Reflections on the GAO Triad Report on the Nuclear Triad
Science and Global Security 6, 383-393 (1997)
(this is the sumbitted version, essentially the same as published)
(GAO Tabular conlclusions are in the next file)
 
David Hafemeister
Physics Department
California Polytechnic State University
San Luis Obispo, CA  93407
 
Abstract

 
Senate Hearings on the GAO report on the nuclear triad and on the START 
treaty showed that vulnerabilities of the U.S. triad were vastly 
over-stated, that the performance of new projected strategic systems were 
over-estimated, and that the performance of existing U.S. strategic 
systems was under-estimated.  These exaggerations enhanced the 
psychological (Freud) aspects of the Cold War and compromised logic (Newton).
 
With the end of the Cold War it is imperative that we re-examine the 
basic premises that guided the choices of the strategic nuclear systems.  
The initial bottom line is that these systems were successful in that 
they did deter nuclear war without destroying either or both 
superpowers.  However, now that the emotion of the conflict has passed, 
the effectiveness of the nuclear triad should be examined to determine 
how much was enough and which technical conventional wisdoms were incorrect.
 
As this paper documents, incorrect technical estimates took place.  These 
errors (and/or exaggerations) caused the U.S. to greatly increase the 
capabilities of its nuclear triad.  At a minimum, these errors were 
fiscally wasteful, and at a maximum they could have endangered the 
stability between the countries.  This paper examines technical aspects 
of the robustness of the triad, rather than the psychological causes and 
consequences of worst-case analysis.  When more historical data becomes 
available, other authors should examine the effects of the U.S. nuclear 
build up on Soviet behavior (e.g., why Gorbachev was willing to let the 
Berlin wall fall and reduce the Warsaw Pact forces by 60% without 
requiring NATO reductions).  The technical conclusions of this paper are 
based on the author's staffing 20 hearings [1] on the START treaty before 
the Senate Foreign Relations Committee, and the Senate Governmental 
Affairs Committee hearing [2] on the General Accounting Office report, 
The U.S. Nuclear Triad:  GAO's Evaluation of the Strategic Modernization 
Program.  The GAO study produced a massive eight volume classified report 
which, according to the GAO, was the most complete examination of 
strategic nuclear forces in the past three decades.
 
The Triad as a Whole
 
The strategic triad, with its three legs of land-, sea- and air-based 
nuclear weapons, was both the capstone of the possible conflict and the 
deterrent that separated the two dance-partners.  Were the conventional 
wisdoms correct?  Was each leg of the triad so vulnerable that the other 
two legs were necessary to truly deter?  Was the Cold War driven by 
"worst-case" analysis?  The GAO concluded that, indeed, the U.S. had used 
worst-case analysis by:
 
	(1) overstating the Soviet threat to the U.S. triad,
	(2) underestimating the performance of existing U.S. systems, and
	(3) overestimating the performance of new U.S. systems.
 
In response to the hearing on the GAO report, former Secretary of defense 
Casper Weinberger agreed [3] that worst-case analysis had been used:
 
"Yes, we used a worst-case analysis.  You should always use a worst-case 
analysis in this business.  You can't afford to be wrong. In the end, we 
won the Cold War, and if we won by too much, if it was overkill, so be it."
 
Thus, there is no disagreement between GAO and Weinberger that worst-case 
analysis was used in winning the Cold War.  Because of uncertainties in 
intelligence information and because leaders are motivated to justify and 
rationalize their military investments, it is expected that worst-case 
analysis will be used.  If this is true, how will the Congress know what 
to believe when appropriating billions and trillions of dollars?  Given 
that governments can not fiscally survive if their trusted advisors 
always used worst-case analysis, it is necessary to restrain the 
psychological and economic motives for exaggeration.
 
Silo Vulnerability
 
There is no disagreement that 33,000 U.S. warheads and 45,000 Soviet 
warheads was colossal over-kill when one considers the gross 
vulnerability of cities.  There has always been disagreement on the need 
for counterforce ability to attack the other side's silo-based missiles.  
In 1981, Richard Perle, then Assistant Secretary of defense, gave his 
opinion [4] on the vulnerability of U.S. missile silos and the inability 
of the U.S. to destroy Soviet silos:
 
"Yet as we examine the SALT II treaty produced during the 1970s, the 
constraints it entailed would have permitted the Soviet Union to continue 
expanding and refining its offensive forces so that sometime during the 
life of the treaty it would have had the capacity to strike a knock-out 
blow against the land-based missile leg of the U.S. strategic triad.  And 
it seem equally certain that the U.S. missile force would have been 
unable to threaten the Soviet ICBM force."
 
