Caption: Illustrated are 3 types of horizon: true horizon, visible horizon, and celestial horizon (AKA astronomical)---which we usually just call horizon for short.
Features:
The celestial horizon is the limit of the true horizon when the observing point is actually on the sphere for which it is specified.
The celestial horizon is usually what we always mean in astronomy---except when we don't.
Let the radius of a exactly spherical astronomical object be R, the height of the observation point be h, and the line-of-sight distance to the true horizon be d.
The angle at the horizon between a radius and the line of sight is a right angle since the line of sight is tangent to the spherical surface of the astronomical object at the true horizon by definition.
From the Pythagorean theorem, we find
Using the approximate formula we find the following reference results given in Table: True Horizon Reference Distances below for a fiducial tall human of observing from h = 2 meters.
In fact, it seems almost impossible to notice the curvature of the
Earth just from looking
at the visible horizon
even over calm bodies of water
where the visible horizon
approximates the true horizon.
However, a person with good eye sight
under clear, calm
weather conditions can see ships, etc., rise or
sink below the visible horizon
over bodies of water.
This sinking-below-the-horizon effect
was recorded first by Strabo (63? BCE--24? CE)
in Classical antiquity
(see Wikipedia: Spherical Earth: Roman Empire).
He thought effect had been well know for
centuries by
sailors and other well informed people.
It has probably been known well into prehistory
by many people who never bothered to interpret it as implying a
spherical Earth.
In the many years of looking over
Lake Erie, yours truly---and
yours truly was born right on the shoreline---staring at
boats (and that
country
yours truly used to refer to when
very young as the United Steaks)
yours truly
has completely failed to notice
the sinking-below-the-horizon effect.
Looking at images taken on the
lunar surface,
it seems as if the visible horizon
is noticeably
closer than what we see on Earth
(see
Wikimedia commons:
Category: Astronaut photography on lunar surface).
This may be partially due to the
shorter distance to
the true horizon, but it may also
be partially due to the sharpness of images because of lack of
an atmosphere
and the stark black of the
lunar sky.
See the discussion at
Wikipedia: Examination of Apollo Moon photographs:
Apparently identical backgrounds in Apollo 15 photographs taken at different locations, but it doesn't seem definitive to me.
Table: True Horizon Reference Distances
Astronomical Mean Radius True Horizon Distance True Horizon Distant
Object Zeroth Order Formula Exact Formula
with h = 2 m with h = 2 m
(km) (km) (km)
General R 5.0482*sqrt(R/R_⊕) 5.0482*sqrt(R/R_⊕)*sqrt(1+(0.002 km)/(2R))
Earth 6371.0 5.0482 5.0482
Mars 3389.5 3.6821 3.6281
Moon 1737.1 2.6360 2.6360
Ceres 473 1.38 1.38
1950 DA 1.1 0.066332 0.066363
Notes: