Compared to the cost of building/certifying/maintaining/launching a shuttle based nuclear capability - the relative cost of building redundant landing sites is

Compared to the cost of building/certifying/maintaining/launching a shuttle based nuclear capability - the relative cost of building redundant landing sites is trivial. Although I am surprised that the salt-lakes or groom lake runway couldn't be used for a shuttle landing site on a polar orbit. 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.