The GAO's findings in Table 1 refute Perle's charge that the Soviets 
"would have had the capacity to strike a knock-out blow against the 
land-based missile leg" and that the D-5/MX "U.S. missile force would 
have been unable to threaten the Soviet ICBM force."  In her testimony 
before the Governmental Affairs Committee, Assistant Comptroller General 
Eleanor Chelimsky stated [2] that the silos were not nearly as vulnerable 
as the government had claimed:
 
"In the case of the land leg, we found that the claimed 'window of 
vulnerability' caused by improved Soviet missile capability against our 
silo-based ICBMs was overstated on three counts.  First, it did not 
recognize the existence of sea and air leg deterrence -- that is, the 
likelihood that the Soviets would hesitate to launch an all-out attack on 
the ICBM silos, given their inability to target submerged U.S. SSBNs or 
on-alert bombers and their thousands of warheads that could be expected 
to retaliate.  Second, the logic behind the claim assumed only the 
highest estimates for such key Soviet missile performance dimensions as 
accuracy, yield and reliability, while at the same time discounting very 
substantial uncertainties about performance that could not have been 
resolved short of nuclear conflict.  Third, it ignored the ability of 
U.S. early warning systems to detect a Soviet ICBM attack, and thereby, 
allow a reasonably rapid response."
 
During the hearings on the START treaty, the highest U.S. military 
figure, General Colin Powell, Chairman of the Joint Chiefs of Staff, 
seemed to concur with this conclusion [1] when he responded to a 
hypothesized question on possible massive Soviet cheating:  "....  but 
even if they had 20,000 SS-25s, I am not sure what that truly does for 
them.  As long as I have survivable systems at sea, for example, what 
would they do with these?  What incentive is there for them to move in 
that direction?"
 
Powell's conclusion was consistent with the National Academy of Science's 
conclusion [5]:  "In the agreements that follow the START treaty, the 
United States and the Soviet Union should reduce the number of nuclear 
warheads in their strategic forces to 3,000 to 4,000 actual warheads, a 
reduction of as much as a factor of three below anticipated START 
levels."  During the course of the arms race a number of authors have 
estimated the robustness of the surviving nuclear force by using 
sensitivity analysis [6] with nuclear exchange models ("red attacks 
blue").  These calculations have shown that break-outs from arms control 
treaties by the other Party did not greatly affect the result.  An attack 
can be enhanced by increasing the numbers and yields of warheads, by 
making the warheads more accurate and reliable, and by discovering that 
one's silos were not as hard as estimated.  By varying the various 
parameters, one can show that the U.S. has always had a robust triad and 
the marginal utility of additional warheads beyond START levels is very 
small.
 
Submarine Vulnerability
 
 In a similar fashion, GAO concluded that the threats to the submarines 
had also been exaggerated.  GAO concluded that the threat had been 
overstated in "unsubstantiated allegations about likely future 
breakthroughs in Soviet submarine detection technologies, along with the 
underestimation of the performance and capabilities of our own nuclear 
powered ballistic missile submarines."  These exaggerated threats to the 
SSBNs were then used as a justification for costly modernization in the 
other legs of the triad to cover the possible vulnerabilities to the 
SSBNs. The threats to the SSBNs have been categorized [7] as 
"non-acoustic anti-submarine warfare," which uses radar, laser, or 
infrared detectors on satellites to search out the signatures of the 
SSBNs.  Two submarine signatures that have been discussed are (1) the 
slightly raised ocean surface above a moving submarine (the Bernoulli 
hump) and (2) the V-shaped wave above the moving submarine (the Kelvin 
wave).  In principle, these signatures might be observed from submarines 
near the surface if one knows where to look with synthetic-aperture radar 
accompanied with significant computer capabilities.  The U.S. and Russia 
have carried out joint experiments on these phenomena, but the GAO 
concluded that these experiments do not give evidence for concern for 
survivability of the SSBNs when they are at sea.  In fact, it is very 
difficult to observe the very small oceanographic signals from submerged 
submarines.  When SSBNs observe radar from satellites, they can easily 
diminish their reflective signuature by cruising just a little deeper.  
Even if the submarines were silly enough to cruise too close to the 
surface, the job of coordinating a very large number of satellites to 
observe some 14-18 SSBN submarines, as well as doing the on-board 
computer analysis to obtain near-real time data for targeting would be 
too large and too expensive.  General Powell concurred by stating:  
"based on our first examination of the claims [of a successful 
non-acoustic anti-submarine warfare technology], we do not believe that 
they are accurate .... [but] we will explore this to make sure that we 
are correct, that it is not feasible."  Nonetheless, in their report on 
the START treaty, the Senate Armed Service Committee recommended [1] that 
a condition be added to the START II Treaty that would give "strong 
support for the joint U.S./Russian submarine detection program."
 
