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

516. Memorandum for the Record by Keeny¹

Washington, January 15, 1960

SUBJECT

• DOD Position Paper on Threshold for Underground Tests

1. At the Principals’ meeting on January 12, Secretary Gates presented a staff paper setting forth the DOD position on the threshold for underground nuclear tests. This paper specifically criticizes the threshold proposals, which Dr. Kistiakowsky presented at the Principals’ meeting on January 8 at the request of Secretary Herter and Mr. McCone, as being deficient in that “the number of inspections would be determined arbitrarily (at a comparatively low figure) without appropriate relationship to the scientific facts and the capabilities of the detection system.” The proposals presented by Dr. Kistiakowsky were developed by a representative group (including Drs. H. Brown, S. English, J. Oliver, W. Panofsky, C. Romney, J. Tukey, and General A. Starbird) which considered the proposals to represent a balanced judgment properly related to the scientific facts and system capabilities.

2. The alternate threshold proposals contained in the DOD position paper essentially call for the inspection of all unidentified events above magnitude 5 (i.e., 100 per cent of all unidentified events or 50 per cent of all detected seismic events). The implication contained in the DOD position paper that this proposal is uniquely in accord with scientific facts and the capabilities of the detection system is not correct since it does not take the following factors into account:

a. An effective deterrence does not require 100 per cent coverage of all possible violations. This is particularly true in the case of nuclear tests where a series of violations rather than a single clandestine test would be required to constitute a real security threat.

b. Among the events which are not positively identified as earthquakes (by U.S. criteria), it would be possible to establish the relative degree of suspicion of various events by means of the following information:

(1) Utilization of auxiliary seismological information. There is a large amount of seismic data which, although not at present sufficiently definitive to be codified as criteria, would provide information indicative of whether an event was either an earthquake or possible explosion. This type of information would be particularly valuable in the case of the larger seismic events considered in the various threshold proposals. Although there is no agreed estimate as to the cumulative value of this auxiliary information, it would probably reduce the number of events worthy of suspicion by at least a factor of 2 (and possibly much more) and would probably focus particular attention on a relatively small number of events in the yield ranges above 4.75.

(2) Utilization of covert and overt intelligence. As long as the choice of events to be inspected in the USSR is made by the U.S., we can make use of all sources of covert intelligence as well as common sense based on geographic and geological considerations in picking the seismic events to be inspected. Since the extent of our intelligence capabilities are presumably not known by the USSR with certainty, the unpredictable aspect of this factor would in itself probably constitute a significant deterrent.

3. The DOD position paper directly criticizes the approach to inspection which would require “the East and West to act unilaterally in selecting the events to be inspected in the host countries’ territory,” on the grounds that this would degrade the authority of the International Control organization. It is precisely this ability to chose events for inspection that will provide the maximum deterrent effect since it would permit the unrestricted use of all of the information discussed above (i.e., auxiliary seismic information as well as overt and covert intelligence) and would remove debate on the validity of individual inspection operations from the international organization. The original U.S. proposals on inspection procedures were based on “random” selection since the concept of selection by the other country or side was not considered negotiable. However, this concept of selection has now been advanced by the Soviets themselves.

4. The threshold concept is presumably being considered by the Principals with the hope of discovering a formula which might have some chance of being negotiable with the USSR (reasonable deterrence coupled with low level of inspection), and at the same time, would establish clearly in world opinion that the U.S. is attempting to extend the test ban as far as technical considerations will permit. It must be recognized that any threshold proposal that we advance will probably be attacked by the Soviets as an attempt on our part to continue testing. The DOD proposal, however, would appear to give the Soviets particular opportunity to question our motives since it couples a relatively high threshold with the highest possible level of inspection. The question should, therefore, be asked whether the threshold proposal in this form would actually serve the basic purpose for advancing the threshold concept in connection with the present Geneva Negotiations.

Attachment

Paper Prepared in the Department of Defense

DOD Staff Paper Presented at the January 12 Principals’ Meeting

PROPOSED POSITION OF THE UNDERGROUND THRESHOLD

1. With regard to an agreement on the discontinuance of nuclear weapons testing, the United States Government has taken the position that the level of inspections must bear an appropriate relationship to the scientific facts and the detection capabilities of the Control System. The USSR representatives, on the other hand, have insisted that the number of inspections must be small, despite the scientific data which indicates that the unidentified events will be numerous. The failure of the recent Technical Conference to achieve agreement on the U.S. criteria which would establish eligibility for inspection also prevented agreement as to the number of unidentified events which might occur. Without agreement on criteria, some other basis for initiating an inspection would have to be devised. Dr. Kistiakowsky has set forth three alternatives, two of which would seem to require agreement on criteria and the third susceptible of application without such agreement. All of the alternatives have one deficiency in common—the number of inspections would be determined arbitrarily (at a comparatively low figure) without appropriate relationship to the scientific facts and the capabilities of the detection system.

