There is an excellent illustration of land surface station locations and their change over time on a website at the University of Delaware (thanks to Ross McKittrick for alerting me to this). Cort Willmott of the University of Delaware is the original source for this information.
The information includes
There is a remarkable reduction in station density after about 1990. An appropriate research question is whether this loss of stations is part of the reason for the greater increase in land surface surface air temperatures in the 1990s, since, as documented in Climate Science (e.g. see), a singinficant number of the current stations are poorly sited with respect to monitoring tenths of a degree C per decade trends.