complete the setting for catastrophy. All that is needed is a source of ignition to start a holocaust. Fortunately, these conditions are met rather infrequently and not over the whole country at any one time. During the period from 1825 to 1956 the average interval of time between catastrophic forest fires was about 11 years. Only once were such fires experienced on successive years, the Yacolt fire in the State of Washington in 1902 being followed by the Adirondack fires in New York in 1903. The next shortest interval was 7 years and five intervals of 13 years or longer between such fire disasters occurred. A number of factors have served to reduce the areas burned in catastrophic fires, such as the reduction in large unbroken areas of virgin forest brought about by the clearing of much land for agriculture and other uses as the country becomes settled, better lumbering practices, and organized fire protection services.

Another important thing to remember is that most of these great fires were not stopped by firefighting, but by natural barriers such as lakes, rivers, and deserts, or by a change in weather conditions such as the onset of rain or fog or changes in the winds.

If such fires could occur frequently, the white man on coming to the New World would not have found over half its area covered with vast unbroken areas of virgin forests. A frequent cause of forest fires throughout the United States is lightning, or 1,000 to 2,000 fires annually are caused by lightning. More than two-thirds of the fires in the Rocky Mountain States and one-third in the Pacific Coast States are ignited in this manner.

Since similar numbers of fires must have started in the forests prior to the coming of the Europeans, most of them must have burned out by themselves without turning into great conflagrations. There are evidences of some large fires which can be read in the tree rings of the redwoods that have been growing since before the time of Christ. Also, many of the stands of southern pine and Douglas-fir are thought to be the result of fires which favored these more fire-resistant types. The conclusion appears inescapable that the large free-running catastrophic forest fire would occur very infrequently as the result of the detonation of a nuclear weapon.

PRIMARY IGNITIONS IN FOREST FUELS BY MEGATON WEAPONS

While many of the factors that determine the distances to which primary fires could be started in forest fuels by megaton explosions are similar to those discussed for urban-area fuels, there are some important differences. In the first place, the thermal energy required to ignite the most susceptible dry forest fuels is greater than for dry urban-area trash. Dry, rotted wood or punk appears to be the most easily ignited forest fuel, 4 calories per square centimeter being required for a 20-kiloton weapon and 9 calories per square centimeter for a 10-megaton weapon. Fine, dry grasses require 5 to 10 calories per square centimeter; dry leaves, 6 to 12 calories per square centimeter; and dry pine needles from 6 to 8 calories per square centimeter for 20 kilotons and 14 to 18 calories per square centimeter for 10 megatons.

A second difference is that sources of secondary ignition resulting from blast damage would be very few in forested areas compared

73266-61--24

with urban areas. Also, in areas of very dense forest fuels which correspond to densely built-up urban areas, the forest canopy would shade the more easily ignited fuels on the forest floor from thermal radiation to a greater degree than structures in urban areas.

It is significant that in Nagasaki where the hillsides of the narrow valley were wooded and free of buildings, there was no evidence that the trees and brush were ignited by the direct thermal radiation from the explosion. In some areas trees were scorched, but no spreading brush fires resulted. Also, there was no general spreading of fires from burning stuctures to the wooded hillsides, although a few small areas of brush and grass immediately contiguous to hot fires in buildings were blackened.

The Forest Service of the U.S. Department of Agriculture has published the results of a study designed to aid in estimating the "probable burnout areas" that might result from airburst nuclear weapons of 1, 3, and 10 megatons. The burnout areas are displayed in tabular form for 475 U.S. regions, supposedly characterized by their climatology, natural fire barriers, and areas covered by forest and rangeland, and for each month of the year. For the regions and 3- to 5-month periods of minimum fire hazard, the entries apparently correspond to the areas of primary ignition by thermal radiation for the various bomb yields. For other regions and periods of greater fire hazard, a degree of fire spread is allowed for ranging up to a factor of 41 for 1 megaton and a factor of 12 for 10 megatons in the worst regions and the worst months.