2. Upon review of the threshold concept as applied to underground test ban prepared under Dr. Kistiakowsky’s direction, the DOD is of the opinion that it would be to the best interests of the U.S. to withhold the introduction of this concept until or unless it had been clearly demonstrated that the USSR (and possibly the U.K.) will not agree on the inspection criteria established by the U.S. technical group or accept a treaty confined to atmospheric tests. The following reasons are cited: Under the threshold concept, one of two procedures is possible, i.e., to select events to be inspected on the basis of the “characteristics” of instrument readings (criteria), or on the basis of a percentage of the total events. If the Soviets refuse to accept the U.S. criteria, we will be led into a compromise of our best scientific data and thus establish a bad precedent for future disarmament negotiations as well as invite the prospects of adverse Congressional and public reaction. If we suggest a percentage of the total events in order to eliminate the criteria problem, we will introduce a very undesirable factor, namely, the necessity for the East and West to act unilaterally in selecting the events to be inspected in the host countries’ territory, thus degrading the authority of the International Control organization. Obviously, if there is no agreement on inspection criteria, there would be no agreement on which events should be eliminated as natural occurrences.

3. The Department of Defense believes that the threshold concept in terms of the magnitude of the seismic events contains uncertainties and ambiguities which detract from its desirability for immediate application without further validation. Further, recognizing the uncertainties in natural and possible artificial decoupling, the proposal does not seek to regulate underground testing to any specific maximum yield but only to make it incumbent upon the signators to the treaty to conduct their operations in such a way as to avoid an instrumental response exceeding the agreed magnitude.

4. If, however, it is decided to be in our over-all interest to advance this concept as a proposal, it is considered that the threshold should be set to meet the following objectives:

a. The threshold should be high enough to require only the number of inspections which would be feasible within the practical limitations of the numbers of inspection groups and their operations which can be supported by the System.

b. The number of inspections must bear an appropriate relationship to the scientific facts, taking into account the best scientific data bearing on the problem; and

c. The threshold should be high enough to permit certain developmental tests should the United States at some time decide to undertake such tests.

5. The question, therefore, is what threshold in terms of the magnitude of the seismic waves generated by an earthquake should be adapted to assure that the foregoing objectives would be attained. Accepting the following tabulation as representing the best U.S. estimates of the data shown and adopting the assumption that Communist China will not accept stations on its territories, it is apparent from the fourth column that a magnitude of 5.00 is about the minimum that could be adopted taking account of the considerations stated above.

M	N	Nu	Nu*	YR
5.25	32	3	8	110
5.00	59	8	27	43
4.75	105	25	70	19
4.50	185	95	150	8
4.25	335	245	300	3.3
4.00	580	490	545	1.45

M	Magnitude
N	Total number of earthquakes in the USSR of magnitude greater than M
Nu	Number of unidentified events in the USSR of magnitude greater than M
Nu*	Number of unidentified events if stations are limited to USSR and US
YR	Yield equivalent in KT for Rainer coupling

6. In the light of the foregoing, it is recommended that the following position be established with respect to the inclusion of a threshold for underground tests in a phased treaty:

a. There shall be agreement with the inspection criteria proposed by the U.S. and this criteria shall be written into the treaty.

b. In the first phase of the treaty, no signator nation shall conduct underground tests above a threshold which shall be defined in terms of the magnitude of seismic waves recorded at the locations and in manners specified in the treaty.

c. The initial threshold shall be magnitude of 5.00 as calculated by the Gutenberg and Richter 1956 formula.

d. All events of magnitude greater than 5.00 shall be analyzed in accordance with the agreed inspection criteria and all those events not thus identified as natural events shall be eligible for inspection.

7. In the event that the Soviets refuse to accept the U.S. inspection criteria as presented by Dr. Fisk and his Technical Group, the following modifications of the above should be introduced:

a. Same as Paragraph 6(b) above.

b. Same as Paragraph 6(c) above.

c. 50% of all detected events of magnitude of 5.00 shall be subject to inspection. For those events occurring in the U.S. or in territories under the control of the U.S. or U.K., the USSR shall have the option of designating the events to be inspected; and for events occurring in the USSR the selection shall rest with the U.S.–U.K.

