EXHIBIT B-RAND CORP. STUDY OF NON MILITARY DEFENSE (The following report, submitted by the Rand Corp., is an advance copy subject to possible revision prior to final publication by the Rand Corp.) Preface. REPORT ON A STUDY OF NONMILITARY DEFENSE I. Introduction: Initial premises. II. Population shelters. A. Types of shelters. CONTENTS B. Warning and movement. C. Performance of possible shelter systems under hypothetical attacks. III. Long-term fallout. A. Medical consequences of radiation. B. Long-term fallout levels after hypothetical attacks. IV. Recuperation of the economy. A. Reorganization problems. B. Food. C. Recuperation after a 50-city attack. D. Heavier attacks and industrial shelters. V. Some possible nonmilitary defense programs. B. Two scales of shelter programs. C. Timing and phasing problems. VI. Interactions with other aspects of national defense. B. Active defense. C. Soviet nonmilitary defense. VII. Conclusion: Some policy suggestions. A. A broad research, development, and planning program. PREFACE The study of nonmilitary defense described in this report1 has been supported by the Rand Corp. as part of its program of Rand-sponsored research. In addition to its work for the United States Air Force and other Government agencies, the corporation regularly sponsors with its own funds research projects in areas of importance to national security and public welfare. Rand-sponsored research is considered to be fundamentally the responsibility of the individuals involved in the project, and the conclusions of such projects are not necessarily endorsed by the corporation. Such studies are published in the hope that they may contribute to wider understanding of important national problems. This study of nonmilitary defense was initiated, directed, and formulated in its central features by Herman Kahn. Particular parts of the study were the responsibility of the following individuals, approximately in the order the subjects are mentioned in this report: Leon Gouré, foreign policy implications; Irwin Mann, improvised fallout shelters and other inexpensive measures; Robert Panero (from the staff of Guy B. Panero Engineers), mines and deep rock shelters; John O'Sullivan, conventional shelters and costs of complete shelter sytems; Fred Iklé, strategic evacuation and social problems; Frank Ross, tactical evacuation; Leonard Berkovitz, performance of shelter systems under hypothetical attacks; Harold Mitchell, medical effects of radiation; Jerald Hill, longtime fallout problems; Joseph Carrier, food and agriculture; Paul Clark, economic recuperation after a 50-city attack; Norman Hanunian, heavier attacks and industrial shelters; George Reinhardt, "starter set" and recuperation interactions with active offense; Philip Dadant, interactions with active defense; Richard Moorsteen, Soviet nonmilitary defense capabilities. This summary report was drafted by Paul Clark. 1 This is an advance copy of the report, and further changes may be made in final editing. A number of people in Government agencies have been helpful in furnishing information and orientation. While it would be impossible to list them all, the assistance of the following should be acknowledged: Federal Civil Defense Administration: John Devaney, Gerald Gallagher, Jack Greene, Ralph Spears, Benjamin Taylor. Federal Reserve Board: Roland Robinson. Naval Radiological Defense Laboratory: Walmer Strope, Paul Tompkins. Office of Defense Mobilization: Joseph Coker, Harold Huglin, Burke Horton, Vincent Rock, Charles Sullavan. Science Advisory Committee: Spurgeon Keeney. Of course, none of the above are responsible for any portion of the study. This report is unclassified, and no part of it depends on the use of classified information. In particular, the hypothetical attacks considered in evaluating various nonmilitary defense measures should not be construed as statements of enemy offense capability or United States defense capability. They are simply hypotheses about threats which appear conceivable sometime in the future and which provide a measure of the possible role of nonmilitary defense systems. Moreover, this report has been written as a summary statement for general distribution; technical aspects of the study are not presented in full detail. REPORT ON A STUDY OF NONMILITARY DEFENSE I. INTRODUCTION: INITIAL PREMISES This study was initiated in the belief that nonmilitary defense measures, if they could be made effective in protecting the civilian population, economy and institutions of the United States, might make two significant contributions to the national defense. First, they might alleviate the catastrophe of a nuclear attack, and if military victory were attained, provide a reasonable chance that the United States as a nation could survive. Second, they might increase United States freedom of action in conducting peacetime foreign policy and implementing a broad deterrence strategy. Alleviating the consequences of a nuclear war is an important objective in its own right. Even if a plausible attack a few years from now killed as many as 90 million Americans, it would still leave 90 million alive. However terrible the prospect, it would be worth investigating whether there are measures which might raise the number of survivors from 90 to 120 or 150 million, and which might increase the likelihood that the survivors could in time restore the national economy and democratic institutions. The prospect is terrible enough to make the avoidance of general war-by deterrence or by any measures which might safely permit reduction in tension-the primary objective. However, general war may nonetheless occur and it would be irresponsible to throw up our hands about the postwar world. Moreover, in the years ahead, willingness to make