GAO had access to the classified data on submarine delectability and 
discussed these issues with the intelligence and military communities.  
GAO concluded [2] that "Our specific finding, based on operational test 
results, was that submerged SSBNs are even less detectable than is 
generally understood, and that there appear to be no current or long-term 
technologies that would change this. Moreover, even if such technologies 
did exist, test and operational data show that the survivability of the 
SSBN fleet would not be in question."  (emphasis by GAO)
 
Vulnerability of Penetrating Bombers
 
In a similar vein, the threat to heavy bombers was also exaggerated.  In 
her analysis [8] of the CIA Team B report of 1976, Anne Cahn pointed out 
that the extreme worst-case analysis by Richard Pipe's Team B was a 
leading factor in the political pressure for the U.S. build-up under 
President Carter (MX and B-2) and President Reagan. The Team B report [9] 
on Soviet Low Altitude Air Defense:  An Alternative View concludes that 
"it is not inconsistent with current evidence that the Soviets believe 
they have and may already posses the inherent ability to prevent most, if 
not all, penetrating bombers (of the kind presently in the force, in raid 
sizes of a few hundred) from reaching targets the Soviets value."  This 
conclusion is obviously wrong because it states "most, if not all" 
implies a kill probability of better than 99% which is far beyond 
expectation.  In addition, cruise missiles have been added to the B-52s 
which allow them to attack the Soviet Union while over the ocean.  This 
overly strong assessment on Soviet air defense was based on projections 
of significant improvements in the kill probabilities and reliabilities 
of improved Soviet SA-2s and SA-3s, as well as their very extensive 
deployments.  Although these projected enhancements never came close to 
being realized, they propelled the U.S. into the B-1 and B-2 bomber 
programs, both of which never materialized as projected and were budget 
breakers to boot.  GAO concluded [2] that "the Soviet air defense threat 
that the B-2 had been created to address was never in fact deployed."
 
ICBMs vs. SLBMs
 
U.S. fears of "Minuteman and Peacekeeper (MX) vulnerability" were always 
driven by concerns that the U.S. would not have a prompt, reliable, 
well-coordinated response to aim at Soviet hard targets.  The GAO [2] 
disagreed with this concern by stating that silo vulnerability was 
exaggerated (see "silo vulnerability" above) and by concluding that the 
offensive power of the sea leg was essentially equivalent to that of the 
land leg:  "the sea leg's performance has been understated (or poorly 
understood) on a number of critical dimensions.  Test and operational 
patrol data show that the speed and reliability of day-to-day 
communications to submerged, deployed SSBNs are far better than widely 
believed, and about the equal in speed and reliability of communications 
to ICBM silos.  Yet conventional wisdom gives much higher marks to ICBM 
command and control responsiveness than to that of submarines.  In point 
of fact, SSBNs are in essentially constant communication with national 
command authorities and, depending on the scenario, SLBMs from submarine 
platforms would be almost as prompt as ICBMs in hitting enemy targets.  
Other test data show that the accuracy and reliability of the Navy's D-5 
SLBM are about equal to DOD's best estimates fro the Peacekeeper.  
Further, its warhead has a higher yield than the Peacekeeper's.  In short 
we estimate that the D-5 has a hard target kill capability about equal to 
the Peacekeeper's, while its platforms remain virtually undetectable, 
unlike easily located silos."
 