8. In either of the above propositions the further condition should be added that the Control Commission will maintain a continuing study and analysis of the feasibility of improving the quality of the detection system and shall from time to time report to the signator nations its recommendations as to improvements to be installed and the corresponding threshold to be adopted. The installation of improvements and the adoption of thresholds below or above those specified in the initial agreement shall be subject to the unanimous agreement of the signator nations.

9. If the Soviets do not agree to any of the foregoing proposals involving a threshold, a phased treaty confined initially to atmospheric tests should be proposed.

Attachment

Paper Prepared by Kistiakowsky

Washington, January 8, 1960

THRESHOLD FOR UNDERGROUND NUCLEAR EXPLOSIONS

I. Definition of Threshold

It is technically possible to define a “threshold” for underground nuclear explosions in terms of the magnitudes of the seismic events detected by the Control System. There are several ways in which magnitude might be defined for this purpose. A detailed definition of magnitude based on the use of short period P waves has been prepared (see attachment). This definition could be used operationally in a straightforward manner to define magnitude above about 4.25. Above about magnitude 4.5, the apparent magnitude by this definition corresponding to a given explosion under fixed coupling conditions could vary between ∤0.1 and ∤0.2 magnitude units. This corresponds to about a ∤25 per cent variation in equivalent yield. It would, therefore, probably be possible to test with confidence nuclear explosions with at least 50 per cent of the yield permitted by the threshold in particular coupling conditions. To avoid accidentally exceeding the threshold, a tester could “creep up” to this 50 per cent level of the threshold by means of a series of smaller explosions. It is to be emphasized that it is difficult to assign a kiloton equivalent to a particular magnitude since, (quite aside from the very large decoupling factors possible in large cavities), this value could vary by a factor of two or more depending on particular medium coupling conditions. In addition, it should be noted that the Soviet report to Technical Working Group 2 differs substantially from that of the U.S. in evaluating the seismic magnitude of specific U.S. nuclear tests.

II. Quantitative Analysis of Thresholds

Estimates of the total number of seismic events and the number of unidentified seismic events above various magnitudes in the USSR, based on the earthquake statistics and criteria in the U.S. report to Technical Working Group 2, are summarized in the following Table. (The estimates used are preliminary calculations by the Livermore Laboratory reduced uniformly by 25 per cent in order to conform with previous AFTAC assumptions as to the percentage of world-wide earthquakes which would occur in the USSR.)

TABLE

m	N	Nu	Nu*	YR
5.25	32	3	8	110
5.00	59	8	27	43
4.75	105	25	70	19
4.50	185	95	150	8
4.25	335	245	300	3.3
4.00	580	490	545	1.45

m	Magnitude
N	Total number of earthquakes in the USSR of magnitude greater than M
Nu	Number of unidentified events in the USSR of magnitude greater than M
Nu*	Number of unidentified events if stations are limited to the USSR and US
YR	Yield equivalent in KT for Rainier coupling

In considering this Table, the following points should be noted:

1. The average annual number of earthquakes becomes increasingly uncertain with decreasing magnitude and in the lower magnitude ranges is probably uncertain by a factor of two.

2. The estimates represent the average number of annual events and there are actually significant variations in the number of annual events.

3. The ability of the system to identify events within the USSR is dependent upon the extent to which the Control System is installed outside the USSR and would be significantly reduced if stations were not installed elsewhere in Asia.

4. Soviet and U.S. scientists disagree significantly on the values for the equivalent yield with Rainier coupling (YR). [text not declassified]

III. Level of Threshold

The establishment of a particular magnitude level as a test threshold presents a difficult problem which is not primarily technical in nature. A judgment on this problem must be based on consideration of such factors as the following:

1. If, on the one hand, the threshold is established on too high a level:

(a) Higher yield underground explosions which could probably be controlled with reasonable effectiveness would be legalized.

(b) There would be little or no justification for inspection.

(c) Support of U.S. position on test control issue in world opinion would presumably be weakened.

2. If, on the other hand, the threshold is established at too low a level:

(a) Number of inspections would either be too large or effectiveness of control would be reduced.

(b) System would have little or no capability of reducing number or eligible events in lower range of permitted magnitudes.

(c) The incentive to undertake concealment by large cavity decoupling would be increased.

In balance, it is concluded that magnitude 4.75 constitutes the best compromise among the above factors. In addition, it should be noted that we have direct measurements from the Blanca explosion corresponding to this magnitude under Rainier coupling conditions.

