Anyone know of a discussion - ideally available online - of the
preferred criteria for locating an observing site? (To be clear,
recommendations for the site itself, ie the physical geography, not
for instrument positioning).

There's some obvious things like preferring at least a semi-rural and
not a built-up location, ensuring a suitable distance from trees and
other obstacles. But what I have in mind more is whether there's any
official preference say for a hill-top location where there's
obviously going to be better wind exposure, but which may not be very
representative of the surrounding area. Maybe hilltops should be
avoided? Is it best to avoid locations near to rivers, reservoirs etc,
again as being potentially unrepresentative of the area as a whole?
And so on.

I suspect that many official sites were set up years ago in particular
locations chosen for some rather arbitrary reason such as availability
or convenience. But if you were to have at least some choice for
locating a new site then are there any guidelines for deciding between
possible sites?

John Dann
www.weatherstations.co.uk