THE LIFE AND DEATH OF COSMOS 954

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP85B01152R000200260006-4
• Release Decision: RIPPUB
• Original Classification: S
• Document Page Count: 8
• Document Creation Date: December 21, 2016
• Document Release Date: August 26, 2008
• Sequence Number: 6
• Publication Date: January 12, 1983
• Content Type: REPORT

Approved For Release 2008/08/26: CIA-RDP85B01152R000200260006-4
MEN RANbt1M FOR: Deputy Direc for'Administration
flWe understand that some
dies of it are circulating through the
White House and Executive Office Building.
by who
relates a-n ian concern over getting the
Soviets to pay for cleanup costs virtually
excluded intelligence officers from partic'pating
in the collection of debris from COSMOS 95
12 January 1983
fbA ko--
5-75 T U 1 ED TIONS loos
Date D/A Registry
"" 1 1 "v 1- - ''~' Approved For Release 2008/08/26: CIA-RDP85B01152R000200260006-4
 Approved For Release 2008/08/26: CIA-RDP85BOl 152R000200260006-4
SECRET
NOT RELEASABLE TO
F-0-REIGN NATIONALS
The satellite that came into the cold.
J0
THE LIFE AND DEATH OF COSMOS 954
DIA review
completed.
Cosmos 954 was launched in the Soviet Union on 18 September 1977. By
November, U.S. tracking radars had observed an unusual decay in its orbit. On
6 January 1978 the satellite lost its attitude stabilization system, a terminal ailment. On
24 January Cosmos 954 crashed in the ice and snow near Great Slave Lake, Northwest
Territories, Canada.
This satellite was one of a series scanning the oceans by radar, seeking out large
surface ships. Such satellites use a small nuclear reactor to power a radar and the
equipment needed to report to a ground station. the reactor to be
of the so-called Romashka variety, but no one could say for certain that it was this
type; this was surmise, nicely done, but still circumstantial. Fuel for Romashka is 90
percent enriched Uranium-235, embedded in carbide and surrounded by graphite
moderator, yielding a compound considered distinctly unhealthy to fondle.
Throughout 954's decay and reentry, its reactor, (of whatever type) was alive and hot.
One virtue of 954s life and death is the simplicity of describing the problem it
posed, namely: what does one do about a live nuclear reactor reentering the earth's
atmosphere aboard a Soviet surveillance satellite? A quick scan of literature showed no
textbook answer, nor even a textbook question. It remained for the National Security
Council Staff to put together a group to cope with the problem, and this article is some
of that group's story. It contains elements of tension, humor, self-satisfaction, and some
demonstration of the timely use of intelligence.
For those enamored of methodology, permit me to suggest two problem-solving
approaches:
Type One: Crisis Management. This scheme suggests that untoward circum-
stances can be contained, that reasoned information can be made available and used,
and that calculated risks are there to be taken. Objectives can be set and means for
those objectives spelled out, and out of that array a decision can be reached.
Type Two: Muddling. This implies making up responses as a problem progresses,
and coping as events and information unfold over time. Muddling is very much the
label for a process, and is to be distinguised from "muddling through," which is a
conclusion. Muddling occurs when the decision maker is not sure where he is headed,
but has a good idea where he would not like to end up. A fancy word is "heuristic,"
meaning figuring out how to figure it out as time yields its clues.
