They had 50 minutes before the tsunami. Now, the big 'what if' game is to think whether we would have had a meltdown on the shaking alone. Seems like it. The meltdown was in full play within 5 hours, but they still had plenty of battery power then. Would the generators have flooded on their own from all the broken pipes?
External power was lost just from the shaking, and I've never seen a backup generator arrangement that was seismically rugged. As more news dribbles out, I am convinced we can treat this as a pure earthquake matter.
As I have stated, the ground motions were low. If there weren't hundreds of accelerometer readings, you would also get this by the light damage in towns that weren't swept aside. Do we actually have readings from the nuclear plant? No.
You realize, of course, that an earlier earthquake also damaged a nuclear plant. In that case, all the readings were destroyed, and we never knew the seismic fragility. I came to the conclusion that the ground motions must have been exceptionally strong, but now I have the feeling that they were similarly low. What is going to be done now? A rational approach might be to investigate why these nuclear plants are 'glass jaws', and perhaps something can be done. But you know what will happen....
In the old days, engineering methodologies were always changed by big disasters. Then, they would just say "Oops, Sorry!", and change things. Now, we have all the legal liabilities, and nobody can ever admit they were wrong. I wonder how many more disasters are needed before we admit that we are building things too fragile to live.