Since 1948, I have been employed by the RAND Corp. where I have specialized in the area of nuclear weapon effects and technology.

PROBLEMS OF FIRE IN NUCLEAR WARFARE

INTRODUCTION

Mr. Chairman and members of the committee, I welcome the opportunity to discuss briefly with you the possible impact of fire in the unfortunate event of a nuclear attack on the United States. Let me state at the outset that the opinions I express are my own and do not represent any official position of the RAND Corp., where I am employed. Also my remarks are based on a study which has been in progress for only a few weeks and is not yet complete. Furthermore, my field of specialization is nuclear physics and while I have been concerned with the effects of nuclear weapons for a considerable part of the past 12 years, I make no claim to being an expert in the field of fire protection or prevention. I also wish to emphasize that all my remarks are unclassified.

GENERAL CONSIDERATIONS

However, as the study has progressed, I have become increasingly convinced that, while fire damage which might be caused by a nuclear attack on the United States could be very serious, it need not be catastrophic in the sense of preventing postwar recovery from rather heavy nuclear attacks.

Furthermore, I am convinced that there are many actions which could be taken before such an attack that would greatly reduce the fire damage inflicted. In addition, if appropriate plans and preparations are made beforehand, many things could be done after the attack to minimize the long-term undesirable consequences of the fire damage which might be experienced.

Concern has been expressed that fire from nuclear attacks on various targets might spread far beyond the area of serious damage from blast, thus mutiplying the area of destruction many times and that free-running fires would spread through forest and grasslands which would burn over such wide areas that the ecological consequences of soil erosion and floods might make postwar recovery impossible.

The problem of estimating fire damage from hypothetical nuclear wars involves many difficulties and uncertainties. In very general terms, the procedure would involve making assumptions about the enemy choice of time and targets for attack and the number, yield and altitude of burst of weapons delivered on designated ground zeros. These assumptions would be much the same as have been made in previous studies of the effects of blast and fallout, except that one would have to decide whether or not forest and grasslands would constitute a primary target subsystem to which enemy weapons would be assigned or be regarded as a bonus from attacks on other targets such as urban areas or military installations, et cetera.

To estimate the fire damage, assumptions about a number of additional factors become important. Among these are: meteorological factors such as wind velocity, temperature, relative humidity, visibility, lapse of time since the last precipitation and presence or absence of

inversion layers and cloud cover in the target area; fuel characteristics such as types of combustible materials, their surface density, uniformity of distribution and moisture content; topography, geometrical form, and degree of built-upness in the target area; and finally numbers and distribution of sources of primary and secondary ignition from thermal and blast effects.

The problem of estimating areas within which initial ignitions would occur for given weapon and target characteristics is relatively straightforward, but estimating the spread of fire from these initial ignitions is much more difficult.

Thus far our study has not progressed to the stage of evaluating fire damage for complete nuclear campaigns, but work has been aimed at understanding the significance and interaction of the various factors outlined above.

One of the most important facts to realize is that in an area as large as the United States, not all of the factors influencing the ignition and spread of fire would be at their worst extremes all over the country during the short period of time required to deliver a nuclear attack. For example, the seasonal periods of worst fire danger are different in different areas of the country and the overall conditions can vary marketly from year to year which means that careful study of the variations of these conditions for various climatological areas over long periods of time should be evaluated statistically.

For example, southern California, Nevada, and parts of Arizona have experienced extreme drought for the past 3 years, but large regions in Texas have had such excessive rainfall this year that crops are being severely damaged.

Similarly, the normal periods of maximum fire danger in parts of Maine are July, August, and September, while in parts of Florida, Alabama, and Mississippi the period from October to March is the most dangerous.

Also in the coastal region of southern California during the months of July and August when virtually no rain falls, even in normal years, there are almost daily foggy periods which would tend to limit the areas of initial ignition and the degree of fire spread in the event of a nuclear attack during such periods.

One could multiply these examples for each of the factors influencing fire damage showing wide variations in time and place. The implication is clear that an exaggerated, misleading picture will be obtained unless average conditions are considered as well as extremes.

FIRE DAMAGE IN URBAN AREAS

Since the only experience with fire damage from actual nuclear attacks resulted from the bombings of Hiroshima and Nagasaki, it is instructive to examine the conditions at the time of these attacks and the resulting fire damage. Table H-1 summarizes the bombing data, meteorological conditions and target characteristics at the time of the atomic bomb attacks on Hiroshima and Nagasaki.