Those bloodied by the real world recognize that both types apply-the issue of
which scheme dominates a problem is determined by the properties of the problem
itself. In theory, it would seem necessary to divine the characteristics of a situation
before launching into its resolution (nobody in modern history has ever done this,
given the resolute dominance of the demand for answers over processes needed to
obtain them). In the instance of Cosmos 954, two properties dominated: the NSC
group had some time to think the problem through, and second, the "opposition" was
Newton's Law of Gravitation, later compounded by Bernoulli and the physical effects
of aerodynamic drag on the satellite. The game was against nature rather than against
conscious intelligence. This in turn became reckoning the time and place of 954's
NOT REELEASABLE TO
FOREIGN NATIONALS
S!- n r-
ITOXI
Approved For Release 2008/08/26: CIA-RDP85BOl 152R000200260006-4
 Approved For Release 2008/08/26: CIA-RDP85BO1152R000200260006-4
NOT RELEASABLE TCJ i
FOREIGN. NATIONALS
Cosmos 954 Decays
reentry and the likely consequences of that reentry. To the group, "likely
consequences" was the probability that one human might be injured (the judgment
was one chance in 10,000). For the time and place of reentry, United States tracking
experts produced the correct date-24 January-with a spread of two days that
narrowed as the 24th approached. For the place, one might cautiously describe this
estimate as night baseball with the lights out.
One precondition to problem-solving is in fact to deduce that a problem exists
and to move the issue to a forum where it can be worked. The Air Defense Command
inferred decay in November, and by formal and random ways sufficient interest
boiled up so that by early December= had its own interagency group in session. At
that time reentry was judged to be for the late spring or summer of 1978, given that
the satellite maintained its attitude stabilization. Loss of stabilization would produce
tumbling and early reentry, a process which indeed began on 6 January. But an open
question in December was the existence of a fail-safe system on 954: if one was on
board, it would sense an abnormal condition and automatically boost the reactor
portion of the vehicle to a higher altitude parking orbit, where the reactor would cool
harmlessly. Judgments varied over this issue, and prudence dictated that the analysis
proceed without any presumption of a fail-safe system (note that the eventual crash of
954 still did not preclude the presence of such a system, as it could itself have failed).
group concluded there was a small chance of a very serious problem. It
recognized that the issues posed by 954 spread across many agencies, a circumstance
anticipated by the Department of State representative who had already drafted a
letter to NSC suggesting that the problem be taken over by the NSC Staff. It was sent.
The Department of Energy representative had nearly finished his statistical study of
the outcomes and their probabilities; this study proved to be the key analytic piece and
upon it was premised the one-in-10,000 chance of harm to a human. The tone of the
group was stoic in that we felt captured by a no-win situation (i.e., a negative sum
game to theorists). A colleague suggested the outcome of 954 would be akin to
determining the winner of a train wreck.
The NSC group formed on 19 December, fetchingly calling itself the Ad Hoc
Committee on Space Debris. It was put together by active recruiting on the part of its
chairman from the NSC Staff, rather than the usual procedure of agency nominations.
The facts were recounted, and NSC directed the preparation of, if you will, pre-
contingency plans. These included the availability of tracking resources and the
assurance that the proper agencies and representatives were receiving information
from collection resources. Operation MORNING LIGHT was born.
By 6 January the NSC representative and the respective members of MORNING
LIGHT had briefed their principals, and the prospective reentry date was judged to
be April. Then, on the 6th, the Air Defense Command reported 954 out of control.
The telephones of MORNING LIGHT members summoned them to NSC, and what
had beep pre-contingencies became an inexorable fact: the satellite was coming down
carrying a live reactor, with the best reentry guess 24 January. (This again validated
the policy maker's rule of probability, i.e., the simple fact that an event is unlikely
does not stop it from happening. For instance, Cosmos 954 landed on its 2,060th
revolution, one that had only an- 8 percent chance of any land impact.)
NSC gave responsibility to the Department of Energy for domestic contingencies,
and assigned it to State in the event of a foreign reentry. Defense was to provide
support as required, and I found myself promising airplanes, reconnaissance, and all
manner of gadgets to these agencies, should they be needed. At the time, I was not
sure I had the authority to do this, but such was the sticky wicket of MORNING
LIGHT members. NOT RELEASABLE TO
LM ON NATIONALS
SECRET
Approved For Release 2008/08/26: CIA-RDP85BO1152R000200260006-4
 Approved For Release 2008/08/26: CIA-RDP85BO1152R000200260006-4
NOT RELEASABLE TCS,
Cosmos 954 Decays LO-REIGN NATIONALS
Arrangements for contingencies are easy; handling the world after the
contingency is yet another proposition. Although the risk of harm was judged low, still,
the possible harm to populated places could be so severe that its low probability could
not allow one just to hope that the satellite could be ignored. And it was here that
MORNING LIGHT became a sobering experience for those navigating the problem.