foreign policy decisions carrying a risk of war may be important to meet major Soviet challenges that would threaten United States security. The more effective the defense of civilian society, the easier it would be for United States leaders to take such decisions. Deterrence of extremely provocative enemy behavior other than a direct attack on the United States might thus be maintained as a credible national policy. If nonmilitary defense measures caused Soviet leaders to believe that aggressive moves would meet firm resistance, they would be less likely to take such provocative actions. Deterrence of aggressions against countries other than the United States might also be accomplished by strengthening United States capability to meet limited aggression in a limited way, and we believe it is important to do so. However, it is possible that some aggression may be hard to deter or meet except by a credible threat of all-out United States resistance. It is true that the likelihood of direct Soviet attack on the United States would be measurably increased, particularly in the case of an implicit or explicit United States threat. Moreover, the level of destruction if deterrence failed would still be a subject for grave consideration by United States leaders. But these difficulties seem inherent in a foreign policy prepared to meet the range of possible Soviet threats. It should be recognized that all-out nuclear war could start in many ways, other than by a premeditated Soviet attack. A local war might become so invested with national interests and prestige that Soviet leaders, if faced with decisive defeat, would choose to counter with an all-out attack. This danger has probably increased because Khrushchev seems less cautious than Stalin, less secure in his grasp of power, yet freer to exercise his diplomacy on a global scale. War might occur because of miscalculation of United States intentions; in a period of acute tension, verbal and even military indicators would be difficult to interpret, and the premium on a first strike might well tempt the U. S. S. R. to launch a preemptive attack. War might even begin by accident, triggered by a chance release of weapons, and carried on because both sides are poised in a high state of alert for quick and nearly automatic retaliation. Finally, as just mentioned, we cannot rule out the possibility that the United States, faced with a major Soviet challenge, might sometime find even the risk of devastation preferable to the consequences of accepting defeat. It may also be noted that nonmilitary defense measures could be more effective if war began in one of these other ways. A key factor in determining their effectiveness would be the ability of United States military forces to stop Soviet air attacks fairly quickly. Control of the military situation could be more quickly seized if the Soviet attack failed to achieve surprise, or if it were poorly executed, or if the United States were compelled to launch an attack. Nonmilitary defense measures themselves would also work better with more warning, as more fully discussed later in this report. Such wars appear sufficiently probable to warrant careful investigation of the potentialities of nonmilitary defense in these cases, as well as in the event of a premeditated Soviet attack. The dependence of the defense of civilian society upon the effectiveness of United States strategic offense and active defense capabilities should be stressed. Nonmilitary defense measures must be evaluated not only with respect to feasibility, but also in their interaction with other aspects of national defense. They should not carry such high economic costs that United States strategic offense, air defense, or local war forces would be dangerously weakened. Such an overall evaluation of the place of nonmilitary measures in the entire field of national defense has not been attempted in this study. On the basis of such initial considerations, the premise on which this study was started is that we should at least examine the feasibility of nonmilitary defense measures in a nuclear war. Of course, after investigation we may conclude that defense of civilian society isn't practicable. The destructive power of a single nuclear weapon, the delivery capabilities of high-speed bombers and ballistic missiles, the seeming inability of any current or proposed air defense system to prevent the delivery by an intelligent attacker of at least a considerable number of nuclear weapons, the widespread vulnerability of cities, the slow reaction time of large civilian populations-all these factors may well mean that effective nonmilitary defense can't be obtained at an acceptable cost. But the issues seem so important that they should be seriously investigated. This study is certainly not a definitive treatise on nonmilitary defense. Rather, the study was designed to provide an initial broad overview. It has seemed preferable to consider a large number of aspects of nonmilitary defense, and to examine their interrelations, rather than to go more deeply into a few questions. The work has been done on a part-time basis by a rather large number of people from different disciplines, and all of the different pieces do not fit perfectly together. The pieces also differ in the research upon which they are based some involve quantitative calculations of the performance of possible nonmilitary defense systems, others involve technical innovations and surveys of technical possibilities, while still others are necessarily based primarily on reflection about nuclear war and national defense strategy. Questions of the psychological reaction of the American people to a nuclear war and its aftermath remain largely unanswered. However, in the view of participants in the