Table H-4 shows the minimum areas listed and the thermal energies to which they correspond for airburst weapons and unlimited visibility.

TABLE H-4.-Fire spread from thermonuclear weapons

The maximum areas of spread listed are also shown. The areas were apparently chosen as those which would result if the prevailing level of fire danger approximates the average peak fire danger attained two to four times during the month.

These estimates of fire spread, while taking into account variations in fire hazard from month to month and from region to region, in my opinion, still overestimate the burned areas, because they represent the 2 to 4 days during each month when the average peak fire danger is greatest. They correspond to conditions which would be experienced on the average only 10 percent of the time.

Also the estimates showing the largest fire spread give greater burned areas than have ever been observed in the most catastrophic forest fires in the past. When one considers that the greatest of these burned over 5,900 square miles, estimates of 8,000 to 9,000 square miles under conditions where the forest cover is far less continuous than it

was in 1871 in northern Michigan and Wisconsin, appear very unlikely. It is also important to understand that the great fires during the 1800's were not the result of spread from single ignition points, but really were many independent fires which coalesced under extremely hazardous fire conditions. During these periods it was customary to pay little attention to forest fires if they did not endanger settled communities. Fires were burning in the debris resulting from lumbering and land-clearing operations most of the time. Forests were regarded as inexhaustible resources, or in many cases, as a nuisance. Probably the largest areas that would burn over now would be more characteristic of the catastrophic fires experienced since 1910, or less than 500 square miles, and these would occur with a very low probability.

Dr. Mitchell has discussed the ecological problems that can result from forest fires and the various methods that are available to deal with them during postwar recovery. In this connection, if one takes the lower estimates of burned areas shown in table 4 as being more reasonable than assuming large factors of spread, 405 1-megaton bombs, 213 3-megaton bombs, or 101 10-megaton aimed specifically at the forests would do damage comparable to that experienced in each of the years 1930 and 1931. Furthermore, since the ecological consequences of forest fires extend over many years, twice as many bombs as mentioned above would be comparable in ecological consequences to the combined effects of the 1930 and 1931 forest fire experience. While such forest damage as was experienced in 1930 and 1931 was undoubtedly serious, it is clear that recovery from such damage and the damage experienced by fire year after year probably would not be sufficient to prevent postwar recovery.

Another point of interest is that in saw timber stands, which have been devastated by fire, much of the lumber can be salvaged during a period of years following the fire and before insect damage destroys its usefulness for lumber. For example, between 1933 and 1952, 8 to 10 billion board feet of lumber were salvaged from the Tillamook burn. Since much lumber would be needed for postwar reconstruction, it would be important to plan for such salvage operations. In addition, this procedure improves conditions for the recovery of the forest and makes future fire damage less probable.

PREVENTIVE AND CORRECTIVE MEASURES FOR FOREST PROTECTION

There are a number of things that could be done during the years preceding a nuclear attack which could reduce the fire damage to forestland in the event of such an attack.

First the practice of good silviculture, which keeps the forests free of dead snags and floor clutter, would reduce the probability of primary ignitions from thermal radiation and reduce the intensity of any fires occurring. This, in turn, reduces the probability of crown fires, which are the most damaging to mature trees.

Increasing and improving the care of firebreaks would help limit the spread of fire and increasing the numbers of access roads would make firefighting easier.

Improving and expanding the fire protection services would reduce peacetime forest fire damage which at the same time would reduce the danger of fire during nuclear attack. Forest areas which have

been severely burned become more susceptible to fire because the cover that grows during the first few years after a fire is more easily ignited and the dead snags left from previous fires burn more readily. For example, most of the Tillamook burn area burned over again in 1939 and in 1945, and a smaller fire covering 38,000 acres burned in 1951. Research and development aimed at improvement of methods of . fighting forest fires, particularly in the presence of fallout, might lead to a real capability to combat fires following a nuclear attack. The use of aircraft to fight forest fires has increased during recent years and further development of these techniques could be of great importance.