This wrenching was not at all helped by worry that a sensationalized leak would
disturb the public in unforeseeable ways.
The MORNING LIGHT group had to cope with a set of arrangements and
decisions:
Contingency Plans for Cleanup of Radioactive Materials: This estimable task
entails finding radioactive sources, decontaminating land areas, and attending to
persons found within a harmful distance. It turns out that locating active materiel on
the ground is far from a simple mission. The hunt for radioactive pieces surviving
reentry has to begin with large search areas, prompted by reentry plots furnished from
tracking radar, backed up by any visual sightings. Sensor aircraft would patiently
refine the ground search, while high altitude U-2s would seek out.the debris cloud left
by burned and dispersed uranium. The Departments of Energy and Defense shared
this unappealing assignment, but the planning was such that the men and equipment
could be launched when the President said to do so.
Ask the soviets for information: Because MORNING LIGHT had to infer the
type of reactor on 954, we were in some measure presuming our design to the Soviet
reactor, substantially complemented by *tails about Romashka. Little
seemed to be at risk in our asking the Soviets for information, and it was the general
view that no good answer could be formulated to a postmortem inquiry asking why
the United States did not seek data from the owners of the satellite. These are the
questions posed to the Soviets, and a paraphrase of their response:
- According to information available to us, Cosmos 954 appears to be quickly
decaying into the earth's atmosphere. We estimate that it will reenter the
atmosphere any time within the next month.
- We are concerned that Cosmos 954 may be powered by a nuclear reactor and
that its reentry into the atmosphere thus may represent a potential for nuclear
contamination. If the debris falls on or near a populated area, there is the
obvious possibility of a serious hazard to the public.
- In view of these serious possibilities and in the spirit of cooperation called for
by the Outer Space Treaty, we would like to hear your view of the problem, as
well as any additional information you can give us.
- In particular, if there is a nuclear reactor on board Cosmos 954, we would like
to know whether it is designed to disintegrate during reentry or whether there
is a significant probability of impact of the nuclear fuel. We need this
information to assist in determining what steps to take in the United States to
protect our public.
- Given the urgency of this question, we request an expeditious answer.
The Soviets replied that:
The small power plant at the satellite "Cosmos 954" operates only on U-
235 fuel. As we have already said, it is explosive-proof because the
accumulation of a critical mass is ruled out. Besides that, the design of the
plant provides for its destruction and burning upon entering denser layers of
atmosphere.
NOT RELEASABLE TO
F REIGN NATIONALS
Approved For Release 2008/08/26: CIA-RDP85BO1152R000200260006-4
 - Approved For Release 2008/08/26: CIA-RDP85BO1152R000200260006-4
SECRET
NOT RELEASABLE Tta 1
REIGN NATIONALS
Cosmos 954 Decays
However, in view of the accident aboad the satellite [depressurization],
it cannot be ruled out that some destroyed parts of the plant still would reach
the surface of the earth. In that case an insignificant local contamination
may occur in the places of impact with earth which would require limited
usual measures of cleaning up.
A colleague remarked that he wasn't too sure what "usual measures of cleaning
up" a reactor crashing in from outer space might be, and there was also some
ambiguity in the meaning of "explosive-proof". ? But some considerable relief was
expressed when MORNING LIGHT was told the reactor had been designed to burn
up during reentry. (I observed that any seemingly hard information is gratefully
received during a crisis, and also noted how vulnerable one can be to that
information.)
