NavList:

Frank, I guess that the horizon is even further above the earth's surface than mountains or clouds. It is likely that height of the tangent point above the horizon where a ray, crossing all of the atmosphere (two times: incoming and outgoing), is totally extincted. If an object is not sufficient bright we can't even see it through (one time) the atmospheric thickness: At sea level we some times can't even see the (full) moon setting at the horizon because the extinction from the atmosphere is so high that it dimms off at some altitude above the horizon. Now in the photo the object (horizon) is not bright, it's dark. The only bright feature is the blue from the illuminated atmosphere. This requires considerably less to be totally extincted. It would actually be interesting to know at which height this total extinction occurs. It is probably different when looking at the illuminated or at the "dark" side of the earth. Does this photo provide an estimate for it?

Regarding refraction: I hope to find some time later today to explain that there is a corrective contribution to be added to the geometrical dip. This is a geometrical correction. The refraction however has to be subtracted.

Marcel