Reg, G4FGQ wrote:
"How is it done?"

We have to do it within the USA broadcast frequencies with the following
method.

The site of the transmitting antenna is plotted on a very accurate map.

Pick map sites along radiaal lines from the antenna which are accessible
and free from possible reradiation sources (hard to do within a city)
but many sites along a radial will work.

For a single-tower, the nearest measurement site should be at least 5x
the tower height away. For a directional array, the nearest measurement
site should be at least 10x the widest gap between towers in the array.
You need to be far enough away so the antenna system appears to be a
point source.

You need to make a log of the measurements you make, showing the site
distance from the transmitter, measured field strength, time and
conditions which influence the measurement. You need to be able to
duplicate the measurements. You would prefer to make the first set of
measurements with the antenna operating in a nondirectional mode even if
it normally does not operate nondirectionally, so you can determine
efficiency very simply.

The more sites and measurements, the better. 25 measurements along each
radial is often considered enough for a nondirectional antenna. 40 or 50
would be required 9in a directional array, as the number of radial
measurements needed depends on the complexity of antenna system and its
pattern.

After completing measurements along a single radial, they should be
analyzed to determine the effective field at one mile from the antenna,
and the effective ground conductivity.

Fortunately, the FCC publishes charts are made a part of the rules in
Part 73 of the FCC Rules. You have likely seen reproductions in many
textbooks. I have an old copy of all the groundwave field intensity
versus conductivity charts which divide the AM broadcast band into
frequency segments.

These FCC charts contain more information than we can use, but they also
have what we need.

At the top of the chart is a straight line that shows how the signal
would be attenuated over perfectly conducting earth. The field strength
value at one mile is 100 mV / m. At 2 miles, it`s 50 millivolts / m, and
so on. This is as expected as over perfedt earth the signal varies
inversely with distance from the transmitter.

Beliw the straight line on the chart is a family of curves, each
dedicated to a particular soil conductivity. There is a curve for sea
water, 5.000 millisiemens (millimhos) and there is a curve for about as
nonconductive soil as is found (0.5 millisiemens), and there are several
curves in between those extremes.

All of the FCC curves are based on 100 mV / m at 1 mile, but can be
scaled. If your transmitter delivers 500 mV / m at 1 mile, aimply
multiply all points on the curve by 5.

We want to find the conductivity of our earth. It can be different on
every radial parh from the antenna.We find conductivity by plottibg our
measured field intensities on translucent graph paper with grid lines
which match the fcc graph. Then we line them up and place them over a
light source. We can see which of the FCC curves our points most closely
follow. It`s labeled ewith its conductivity.

Best regards, Richard Harrison, KB5WZI