Notification of Other Countries: The Soviets had not told anybody of 954's
impending reentry. The United States had the information, a sensitive problem was
nearing full brew, we had no idea where 954 would come down, a lurid leak seemed
inevitable, and the satellite belonged to the Soviets.
Who should be informed by the United States, if anybody? What were our
responsibilities to our allies and to the world for a problem which was not of our
making but about which we knew? Skipping pros and cons, lists of countries, and the
imponderable factor that the more nations informed, the greater the chance of a leak,
the notification problem was surely disturbing. Mortals, with notorious shortcomings,
should not have to make these judgments. But it had to be done, and we went at it on
the general approach of notifying our allies and some other countries with which we
share such special relationships as tracking facilities. From those countries notified,
State received uniform gratitude, but some governments were intensely upset over the
slightest prospect that Cosmos 954 might land on their soil.
Notification of U.S. Citizens: Another scrimmage. Congressional leadership was
briefed. In MORNING LIGHT contingency planning, the Federal Preparedness
Agency was primed to provide state and local civil defense officials with pertinent
information after a reentry in the United States. At issue again was what to say before
anything happened. Judging the imponderables, the plan was to notify local
authorities if there was something tangible to notify them about. In retrospect, the
steps taken proved right, but 16 hours before reentry, the MORNING LIGHT group
was still debating public announcement; we argued through the circumstances
pertinent at that late hour and decided to stay on our original approach. The effort
had come nearly to reentry time without a break in security, and the United States
had been encouraging other countries to contain public comment. A reversal at that
hour would have been, at best, awkward.
As the Day of Reentry neared, computers plotted the predicted ground trace.
Gazing at the traces, a colleague observed that only one pass would cut across the
Soviet Union and that a number of ascending and descending paths traversed Canada
(after it w l over, we noted the marvels of retrospective clarity). At 0500 on the
24th, Current Operations Center called to announce imminent reentry, with
a win ow equal to one complete orbit. At 0653 EST the satellite was down in Canada.
The Center has its complement of maps, red and green telephones, TV monitors,
and flashing lights: I didn't have all the right badges (two were required, and at one
point some earnest young officer asked why I didn't have all the right badges, at a
? Three senior government officials when briefed wondered if the reactor could detonate like a nuclear
weapon. The lesson applies to those who live so close to technical matters that they tend to presume
widespread knowledge of engineering or physics to officials trained or experienced in other fields.
NOT RELEASABLE TO
A FQREIGN NATIONALS
SECRET
Approved For Release 2008/08/26: CIA-RDP85BO1152R000200260006-4
 Approved For Release 2008/08/26: CIA-RDP85BO1152R000200260006-4
SECRET.
Cosmos 954 Decays NOT RELEASABLE TO
Rtox NATIONALS
moment of high crisis, his inquiry about badges proved most helpful). The CRITIC
message failed to reach the responsible State Department officer; he called to find out
what had happened. Another officer from State inquired about the satellite impact in
the Azores. The computer connecting the DIA center and NSA went down, so a
backup system had to be brought up. Later on, a commercial power reactor in
Colorado experienced a valve malfunction and some release of radioactivity; at first
the release was reported as very dangerous, so intense that it would require the
diversion of a nuclear emergency team from the Canadian mission to Colorado.
Because of a mix-up, the first report never reached me; by the time the mix-up was
repaired, it had been discovered that someone in Colorado misread the radiation
sensors and that there was never a problem in the first place. Despite these tiny upsets,
the necessary airplanes and search teams were ordered to Canada by the President
and did their job.
Later a well-known television journalist called, breathlessly announcing that his
network management in New York was "about to be tipped off" that the mysterious
East Coast sonic booms were really American ABM attempts to shoot down the errant
satellite. Could I confirm before the story went on the air? With an effort worthy of
Zeus, I resisted the temptation to respond that the booms were really the Navy's X-25
hypersonic underwater earthquake generator. (Isn't it strange that a reporter might
ask for verification of a story about which he was about to be tipped off? But then, one
can't bother too much about logical sequences.)
