The irrational reason is that it's been 60+ years since atomic weapons were deployed, and we are confident the danger has passed. But really, there's plenty of weapons out there: http://www.motherjones.com/politics/2011/11/map-nuclear-bomb...
A serious, concrete underground "hideout" will save you from anything but a really close hit; I can look up numbers if you want, but I maybe remember a good metric is 1MT and one mile of separation, requiring a less intense shelter. And a large fraction of the population can get by with much less, carpet bombing suburbia and exurbia was never going to happen, especially due to counterforce requirements and the Soviets believing that MAD was profoundly immoral and never buying into it.
Also, I think most shelters would be far enough from blast areas to provide some protection.
Eizo Nomura (who, amazingly, doesn't have a Wikipedia page) was 170 meter from ground zero in Hiroshima, and survived (http://en.m.wikipedia.org/wiki/Atomic_bombings_of_Hiroshima_...)
Akiko Takakura was above ground at less than 300m and lived for 60 years.
Even though later bombs were much more powerful that makes it likely that shelters at the edge of the blast radius can save lives.
Not all of them, but any program to protect civilians is a statistics game.
All the construction in SF isn't 100% effective against earthquakes, either.
What you need is plenty of mass between sources of radiation, i.e. fallout particles where they settle, and from "skyshine". In a sufficiently bad situation, which would be pretty common, a simple ditch where you made sure the fallout kept out and didn't accumulate on, say, the tarp you were using for that, you'd still get zapped by gamma rays that bounce off atoms in the air and come down into the ditch.
So the most basic expedient fallout shelter is a ditch with hollow core interior doors over it or the like, and foot of dirt on top, which through arching holds most of its weight.
An above-ground building is iffy, because you want, say, 3 feet of dirt all around you, and either a foot on top or if the roof is high enough that's not as much of an issue. Middle stories of a tall building could suffice.
The lower 48 states are only 3,119,884.69 square miles if you inclde waterways. With just land it's only 2,959,064.44 square miles.
So, 10k nukes let's you carpet bomb the lower 48 and have one within 10 miles of just about every point. Focus on population centers and I suspect you could easily get 95+% of the population within 5 miles of a detonation.
Mostly, I was just pointing out that fairly soon they could branch out and target just about any population center. I don't think they ever actually targeted death valley for example.
Anyway, the important thing to remember is by the mid 1970's many of these bomb shelters had been unused for 20+ years. It's very natural to look at something like that and say spending more on this is probably pointless.
A well built bomb shelter would be very effective at reducing the lethal range of nuclear weapons.
Most of the stockpiles during the peak of the cold war were tactical nuclear weapons with low yields designed for the battle field.
And most warheads are less powerful (see my other comments); using my handy "RAND" nuclear effects calculator from the back of my copy of The Effects of Nuclear Weapons, 15 PSI for the following weapon air bursts (targeting cities, not ground bursts against ICBMs etc.):
500 Kt: 1.8 miles
300 Kt: 1.5 miles
100 Kt: a bit over 1 mile
50 Kt: 0.85 miles
The rational reason we don't build nuclear shelters anymore (which is not actually true by the way, Switzerland still builds them although they're not required anymore for private residences) is because there's very little chance of a massive nuclear war which is what they were for.
> As weapon yields increased, it's become apparent that a concrete, underground hideout is not going to save you.
Weapon yield has decreased, not increased. In the 50s, delivery was through bombers, you wanted big bombs because many bombers were going to be shot down so each nuke delivered had to pack as much punch as it could. Early ICBM had similar-ish issue, you had few inaccurate rockets and they had a warhead each so each warhead had to count, you built a big rocket and a big warhead on top of it, and the ones that didn't fail and weren't shot down razed a city even when they missed it by tens of miles. That's where you had multi-megaton designs
With the 60s and MIRV multi-megaton went online (they'd been designed in the 50s) but systems designed in the 60s and deployed in the 70s for the exact same role all went sub-megaton, half a megaton at most, usually less, for smaller and more precise delivery platforms and MIRV systems.
The most numerous warhead in the US nuclear arsenal is the 100kT W76.
Because both leaderships of US and USSR knew starting WW3 would mean end of civilization on the planet, they controlled themselves.
MAD was insanity cooked up by Robert Strange McNamara and company of Vietnam infamy in the early-mid-60s. The Soviets never bought into it, and thought, correctly, that a plan to burn alive children in their homes was irredeemably evil, and a sign the US was very evil. And at this level it was and still is.
Although emphasize could, it was hardly certain. (And if this strikes you as crazy, well, remember that since a little before 1920 they explicitly were planning on taking over the world, and got uncomfortably close to achieving that goal.)
This was one of the reasons SDI, even on paper/in the lab, was so terribly effective. Against an even partly effective ABM system a counterforce first strike becomes impossible, since you can't choose which warheads or boosters get intercepted.
Counterforce is an attack on your opponent's strategic forces, aimed at preventing them from retaliating. That's one of the reasons we had to build the expensive "Triad" of ICBMs, SLBMs, and manned bombers. Reliably taking all three out is very difficult.