However, I am looking at the Wikipedia list of runways in service today. I'm not about to embark on looking at historical data.
The core shuttle design to have rocket engines on the orbiter makes substantially more sense if there was at one point a planned orbital delta-v component. Orbital delta-v really only makes sense for military applications involving satellite capture/servicing or nuclear strike. As all of these can be better done with missiles/robots, there isn't much advantage in a space shuttle except for the case where the payload must be returned from orbit. On a polar trajectory the only use cases where this would make sense are spy satellite returns or weapons missions. Replacing the spy satellite entirely is obviously cheaper than returning and refurbishing it - which leaves a weapons use case which was likely discarded sometime late in the program when it was determined to be logistically infeasible or difficult to pull off.
Yes, most (maybe all) shuttle rollouts were less than the 10000 length of the Machrihanish runway, but that's really cutting it close. SLF Florida is 15000 feet. But I guess if the situation is so bad that the shuttle needs to come down outside the US, it's probably not going to be used again so why not mow the grass at the end of the runway.