1. This memorandum responds to a request by the Assistant Secretary of
Defense (ISA), I–21111/63, dated 14
February 1963, to provide comments on the Arms Control and Disarmament
Agency (ACDA) position paper, dated 14
February 1963, subject as above. It is understood that the Committee of
Principals will meet with the President on 16 February 1963 to discuss
this matter.
2. In the opinion of the Joint Chiefs of Staff, there are certain
cardinal principles which must govern any nuclear test ban treaty. These
are:
3. The ACDA position paper recommends
that the United States be prepared to negotiate a precise (though
unspecified) quota of on-site inspections which would constitute a
“reasonable deterrent.” As ACDA has
correctly noted, the effectiveness of the deterrent, and hence the
degree of assurance of Soviet compliance, is a function of more than the
number of on-site inspections. However, under the ACDA proposal the US detection and
verification capabilities would be further circumscribed from previously
held US positions by the following:
5. ACDA contemplates a seismic system
having a threshold of about magnitude 4.0. Below that figure there would
be what amounts to an unpoliced moratorium. Much significant technical
and military progress can be made by low-yield testing below the
threshold of this system (paragraph 1 to the Appendix hereto). The Joint
Chiefs of Staff reaffirm their view that any test ban agreement should
authorize testing below the detection threshold.
6. In conclusion, from a national security viewpoint, the Joint Chiefs of
Staff are gravely concerned by proposed departures from principles which
should govern any nuclear test ban treaty (paragraph 2, above). First,
it has not been demonstrated that the system provides a reasonable
chance for detecting evasion.2 Second, in effect an
unpoliced moratorium would exist below the threshold of detection. And
third, prohibition from withdrawal from the treaty for three years would
be counter to our interest if we found the implementation of the treaty
unsatisfactory.
7. The Joint Chiefs of Staff believe that the current ACDA proposal should be modified to align
it with the principles considered essential for an adequate nuclear test
ban treaty.
Appendix
JCS
COMMENTS ON ACDA
RECOMMENDATIONS FOR US POSITION REGARDING A NUCLEAR TEST BAN
TREATY
Detection Systems
1. ACDA
recommends basic reliance for detection and, where
possible, identification to be placed on a national detection system
(US & UK) which could begin
operation immediately after any treaty was signed. It is highly
desirable that this be unclassified.
[Typeset Page 448]
Comment
a. Detection System. The US Air Force
Technical Applications Center estimated in its study dated 11
February 1963 that it would be possible to establish an unclassified
system with seismic detection capability comparable to the AEDS (about 4.0 seismic magnitude) in
30 months. In six months an interim system having a detection
capability of about 4.2 seismic magnitude could become operable. The
number 4.0 on the seismic scale represents a detection threshold of
about 11–22 KT in alluvium, 2 KT in tuff, 1 KT in granite, and perhaps 50–100 KT decoupled. A 4.2 detection threshold
would permit undetected tests of several times those magnitudes.
For the final unclassified system comparable to the AEDS, AFTAC estimates an average of 170 shallow events can be
anticipated in the USSR yearly, of
which 20 can be identified as earthquakes by first motion. There
will be about 150 events AEDS
cannot reliably identify as earthquakes, of which 75 will be found
to have depth indications of varying reliability. There will be no
indicators of natural origin of the event for the remaining 75. All
150 events will be suspicious. In determining which of these events
should be investigated by means of on-site inspection, criteria such
as geography, nearness to population [illegible in the original]
remoteness, [illegible in the original].
[Facsimile Page 5]
It would not, however, be justifiable to eliminate any events on the
basis of these criteria from consideration for inspection or from
consideration in determining the number of inspections required. The
reason is that these events remain inherently suspicious and that
these criteria would be of substantial value in determining the
likelihood of an event being a nuclear explosion only if the
unrealistic assumption were made that the Soviets were not
attempting to evade. If the United States, for example, were to
consider events in the Arctic above suspicion because testing there
would present unusual difficulties, this fact would constitute an
invitation to the Soviets to test in the Arctic without fear of
detection. The Joint Chiefs of Staff cannot concur in the use of
subjective criteria to reduce the number of events considered to be
suspicious.
b. Threshold. Recently little consideration
has been accorded underground testing below the capabilities of
detection systems to detect. Tests beneath the threshold can provide
significant technological advances beneficial to the nation that
tests.
The AEDS detection capability is
approximately 4.0 seismic magnitude. It is probable that, under a
test ban agreement much more stringent than ACDA envisages, tests could be
conducted in the low KT range
without fear of detection.
The importance of low yield testing has been pointed out by DOD and AEC in July, 1962, and again more recently. The
consensus was
[Typeset Page 449]
that most of the important scientific principles involved in
weapons can be studied effectively below 3 KT. These include:
(1) Development of primaries for low-weight, two-stage systems
(such as Minuteman);
(2) Development of low-yield tactical systems;
(3) Tests of integrated primary, bomb case, and sparkplugs for
two-stage devices;
(4) Development of very cheap, relatively clean weapons requiring
only [illegible in the original];
[Facsimile Page 6]
(5) Possible development of all-fusion weapons requiring no
fissionable material and having many military applications;
(6) Tests of hardened primaries as well as measurements of their
vulnerability to prompt radiations;
(7) Some important weapons effects experiments which can be
performed underground.
