Mr. HANUNIAN. Yes. Here [pointing to the uppermost segments] there is no instruction, no knowledge, heedless behavior essentially.

Over here [pointing to the lower segments] there is instruction, there is indoctrination, in fact the population is assumed to heed the instruction, take maximum advantage of the basements that are available, yes.

Mrs. GRIFFITHS. May I say again, make clear, I think the top of the diagram which shows no knowledge, is sort of a false picture, because it really is no knowledge of just one item. It is not no knowledge at all, but if on one thing you failed, you can die.

Mr. HANUNIAN. Yes. The criticism has some validity, I must confess. However, the gamma radiation is the major threat. That must be overcome. Then, after you think you see a way of overcoming that, then there are other things that you must also do, I agree.

EFFECT OF ATTACK VARIABLES ON CASUALTIES

Well, to resume my statement, it is apparent that fallout might very substantially increase the level of fatalities. If surface bursts were used exclusively, and that is assumed here, fallout's contribution might well be much as this chart suggests. Fallout has the potential of adding at least 10 percentage points to the percentage of the population killed by blast in each attack depicted.

This amounts to raising each total by at least 20 million dead. Even when blast damage is itself appallingly heavy and fallout accounts for only a modest proportion of all deaths, as here in the attacks on cities, then we find that the absolute numbers of deaths due to fallout radiation is nonetheless large, 20 million, as I suggested.

Second, it is clear that even when fallout is taken into account, the uncertainty in predictions of the number of dead that enemy attacks might produce still arises in very large part from ignorance about the megatonnage he will deliver to population centers.

The differences between the city bombing bars and the military target bars are the most prominent feature of the chart, either with the large attack-the 3,000-megaton attack-or the small attack-300 megatons.

Third, the effectiveness with which blast survivors exploit what shielding is available to them makes an important difference to the final levels of deaths. Population behavior can quite conceivably alter survival by 10 percentage points even though civil defense shelters may not have been built.

Fourth, our uncertainty about certain weapon characteristics I mentioned earlier, fission fraction and radioactivity level, is revealed by these estimates to be only a little less important than population behavior, civil defense-wise.

Fifth, we may note that total deaths are rather insensitive to the yield of individual weapons, if attacks are directed against military installations. That is to say, looking at the 3,000-megaton attack against military installations, we find the tops of the bars, for example-or the bottom of the yellow segment, for another-are just about at the same level.

Much the same is true here at 300 megatons.

But, perhaps, the most important inference to be drawn from this chart is that the casualty levels from plausible attacks would all be op

pressively high. We ought to explore every possibility for reducing them.

One of the critical factors is the amount of shelter likely to be available to our population. I shall now address myself to this aspect of the problem.

AVAILABILITY OF SHIELDING

As a nation, we have not so far made any great strides in improving the shielding available to the general population or even in identifying those structures whose nature would make them passable fallout shelters. Nor have we managed to impress upon many people how they might sensibly behave if faced with nuclear attacks.

Thus, until developments of recent weeks, it seemed entirely possible that an attack, say, in 1963 would find the U.S. population still unprepared and making rather ineffective use of what shelter existed.

INDOOR-OUTDOOR BEHAVIOR PATTERNS

Even then, however, people would surely not expose themselves to the maximum possible dose. To do so would require absurd behavior, staying outdoors around the clock, day after day.

Actually, people outside the areas where structural damage was significant would presumably spend at least their sleeping hours indoors, benefiting from whatever shielding their quarters happened to afford. Surely part of the remainder of each day would also be spent indoors. If most people would behave in some such fashion it would turn out that the general population would fare little worse than it would have if, unaware of the fallout hazard or indifferent to it, everyone resumed peacetime behavior patterns immediately on sensing the attack's end.

The casualty estimates I label "incidental protection" here (the yellow bars talked about) have been made using the shielding factors appropriate to peacetime behavior patterns. The population ought to do no worse. Account has been taken of the fractions of time that sectors of the population would normally spend out of doors and in various kinds of buildings.

On the average, in this incidental protection case indicated by the yellow, about 60 percent of all radiation is effectively eliminated by environmental shielding. By way of a reasonableness check, the 40 percent residual can be compared with residuals of 70 percent for a person in the open, 30 to 60 percent in a frame house and 10 percent in a factory. These are generalized sorts of figures, of course.

USE OF EXISTING BASEMENTS

The other case considered so far assumes as I have indicated, that equal parts of the population receive 5, 10, and 20 percent of the nominal radiation. This is consistent with planning factors for continuous occupancy of ordinary basements and foxholes. It would also be consistent with part-time occupancy of improved basements, basements with sandbagged windows, for example; or, better, with sandbagged walled and roofed shelters built in corners of the basements. The fatalities estimates, however, are simply labeled "basement protection."

