Dear Reg:

You have articulated one of the many reasons why HF propagation is
described in stochastic terms. As you know very well, measurements or
predictions comprise at least two numbers: the best estimate of the number
and an estimate of the uncertainty of the first number. Present models of
HF propagation, which include antenna characteristics, provide both numbers.
Early models of HF propagation tended to be somewhat deficient in providing
the second number.

However, I remember using the early models to predict (extrapolate) in
real-time how much longer a certain frequency was likely to remain usable
from noting the drop-out of a higher frequency. The physics involved has
been understood for many years. It takes a long period of data gathering to
be able to do a good job with the second number.

In the early days (post WW2) of radio astronomy, the uncertainties of
some important measurements were greater than the estimate. That did not
last.

73 Mac N8TT

P.S. Nice to know that W2DU is back.
--
J. Mc Laughlin; Michigan U.S.A.
Home:
"Reg Edwards" wrote in message
...

But is it actually reflection?

no, it is actually a refraction. but it is useful sometimes to
model it as
a reflection from a slightly higher level. that makes computation
of angles
of incidence and height a bit easier.

============================================
The trigonometry is quite simple. Things become complicated when the
reflecting layer is not horizontal, ie., the layer is tilted.

The angle and direction of tilt are very difficult to predict.
Consequently, where on the surface of the Earth a ray returns is
anybody's guess.

This makes the vertical take-off angle, reported by antenna modelling
programs, even less useful.
----
Reg.