The way I see it, if you have two stations, one on the ground at 350
feet MSL, and the other on the ground (on top of a mountain) at 5000
feet MSL, and they are both "nearby", then the one that is actually at
5000 feet MSL will give an altimeter setting that will be more accurate
for an airplane that's flying at 5000 feet MSL.

The altimeter setting from the 350' station will be correct at that 350'
elevation, but the altitude indicated by using that setting in the
window will be a guess (based on standard lapse rate and other such) for
an airplane at 5000 feet. Granted, usually a pretty good guess, but
altimeter altitude is still an indirect inference from other parameters
(pressure and such). If the actual atmosphere that day does not follow
the theoretical average curves, the actual altitude of an airplane that
is indicating 5000 feet using an altimeter setting from 350' will be
somewhat off.

Of course, the airplane that uses the 5000' station's altimeter setting
while at 5000 feet, and proceeds to land at the 350' high airstrip, will
likely find the indicated altitude once on the ground to be different
from 350' (by about the same amount)

Given this, it makes sense to me (depending on the accuracy needed and
the actual difference from standard lapse rate) that flights =through=
an area might use one setting, and flights =to= an area might use
another one - at least when setting up to land.

Jose
--
Get high on gasoline: fly an airplane.
for Email, make the obvious change in the address.