The majority of GPS satellites do not have one rubidium clock on board, they have four. For a discussion of the experimental time sync system planned for QZSS see: http://web.archive.org/web/20130411122243/http://unsworks.un...

I am curious before reading my comment did you think that GPS would continue to work without navy.mil and af.mil?

According to this "Fact Sheet", the latest GPS generation currently in orbit (Block IIF) has two Rb and one Cs.

http://www.losangeles.af.mil/library/factsheets/factsheet.as...

The next generation (III) will take away the Cs and only rely on (three) Rb:

http://www.losangeles.af.mil/library/factsheets/factsheet.as...

The current GPS constellation status can be seen here, Plane/Slot describes the position in orbit, SVN is the "Space Vehicle Number", "PRN" is typically the "Sattelite Number" as displayed on a GPS receiver, because it describes the "Pseudorandom number (sequence)" used to decode signals of this particular satellite. Clock is just for information what the satellite right now uses as a timebase. As of 2014-04-27

http://www.navcen.uscg.gov/?Do=constellationStatus

Hell yes I did. Whole point of Rubidium clocks is you dont have to correct them.

I was hoping GPS would stay up for at least 5-10 years - I assumed all they needed was periodic orbit corrections.

Well, there goes my GPS to the rescue Zombie survival plan :/. Time to learn how to read stars, or better yet - find android app that does it automagically from a picture :)

Or you could learn to use a map and compass. I know, I know, antique and all that, but reliable.

That is going to take the party out of Tactical Air Control Party.

I would have done. Having once convinced myself that four satellites shouting out the (accurate) time is sufficient to locate oneself in three dimensions, I assumed that was all there is to it.

A really big problem with satellite technology are the harsh environments they go through. You're dealing with something like 10-20g's of vibration peak-to-peak on ascent, and then you subject it to the freezing colds and searing heat of space, all while doing this in vacuum.

Since those clocks need to be super precise, it's easier to have an accurate reference on the ground where everything's not getting shaken and baked to bits.

Not right away, it will just be slightly less accurate. Ignoring all other sources of errors, for extremely high precision work you still might want to correct for the relativistic difference in time between the satellite and the surface; but in a post-apocalypse setting, I don't think you'll mind the small drifts.

There's still plenty of other maintenance, like delta-v maneuvers that need to get done or the system will eventually fail. Likely within a year.

In japan, in the typical post-apocalypse setting, small drifts are indeed not a problem. Who cares if your GPS has drifted by a few centimeter? That apocalypse may have moved the country by meters.

This system will probably be useful for quickly determining what moved in an earthquake.

Rubidium clocks aren't magically perfect at all timescales. They're better than quartz at long timescales, but they still need periodic recalibration.