IV. Number vs. Percentage Quota

The eligibility of events for inspection under any magnitude threshold can be determined on either a numerical quota or percentage quota basis. The comparative advantages of these two methods are as follows:

1. Percentage Quota

a. Covers uncertainties in estimates and yearly fluctuations in number of seismic events.

b. If based on unidentified events:

(1) Presents much greater incentive than numerical quota for Soviets to improve technical capabilities of system;

(2) Presents stronger incentive to Soviets to include China in system in order to reduce number of eligible events in USSR;

(3) Eliminates possible Soviet argument that there are too few unidentified events for a particular numerical quota; and

(4) Same percentage can apply to different thresholds.

2. Number Quota

a. Easier to negotiate since essentially accepts Soviet quota concept and indicates clearly level of inspection to be conducted.

b. Independent of changes in criteria of eligibility for inspection or could be applied against all events independent of criteria.

c. Somewhat easier to apply.

d. Possibly more effective from point of view of public relations as representing simple, understandable proposal.

V. Suggested Position

Technically, there appear to be three reasonable approaches to a threshold proposal based on magnitude 4.75 seismic events:

1. Require inspection of 10 per cent of all located seismic events with magnitude above 4.75. Since the number of located seismic events of this size in the Soviet Union is 105, this would result on the average in approximately 10 on-site inspections per year. This proposal could bypass the lack of agreement on the criteria for eligibility of inspection.

2. If the Soviets are willing to accept the U.S. criteria for determining the eligibility of seismic events for inspection, require inspection of 20% of all located seismic events with magnitude above 4.75 which are deemed eligible for inspection. Since the number of unidentified seismic events above magnitude 4.75 is estimated to be between 25 and 70, depending on the extent of installation of the system on a world-wide basis, this proposal would result in an average annual level of inspection between 5 and 14.

3. Establish a quota of 10 inspections per year which could be applied against either the total number of located seismic events or the number of unidentified events with magnitudes greater than 4.75.

Technically, the proposals based on the use of percentages, particularly the first proposal (i.e., 10 per cent of all located seismic events above magnitude 4.75), appear to be the safer course in view of the uncertainties in the estimates of the number of seismic events. However, the quota proposal would probably give the same general level of deterrence and might have advantages from the point of view of negotiability and public relations.

Attachment

SUGGESTED THRESHOLD DEFINITION

1. A seismic event shall qualify for consideration by the Control System if its seismic waves are as large as, or larger than, the seismic waves recorded from an earthquake of magnitude 4.75 as determined by the procedure given in paragraph 2 below.

2. The method for determining magnitude is as follows:

a. From the seismograms of each control post which detects signals from the event in question for which A is measurable as defined below and for which the epicentral distance is greater than 16° and less than 36° to 90°², determine whether or not the magnitude m, according to:

is greater than 4.75

In the above formula, the symbols are defined as follows:

A = half of the maximum peak positive to negative amplitude (displacement), measured in microns, in the first three cycles of the P waves, as recorded by a vertical seismograph of the type described in paragraph 3 below. For the purpose of this computation, signals are considered to be detectable and measurable to sufficient accuracy if A is 3 times the peak noise amplitude during the preceding few minutes.

T = the time, measured in seconds, between the peak displacement (used in determining A) and the next following peak.

G = the steady-state magnification of the seismograph at period T.

Q is given as a function of distance in Table 1.

b. If the apparent magnitude exceeds 4.75 at half, or more than half, of the control posts specified in 2a above, the event is judged to have a magnitude of 4.75 or more.

3. The vertical seismograph used for measuring A, as defined in paragraph 2a above, shall be a short period seismograph with characteristics which will permit operation of single seismographs at quiet stations with magnification greater than 106 at the frequency of peak response. It may, in particular, conform to the specifications in GEN/DNT/TWG. 2/9, Annex IV, page 12.

Table I

(Sample)
Distance	G
16°	5.92
17°	5.95
18°	5.98
19°	6.02
20°	6. 05
.	.
.	.
.	.
90°	7.03
Note: 1° =111. 11 Km

1. Source: Comments on Department of Defense staff paper (attached) on the threshold for underground tests. Secret. 18 pp. Eisenhower Library, White House Office Files, Additional Records of the Office of the Special Assistant for Science and Technology, Disarmament, NT, Threshold.
2. (The most desirable cut-off distance for data to be used in determining magnitude under this definition is still under consideration.) [Footnote is in the original.]