The Use of Intelligence in Operation MORNING LIGHT: A pet definition of
intelligence is simply information that helps people making decisions think ahead. But
"information" is a loaded word. There are facts or data, but then there is creeping
interpretation (opinions about facts), followed by hypothesis, conjecture, and theory.
Hard work must be done to keep from scrambling these terms and acting on one when
it is in the guise of another. Facts and data are observational (at least some of the
time), while information is a testing of observation and an arranging of data by some
thoughtful procedure. Information gets at the question, "what do the data mean?" and
so must lend itself to creeping interpretation. This sets up a dilemma, in that there are
too many pieces of data for any decision maker to handle, but going beyond raw data
requires the use of judgment by the person doing the organizing. The First Rule of
Intelligence is-forgive me-if you want to know something, ask, but be careful
whom you ask and how you do it. Hear and listen, see and observe, and go back and
do it again, while trying mightily to diversify sources of both data and meaning.
The ear should be tuned to the uncertainty latent in any interpretation, and
special efforts have to be made to isolate the variance of a projection, extrapolation
beyond known facts, and, perhaps most of all, the compound or aggregate
consequences of the separate unknowns bearing on a particular problem. For the re-
entry of Cosmos 954, the MORNING LIGHT group had to contend with its fair
allotment of queasy questions (some of which could not be termed intelligence issues):
- Did 954 have a fail-safe system?
- What sort of power supply?
- Health hazard?
- Landing when and where?
- Could the Soviet response be believed? Were they
scheme to throw us off?
- Consequences of a leak on public psychology?
NOT RELEASABLE TO -
L'OREIGN NATIONALS
SECEIT.
Approved For Release 2008/08/26: CIA-RDP85BO1152R000200260006-4
 Approved For Release 2008/08/26: CIA-RDP85BO1152R000200260006-4
[SECRET
NOT RELEASABLE TO
EOREI GN - NATIONALS
Cosmos 954 Decays
- Response of foreign governments to our telling them, and responses of those
governments not cut in?
For instance, the conviction about the fail-safe device weakened about in
proportion to the length of time 954 was in orbit decay. References to the system
shifted from "the system," to "if there is a system," to 50-50 odds. Cross-checking
suggested that one observation and the notion of sound engineering practice, that is,
our own American perspectives, were the premises of the first judgment. Uncertainties
became more apparent as time went on, but early checking among sources had made
it clear that this would probably happen. This experience underscores the Second
Principle of Analysis, that is, carefully observe the difference between the second
estimate and the first, for that difference gives a good clue to emerging uncertainty.
Surely there must be an epistemology of variable constants, and a price, that is, the
longer the waiting period for information, the more circumstances constrain the range
of response.
As for public reaction, one reporter said, "What I could have done with that story
if I had it a day early!" You can visualize the headline, and the MORNING LIGHT
working group took it for granted that any publicity before impact would produce
florid reporting and a tense public response. After it was over, a social psychologist at
the Center for Disaster Research observed that "people prepare" for earthquakes,
floods, fires, and hurricanes, and in general make do. One would surely like to see the
research footnotes on public reaction to the first live nuclear object (spewing deadly
nuclear radiation) tumbling in from the cold depths of outer space. Intelligence was
not asked to judge likely public response or that of governments to notification, but a
compilation of post-event reactions was put together.
In retrospect, intelligence stuck to its knitting. The MORNING LIGHT team had
a good grasp for that which was known and that which was conjecture. Creeping
interpretation never became a problem: the CIA representative gave his reports by
way of what we knew, what we did not know, followed by this or that could happen.
By 24 January the compound uncertainties had become an almost agreeable way of
life; there is some comfort in knowing what is in the realm of the factual and what
resides in the realm of pure chance. MORNING LIGHT knew what it couldn't know
and prepared accordingly.