project, the study as a whole does provide a comprehensive orientation to nonmilitary defense problems which is fundamentally sound. II. POPULATION SHELTERS The first big question which must be raised about nonmilitary defense is whether people can in fact be protected from modern nuclear weapons. Protection involves not only provision of shelters capable of withstanding blast and fallout effects, but also arrangements for getting people into the shelters in response to different kinds of warnings. It should be stated at the beginning that it is impossible to provide reliable protection for all the population, and that the fraction of the population effectively protected depends greatly on the essentially uncertain nature of the enemy attack. There appear to be a number of possibilities for protective systems, however, and under plausible assumptions about the enemy attack and the civilian response, significant-and in some cases dramatic-reductions in civilian casualties appear to be obtainable. A. TYPES OF SHELTERS Improvised fallout shelters, even if only capable of reducing radiation to one-twentieth or one-thirtieth of the radiation outside, could have a significant effect in reducing casualties among people outside the areas of blast damage. There seem to be many possibilities of identifying and preparing such shelters in existing buildings in small cities and towns. For example, a location in the center of the basement of a 40,000-square-foot building (a typical large office, store, or school building) may have an attenuation factor of about 80. Moreover, a foot of earth gives a reduction factor of about 30, and sandbags distributed in advance could be quickly filled and placed to provide this type of shielding. Even buildings whose structural characteristics provide smaller attenuation factors could be quite useful, with arrangements for washing down or sweeping the roofs and surrounding areas (exposure to carry out the decontamination being rationed among the shelter inhabitants). An essential element in the use of such improvised fallout shelters would be radiation meters. The meters would indicate how long outside activity could continue until heavy fallout arrived, would guide immediate decontamination work, would show when it was safe to emerge from the shelters, and would continue to be needed in postwar reorganization. Two main types of meters are available—a dosimeter, which measures cumulative radiation exposure over an interval of time and might cost $1 to $5, depending on the model, and a dose-rate meter, which would be more convenient in some operations but might cost $15 to $20. Predistributing 1 dosimeter for every 5 persons in the country and 1 dose-rate meter for every 50 might thus cost $150 million to $250 million spread over several years. An often-neglected possibility is the use of suitably located mines for both fallout and blast protection. Mines for low-priced ores, such as limestone, sandstone, rock salt, and gypsum, typically consist of a regular pattern of rooms with level floors and 10- to 12-foot ceilings, completely self-supporting and dry. An engineering calculation prepared as part of the study indicated that a limestone mine at West Winfield, Pa., could be prepared for emergency 7-day occupancy at a cost of $25 to $35 per person. Such a mine would be provided with water tanks, latrines, utilities, and some air-conditioning equipment, and would be stocked with a bedroll for each person, cold processed rations, and some medical supplies. A wide range of shelter designs providing blast protection of 50 to 200 p. s. i. (pounds per square inch above normal atmospheric pressure) seem to be possible using conventional construction techniques-shallow underground location, reinforced concrete or corrugated steel material, and heavy airtight blast doors. There are still technical engineering uncertainties about many aspects of these designs, but corrugated steel shelters buried fairly deep look promising and are much cheaper than reinforced concrete. There is also uncertainty about the number of square feet to be provided per person, ranging from the 5 square feet in European shelters for short-time occupancy, to the 20 square feet in the Manhattan design discussed below. However, a reasonable guess is that a bunkroom type of accommodation for 90-day occupancy could be provided for something in the order of $300 to $400 per person. Such "medium" shelters might be appropriate for suburban fringes of large cities and for small cities which are presumed to be lower priority enemy targets. "Light" shelters, primarily for fallout protection, would of course be cheaperperhaps $150 per person. Constructed shelters should be designed to provide much greater attenuation of surrounding radiation than improvised shelters; fortunately, 3 feet of earth provides an attenuation factor of about 1,000. Moreover, most fallout shelters ought to be planned and placed deep enough so that they could later be made into blast shelters in the 20-100 p. s. i. range, primarily by adding better blast doors. With appropriate evacuation procedures, such "light" shelters in towns and rural areas might be used by more people than any other type. Finally, deep rock shelters created by mining rather than construction techniques appear to be the most promising approach (where sound rock is available) for blast protection in the range of 500 to several thousand p. s. i. Such "heavy" shelters would be indicated if a serious attempt were made to provide protection, other than by evacuation, for residents of large cities. Shelters with lower levels of protection might become partially obsolete soon after they were built. An engineering calculation of a system of deep rock shelters under Manhattan Island |