Here is a thread on the "Recovery Manual for Civilization," which I thought is a useful addition to your list: http://lesswrong.com/lw/l6r/manual_for_civilization/

And here was (most of) my comment in that thread:

My first conclusion was that there are all kinds of events that could lead to a collapse of civilization without exterminating humanity directly. But it may be impossible for humanity to rise back from the ashes if it stays there too long. Humanity can't take the same path it took to get to where it is now. For example, humanity developed different forms of energy as prices of previous forms rose. For example, we started digging up shallow coal when population grew too high to use charcoal produced from wood. But many of those resources are no longer near the surface. All coal, oil, and gas near the surface has been extracted. So humanity, if it rose again, would have to find a different path.

My conclusion was that I would not want to invest enough time into preparing for the end of the world to get personally involved in preparing for the end of civilization. But some people already do: all the survivalists and similar people. So, if I want to help civilization recover after a collapse, my best alternative is not to try to store the information myself. Instead, it is to reduce the information to a useful form and give it to these people. This is far more robust that trying to think of the best way of having the information survive. Give it to a few thousand people in scattered places with different survival strategies, and it is much more likely that the information survives.

These two facts together led me to conclude that a collaboration to prepare some kind of big, easily read manual of technology would be a valuable contribution to the survival of humanity. It would need to build on itself, somewhat like tech trees in strategy games.

Plus, it would be really amusing for the LW community to decide that one of its bets for saving humanity is to help outfit survivalists with technical know-how.

I'd add that a case that fits within the general idea of a disaster that destroys civilization, but does not extinguish humanity would be a few (possibly as few as one or two) electro-magnetic pulse detonations over the eastern or western seaboard of the United States. I can see the follow-on-effects bringing civilization down. I would think that would get worse in the next 50 years or so, as India and China catch up to the level of computerization prevalent in the US.