TABLE H-1.—Significant target parameters in the atomic bomb attacks on Hiroshima and Nagasaki

It is significant to note that the bomb yield, altitude of burst, meteorological conditions and target fuels were very similar in the two cities at the time of the attacks.

The major differences were the positions of the ground zeros relative to the more densely built-up areas of the cities, their geometrical shapes and their topography.

The fire experience in the two cities is summarized in table H-2. TABLE H-2.-Fire damage in Hiroshima and Nagasaki

Examination of tables H-1 and H-2 indicates that, despite the favorable burning conditions of clear, dry weather, highly combustible fuels and densely built-up targets, the burned over areas were largely restricted to areas of serious blast damage in both cities and were much less than would be predicted by the primary ignition experiments in the Nevada tests.

The probable reasons for the smaller-than-expected areas of initial ignition and the negligible fire spread are different in the two cities. In Hiroshima, despite the fact that the densely built-up area greatly exceeded the burned over area; the fire did not spread appreciably beyond the area of initial ignition because of the rapid development of a "fire storm."

"FIRE STORMS"

A fire storm is characterized by strong to gale force winds blowing toward the fire everywhere around the fire perimeter and results from the rising column of hot gases over an intense, mass fire drawing in the cool air from the periphery. These winds blow the firebrands into the burning area and tend to cool the unignited fuel outside so that ignition by radiated heat is more difficult, thus limiting fire spread.

The conditions which give rise to a fire storm appear to be low natural wind velocity, flat terrain and a uniform distribution of high-surface density, highly combustible fuels which burn rapidly, coalescing individual fires into one burning mass within the fire perimeter.

Such fire storms have been observed in forest fires and were frequently experienced in the mass incendiary air raids in Europe and Japan during World War II. In fact, such fire storms were the most frequent type observed in Japan during mass incendiary raids. It was typical in such cases that the fire was mainly confined to the areas initially seeded with incendiary bombs, but within these areas fire destruction was virtually complete.

In Hiroshima, hundreds of fires were burning throughout the area ultimately burned over within 10 minutes after the bomb exploded. Each of these spread rapidly to adjacent structures during the first half-hour, by which time the fire storm was well developed. Practically all fire spread had ceased after 2 hours at which time the fire storm was approaching its peak intensity, with centrally directed winds of 30-40 miles per hour.

In Nagasaki, in spite of the similar yield, altitude of burst and weather conditions, a fire storm did not develop, probably because of the uneven terrain, the irregular layout of the city and the location of ground zero in a long relatively narrow river valley north of the center of the city. Here, such spread of the fire beyond the area of initial ignition as was observed, was to the southeast against the wind direction at the time of the explosion. Because the rate of spread was slower, the fire burned longer. Here also, the combination of terrain, city layout, position of ground zero and wind direction limited the spread of fire primarily to areas seriously damaged by blast.

The fact that the areas of primary ignition from thermal radiation were significantly smaller than the area predicted for dry, combustible, light fuels (newspaper, et cetera) in both cities was probably due to a number of factors. In the first place, the ignition energies measured at the Nevada tests were made under conditions of very low humidity typical of the desert. Since Hiroshima and Nagasaki are both seacoast cities cut by numerous bays and rivers, the relative humidity was probably considerably higher than in the Nevada desert, thus increasing the ignition energies markedly. Also not all potential sources ignite because many are shadowed from the thermal radiation and of those which do ignite, many are not close enough to heavier fuels to ignite them and quickly burn out or are blown out by the blast wind. The fact that initial fires in both cities were confined to areas of substantial blast damage suggests that most of them, at least at the outer limits of these areas, were the result of secondary ignitions caused by blast damage. This was borne out by the testimony of survivors.

In Nagasaki, many areas were protected from both blast and thermal radiation by being in the shadows of hills and ridges and this, coupled with the long narrow shape of the built-up area in the valley around ground zero and the much lower built-upness at the north end of the valley, limited both the blast and fire damage even more than the development of the fire storm in Hiroshima.

All this does not deny that under exceptional conditions of high wind, very low fuel moisture and a high degree of builtupness fire