In article id,
Dr J R Stockton writes:
I assume the atmospheric effects are almost entirely on the deflection
of the sunlight, and only slightly on the actual rotation of the Earth.

That's correct. The orientation of the Earth in space is defined by
"UT1"(and to a much smaller extent by polar motion and precession/
nutation). UT1 is _predicted_ to some amazing accuracy (I think it's
microseconds but could be wrong.) and known after the fact much
better than that. Details are no doubt given at

http://www.iers.org

but I don't find them in an instant and don't have patience to search
the site more diligently. Measurements come from VLBI monitoring of
a network of quasars.

As Mike mentioned, visible sunrise and sunset times are sensitive to
local atmospheric conditions (mainly density, which depends on
pressure and humidity) along the line of sight to the horizon.

--
Help keep our newsgroup healthy; please don't feed the trolls.
Steve Willner Phone 617-495-7123
Cambridge, MA 02138 USA