Of the two decision approaches, Crisis Management and Muddling, it is clear
both were used, however unconsciously they may have been incorporated in the
proceedings. Policy work requires the use of any procedure that helps lead to sound
decisions, and only the unwise lock themselves into any single approach. The 954
project was blessed with perceptive leadership from the NSC (get a strong team
together, listen and fight the issues, but force decisions and get those to officials who
can put them into effect). The inexorable approach of 24 January surely provided an
incentive not to dawdle over methodological niceties, but the deadly deadline was not
by itself the reason MORNING LIGHT worked well.
Early on, the fate of Cosmos 954 was largely of statistical interest; but after
tumbling, arithmetic calculation gave way to more intense considerations. Surprises,
when they arrive, are usually unpleasant (this may be especially true in technology).
History has shown there is a vast market for oracles, but vision is not a gift nature
dispenses generously, so mortals are left to ponder what-ifs and prepare for them as
best they can. When 954 tumbled out of control, MORNING LIGHT was able to
move quickly, mainly because much of its homework had been done, this thanks in
good measure to timely and persuasive intelligence work.
NOT RELEASABLE T01
tQ6El_Oll NATIONALS
Approved For Release 2008/08/26: CIA-RDP85BO1152R000200260006-4
 Approved For Release 2008/08/26: CIA-RDP85BO1152R000200260006-4
P
~, }
EG 1,
Cosmos 954 Decays NOT RELEASABLE TO
FOREIGN NATIONALS
OPERATION MORNING LIGHT
CHRONOLOGY OF EVENTS, COSMOS 954
-September 18, 1977 Soviets launched Cosmos 954.
-November 1977 Intelligence sources determined Soviets experiencing
difficulties.
-December 1, 1977 Deputy Under Secretary of Defense for Policy (Admiral
Murphy) alerted by his staff.
-December 19, 1977 NSC interagency task force on space debris established
(State, Defense, NASA, CIA, Energy and OSTP).
-January 6, 1978 Through intelligence sources, determined Soviets had lost
control of satellite; reentry date established January 23 or -24, 1978.
-January 12, 1978 Approached Soviets on the issue.
-January 14, 1978 Soviets' response confirmed that nuclear power source was on
board and that they had lost control of satellite. Projected reentry on January 24, 1978.
-January 17, 18, 1978 Informed key Congressional leaders (Byrd, O'Neill,
Baker, Rhodes, Inouye and Boland) of problem and steps we were taking in event the
satellite landed in the United States.
-January 17, 1978 Dr. Brzezinski memorandum making Energy responsible for
clean-up and other safety considerations relating to possible reentry in the United
States, receiving support from Defense and from other agencies as appropriate. State
was made responsible for foreign requests for assistance, calling on Energy and
Defense as appropriate.
-January 17, 1978 Approached Soviets for clarification.
-January 18, 1978 Informed our allies and other countries with which we have a
special relationship, e.g., tracking facilities.
-January 19, 1978 Soviets responded, indicating the reactor would not go critical
and that it was designed to disintegrate during reentry.
-January 20, 1978 The Department Energy Task Force went into operation.
-January 22, 1978 Approached the Soviets and asked if there were any new
developments.
-January 23, 1978 Soviet response indicated that reentry was projected for
January 24, 1978.
-January 24, 1978 Satellite reentered atmosphere at 6:53 a.m. EST over Queen
Charlotte Island and impacted last at Great Slave Lake.
-January 24, 1978 Prime Minister Trudeau was immediately informed that the
satellite had landed in Canada. United States offered assistance in locating fallen
debris and in cleanup. Trudeau accepted by a return telephone call to the President.
-January 24, 1978 Dr. Brzezinski informed the Soviets that the satellite had
landed in Canada.
-January 24, 1978 aircraft and Nuclear Energy Search Teams
d
d spatcneu to Cana
a. NOT RELEASABLE TO
FOREIGN NATIONALS
sECF ET
Approved For Release 2008/08/26: CIA-RDP85BO1152R000200260006-4