Mark wrote:

I always think in terms of wavelength when calculating
the approximate efficiency of an elevated radial set.
For instance, three radials at 1/2 wave up will be pretty
much equal to about 120 on the ground.
Three at 1/4 wave will be equal to about 50-60 on the ground.
Three at 1/8 wave might be equal to 15-20 on the ground.
Three at cigarette pack height will be equal to about twice
as many as actually used at best. "slightly guessing
on that one, but my real world tests seem to pretty much
agree".
So being as the increase is fairly small at such low heights
in wavelength, it is probably practical to just bury them so
people won't trip over them.. :/

I have explored what you have said in an NEC4 model of a quarter wave
monopole with three quarter wave radials at varying heights over 'average
ground'. The results are summarised at
http://www.vk1od.net/lost/Clip053a.png . The reference for the graph is
the efficiency of the same antenna with 120 buried radials in the same
soil type.

If the models are correct, laying just a few radials on or very close to
the ground (eg the popular method of pinned into the turf) would appear
to be a very poor option. The model indicates efficiency improves with a
very small increase in height above the dirt, just 30mm is a 6dB
improvement of lying on the dirt, just half a metre achieves 90% of the
available efficiency.

Owen