FIGURE NH-4.-PERCENTAGE OF POPULATION HAVING READY ACCESS TO DWELLING BASEMENTS, BY REGION

In this connection, we should note that basements are by no means universally available. Only about 60 percent of the U.S. population has ready access to dwelling basements, and the percentage varies regionally. (See fig. NH-4.)

Almost all the people in the Northeast, almost all dwellings there, have basements, 93 percent of the people have ready access to basements in the northeastern section of the United States.

Mr. ROBACK. Mr. Hanunian, is that information based upon the housing census?

Mr. HANUNIAN. This information is not based upon the census taken last year. That information was not available when I made these estimates and, so far as I know, it still is not.

The number of basements included as part of the basement census is based rather on the Bureau of Labor Statistics work showing the increments to dwellings, the new dwellings built, and their nature.

Those surveys by the Bureau of Labor Statistics were sample surveys. There is a sampling error involved, but I gather from talking to population census people that the overall figure of about 60 percent or 55 percent is in the right ball park.

Mr. ROBACK. Sixty percent; what is the total number of dwelling units, about 60 million?

Mr. HANUNIAN. I do not know that.

Mr. HOLIFIELD. I think the previous hearings showed about 45 million, if I remember correctly.

Mr. ROBACK. I believe the figure is higher, but if I recall the figures, Mr. Chairman, as many as 20 million units would be tenant occupied, in which case there might not be as much incentive to do anything about improving the opportunity for shelter.

However, Mr. Hanunian is not testifying on that point.

Mr. HANUNIAN. Well, because of this uneven distribution of basements, because not everyone, by any means, has basements or ready access to basements, there is some point in considering a case in which those persons who have ready access to basements make use of them while the rest of the population benefits from no more than the incidental protection described earlier.

MORE BASEMENTS IN DENSELY POPULATED AREAS

Mr. HOLIFIELD. Fortunately, the population concentration seems to be in the area where you do have basements.

Mr. HANUNIAN. Yes, that is right. The location of the basements is doubly fortunate, in fact. It is fortunate because basements are common in the most densely populated regions.

It is also fortunate in a different sort of way, in that basements are scarce in those regions where the amount of shelter they provide would probably be inadequate anyway; that is to say, near the densely targeted missile installations.

But to go on, the three degrees of fallout protection I have described so far reflect, at most, only very modest civil defense protection, if any at all. None imposes a requirement for construction of anything more elaborate than a foxhole, and that much is required only when persons who lack access to basements try to achieve equivalent shelter. However, it now seems that the United States is about to begin a more ambitious civil defense program, one involving the creation of special fallout shelters of some sort. To reflect a completed program of this general sort (though certainly not specifically of the kind the officials have in mind-I do not yet know what that is exactly),

but a modest shelter program-to reflect a completed program of that type, there is a fourth shelter case I want to consider.

ASSUMED FALLOUT SHELTER POTENTIAL

In this fourth case the nuclear residual is 2 percent, that is to say, 2 percent of the radiation outside reaches the population which is inside some kind of shelter. And thirds of a population again receive differing fractions: One-third receiving 1 percent; one-third receiving 2 percent, and one-third receiving 4 percent.

At the very least, these residuals imply virtually continuous occupancy of special shelters in basements, something like the sandbag walled and roofed huts I described a moment ago; however, part-time occupancy of fallout shelters that also provide modest blast protection is implicit in the casualty estimate I am going to show you. This next chart summarizes these four types of shelters. (See table NH-1.)

TABLE NH-1.-Levels of residual radiation

1 Radiation received, as a percentage of the ambient dose.

The incidental protection (which is no shelter at all, really, except that incidental to ordinary behavior), in which the population receives about 40 percent of the dose available to them, the ambient dose.

The basement protection case, the one I discussed second, in which they get a median value of 10 percent of the dose; then a mixture of the incidental and basement cases (to take account of the varying availability of basements), in which, of course, the protection mounts and they receive between 10 and 40 percent; and then the special fallout shelter where the median residual is 2 percent.

Before leaving this topic I should point out that it may be important for shelter occupants to remain in their shelters continuously, or nearly so, for much more than a day or so. To be sure, the rate of radiation falls rapidly over time. Two days after detonation it will be down to 1 percent of the rate at 1 hour after detonation; but even a week or more later, the level may still be high enough to be unacceptable.

The damage estimates I am presenting today are based on maximum biological doses. These were computed with attenuation of the radiation assumed to be constant over time. This may be interpreted as meaning either that the behavior of the population does not change (for example the shelters are occupied continuously), or that abandonment of the shelters is accompanied by local decontamination efforts, or that areas that are still dangerous are evacuated when the shelters are abandoned-which of these interpretations it is appropriate to put on it depending upon which kind of shelter you are considering. It certainly is possible for the population to continue "peacetime" behavior indefinitely until it dies, at least in the incidental protection