I. Analysis of Long-Period Surface Waves

Long-period seismograph data (periods greater than 5 seconds) available from HARDTACK II and natural earthquakes suggest additional criteria for the identification of seismic events as natural earthquakes. In the yield range 5KT–23 KT, stations at distances up to at least 3500 km can provide the necessary data, in the absence of microseismic storms. The capability of the Geneva System for the identification as earthquakes of seismic events equivalent to 5 kilotons or larger by the analysis of long period surface waves is conservatively estimated at 25% and may be much larger. Estimates of capabilities are based on observations with instruments not designed for this purpose. The use of specifically designed equipment should further improve the estimates of capabilities.

On the basis of present technology, it is concluded that the following techniques are available:

a) Love-Rayleigh wave amplitude ratio. A preliminary study of amplitude ratio of Love waves to Rayleigh waves for periods greater than about 10 sec. from earthquakes and underground explosions has been conducted at the Lamont Geological Observatory at Palisades, New York. The results indicate that in the equivalent magnitude range 5 KT–23 KT a single station at a distance of 3500 km or less can identify about 10% of seismic events as natural earthquakes. Data from stations in appreciable different azimuths are relatively independent and so increase the probability of identification significantly but not above some as yet undetermined limit.

b) Spectra of surface waves recorded on long period, horizontal component seismographs. Data from Palisades, Pasadena, and Berkley indicate a systematic difference in the long period spectra of earthquakes and underground nuclear explosions. Although the effect is striking, it is not possible to quote the capability of this method at this time except to estimate that identification of earthquakes from a network of stations is no worse than 10% and the upper limit is open.

The possibility exists that further results can be obtained in the immediate future by additional studies of the amplitude ratio for Love-Rayleigh waves and its aximuthal dependence for earthquakes. This will provide additional data to verify the estimate of 10% identification of earthquakes and examine the possibility of increasing this figure.

II. Network of Unmanned Auxiliary Seismic Stations

Interpolation of unmanned automatic seismic stations into the grid of control posts of the Geneva plan gives promise of providing significantly greater information on weak seismic events, corresponding in intensity to one kiloton. A triangular network of such unmanned stations is suggested, spaced 170 km apart between the stations of the 1000 km grid proposed in Geneva for the seismic regions of the world. In such a network, a one kiloton shot coupled seismically to the same degree as Rainier, Logan, and Blanca would give 50 millimicron amplitudes or higher for first motion of P-waves at nine stations on the average. Thus data from single vertical-component seismometers disposed in such a network would suffice to detect first motion with reasonable certainty.

The practical problems of installing such a network may be estimated by noting that the spacings mentioned above imply having 35 auxiliary stations per main station and that the maximum communication distance between a main and auxiliary station is 600 km. Each auxiliary station would require a seismometer, a recording device (for providing a permanent record), a clock, a radio transmitter, a source of electric power, and probably a data storage device which will permit compressed data transmission as required. The cost of the technical apparatus needed for a single station, when manufactured in large quantities, might be ten to thirty thousand dollars. Access roads (or helicopter pads), installation and related costs may run the total costs up considerably higher, depending on local conditions. Periodic maintenance and record retrieval would probably be required at 30 to 60 day intervals.

The need for this network of auxiliary stations is clearly greatest in the areas of the world where earthquakes occur with high frequency. It might be acceptable to limit application of the network only to these areas, possibly 20% of the land surface of the world. Also one could tailor power and communication arrangements (possibly using wire lines in some places) according to the local facilities available.

If this network could be installed and effectively operated, it would very greatly increase the capability of the Geneva System for identifying those seismic events which are earthquakes, occurring in interior areas and corresponding in intensity to underground one kiloton or larger shots. The practical difficulties of installation, maintenance, and operation of the system, especially in remote areas, should not, however, be underestimated, and the possibilities for occasional spoofing must be recognized.

III. Larger Arrays of Seismometers at Manned Control Posts.

On the basis of present knowledge, replacing 10 distributed seismometers in a 3 km × 3 km square as recommended for each control post in the Geneva System by 100 distributed seismometers in the same square would affect the signal to noise ratio at frequencies near 1 cycle per second as follows using specifiable techniques:

(1) It is reasonably certain that an improvement at most stations by a factor of between 1.5 to 2.5 will be obtained;

(2) There is reason to hope for improvements in the range from 2.0 to 2.5;

(3) Increases from 10 to 10k stations should provide improvement by a factor from k0.2 to k0.4 for k = 10.

(4) Data which can be obtained within one month will provide a much firmer estimate of what may be expected. The data to be obtained include, most importantly, data on coherence of noise at station separations of 150 to 1500 feet, and secondarily, data on dependence of typical noise levels on wind velocity.

(5) When more is known about noise characteristics, it may well be possible to gain further improvement by applying other analytical techniques to a 100 seismometer array.

1. Source: Suggests improvements to detection systems. Confidential. 6 pp. Eisenhower Library, White House Office Files, Additional Records of the Office of the Special Assistant for Science and Technology, Disarmament.