The development of more tree farms and improved methods of reseeding or replanting burned forest areas could also be of great service in repairing forest damage during the recovery period following a nuclear attack.

Most of these measures would greatly reduce our peacetime forest fire damage and would not be wasted in the event, as we all fervently hope, that no nuclear attack is experienced.

CONCLUSIONS ON MASS FIRE THREAT

Fire damage to urban and forested areas from a nuclear attack is frequently estimated by taking the most pessimistic values for all factors involved. This leads to gross overestimates of the damage likely to be experienced. By making the situation appear hopeless, such estimates do a great disservice by preventing actions which could do much to reduce the damage from a nuclear attack and help speed recovery during the postwar period.

Preliminary study indicates that fire damage to urban areas is likely to be confined largely to areas seriously damaged by blast. In relatively infrequent weather situations, fire may spread beyond the areas of blast damage, but even in these cases, increase in damage area due to fire spread is unlikely to exceed a factor of two.

Estimates that conclude that fire would destroy the greater part of our forest and range lands are probably very erroneous, because the enemy would use his weapons to better advantage by assigning them to military or urban targets and spread of fire from such targets to forest areas is unlikely to occur for the major portion of these targets. Many measures can be taken before attack that would reduce the fire damage, if an attack should occur, and also aid in postattack recovery. Such measures require study, research, and development. Those that show real merit should be incorporated into civil defense systems.

Thank you, Mr. Chairman.

Mr. HOLIFIELD. Thank you Dr. Hill. You have certainly given us a very comprehensive paper on this particular subject. It is good to have it in the record because I don't believe it has been thoroughly done before.

Any questions?

Mrs. GRIFFITH. No questions. It was very interesting.

Dr. HILL. Thank you.

Mr. HOLIFIELD. Mr. Morse.

Mr. MORSE. No questions. I would like to congratulate Dr. Hill on his scholarly treatment.

RELATIVE EFFECTS OF FIRES

Mr. ROBACK. I wonder if Dr. Hill would comment on the relationship between burning and other weapon effects from the standpoint of civil defense programs.

Dr. HILL. You mean the amount of effort that has been put into studying it?

Mr. ROBACK. In terms of the fire hazards, let us say, of thermonuclear attack in relation to the civil defense program that we heard testimony on today.

Dr. HILL. All I can say is that it is my general conclusion that the damage resulting from fire will be rather of a third order effect compared to the potential damage from blast and radiation effects. However, it is not to be denied that in some instances there may be real spreading conflagrations. There is a small probability of it but there is a chance it might occur in some target areas.

I am not acquainted with any efforts going on, proposed in the new program, toward this. The reference which I quoted which is being done by the Forestry Service in the Department of Agriculture under contract with the Broadview Corp. has been done for some little time and whether there is any idea of extending this work or not, I don't know.

Maybe Dr. Mitchell knows something about that. That is the only comprehensive Government program that I know anything about in the forest fire area.

I think it is very important in connection with the idea of personnel shelters. If one plans to make shelters, as Mr. Strope mentioned yesterday, in the areas where you get significant protection from blast and therefore may have significant blast damage, the experiment which he understook involving the fire effect was very important. However, I think that the assumption that shuch shelters would be in rather open areas where the proximity of the intense fire would not be as great, might allow him to get away without closing off the ventilation system, I mean this wouldn't be true for big shelters under buildings and so forth, as he mentioned. I think this requires more research. It should be looked into.

SPREAD AND LENGTH OF FIRES

Mr. HOLIFIELD. What about shelters in residential areas where you had the usual 1- and 2-story buildings. Would there be enough material there that would ignite to cause a fire storm?

Dr. HILL. I think it could happen, particularly in fairly densely built-up, highly combustible areas.

Now, in very good residential areas with not a very large built-upness, it would be less likely to happen. But I must say I would be very unhappy to build a shelter in close proximity to my house as it would have to be, without having some way of shutting off outside ventilation and providing for the purification of air for a period of a half-day or a day, whatever it may be.

Mr. ROBACK. Would you think that is the length of toxic heat effects?