368. Memorandum From the President’s Assistant for National Security
Affairs (Brzezinski) to
President Carter1
Washington, January 9, 1980
SUBJECT
- Possible Nuclear Explosion in the South Atlantic (U)
A mini-SCC meeting chaired by Henry Owen today reviewed the
conclusions of a panel of outside scientific experts concerning [less than 1 line not declassified] the South
Atlantic on September 22.2 The panel concluded
that:
—There is no evidence to date which would clearly corroborate a nuclear
explosion;
—The September 22 signal closely resembles those obtained from known
nuclear explosions, but reveals a discrepancy “sufficient to raise some
doubt;”
—All other possible causes were ruled out except one: the possibility
that sunlight reflected from a small meteoroid or space debris [less than 1 line not declassified] could have
closely duplicated a nuclear signal;
—The probability of a meteoroid with just the right properties to produce
such reflections is of course low; but so is the probability that a
nuclear explosion would fail to produce any corroborative data. Yet one
of these improbable events appears to have occurred;
—Thus, the panel could not determine whether the signal was generated by
a nuclear explosion. (S)
In the light of these findings, the mini-SCC reached the following conclusions: (U)
Following notification to key Members of Congress and our Allies
tomorrow, we should inform the South Africans of the outcome of our
analysis, and urge them to resume promptly negotiation of the proposals
we made earlier (resumption of US fuel supply in return for South
African adherence to the NPT and
acceptance of interim safeguards on their enrichment plant). A public
statement summarizing the panel’s conclusions would then be issued; a
contingency statement is being prepared in the event of an early leak.
(S)
The panel’s conclusions may have important implications concerning the
adequacy of our detection systems to adequately monitor the
[Page 1101]
Limited Test Ban Treaty in
areas remote from the Soviet Union, and to support our non-proliferation
policy. Recommendations on this aspect of the matter will be forwarded
to you separately.3 (S)
The panel’s report is at Tab A. (U)
Tab A
Paper Prepared in the Office of Science and
Technology Policy4
Washington, January 7, 1980
POSSIBLE NUCLEAR EXPLOSION PANEL
FINDINGS AND CONCLUSIONS
1. Despite extensive analysis of all data available, the Panel cannot
determine whether or not the light signal that was recorded by one
VELA satellite (No. 6911) over the South Atlantic on September 22,
1979 was generated by a nuclear explosion. (S)
2. The light signal does have the general character of signals
generated by nuclear explosions. It resembles a nuclear explosion
signal more than any one previously recorded except, of course, for
those known to have been from nuclear explosions. However, a
detailed comparison of the September 22 signal with those observed
from other nuclear explosions reveals a discrepancy, sufficient to
raise some doubt about whether the signal was of nuclear origin.
(S)
3. Also, as of this date, there is no evidence which would clearly
corroborate the occurrence of a nuclear explosion on September 22.
All previously recorded signals which had the character of those
from nuclear explosions were confirmed by other evidence to have
been from nuclear explosions. Ionospheric observations made by the
Arecibo radar in Puerto Rico, which is currently unique in its
sensitivity, provided the only suggestive additional evidence for a
nuclear event. But ionospheric signals are not generally well
understood. Also, there were ambiguities in interpretation of the
signal. As a result the Arecibo data were not persuasive as
corroborative evidence for a nuclear event. (S)
4. In over ten years of operation, the recording devices of different
bhangmeters on VELA satellites have been “triggered” about 400,000
times—mostly by signals attributed to lightning (which triggers both
bhangmeters on a satellite) and to nuclear particles (which trigger
only
[Page 1102]
one of the
bhangmeters onboard a satellite). In addition, every known
atmospheric nuclear explosion occurring in the time the VELA
satellite system has been in operation has been recorded by at least
one bhangmeter in one of the VELA satellites. Also included in the
set of bhangmeter recordings are a group of signals which were
tentatively attributed to sunlight reflection from a meteoroid
passing in the field of view of one of the VELA satellites. (S)
5. The Panel has reviewed a number of known alternative natural
sources which might have caused the September 22 signal and has
ruled them out with one exception, and that is reflection of
sunlight from a small meteoroid (or a small piece of satellite
debris) passing near the satellite sensors that detected the event.
The Panel also ruled out the possibility of a technical malfunction
of the satellite being the source of the signal. (S)
6. Computer models have confirmed that the September 22 signal could
have been generated by a small meteoroid with specified properties
passing near the satellite sensors. The necessary shape, size, and
velocity of the meteoroid—characterized as a “fractured
ball-bearing”—that could cause this signal are reasonable. Small
meteoroids of this size are numerous as evidenced by the findings of
the Pioneer 10 probe. (S)
7. It should be recognized that the a priori
probability of the occurrence of a nuclear explosion with no
corroborative data such as nuclear debris is small, as is the
probability of the first signal from a meteoroid or other physical
phenomenon with just the right properties after ten years of
observation. Therefore, we are unable to assign relative
probabilities to whether the signal was generated by a nuclear
explosion or natural phenomenon. (S)
8. Further analysis will be useful in improving our understanding of
the September 22 signal. To this end, we recommend (1) continuing
the search for corroborative data, (2) a comprehensive statistical
analysis by an outside group of the signal characteristics from
nuclear and non-nuclear signals discussed above to improve our
understanding of their physical origin and enable us to evaluate the
likelihood that a member of this class could cause a “false alarm,”
and (3) a thorough review of meteoroid data to examine the
possibility that a meteoroid or piece of space debris could cause
such a signal, including the permissible ranges of values of its
physical parameters and the probability of appearance of such an
object in the vicinity of the VELA satellite. (S)
9. From our study of the September 22 signal, we conclude that in the
absence of corroborative data, signals from current bhangmeters on a
single satellite now in orbit cannot for a single event provide
definitive evidence of the occurrence of a nuclear explosion.
(S)
