I think missile silos are stationary is pretty easy to understand. They're easier to guard and if you look at when they were built rather than their current functional use (its debatable if there is one when looking at SLBM and B-52/B-2 dropped nuclear weapons). In the 1950s, when ICBM projects began, the threat was not thermonuclear weapons, but fission weapons. It was thought that the silos might survive a direct/indirect attack on a facility.
After the advent of the hydrogen bomb, no harden target can really be called safe. Coupled with modern military GPS and you have a super-precise, super-powerful weapon.
The silo parks in USA are still nuke-proof. Because they are far enough apart that you can't get more than one per hit, and it has to be a low or groundburst, but close enough together that the first hit causes a cloud of dust that the subsequent missiles would have to go through to hit their targets. And a nuclear warhead in re-entry has no chance of surviving hitting that dust cloud.
This means that the silos are protecting each other -- you'd need to wait hours between hits to be able to land nukes on all of them. (Or hit them all within the minute.)
So why the need for launch on warning? If you are confident that your missiles are going to be safe from a first strike you could sit there and take the damage.
What you describe sounds more like Dense Pack - which I don't think was ever implemented:
I'm not sure what you mean by this. "Real-world data show that some high-quality GPS SPS receivers currently attain better than 3 meter horizontal position accuracy" [1], and military GPS is even more accurate.
Umm, that back in the time the missile silos were planned, there was no GPS? That combined with the relatively low-yield warheads meant that hardened silos were a viable weapon - they were likely to survive a first-strike.
After the advent of the hydrogen bomb, no harden target can really be called safe. Coupled with modern military GPS and you have a super-precise, super-powerful weapon.