Foreign Relations of the United States, 1961–1963, Volume VII, Arms Control and Disarmament

196. Memorandum From the President’s Deputy Special Assistant for National Security Affairs (Kaysen) to President Kennedy⁰

Washington, July 20, 1962.

SUBJECT

• New Data on Detecting Underground Nuclear Explosions

1. You asked last week how long the Defense Department had known the new figures on the basis of which they made their announcement about the change in our ability to detect underground nuclear tests in the Soviet Union a little while ago.¹ There is no simple answer to this question in terms of time. In one sense the Department of Defense did not know that they had the new information until they were asked by Disarmament (ACDA) on 15 June to give an evaluation of the proposals that the neutrals had been making in Geneva.² These proposals rested on the proposition that we could reliably detect underground nuclear tests in the Soviet Union without requiring seismic stations inside the Soviet Union. But in the course of looking up the answer to this question they looked at all their earlier figures and thought about them anew. This led to a realization, by 3 July,³ that the previous estimates of how many small earthquakes per year of a given size occurred on the average in the Soviet Union were too high by a significant factor.

2. However, the answer that we first knew the new figures on 3 July is an over-simplification. In fact, the scientists in the Air Force who run the detection system for nuclear explosions (AFTAC) suspected as early as 1958 that their current interpretation of their measurements might be incorrect. The Hardtack shot at the end of 1958 raised this suspicion. However, we changed our official estimates in the opposite direction. In the summer of 1961 the first half-year report from one of the new accurately-instrumented and specially organized stations in the Vela program led to further doubt about the validity of our previous estimates. These doubts remained no more than doubts until February of this year. Then this data on the Soviet tests made the suspicion stronger. Our measurements on the French test of 1 May and our measurements on the Aardvark shot in our own underground test series of 12 May changed suspicion to new knowledge.⁴ The inquiry by ACDA crystallized the new knowledge into a new estimate.

3. The essence of the problem lies in the fact that our previous measurements rested on the observations of the world-wide system of existing academic and governmental seismic stations. These stations had two shortcomings. First, their instruments were not carefully calibrated against each other or a known standard. Second, they were chiefly useful in measuring relatively large earthquakes in which seismologists were interested for scientific purposes. When we created the Atomic Energy Detection System of stations operated by the AFTAC, we got the first set of stations with high quality, uniformly calibrated instruments directed toward measuring small underground shocks, such as arise from underground nuclear tests. These stations were designed to measure a different kind of seismic shock than the others, namely the underground or body wave as opposed to the surface wave which was used to measure large earthquakes. The problem has always been: what is the relation of the two kinds of measurements? It is only on the basis of this relation that we are able to interpret our observations to make an estimate of the number of earthquakes of given magnitude in the Soviet Union. What the events I have summarized above did was lead the scientists in AFTAC to conclude that the old formula for relating these two sets of measurements was wrong, and that a new formula was needed.

4. The estimates of the frequency of small shocks⁵ in the Soviet Union which we presented at the Geneva Experts Conference⁶ were challenged by Soviet scientists. Their figures differed from our figures by roughly the amount of the correction factor which we have now accepted. Thus our new figures are about the same as those the Soviets offered in 1958. We still do not understand whether this is a coincidence, which is what the AFTAC scientists say, or whether the Soviets did have a better understanding of the real relation between the two sets of measurements. We cannot simply say that the Soviets had the right formula, because the nature of their argument about the difference between these two sets of figures was not couched in these terms. Frank Press, one of the leading seismologists in the country, who is a member of your Scientific Advisory Committee, does not accept the explanation that the identity of our new result and the Soviet figures in 1958 is just a coincidence. However, he is still unsure of just what the relation between the two sets of figures is. The important practical fact, of course, is that we have changed our figures to look more like the Soviet figures, whatever the reason.

5. It is still too soon to say precisely what the significance of the new figures is for our position on a test ban treaty. The interested people will be discussing this in the coming week, and you will probably get an evaluation by the end of the week or early the following week.

6. I think there is one significant moral to be drawn from the story. It is the fact that the accurate data which AFTAC collected on its new system was highly classified and its existence in detail was unknown to all except a few scientists inside the Government. In these circumstances, it is not surprising that the technical people who had produced the original set of figures, did not quickly leap to the thought that the discrepancies which they had observed, and the questions which others had raised, based on the small amount of public data and the Russian figures, might be explicable in some other way than as mistakes in public or Soviet data. If the AFTAC figures had been made more widely available, or if AFTAC’s activities had been reviewed in detail more frequently by outside scientists, the change in understanding and evaluation might have come a good deal more quickly. It is fair to say, of course, that this is a speculation. How hard and how long a man has to look at new figures before he sees that they are inconsistent with an old theory remains one of the inscrutable mysteries of science.

CK⁷

0. Source: Kennedy Library, National Security Files, Subjects Series, Nuclear Weapons, Testing, 4/5-7/30/62. Secret. Copies were sent to Bundy and Wiesner.
1. See footnote 1, Document 204.
2. This ACDA request has not been further identified.
3. Reference presumably is to an interagency meeting on July 3 at which the Vela findings were discussed in some detail. John McNaughton summarized the consensus at the meeting in a July 3 memorandum to Nitze. (Washington National Records Center, RG 330, OSD/AE Files: FRC 69 A 2243, 97 USP, Nuclear Test Suspension/Geneva (1960-1962)) See the Supplement. See also Document 204.
4. Aardvark was an underground shot with a yield of 40 kilotons in the Nevada test series. (Department of Energy, Announced United States Nuclear Tests, p. 19)
5. The word “shots” has been crossed out and the word “shocks” inserted by hand.
6. Regarding the conclusions of this conference, which met in Geneva July 1-August 21, 1958, see Report of the Conference of Experts To Study the Possibility of Detecting Violations of a Possible Agreement on the Suspension of Nuclear Tests (London: Her Majesty’s Stationery Office), Miscellaneous No. 11 (1958). Additional documentation on the conference is in Foreign Relations, 1958–1960, volume III. 456
7. Printed from a copy that bears these typed initials.