With a limitation of 10 KT,
development can be carried considerably further. For example, tests
under 10 KT can lead to specific
weapons such as low yield weapons for either tactical, air defense
or anti-submarine warfare use; most primaries could be completely
developed underground; mock-up larger weapons tests, utilizing these
primaries and secondary “sparkplugs,” could lead to major advance in
thermonuclear weapon technology.
c. Use of Classified Data. ACDA would require establishment of
epicenters by means of seismic data submitted to an international
commission. This would preclude use of supplementary data derived
from the AEDS unless the AEDS were declassified and would
preclude the use of unilateral intelligence information in selecting
precise areas to be inspected. In the 30 months before an
unclassified detection system could attain the seismic capability of
the AEDS, the United States would
be particularly hampered. It is considered essential that the United
States be able to demand inspection of a suspicious event without
being obliged to prove an epicentric location on the basis of
unclassified, inherently less accurate, criteria. Declassification
of AEDS remains distinctly
undesirable because it would not only create political problems and
compromise intelligence collection facilities but also permit the
Soviets to have precise knowledge of US [illegible in the original]
that [illegible in the original].
[Facsimile Page 7]
Summary. The national system described above
has never been fully coordinated with the JCS or DOD. It would
offer less comprehensive coverage than would be provided by the
“national stations internationally supervised” proposed by the
United States in August 1962 and less than the internationally
manned stations on Soviet soil. The Joint Chiefs of Staff reaffirm
their view that any test ban treaty proscribe only those tests which
the detection system is capable of detecting.
[Typeset Page 450]
On-site Inspection Quota
2. On-site Inspection
Quota
a. ACDA recommends: The United
States should be prepared to negotiate a precise (though
unspecified) quota of on-site inspections. ACDA asserts that national detection systems
supplemented by unmanned seismic stations, considerations of
geographical factors, and unilateral intelligence would reduce the
number of suspicious seismic events to the point where “a quota of
on-site inspections within the number range now being considered by
the US will constitute a substantial deterrence to clandestine
underground testing.” In support of this view, ACDA claims there would be “a good
probability of detecting at least one of a series of tests
interspersed among a group of natural events.”
b. Comment: The JCS have never
approved a reduction below 12–20 inspections. This number was based
on a ratio of 20% of suspicious events. The United States has
already committed itself publicly to acceptance of 8–10 on-site
inspections, and it is understood ACDA is prepared to fall back to 5 or 6 inspections.
These figures are inconsistent with the ACDA concept of a “reasonable deterrent” which requires
that there be a fair expectation of detecting violations. AFTAC probability calculations
indicate there would be small likelihood of selecting for inspection
even one of a series of detectable clandestine tests. Any number of
inspections [illegible in the original] which the [illegible in the
original].
[Facsimile Page 8]
Moreover, for on-site inspections to be of any use it is necessary to
have workable terms and conditions for their conduct. Any control
commission must not have authority to limit or prevent inspection. A
party must be able to request inspection without seismic evidence of
a specific epicenter in order that national intelligence may be used
as a basis of decision to request inspection. An inspection team
must be logistically self-sufficient, permitted to persist in
inspection for a period possibly as long as six months, and so
manned and equipped that it could have at least a fair probability
of finding evidence of a well concealed clandestine test. The ACDA position shows little
appreciation of the physical and technical difficulties involved in
making on-site inspection an effective procedure.
Exclusion of Sensitive Defense
Installations
3. Exclusion of
Sensitive Defense Installations
a. ACDA recommends: A party could
exclude from inspection especially sensitive defense installations
located within the area to be inspected.
b. Comment: A denial of inspection on security grounds is tantamount
to a veto and would vitiate any inspection provisions if the
[Typeset Page 451]
desired
area is of appreciable size. When coupled with other limitations on
our detection and verification capabilities, e.g., limited
capability to distinguish between earthquakes and nuclear
explosions, inability to detect below seismic threshold and limited
inspection quota, this provision would so degrade the deterrent
value of the treaty as to create an unacceptable risk of Soviet
evasion. However, if exclusion from access is limited to very small
areas, the restriction would not affect inspection. For instance,
access to an installation could be denied if [illegible in the
original] permitted.
[Facsimile Page 9]
[illegible in the original] sufficiently
close to ascertain whether or not a nuclear test has been conducted.
Even if complete inspection were permitted, one side might not take
undue advantage of the right because of fear of retaliation in
kind.
Size of Inspection Area
4. Size of Inspection Area
a. ACDA recommends the outward
limits of the size of the area to be inspected would be 500 square
kilometers where the number of reliable reporting stations do not
make good travel time calculations available, and where available
the area would be 300 square kilometers.
b. Comment: AFTAC estimates a
two-thirds probability of an epicenter falling within a 700 square
kilometer area if the epicenter is located by well distributed
stations. Consequently, the area proposed by ACDA would by itself commit more than
half of any inspections undertaken of the sites of actual nuclear
explosions to failure.
Countries Not Party To
Treaty
5. a. ACDA Recommendation: The US
should not insist France and China be initial parties to the treaty
but withdrawal procedures should be included.
b. Comment: It is probable that certain states not now negotiating in
Geneva nor intending to negotiate must eventually be parties to the
treaty if there is to be an effective test world-wide test ban. This
includes both France and Communist China. Any treaty must
incorporate simple, uncomplicated withdrawal procedures for
termination in the event a state, not a party to the treaty, is
found to have conducted a nuclear test. The treaty should also
contain a provision which recognizes that timely inclusion of
certain states is an essential condition of a test ban treaty. It is
noted that ACDA suggests we
explore with the Soviets a nonwithdrawal clause effective for three
years from the [illegible in the original].