Conclusion
 
We have summarized some of the technical findings on the U.S. strategic 
triad and in examining these technical facts it is clear that the U.S. 
exaggerated the threats to its own triad.  These exaggerated facts and 
projections were the main driving force for the large scale modernization 
of the U.S. triad.  In hindsight we see that the projected threat to U.S. 
strategic submarines was not credible, the vulnerability of U.S. 
penetrating bombers was overstated, and the vulnerability of our silos 
was exaggerated by using worst-case parameters.  In the future if the 
U.S. ever faces new strategic build-ups by evil empires, the technical 
history of the strategic triad should be re-examined to avoid these 
errors.  In answer to the lead question of this paper, exaggerations of 
U.S. vulnerability to Soviet threats probably resulted more from the need 
to rationalize arms race policies than from rational analysis of military 
facts.  The ultimate total cost of U.S. nuclear weapons systems was about 
$4 trillion [11].  The size and scope of the U.S. strategic nuclear triad 
was largely determined by these worst-case exaggerations and by 
psychological theories [10] about effects of arms race policies on U.S. 
morale and Soviet perceptions. It is my strong conlcusion that Freud 
defeated Newton.
_________________________________________________________
David Hafemeister is Professor of Physics at the California Polytechnic 
State University.  He was formerly Professional Staff Member at the 
Senate Foreign Relations Committee (1990-92) and the Senate Governmental 
Affairs Committee (1992-93), the Office of Senator John Glenn (1975-77) 
and the State Department (1977-79, 1987), working on arms control matters.
_________________________________________________________
References
 
 
1.  U.S. Senate Foreign Relations Committee, The START Treaty, S.Hrg. 
102-607 (Parts 1-2), S.Exec. Report 102-53, 1992.
 
2.  U.S. Governmental Affairs Committee, Evaluation of the U.S. Strategic 
Triad, S.Hrg 103-457, 1994.  U.S. General Accounting Office, The U.S. 
Nuclear Triad:  GAO's Evaluation of the Strategic Modernization Program 
(plus 8 classified volumes), GAO/T-PEMD-93-5, 1993.
 
3.  C. Weinberger, New York Times, June 28, 1993.
 
4.  R. Perle, introduction to Beyond the SALT II Failure, J. Lehman and 
S. Weiss, Praeger, New York, 1981.
 
5.  National Academy of Sciences, The Future of the U.S.-Soviet Nuclear 
Relationship, National Academy Press, 1991.
 
6.  D. Hafemeister, "Breakout from Arms Control Treaties:  A Sensitively 
Analysis of the Threat to National Security," in Arms Control 
Verification, ed. by K. Tsipis, D. Hafemeister and P. Janeway, 
Pergamon-Brassey's, Washington, DC, 1985.  M. May, G. Bing and J. 
Steinbrunner, "Arsenals after START:  The Implications of Deep Cuts," 
International Security 13, no. 1, 90-133 (1988).
 
7.  T. Stefanick, "The Nonacoustic Detection of Submarines," Scientific 
American 258, no. 3, 41-47 (1988) and 268, no. 2, 32 (1993).
 
8.  A Cahn, "Team B:  The Trillion Dollar Experiment," Bull. At. 
Scientists 49, no. 3, 22-27 (1993).
 
9.  C. Lerch and CIA Team B, "Soviet Low Altitude Air Defense:  An 
Alternative View," In Intelligence Community Experiment in Competitive 
Analysis, December 1976 (partially declassified Sept. 16, 1992).
 
10.  S. Kull, Minds at War:  Nuclear Reality and the Inner Conflicts of 
Defense Policy Makers, Basic Books, NY, 1988.  L. Nelson and G. 
Beardsley, "Towards an Interdisciplinary Model of Barriers to Nuclear 
Arms Control," Social Science Journal 24, 375-388 (1987).
 
11.  S. Schwartz, et al, "Four Trillion Dollars and Counting," Bull. At. 
Sci. 51, 32-52 (Nov. 1995).  The U.S. Nuclear Weapons Cost Study Project, 
Atomic Audit: The Costs and Consequences of U.S. Nuclear Weapons, 
1940-1995, Brookings, Washington, D.C., to be published, 1997.